Show More
@@ -1,621 +1,621 b'' | |||
|
1 | 1 | # encoding: utf-8 |
|
2 | 2 | """A base class for a configurable application.""" |
|
3 | 3 | |
|
4 | 4 | # Copyright (c) IPython Development Team. |
|
5 | 5 | # Distributed under the terms of the Modified BSD License. |
|
6 | 6 | |
|
7 | 7 | from __future__ import print_function |
|
8 | 8 | |
|
9 | 9 | import json |
|
10 | 10 | import logging |
|
11 | 11 | import os |
|
12 | 12 | import re |
|
13 | 13 | import sys |
|
14 | 14 | from copy import deepcopy |
|
15 | 15 | from collections import defaultdict |
|
16 | 16 | |
|
17 |
from |
|
|
17 | from decorator import decorator | |
|
18 | 18 | |
|
19 | 19 | from IPython.config.configurable import SingletonConfigurable |
|
20 | 20 | from IPython.config.loader import ( |
|
21 | 21 | KVArgParseConfigLoader, PyFileConfigLoader, Config, ArgumentError, ConfigFileNotFound, JSONFileConfigLoader |
|
22 | 22 | ) |
|
23 | 23 | |
|
24 | 24 | from IPython.utils.traitlets import ( |
|
25 | 25 | Unicode, List, Enum, Dict, Instance, TraitError |
|
26 | 26 | ) |
|
27 | 27 | from IPython.utils.importstring import import_item |
|
28 | 28 | from IPython.utils.text import indent, wrap_paragraphs, dedent |
|
29 | 29 | from IPython.utils import py3compat |
|
30 | 30 | from IPython.utils.py3compat import string_types, iteritems |
|
31 | 31 | |
|
32 | 32 | #----------------------------------------------------------------------------- |
|
33 | 33 | # Descriptions for the various sections |
|
34 | 34 | #----------------------------------------------------------------------------- |
|
35 | 35 | |
|
36 | 36 | # merge flags&aliases into options |
|
37 | 37 | option_description = """ |
|
38 | 38 | Arguments that take values are actually convenience aliases to full |
|
39 | 39 | Configurables, whose aliases are listed on the help line. For more information |
|
40 | 40 | on full configurables, see '--help-all'. |
|
41 | 41 | """.strip() # trim newlines of front and back |
|
42 | 42 | |
|
43 | 43 | keyvalue_description = """ |
|
44 | 44 | Parameters are set from command-line arguments of the form: |
|
45 | 45 | `--Class.trait=value`. |
|
46 | 46 | This line is evaluated in Python, so simple expressions are allowed, e.g.:: |
|
47 | 47 | `--C.a='range(3)'` For setting C.a=[0,1,2]. |
|
48 | 48 | """.strip() # trim newlines of front and back |
|
49 | 49 | |
|
50 | 50 | # sys.argv can be missing, for example when python is embedded. See the docs |
|
51 | 51 | # for details: http://docs.python.org/2/c-api/intro.html#embedding-python |
|
52 | 52 | if not hasattr(sys, "argv"): |
|
53 | 53 | sys.argv = [""] |
|
54 | 54 | |
|
55 | 55 | subcommand_description = """ |
|
56 | 56 | Subcommands are launched as `{app} cmd [args]`. For information on using |
|
57 | 57 | subcommand 'cmd', do: `{app} cmd -h`. |
|
58 | 58 | """ |
|
59 | 59 | # get running program name |
|
60 | 60 | |
|
61 | 61 | #----------------------------------------------------------------------------- |
|
62 | 62 | # Application class |
|
63 | 63 | #----------------------------------------------------------------------------- |
|
64 | 64 | |
|
65 | 65 | @decorator |
|
66 | 66 | def catch_config_error(method, app, *args, **kwargs): |
|
67 | 67 | """Method decorator for catching invalid config (Trait/ArgumentErrors) during init. |
|
68 | 68 | |
|
69 | 69 | On a TraitError (generally caused by bad config), this will print the trait's |
|
70 | 70 | message, and exit the app. |
|
71 | 71 | |
|
72 | 72 | For use on init methods, to prevent invoking excepthook on invalid input. |
|
73 | 73 | """ |
|
74 | 74 | try: |
|
75 | 75 | return method(app, *args, **kwargs) |
|
76 | 76 | except (TraitError, ArgumentError) as e: |
|
77 | 77 | app.print_help() |
|
78 | 78 | app.log.fatal("Bad config encountered during initialization:") |
|
79 | 79 | app.log.fatal(str(e)) |
|
80 | 80 | app.log.debug("Config at the time: %s", app.config) |
|
81 | 81 | app.exit(1) |
|
82 | 82 | |
|
83 | 83 | |
|
84 | 84 | class ApplicationError(Exception): |
|
85 | 85 | pass |
|
86 | 86 | |
|
87 | 87 | class LevelFormatter(logging.Formatter): |
|
88 | 88 | """Formatter with additional `highlevel` record |
|
89 | 89 | |
|
90 | 90 | This field is empty if log level is less than highlevel_limit, |
|
91 | 91 | otherwise it is formatted with self.highlevel_format. |
|
92 | 92 | |
|
93 | 93 | Useful for adding 'WARNING' to warning messages, |
|
94 | 94 | without adding 'INFO' to info, etc. |
|
95 | 95 | """ |
|
96 | 96 | highlevel_limit = logging.WARN |
|
97 | 97 | highlevel_format = " %(levelname)s |" |
|
98 | 98 | |
|
99 | 99 | def format(self, record): |
|
100 | 100 | if record.levelno >= self.highlevel_limit: |
|
101 | 101 | record.highlevel = self.highlevel_format % record.__dict__ |
|
102 | 102 | else: |
|
103 | 103 | record.highlevel = "" |
|
104 | 104 | return super(LevelFormatter, self).format(record) |
|
105 | 105 | |
|
106 | 106 | |
|
107 | 107 | class Application(SingletonConfigurable): |
|
108 | 108 | """A singleton application with full configuration support.""" |
|
109 | 109 | |
|
110 | 110 | # The name of the application, will usually match the name of the command |
|
111 | 111 | # line application |
|
112 | 112 | name = Unicode(u'application') |
|
113 | 113 | |
|
114 | 114 | # The description of the application that is printed at the beginning |
|
115 | 115 | # of the help. |
|
116 | 116 | description = Unicode(u'This is an application.') |
|
117 | 117 | # default section descriptions |
|
118 | 118 | option_description = Unicode(option_description) |
|
119 | 119 | keyvalue_description = Unicode(keyvalue_description) |
|
120 | 120 | subcommand_description = Unicode(subcommand_description) |
|
121 | 121 | |
|
122 | 122 | # The usage and example string that goes at the end of the help string. |
|
123 | 123 | examples = Unicode() |
|
124 | 124 | |
|
125 | 125 | # A sequence of Configurable subclasses whose config=True attributes will |
|
126 | 126 | # be exposed at the command line. |
|
127 | 127 | classes = [] |
|
128 | 128 | @property |
|
129 | 129 | def _help_classes(self): |
|
130 | 130 | """Define `App.help_classes` if CLI classes should differ from config file classes""" |
|
131 | 131 | return getattr(self, 'help_classes', self.classes) |
|
132 | 132 | |
|
133 | 133 | @property |
|
134 | 134 | def _config_classes(self): |
|
135 | 135 | """Define `App.config_classes` if config file classes should differ from CLI classes.""" |
|
136 | 136 | return getattr(self, 'config_classes', self.classes) |
|
137 | 137 | |
|
138 | 138 | # The version string of this application. |
|
139 | 139 | version = Unicode(u'0.0') |
|
140 | 140 | |
|
141 | 141 | # the argv used to initialize the application |
|
142 | 142 | argv = List() |
|
143 | 143 | |
|
144 | 144 | # The log level for the application |
|
145 | 145 | log_level = Enum((0,10,20,30,40,50,'DEBUG','INFO','WARN','ERROR','CRITICAL'), |
|
146 | 146 | default_value=logging.WARN, |
|
147 | 147 | config=True, |
|
148 | 148 | help="Set the log level by value or name.") |
|
149 | 149 | def _log_level_changed(self, name, old, new): |
|
150 | 150 | """Adjust the log level when log_level is set.""" |
|
151 | 151 | if isinstance(new, string_types): |
|
152 | 152 | new = getattr(logging, new) |
|
153 | 153 | self.log_level = new |
|
154 | 154 | self.log.setLevel(new) |
|
155 | 155 | |
|
156 | 156 | _log_formatter_cls = LevelFormatter |
|
157 | 157 | |
|
158 | 158 | log_datefmt = Unicode("%Y-%m-%d %H:%M:%S", config=True, |
|
159 | 159 | help="The date format used by logging formatters for %(asctime)s" |
|
160 | 160 | ) |
|
161 | 161 | def _log_datefmt_changed(self, name, old, new): |
|
162 | 162 | self._log_format_changed('log_format', self.log_format, self.log_format) |
|
163 | 163 | |
|
164 | 164 | log_format = Unicode("[%(name)s]%(highlevel)s %(message)s", config=True, |
|
165 | 165 | help="The Logging format template", |
|
166 | 166 | ) |
|
167 | 167 | def _log_format_changed(self, name, old, new): |
|
168 | 168 | """Change the log formatter when log_format is set.""" |
|
169 | 169 | _log_handler = self.log.handlers[0] |
|
170 | 170 | _log_formatter = self._log_formatter_cls(fmt=new, datefmt=self.log_datefmt) |
|
171 | 171 | _log_handler.setFormatter(_log_formatter) |
|
172 | 172 | |
|
173 | 173 | |
|
174 | 174 | log = Instance(logging.Logger) |
|
175 | 175 | def _log_default(self): |
|
176 | 176 | """Start logging for this application. |
|
177 | 177 | |
|
178 | 178 | The default is to log to stderr using a StreamHandler, if no default |
|
179 | 179 | handler already exists. The log level starts at logging.WARN, but this |
|
180 | 180 | can be adjusted by setting the ``log_level`` attribute. |
|
181 | 181 | """ |
|
182 | 182 | log = logging.getLogger(self.__class__.__name__) |
|
183 | 183 | log.setLevel(self.log_level) |
|
184 | 184 | log.propagate = False |
|
185 | 185 | _log = log # copied from Logger.hasHandlers() (new in Python 3.2) |
|
186 | 186 | while _log: |
|
187 | 187 | if _log.handlers: |
|
188 | 188 | return log |
|
189 | 189 | if not _log.propagate: |
|
190 | 190 | break |
|
191 | 191 | else: |
|
192 | 192 | _log = _log.parent |
|
193 | 193 | if sys.executable.endswith('pythonw.exe'): |
|
194 | 194 | # this should really go to a file, but file-logging is only |
|
195 | 195 | # hooked up in parallel applications |
|
196 | 196 | _log_handler = logging.StreamHandler(open(os.devnull, 'w')) |
|
197 | 197 | else: |
|
198 | 198 | _log_handler = logging.StreamHandler() |
|
199 | 199 | _log_formatter = self._log_formatter_cls(fmt=self.log_format, datefmt=self.log_datefmt) |
|
200 | 200 | _log_handler.setFormatter(_log_formatter) |
|
201 | 201 | log.addHandler(_log_handler) |
|
202 | 202 | return log |
|
203 | 203 | |
|
204 | 204 | # the alias map for configurables |
|
205 | 205 | aliases = Dict({'log-level' : 'Application.log_level'}) |
|
206 | 206 | |
|
207 | 207 | # flags for loading Configurables or store_const style flags |
|
208 | 208 | # flags are loaded from this dict by '--key' flags |
|
209 | 209 | # this must be a dict of two-tuples, the first element being the Config/dict |
|
210 | 210 | # and the second being the help string for the flag |
|
211 | 211 | flags = Dict() |
|
212 | 212 | def _flags_changed(self, name, old, new): |
|
213 | 213 | """ensure flags dict is valid""" |
|
214 | 214 | for key,value in iteritems(new): |
|
215 | 215 | assert len(value) == 2, "Bad flag: %r:%s"%(key,value) |
|
216 | 216 | assert isinstance(value[0], (dict, Config)), "Bad flag: %r:%s"%(key,value) |
|
217 | 217 | assert isinstance(value[1], string_types), "Bad flag: %r:%s"%(key,value) |
|
218 | 218 | |
|
219 | 219 | |
|
220 | 220 | # subcommands for launching other applications |
|
221 | 221 | # if this is not empty, this will be a parent Application |
|
222 | 222 | # this must be a dict of two-tuples, |
|
223 | 223 | # the first element being the application class/import string |
|
224 | 224 | # and the second being the help string for the subcommand |
|
225 | 225 | subcommands = Dict() |
|
226 | 226 | # parse_command_line will initialize a subapp, if requested |
|
227 | 227 | subapp = Instance('IPython.config.application.Application', allow_none=True) |
|
228 | 228 | |
|
229 | 229 | # extra command-line arguments that don't set config values |
|
230 | 230 | extra_args = List(Unicode) |
|
231 | 231 | |
|
232 | 232 | |
|
233 | 233 | def __init__(self, **kwargs): |
|
234 | 234 | SingletonConfigurable.__init__(self, **kwargs) |
|
235 | 235 | # Ensure my class is in self.classes, so my attributes appear in command line |
|
236 | 236 | # options and config files. |
|
237 | 237 | if self.__class__ not in self.classes: |
|
238 | 238 | self.classes.insert(0, self.__class__) |
|
239 | 239 | |
|
240 | 240 | def _config_changed(self, name, old, new): |
|
241 | 241 | SingletonConfigurable._config_changed(self, name, old, new) |
|
242 | 242 | self.log.debug('Config changed:') |
|
243 | 243 | self.log.debug(repr(new)) |
|
244 | 244 | |
|
245 | 245 | @catch_config_error |
|
246 | 246 | def initialize(self, argv=None): |
|
247 | 247 | """Do the basic steps to configure me. |
|
248 | 248 | |
|
249 | 249 | Override in subclasses. |
|
250 | 250 | """ |
|
251 | 251 | self.parse_command_line(argv) |
|
252 | 252 | |
|
253 | 253 | |
|
254 | 254 | def start(self): |
|
255 | 255 | """Start the app mainloop. |
|
256 | 256 | |
|
257 | 257 | Override in subclasses. |
|
258 | 258 | """ |
|
259 | 259 | if self.subapp is not None: |
|
260 | 260 | return self.subapp.start() |
|
261 | 261 | |
|
262 | 262 | def print_alias_help(self): |
|
263 | 263 | """Print the alias part of the help.""" |
|
264 | 264 | if not self.aliases: |
|
265 | 265 | return |
|
266 | 266 | |
|
267 | 267 | lines = [] |
|
268 | 268 | classdict = {} |
|
269 | 269 | for cls in self._help_classes: |
|
270 | 270 | # include all parents (up to, but excluding Configurable) in available names |
|
271 | 271 | for c in cls.mro()[:-3]: |
|
272 | 272 | classdict[c.__name__] = c |
|
273 | 273 | |
|
274 | 274 | for alias, longname in iteritems(self.aliases): |
|
275 | 275 | classname, traitname = longname.split('.',1) |
|
276 | 276 | cls = classdict[classname] |
|
277 | 277 | |
|
278 | 278 | trait = cls.class_traits(config=True)[traitname] |
|
279 | 279 | help = cls.class_get_trait_help(trait).splitlines() |
|
280 | 280 | # reformat first line |
|
281 | 281 | help[0] = help[0].replace(longname, alias) + ' (%s)'%longname |
|
282 | 282 | if len(alias) == 1: |
|
283 | 283 | help[0] = help[0].replace('--%s='%alias, '-%s '%alias) |
|
284 | 284 | lines.extend(help) |
|
285 | 285 | # lines.append('') |
|
286 | 286 | print(os.linesep.join(lines)) |
|
287 | 287 | |
|
288 | 288 | def print_flag_help(self): |
|
289 | 289 | """Print the flag part of the help.""" |
|
290 | 290 | if not self.flags: |
|
291 | 291 | return |
|
292 | 292 | |
|
293 | 293 | lines = [] |
|
294 | 294 | for m, (cfg,help) in iteritems(self.flags): |
|
295 | 295 | prefix = '--' if len(m) > 1 else '-' |
|
296 | 296 | lines.append(prefix+m) |
|
297 | 297 | lines.append(indent(dedent(help.strip()))) |
|
298 | 298 | # lines.append('') |
|
299 | 299 | print(os.linesep.join(lines)) |
|
300 | 300 | |
|
301 | 301 | def print_options(self): |
|
302 | 302 | if not self.flags and not self.aliases: |
|
303 | 303 | return |
|
304 | 304 | lines = ['Options'] |
|
305 | 305 | lines.append('-'*len(lines[0])) |
|
306 | 306 | lines.append('') |
|
307 | 307 | for p in wrap_paragraphs(self.option_description): |
|
308 | 308 | lines.append(p) |
|
309 | 309 | lines.append('') |
|
310 | 310 | print(os.linesep.join(lines)) |
|
311 | 311 | self.print_flag_help() |
|
312 | 312 | self.print_alias_help() |
|
313 | 313 | print() |
|
314 | 314 | |
|
315 | 315 | def print_subcommands(self): |
|
316 | 316 | """Print the subcommand part of the help.""" |
|
317 | 317 | if not self.subcommands: |
|
318 | 318 | return |
|
319 | 319 | |
|
320 | 320 | lines = ["Subcommands"] |
|
321 | 321 | lines.append('-'*len(lines[0])) |
|
322 | 322 | lines.append('') |
|
323 | 323 | for p in wrap_paragraphs(self.subcommand_description.format( |
|
324 | 324 | app=self.name)): |
|
325 | 325 | lines.append(p) |
|
326 | 326 | lines.append('') |
|
327 | 327 | for subc, (cls, help) in iteritems(self.subcommands): |
|
328 | 328 | lines.append(subc) |
|
329 | 329 | if help: |
|
330 | 330 | lines.append(indent(dedent(help.strip()))) |
|
331 | 331 | lines.append('') |
|
332 | 332 | print(os.linesep.join(lines)) |
|
333 | 333 | |
|
334 | 334 | def print_help(self, classes=False): |
|
335 | 335 | """Print the help for each Configurable class in self.classes. |
|
336 | 336 | |
|
337 | 337 | If classes=False (the default), only flags and aliases are printed. |
|
338 | 338 | """ |
|
339 | 339 | self.print_description() |
|
340 | 340 | self.print_subcommands() |
|
341 | 341 | self.print_options() |
|
342 | 342 | |
|
343 | 343 | if classes: |
|
344 | 344 | help_classes = self._help_classes |
|
345 | 345 | if help_classes: |
|
346 | 346 | print("Class parameters") |
|
347 | 347 | print("----------------") |
|
348 | 348 | print() |
|
349 | 349 | for p in wrap_paragraphs(self.keyvalue_description): |
|
350 | 350 | print(p) |
|
351 | 351 | print() |
|
352 | 352 | |
|
353 | 353 | for cls in help_classes: |
|
354 | 354 | cls.class_print_help() |
|
355 | 355 | print() |
|
356 | 356 | else: |
|
357 | 357 | print("To see all available configurables, use `--help-all`") |
|
358 | 358 | print() |
|
359 | 359 | |
|
360 | 360 | self.print_examples() |
|
361 | 361 | |
|
362 | 362 | |
|
363 | 363 | def print_description(self): |
|
364 | 364 | """Print the application description.""" |
|
365 | 365 | for p in wrap_paragraphs(self.description): |
|
366 | 366 | print(p) |
|
367 | 367 | print() |
|
368 | 368 | |
|
369 | 369 | def print_examples(self): |
|
370 | 370 | """Print usage and examples. |
|
371 | 371 | |
|
372 | 372 | This usage string goes at the end of the command line help string |
|
373 | 373 | and should contain examples of the application's usage. |
|
374 | 374 | """ |
|
375 | 375 | if self.examples: |
|
376 | 376 | print("Examples") |
|
377 | 377 | print("--------") |
|
378 | 378 | print() |
|
379 | 379 | print(indent(dedent(self.examples.strip()))) |
|
380 | 380 | print() |
|
381 | 381 | |
|
382 | 382 | def print_version(self): |
|
383 | 383 | """Print the version string.""" |
|
384 | 384 | print(self.version) |
|
385 | 385 | |
|
386 | 386 | def update_config(self, config): |
|
387 | 387 | """Fire the traits events when the config is updated.""" |
|
388 | 388 | # Save a copy of the current config. |
|
389 | 389 | newconfig = deepcopy(self.config) |
|
390 | 390 | # Merge the new config into the current one. |
|
391 | 391 | newconfig.merge(config) |
|
392 | 392 | # Save the combined config as self.config, which triggers the traits |
|
393 | 393 | # events. |
|
394 | 394 | self.config = newconfig |
|
395 | 395 | |
|
396 | 396 | @catch_config_error |
|
397 | 397 | def initialize_subcommand(self, subc, argv=None): |
|
398 | 398 | """Initialize a subcommand with argv.""" |
|
399 | 399 | subapp,help = self.subcommands.get(subc) |
|
400 | 400 | |
|
401 | 401 | if isinstance(subapp, string_types): |
|
402 | 402 | subapp = import_item(subapp) |
|
403 | 403 | |
|
404 | 404 | # clear existing instances |
|
405 | 405 | self.__class__.clear_instance() |
|
406 | 406 | # instantiate |
|
407 | 407 | self.subapp = subapp.instance(config=self.config) |
|
408 | 408 | # and initialize subapp |
|
409 | 409 | self.subapp.initialize(argv) |
|
410 | 410 | |
|
411 | 411 | def flatten_flags(self): |
|
412 | 412 | """flatten flags and aliases, so cl-args override as expected. |
|
413 | 413 | |
|
414 | 414 | This prevents issues such as an alias pointing to InteractiveShell, |
|
415 | 415 | but a config file setting the same trait in TerminalInteraciveShell |
|
416 | 416 | getting inappropriate priority over the command-line arg. |
|
417 | 417 | |
|
418 | 418 | Only aliases with exactly one descendent in the class list |
|
419 | 419 | will be promoted. |
|
420 | 420 | |
|
421 | 421 | """ |
|
422 | 422 | # build a tree of classes in our list that inherit from a particular |
|
423 | 423 | # it will be a dict by parent classname of classes in our list |
|
424 | 424 | # that are descendents |
|
425 | 425 | mro_tree = defaultdict(list) |
|
426 | 426 | for cls in self._help_classes: |
|
427 | 427 | clsname = cls.__name__ |
|
428 | 428 | for parent in cls.mro()[1:-3]: |
|
429 | 429 | # exclude cls itself and Configurable,HasTraits,object |
|
430 | 430 | mro_tree[parent.__name__].append(clsname) |
|
431 | 431 | # flatten aliases, which have the form: |
|
432 | 432 | # { 'alias' : 'Class.trait' } |
|
433 | 433 | aliases = {} |
|
434 | 434 | for alias, cls_trait in iteritems(self.aliases): |
|
435 | 435 | cls,trait = cls_trait.split('.',1) |
|
436 | 436 | children = mro_tree[cls] |
|
437 | 437 | if len(children) == 1: |
|
438 | 438 | # exactly one descendent, promote alias |
|
439 | 439 | cls = children[0] |
|
440 | 440 | aliases[alias] = '.'.join([cls,trait]) |
|
441 | 441 | |
|
442 | 442 | # flatten flags, which are of the form: |
|
443 | 443 | # { 'key' : ({'Cls' : {'trait' : value}}, 'help')} |
|
444 | 444 | flags = {} |
|
445 | 445 | for key, (flagdict, help) in iteritems(self.flags): |
|
446 | 446 | newflag = {} |
|
447 | 447 | for cls, subdict in iteritems(flagdict): |
|
448 | 448 | children = mro_tree[cls] |
|
449 | 449 | # exactly one descendent, promote flag section |
|
450 | 450 | if len(children) == 1: |
|
451 | 451 | cls = children[0] |
|
452 | 452 | newflag[cls] = subdict |
|
453 | 453 | flags[key] = (newflag, help) |
|
454 | 454 | return flags, aliases |
|
455 | 455 | |
|
456 | 456 | @catch_config_error |
|
457 | 457 | def parse_command_line(self, argv=None): |
|
458 | 458 | """Parse the command line arguments.""" |
|
459 | 459 | argv = sys.argv[1:] if argv is None else argv |
|
460 | 460 | self.argv = [ py3compat.cast_unicode(arg) for arg in argv ] |
|
461 | 461 | |
|
462 | 462 | if argv and argv[0] == 'help': |
|
463 | 463 | # turn `ipython help notebook` into `ipython notebook -h` |
|
464 | 464 | argv = argv[1:] + ['-h'] |
|
465 | 465 | |
|
466 | 466 | if self.subcommands and len(argv) > 0: |
|
467 | 467 | # we have subcommands, and one may have been specified |
|
468 | 468 | subc, subargv = argv[0], argv[1:] |
|
469 | 469 | if re.match(r'^\w(\-?\w)*$', subc) and subc in self.subcommands: |
|
470 | 470 | # it's a subcommand, and *not* a flag or class parameter |
|
471 | 471 | return self.initialize_subcommand(subc, subargv) |
|
472 | 472 | |
|
473 | 473 | # Arguments after a '--' argument are for the script IPython may be |
|
474 | 474 | # about to run, not IPython iteslf. For arguments parsed here (help and |
|
475 | 475 | # version), we want to only search the arguments up to the first |
|
476 | 476 | # occurrence of '--', which we're calling interpreted_argv. |
|
477 | 477 | try: |
|
478 | 478 | interpreted_argv = argv[:argv.index('--')] |
|
479 | 479 | except ValueError: |
|
480 | 480 | interpreted_argv = argv |
|
481 | 481 | |
|
482 | 482 | if any(x in interpreted_argv for x in ('-h', '--help-all', '--help')): |
|
483 | 483 | self.print_help('--help-all' in interpreted_argv) |
|
484 | 484 | self.exit(0) |
|
485 | 485 | |
|
486 | 486 | if '--version' in interpreted_argv or '-V' in interpreted_argv: |
|
487 | 487 | self.print_version() |
|
488 | 488 | self.exit(0) |
|
489 | 489 | |
|
490 | 490 | # flatten flags&aliases, so cl-args get appropriate priority: |
|
491 | 491 | flags,aliases = self.flatten_flags() |
|
492 | 492 | loader = KVArgParseConfigLoader(argv=argv, aliases=aliases, |
|
493 | 493 | flags=flags, log=self.log) |
|
494 | 494 | config = loader.load_config() |
|
495 | 495 | self.update_config(config) |
|
496 | 496 | # store unparsed args in extra_args |
|
497 | 497 | self.extra_args = loader.extra_args |
|
498 | 498 | |
|
499 | 499 | @classmethod |
|
500 | 500 | def _load_config_files(cls, basefilename, path=None, log=None): |
|
501 | 501 | """Load config files (py,json) by filename and path. |
|
502 | 502 | |
|
503 | 503 | yield each config object in turn. |
|
504 | 504 | """ |
|
505 | 505 | |
|
506 | 506 | if not isinstance(path, list): |
|
507 | 507 | path = [path] |
|
508 | 508 | for path in path[::-1]: |
|
509 | 509 | # path list is in descending priority order, so load files backwards: |
|
510 | 510 | pyloader = PyFileConfigLoader(basefilename+'.py', path=path, log=log) |
|
511 | 511 | jsonloader = JSONFileConfigLoader(basefilename+'.json', path=path, log=log) |
|
512 | 512 | config = None |
|
513 | 513 | for loader in [pyloader, jsonloader]: |
|
514 | 514 | try: |
|
515 | 515 | config = loader.load_config() |
|
516 | 516 | except ConfigFileNotFound: |
|
517 | 517 | pass |
|
518 | 518 | except Exception: |
|
519 | 519 | # try to get the full filename, but it will be empty in the |
|
520 | 520 | # unlikely event that the error raised before filefind finished |
|
521 | 521 | filename = loader.full_filename or basefilename |
|
522 | 522 | # problem while running the file |
|
523 | 523 | if log: |
|
524 | 524 | log.error("Exception while loading config file %s", |
|
525 | 525 | filename, exc_info=True) |
|
526 | 526 | else: |
|
527 | 527 | if log: |
|
528 | 528 | log.debug("Loaded config file: %s", loader.full_filename) |
|
529 | 529 | if config: |
|
530 | 530 | yield config |
|
531 | 531 | |
|
532 | 532 | raise StopIteration |
|
533 | 533 | |
|
534 | 534 | |
|
535 | 535 | @catch_config_error |
|
536 | 536 | def load_config_file(self, filename, path=None): |
|
537 | 537 | """Load config files by filename and path.""" |
|
538 | 538 | filename, ext = os.path.splitext(filename) |
|
539 | 539 | loaded = [] |
|
540 | 540 | for config in self._load_config_files(filename, path=path, log=self.log): |
|
541 | 541 | loaded.append(config) |
|
542 | 542 | self.update_config(config) |
|
543 | 543 | if len(loaded) > 1: |
|
544 | 544 | collisions = loaded[0].collisions(loaded[1]) |
|
545 | 545 | if collisions: |
|
546 | 546 | self.log.warn("Collisions detected in {0}.py and {0}.json config files." |
|
547 | 547 | " {0}.json has higher priority: {1}".format( |
|
548 | 548 | filename, json.dumps(collisions, indent=2), |
|
549 | 549 | )) |
|
550 | 550 | |
|
551 | 551 | |
|
552 | 552 | def generate_config_file(self): |
|
553 | 553 | """generate default config file from Configurables""" |
|
554 | 554 | lines = ["# Configuration file for %s."%self.name] |
|
555 | 555 | lines.append('') |
|
556 | 556 | lines.append('c = get_config()') |
|
557 | 557 | lines.append('') |
|
558 | 558 | for cls in self._config_classes: |
|
559 | 559 | lines.append(cls.class_config_section()) |
|
560 | 560 | return '\n'.join(lines) |
|
561 | 561 | |
|
562 | 562 | def exit(self, exit_status=0): |
|
563 | 563 | self.log.debug("Exiting application: %s" % self.name) |
|
564 | 564 | sys.exit(exit_status) |
|
565 | 565 | |
|
566 | 566 | @classmethod |
|
567 | 567 | def launch_instance(cls, argv=None, **kwargs): |
|
568 | 568 | """Launch a global instance of this Application |
|
569 | 569 | |
|
570 | 570 | If a global instance already exists, this reinitializes and starts it |
|
571 | 571 | """ |
|
572 | 572 | app = cls.instance(**kwargs) |
|
573 | 573 | app.initialize(argv) |
|
574 | 574 | app.start() |
|
575 | 575 | |
|
576 | 576 | #----------------------------------------------------------------------------- |
|
577 | 577 | # utility functions, for convenience |
|
578 | 578 | #----------------------------------------------------------------------------- |
|
579 | 579 | |
|
580 | 580 | def boolean_flag(name, configurable, set_help='', unset_help=''): |
|
581 | 581 | """Helper for building basic --trait, --no-trait flags. |
|
582 | 582 | |
|
583 | 583 | Parameters |
|
584 | 584 | ---------- |
|
585 | 585 | |
|
586 | 586 | name : str |
|
587 | 587 | The name of the flag. |
|
588 | 588 | configurable : str |
|
589 | 589 | The 'Class.trait' string of the trait to be set/unset with the flag |
|
590 | 590 | set_help : unicode |
|
591 | 591 | help string for --name flag |
|
592 | 592 | unset_help : unicode |
|
593 | 593 | help string for --no-name flag |
|
594 | 594 | |
|
595 | 595 | Returns |
|
596 | 596 | ------- |
|
597 | 597 | |
|
598 | 598 | cfg : dict |
|
599 | 599 | A dict with two keys: 'name', and 'no-name', for setting and unsetting |
|
600 | 600 | the trait, respectively. |
|
601 | 601 | """ |
|
602 | 602 | # default helpstrings |
|
603 | 603 | set_help = set_help or "set %s=True"%configurable |
|
604 | 604 | unset_help = unset_help or "set %s=False"%configurable |
|
605 | 605 | |
|
606 | 606 | cls,trait = configurable.split('.') |
|
607 | 607 | |
|
608 | 608 | setter = {cls : {trait : True}} |
|
609 | 609 | unsetter = {cls : {trait : False}} |
|
610 | 610 | return {name : (setter, set_help), 'no-'+name : (unsetter, unset_help)} |
|
611 | 611 | |
|
612 | 612 | |
|
613 | 613 | def get_config(): |
|
614 | 614 | """Get the config object for the global Application instance, if there is one |
|
615 | 615 | |
|
616 | 616 | otherwise return an empty config object |
|
617 | 617 | """ |
|
618 | 618 | if Application.initialized(): |
|
619 | 619 | return Application.instance().config |
|
620 | 620 | else: |
|
621 | 621 | return Config() |
@@ -1,965 +1,965 b'' | |||
|
1 | 1 | # -*- coding: utf-8 -*- |
|
2 | 2 | """Display formatters. |
|
3 | 3 | |
|
4 | 4 | Inheritance diagram: |
|
5 | 5 | |
|
6 | 6 | .. inheritance-diagram:: IPython.core.formatters |
|
7 | 7 | :parts: 3 |
|
8 | 8 | """ |
|
9 | 9 | |
|
10 | 10 | # Copyright (c) IPython Development Team. |
|
11 | 11 | # Distributed under the terms of the Modified BSD License. |
|
12 | 12 | |
|
13 | 13 | import abc |
|
14 | 14 | import inspect |
|
15 | 15 | import json |
|
16 | 16 | import sys |
|
17 | 17 | import traceback |
|
18 | 18 | import warnings |
|
19 | 19 | |
|
20 |
from |
|
|
20 | from decorator import decorator | |
|
21 | 21 | |
|
22 | 22 | from IPython.config.configurable import Configurable |
|
23 | 23 | from IPython.core.getipython import get_ipython |
|
24 | 24 | from IPython.lib import pretty |
|
25 | 25 | from IPython.utils.traitlets import ( |
|
26 | 26 | Bool, Dict, Integer, Unicode, CUnicode, ObjectName, List, |
|
27 | 27 | ForwardDeclaredInstance, |
|
28 | 28 | ) |
|
29 | 29 | from IPython.utils.py3compat import ( |
|
30 | 30 | with_metaclass, string_types, unicode_type, |
|
31 | 31 | ) |
|
32 | 32 | |
|
33 | 33 | |
|
34 | 34 | #----------------------------------------------------------------------------- |
|
35 | 35 | # The main DisplayFormatter class |
|
36 | 36 | #----------------------------------------------------------------------------- |
|
37 | 37 | |
|
38 | 38 | |
|
39 | 39 | def _safe_get_formatter_method(obj, name): |
|
40 | 40 | """Safely get a formatter method |
|
41 | 41 | |
|
42 | 42 | - Classes cannot have formatter methods, only instance |
|
43 | 43 | - protect against proxy objects that claim to have everything |
|
44 | 44 | """ |
|
45 | 45 | if inspect.isclass(obj): |
|
46 | 46 | # repr methods only make sense on instances, not classes |
|
47 | 47 | return None |
|
48 | 48 | method = pretty._safe_getattr(obj, name, None) |
|
49 | 49 | if callable(method): |
|
50 | 50 | # obj claims to have repr method... |
|
51 | 51 | if callable(pretty._safe_getattr(obj, '_ipython_canary_method_should_not_exist_', None)): |
|
52 | 52 | # ...but don't trust proxy objects that claim to have everything |
|
53 | 53 | return None |
|
54 | 54 | return method |
|
55 | 55 | |
|
56 | 56 | |
|
57 | 57 | class DisplayFormatter(Configurable): |
|
58 | 58 | |
|
59 | 59 | # When set to true only the default plain text formatter will be used. |
|
60 | 60 | plain_text_only = Bool(False, config=True) |
|
61 | 61 | def _plain_text_only_changed(self, name, old, new): |
|
62 | 62 | warnings.warn("""DisplayFormatter.plain_text_only is deprecated. |
|
63 | 63 | |
|
64 | 64 | Use DisplayFormatter.active_types = ['text/plain'] |
|
65 | 65 | for the same effect. |
|
66 | 66 | """, DeprecationWarning) |
|
67 | 67 | if new: |
|
68 | 68 | self.active_types = ['text/plain'] |
|
69 | 69 | else: |
|
70 | 70 | self.active_types = self.format_types |
|
71 | 71 | |
|
72 | 72 | active_types = List(Unicode, config=True, |
|
73 | 73 | help="""List of currently active mime-types to display. |
|
74 | 74 | You can use this to set a white-list for formats to display. |
|
75 | 75 | |
|
76 | 76 | Most users will not need to change this value. |
|
77 | 77 | """) |
|
78 | 78 | def _active_types_default(self): |
|
79 | 79 | return self.format_types |
|
80 | 80 | |
|
81 | 81 | def _active_types_changed(self, name, old, new): |
|
82 | 82 | for key, formatter in self.formatters.items(): |
|
83 | 83 | if key in new: |
|
84 | 84 | formatter.enabled = True |
|
85 | 85 | else: |
|
86 | 86 | formatter.enabled = False |
|
87 | 87 | |
|
88 | 88 | ipython_display_formatter = ForwardDeclaredInstance('FormatterABC') |
|
89 | 89 | def _ipython_display_formatter_default(self): |
|
90 | 90 | return IPythonDisplayFormatter(parent=self) |
|
91 | 91 | |
|
92 | 92 | # A dict of formatter whose keys are format types (MIME types) and whose |
|
93 | 93 | # values are subclasses of BaseFormatter. |
|
94 | 94 | formatters = Dict() |
|
95 | 95 | def _formatters_default(self): |
|
96 | 96 | """Activate the default formatters.""" |
|
97 | 97 | formatter_classes = [ |
|
98 | 98 | PlainTextFormatter, |
|
99 | 99 | HTMLFormatter, |
|
100 | 100 | MarkdownFormatter, |
|
101 | 101 | SVGFormatter, |
|
102 | 102 | PNGFormatter, |
|
103 | 103 | PDFFormatter, |
|
104 | 104 | JPEGFormatter, |
|
105 | 105 | LatexFormatter, |
|
106 | 106 | JSONFormatter, |
|
107 | 107 | JavascriptFormatter |
|
108 | 108 | ] |
|
109 | 109 | d = {} |
|
110 | 110 | for cls in formatter_classes: |
|
111 | 111 | f = cls(parent=self) |
|
112 | 112 | d[f.format_type] = f |
|
113 | 113 | return d |
|
114 | 114 | |
|
115 | 115 | def format(self, obj, include=None, exclude=None): |
|
116 | 116 | """Return a format data dict for an object. |
|
117 | 117 | |
|
118 | 118 | By default all format types will be computed. |
|
119 | 119 | |
|
120 | 120 | The following MIME types are currently implemented: |
|
121 | 121 | |
|
122 | 122 | * text/plain |
|
123 | 123 | * text/html |
|
124 | 124 | * text/markdown |
|
125 | 125 | * text/latex |
|
126 | 126 | * application/json |
|
127 | 127 | * application/javascript |
|
128 | 128 | * application/pdf |
|
129 | 129 | * image/png |
|
130 | 130 | * image/jpeg |
|
131 | 131 | * image/svg+xml |
|
132 | 132 | |
|
133 | 133 | Parameters |
|
134 | 134 | ---------- |
|
135 | 135 | obj : object |
|
136 | 136 | The Python object whose format data will be computed. |
|
137 | 137 | include : list or tuple, optional |
|
138 | 138 | A list of format type strings (MIME types) to include in the |
|
139 | 139 | format data dict. If this is set *only* the format types included |
|
140 | 140 | in this list will be computed. |
|
141 | 141 | exclude : list or tuple, optional |
|
142 | 142 | A list of format type string (MIME types) to exclude in the format |
|
143 | 143 | data dict. If this is set all format types will be computed, |
|
144 | 144 | except for those included in this argument. |
|
145 | 145 | |
|
146 | 146 | Returns |
|
147 | 147 | ------- |
|
148 | 148 | (format_dict, metadata_dict) : tuple of two dicts |
|
149 | 149 | |
|
150 | 150 | format_dict is a dictionary of key/value pairs, one of each format that was |
|
151 | 151 | generated for the object. The keys are the format types, which |
|
152 | 152 | will usually be MIME type strings and the values and JSON'able |
|
153 | 153 | data structure containing the raw data for the representation in |
|
154 | 154 | that format. |
|
155 | 155 | |
|
156 | 156 | metadata_dict is a dictionary of metadata about each mime-type output. |
|
157 | 157 | Its keys will be a strict subset of the keys in format_dict. |
|
158 | 158 | """ |
|
159 | 159 | format_dict = {} |
|
160 | 160 | md_dict = {} |
|
161 | 161 | |
|
162 | 162 | if self.ipython_display_formatter(obj): |
|
163 | 163 | # object handled itself, don't proceed |
|
164 | 164 | return {}, {} |
|
165 | 165 | |
|
166 | 166 | for format_type, formatter in self.formatters.items(): |
|
167 | 167 | if include and format_type not in include: |
|
168 | 168 | continue |
|
169 | 169 | if exclude and format_type in exclude: |
|
170 | 170 | continue |
|
171 | 171 | |
|
172 | 172 | md = None |
|
173 | 173 | try: |
|
174 | 174 | data = formatter(obj) |
|
175 | 175 | except: |
|
176 | 176 | # FIXME: log the exception |
|
177 | 177 | raise |
|
178 | 178 | |
|
179 | 179 | # formatters can return raw data or (data, metadata) |
|
180 | 180 | if isinstance(data, tuple) and len(data) == 2: |
|
181 | 181 | data, md = data |
|
182 | 182 | |
|
183 | 183 | if data is not None: |
|
184 | 184 | format_dict[format_type] = data |
|
185 | 185 | if md is not None: |
|
186 | 186 | md_dict[format_type] = md |
|
187 | 187 | |
|
188 | 188 | return format_dict, md_dict |
|
189 | 189 | |
|
190 | 190 | @property |
|
191 | 191 | def format_types(self): |
|
192 | 192 | """Return the format types (MIME types) of the active formatters.""" |
|
193 | 193 | return list(self.formatters.keys()) |
|
194 | 194 | |
|
195 | 195 | |
|
196 | 196 | #----------------------------------------------------------------------------- |
|
197 | 197 | # Formatters for specific format types (text, html, svg, etc.) |
|
198 | 198 | #----------------------------------------------------------------------------- |
|
199 | 199 | |
|
200 | 200 | |
|
201 | 201 | def _safe_repr(obj): |
|
202 | 202 | """Try to return a repr of an object |
|
203 | 203 | |
|
204 | 204 | always returns a string, at least. |
|
205 | 205 | """ |
|
206 | 206 | try: |
|
207 | 207 | return repr(obj) |
|
208 | 208 | except Exception as e: |
|
209 | 209 | return "un-repr-able object (%r)" % e |
|
210 | 210 | |
|
211 | 211 | |
|
212 | 212 | class FormatterWarning(UserWarning): |
|
213 | 213 | """Warning class for errors in formatters""" |
|
214 | 214 | |
|
215 | 215 | @decorator |
|
216 | 216 | def catch_format_error(method, self, *args, **kwargs): |
|
217 | 217 | """show traceback on failed format call""" |
|
218 | 218 | try: |
|
219 | 219 | r = method(self, *args, **kwargs) |
|
220 | 220 | except NotImplementedError: |
|
221 | 221 | # don't warn on NotImplementedErrors |
|
222 | 222 | return None |
|
223 | 223 | except Exception: |
|
224 | 224 | exc_info = sys.exc_info() |
|
225 | 225 | ip = get_ipython() |
|
226 | 226 | if ip is not None: |
|
227 | 227 | ip.showtraceback(exc_info) |
|
228 | 228 | else: |
|
229 | 229 | traceback.print_exception(*exc_info) |
|
230 | 230 | return None |
|
231 | 231 | return self._check_return(r, args[0]) |
|
232 | 232 | |
|
233 | 233 | |
|
234 | 234 | class FormatterABC(with_metaclass(abc.ABCMeta, object)): |
|
235 | 235 | """ Abstract base class for Formatters. |
|
236 | 236 | |
|
237 | 237 | A formatter is a callable class that is responsible for computing the |
|
238 | 238 | raw format data for a particular format type (MIME type). For example, |
|
239 | 239 | an HTML formatter would have a format type of `text/html` and would return |
|
240 | 240 | the HTML representation of the object when called. |
|
241 | 241 | """ |
|
242 | 242 | |
|
243 | 243 | # The format type of the data returned, usually a MIME type. |
|
244 | 244 | format_type = 'text/plain' |
|
245 | 245 | |
|
246 | 246 | # Is the formatter enabled... |
|
247 | 247 | enabled = True |
|
248 | 248 | |
|
249 | 249 | @abc.abstractmethod |
|
250 | 250 | def __call__(self, obj): |
|
251 | 251 | """Return a JSON'able representation of the object. |
|
252 | 252 | |
|
253 | 253 | If the object cannot be formatted by this formatter, |
|
254 | 254 | warn and return None. |
|
255 | 255 | """ |
|
256 | 256 | return repr(obj) |
|
257 | 257 | |
|
258 | 258 | |
|
259 | 259 | def _mod_name_key(typ): |
|
260 | 260 | """Return a (__module__, __name__) tuple for a type. |
|
261 | 261 | |
|
262 | 262 | Used as key in Formatter.deferred_printers. |
|
263 | 263 | """ |
|
264 | 264 | module = getattr(typ, '__module__', None) |
|
265 | 265 | name = getattr(typ, '__name__', None) |
|
266 | 266 | return (module, name) |
|
267 | 267 | |
|
268 | 268 | |
|
269 | 269 | def _get_type(obj): |
|
270 | 270 | """Return the type of an instance (old and new-style)""" |
|
271 | 271 | return getattr(obj, '__class__', None) or type(obj) |
|
272 | 272 | |
|
273 | 273 | _raise_key_error = object() |
|
274 | 274 | |
|
275 | 275 | |
|
276 | 276 | class BaseFormatter(Configurable): |
|
277 | 277 | """A base formatter class that is configurable. |
|
278 | 278 | |
|
279 | 279 | This formatter should usually be used as the base class of all formatters. |
|
280 | 280 | It is a traited :class:`Configurable` class and includes an extensible |
|
281 | 281 | API for users to determine how their objects are formatted. The following |
|
282 | 282 | logic is used to find a function to format an given object. |
|
283 | 283 | |
|
284 | 284 | 1. The object is introspected to see if it has a method with the name |
|
285 | 285 | :attr:`print_method`. If is does, that object is passed to that method |
|
286 | 286 | for formatting. |
|
287 | 287 | 2. If no print method is found, three internal dictionaries are consulted |
|
288 | 288 | to find print method: :attr:`singleton_printers`, :attr:`type_printers` |
|
289 | 289 | and :attr:`deferred_printers`. |
|
290 | 290 | |
|
291 | 291 | Users should use these dictionaries to register functions that will be |
|
292 | 292 | used to compute the format data for their objects (if those objects don't |
|
293 | 293 | have the special print methods). The easiest way of using these |
|
294 | 294 | dictionaries is through the :meth:`for_type` and :meth:`for_type_by_name` |
|
295 | 295 | methods. |
|
296 | 296 | |
|
297 | 297 | If no function/callable is found to compute the format data, ``None`` is |
|
298 | 298 | returned and this format type is not used. |
|
299 | 299 | """ |
|
300 | 300 | |
|
301 | 301 | format_type = Unicode('text/plain') |
|
302 | 302 | _return_type = string_types |
|
303 | 303 | |
|
304 | 304 | enabled = Bool(True, config=True) |
|
305 | 305 | |
|
306 | 306 | print_method = ObjectName('__repr__') |
|
307 | 307 | |
|
308 | 308 | # The singleton printers. |
|
309 | 309 | # Maps the IDs of the builtin singleton objects to the format functions. |
|
310 | 310 | singleton_printers = Dict(config=True) |
|
311 | 311 | |
|
312 | 312 | # The type-specific printers. |
|
313 | 313 | # Map type objects to the format functions. |
|
314 | 314 | type_printers = Dict(config=True) |
|
315 | 315 | |
|
316 | 316 | # The deferred-import type-specific printers. |
|
317 | 317 | # Map (modulename, classname) pairs to the format functions. |
|
318 | 318 | deferred_printers = Dict(config=True) |
|
319 | 319 | |
|
320 | 320 | @catch_format_error |
|
321 | 321 | def __call__(self, obj): |
|
322 | 322 | """Compute the format for an object.""" |
|
323 | 323 | if self.enabled: |
|
324 | 324 | # lookup registered printer |
|
325 | 325 | try: |
|
326 | 326 | printer = self.lookup(obj) |
|
327 | 327 | except KeyError: |
|
328 | 328 | pass |
|
329 | 329 | else: |
|
330 | 330 | return printer(obj) |
|
331 | 331 | # Finally look for special method names |
|
332 | 332 | method = _safe_get_formatter_method(obj, self.print_method) |
|
333 | 333 | if method is not None: |
|
334 | 334 | return method() |
|
335 | 335 | return None |
|
336 | 336 | else: |
|
337 | 337 | return None |
|
338 | 338 | |
|
339 | 339 | def __contains__(self, typ): |
|
340 | 340 | """map in to lookup_by_type""" |
|
341 | 341 | try: |
|
342 | 342 | self.lookup_by_type(typ) |
|
343 | 343 | except KeyError: |
|
344 | 344 | return False |
|
345 | 345 | else: |
|
346 | 346 | return True |
|
347 | 347 | |
|
348 | 348 | def _check_return(self, r, obj): |
|
349 | 349 | """Check that a return value is appropriate |
|
350 | 350 | |
|
351 | 351 | Return the value if so, None otherwise, warning if invalid. |
|
352 | 352 | """ |
|
353 | 353 | if r is None or isinstance(r, self._return_type) or \ |
|
354 | 354 | (isinstance(r, tuple) and r and isinstance(r[0], self._return_type)): |
|
355 | 355 | return r |
|
356 | 356 | else: |
|
357 | 357 | warnings.warn( |
|
358 | 358 | "%s formatter returned invalid type %s (expected %s) for object: %s" % \ |
|
359 | 359 | (self.format_type, type(r), self._return_type, _safe_repr(obj)), |
|
360 | 360 | FormatterWarning |
|
361 | 361 | ) |
|
362 | 362 | |
|
363 | 363 | def lookup(self, obj): |
|
364 | 364 | """Look up the formatter for a given instance. |
|
365 | 365 | |
|
366 | 366 | Parameters |
|
367 | 367 | ---------- |
|
368 | 368 | obj : object instance |
|
369 | 369 | |
|
370 | 370 | Returns |
|
371 | 371 | ------- |
|
372 | 372 | f : callable |
|
373 | 373 | The registered formatting callable for the type. |
|
374 | 374 | |
|
375 | 375 | Raises |
|
376 | 376 | ------ |
|
377 | 377 | KeyError if the type has not been registered. |
|
378 | 378 | """ |
|
379 | 379 | # look for singleton first |
|
380 | 380 | obj_id = id(obj) |
|
381 | 381 | if obj_id in self.singleton_printers: |
|
382 | 382 | return self.singleton_printers[obj_id] |
|
383 | 383 | # then lookup by type |
|
384 | 384 | return self.lookup_by_type(_get_type(obj)) |
|
385 | 385 | |
|
386 | 386 | def lookup_by_type(self, typ): |
|
387 | 387 | """Look up the registered formatter for a type. |
|
388 | 388 | |
|
389 | 389 | Parameters |
|
390 | 390 | ---------- |
|
391 | 391 | typ : type or '__module__.__name__' string for a type |
|
392 | 392 | |
|
393 | 393 | Returns |
|
394 | 394 | ------- |
|
395 | 395 | f : callable |
|
396 | 396 | The registered formatting callable for the type. |
|
397 | 397 | |
|
398 | 398 | Raises |
|
399 | 399 | ------ |
|
400 | 400 | KeyError if the type has not been registered. |
|
401 | 401 | """ |
|
402 | 402 | if isinstance(typ, string_types): |
|
403 | 403 | typ_key = tuple(typ.rsplit('.',1)) |
|
404 | 404 | if typ_key not in self.deferred_printers: |
|
405 | 405 | # We may have it cached in the type map. We will have to |
|
406 | 406 | # iterate over all of the types to check. |
|
407 | 407 | for cls in self.type_printers: |
|
408 | 408 | if _mod_name_key(cls) == typ_key: |
|
409 | 409 | return self.type_printers[cls] |
|
410 | 410 | else: |
|
411 | 411 | return self.deferred_printers[typ_key] |
|
412 | 412 | else: |
|
413 | 413 | for cls in pretty._get_mro(typ): |
|
414 | 414 | if cls in self.type_printers or self._in_deferred_types(cls): |
|
415 | 415 | return self.type_printers[cls] |
|
416 | 416 | |
|
417 | 417 | # If we have reached here, the lookup failed. |
|
418 | 418 | raise KeyError("No registered printer for {0!r}".format(typ)) |
|
419 | 419 | |
|
420 | 420 | def for_type(self, typ, func=None): |
|
421 | 421 | """Add a format function for a given type. |
|
422 | 422 | |
|
423 | 423 | Parameters |
|
424 | 424 | ----------- |
|
425 | 425 | typ : type or '__module__.__name__' string for a type |
|
426 | 426 | The class of the object that will be formatted using `func`. |
|
427 | 427 | func : callable |
|
428 | 428 | A callable for computing the format data. |
|
429 | 429 | `func` will be called with the object to be formatted, |
|
430 | 430 | and will return the raw data in this formatter's format. |
|
431 | 431 | Subclasses may use a different call signature for the |
|
432 | 432 | `func` argument. |
|
433 | 433 | |
|
434 | 434 | If `func` is None or not specified, there will be no change, |
|
435 | 435 | only returning the current value. |
|
436 | 436 | |
|
437 | 437 | Returns |
|
438 | 438 | ------- |
|
439 | 439 | oldfunc : callable |
|
440 | 440 | The currently registered callable. |
|
441 | 441 | If you are registering a new formatter, |
|
442 | 442 | this will be the previous value (to enable restoring later). |
|
443 | 443 | """ |
|
444 | 444 | # if string given, interpret as 'pkg.module.class_name' |
|
445 | 445 | if isinstance(typ, string_types): |
|
446 | 446 | type_module, type_name = typ.rsplit('.', 1) |
|
447 | 447 | return self.for_type_by_name(type_module, type_name, func) |
|
448 | 448 | |
|
449 | 449 | try: |
|
450 | 450 | oldfunc = self.lookup_by_type(typ) |
|
451 | 451 | except KeyError: |
|
452 | 452 | oldfunc = None |
|
453 | 453 | |
|
454 | 454 | if func is not None: |
|
455 | 455 | self.type_printers[typ] = func |
|
456 | 456 | |
|
457 | 457 | return oldfunc |
|
458 | 458 | |
|
459 | 459 | def for_type_by_name(self, type_module, type_name, func=None): |
|
460 | 460 | """Add a format function for a type specified by the full dotted |
|
461 | 461 | module and name of the type, rather than the type of the object. |
|
462 | 462 | |
|
463 | 463 | Parameters |
|
464 | 464 | ---------- |
|
465 | 465 | type_module : str |
|
466 | 466 | The full dotted name of the module the type is defined in, like |
|
467 | 467 | ``numpy``. |
|
468 | 468 | type_name : str |
|
469 | 469 | The name of the type (the class name), like ``dtype`` |
|
470 | 470 | func : callable |
|
471 | 471 | A callable for computing the format data. |
|
472 | 472 | `func` will be called with the object to be formatted, |
|
473 | 473 | and will return the raw data in this formatter's format. |
|
474 | 474 | Subclasses may use a different call signature for the |
|
475 | 475 | `func` argument. |
|
476 | 476 | |
|
477 | 477 | If `func` is None or unspecified, there will be no change, |
|
478 | 478 | only returning the current value. |
|
479 | 479 | |
|
480 | 480 | Returns |
|
481 | 481 | ------- |
|
482 | 482 | oldfunc : callable |
|
483 | 483 | The currently registered callable. |
|
484 | 484 | If you are registering a new formatter, |
|
485 | 485 | this will be the previous value (to enable restoring later). |
|
486 | 486 | """ |
|
487 | 487 | key = (type_module, type_name) |
|
488 | 488 | |
|
489 | 489 | try: |
|
490 | 490 | oldfunc = self.lookup_by_type("%s.%s" % key) |
|
491 | 491 | except KeyError: |
|
492 | 492 | oldfunc = None |
|
493 | 493 | |
|
494 | 494 | if func is not None: |
|
495 | 495 | self.deferred_printers[key] = func |
|
496 | 496 | return oldfunc |
|
497 | 497 | |
|
498 | 498 | def pop(self, typ, default=_raise_key_error): |
|
499 | 499 | """Pop a formatter for the given type. |
|
500 | 500 | |
|
501 | 501 | Parameters |
|
502 | 502 | ---------- |
|
503 | 503 | typ : type or '__module__.__name__' string for a type |
|
504 | 504 | default : object |
|
505 | 505 | value to be returned if no formatter is registered for typ. |
|
506 | 506 | |
|
507 | 507 | Returns |
|
508 | 508 | ------- |
|
509 | 509 | obj : object |
|
510 | 510 | The last registered object for the type. |
|
511 | 511 | |
|
512 | 512 | Raises |
|
513 | 513 | ------ |
|
514 | 514 | KeyError if the type is not registered and default is not specified. |
|
515 | 515 | """ |
|
516 | 516 | |
|
517 | 517 | if isinstance(typ, string_types): |
|
518 | 518 | typ_key = tuple(typ.rsplit('.',1)) |
|
519 | 519 | if typ_key not in self.deferred_printers: |
|
520 | 520 | # We may have it cached in the type map. We will have to |
|
521 | 521 | # iterate over all of the types to check. |
|
522 | 522 | for cls in self.type_printers: |
|
523 | 523 | if _mod_name_key(cls) == typ_key: |
|
524 | 524 | old = self.type_printers.pop(cls) |
|
525 | 525 | break |
|
526 | 526 | else: |
|
527 | 527 | old = default |
|
528 | 528 | else: |
|
529 | 529 | old = self.deferred_printers.pop(typ_key) |
|
530 | 530 | else: |
|
531 | 531 | if typ in self.type_printers: |
|
532 | 532 | old = self.type_printers.pop(typ) |
|
533 | 533 | else: |
|
534 | 534 | old = self.deferred_printers.pop(_mod_name_key(typ), default) |
|
535 | 535 | if old is _raise_key_error: |
|
536 | 536 | raise KeyError("No registered value for {0!r}".format(typ)) |
|
537 | 537 | return old |
|
538 | 538 | |
|
539 | 539 | def _in_deferred_types(self, cls): |
|
540 | 540 | """ |
|
541 | 541 | Check if the given class is specified in the deferred type registry. |
|
542 | 542 | |
|
543 | 543 | Successful matches will be moved to the regular type registry for future use. |
|
544 | 544 | """ |
|
545 | 545 | mod = getattr(cls, '__module__', None) |
|
546 | 546 | name = getattr(cls, '__name__', None) |
|
547 | 547 | key = (mod, name) |
|
548 | 548 | if key in self.deferred_printers: |
|
549 | 549 | # Move the printer over to the regular registry. |
|
550 | 550 | printer = self.deferred_printers.pop(key) |
|
551 | 551 | self.type_printers[cls] = printer |
|
552 | 552 | return True |
|
553 | 553 | return False |
|
554 | 554 | |
|
555 | 555 | |
|
556 | 556 | class PlainTextFormatter(BaseFormatter): |
|
557 | 557 | """The default pretty-printer. |
|
558 | 558 | |
|
559 | 559 | This uses :mod:`IPython.lib.pretty` to compute the format data of |
|
560 | 560 | the object. If the object cannot be pretty printed, :func:`repr` is used. |
|
561 | 561 | See the documentation of :mod:`IPython.lib.pretty` for details on |
|
562 | 562 | how to write pretty printers. Here is a simple example:: |
|
563 | 563 | |
|
564 | 564 | def dtype_pprinter(obj, p, cycle): |
|
565 | 565 | if cycle: |
|
566 | 566 | return p.text('dtype(...)') |
|
567 | 567 | if hasattr(obj, 'fields'): |
|
568 | 568 | if obj.fields is None: |
|
569 | 569 | p.text(repr(obj)) |
|
570 | 570 | else: |
|
571 | 571 | p.begin_group(7, 'dtype([') |
|
572 | 572 | for i, field in enumerate(obj.descr): |
|
573 | 573 | if i > 0: |
|
574 | 574 | p.text(',') |
|
575 | 575 | p.breakable() |
|
576 | 576 | p.pretty(field) |
|
577 | 577 | p.end_group(7, '])') |
|
578 | 578 | """ |
|
579 | 579 | |
|
580 | 580 | # The format type of data returned. |
|
581 | 581 | format_type = Unicode('text/plain') |
|
582 | 582 | |
|
583 | 583 | # This subclass ignores this attribute as it always need to return |
|
584 | 584 | # something. |
|
585 | 585 | enabled = Bool(True, config=False) |
|
586 | 586 | |
|
587 | 587 | max_seq_length = Integer(pretty.MAX_SEQ_LENGTH, config=True, |
|
588 | 588 | help="""Truncate large collections (lists, dicts, tuples, sets) to this size. |
|
589 | 589 | |
|
590 | 590 | Set to 0 to disable truncation. |
|
591 | 591 | """ |
|
592 | 592 | ) |
|
593 | 593 | |
|
594 | 594 | # Look for a _repr_pretty_ methods to use for pretty printing. |
|
595 | 595 | print_method = ObjectName('_repr_pretty_') |
|
596 | 596 | |
|
597 | 597 | # Whether to pretty-print or not. |
|
598 | 598 | pprint = Bool(True, config=True) |
|
599 | 599 | |
|
600 | 600 | # Whether to be verbose or not. |
|
601 | 601 | verbose = Bool(False, config=True) |
|
602 | 602 | |
|
603 | 603 | # The maximum width. |
|
604 | 604 | max_width = Integer(79, config=True) |
|
605 | 605 | |
|
606 | 606 | # The newline character. |
|
607 | 607 | newline = Unicode('\n', config=True) |
|
608 | 608 | |
|
609 | 609 | # format-string for pprinting floats |
|
610 | 610 | float_format = Unicode('%r') |
|
611 | 611 | # setter for float precision, either int or direct format-string |
|
612 | 612 | float_precision = CUnicode('', config=True) |
|
613 | 613 | |
|
614 | 614 | def _float_precision_changed(self, name, old, new): |
|
615 | 615 | """float_precision changed, set float_format accordingly. |
|
616 | 616 | |
|
617 | 617 | float_precision can be set by int or str. |
|
618 | 618 | This will set float_format, after interpreting input. |
|
619 | 619 | If numpy has been imported, numpy print precision will also be set. |
|
620 | 620 | |
|
621 | 621 | integer `n` sets format to '%.nf', otherwise, format set directly. |
|
622 | 622 | |
|
623 | 623 | An empty string returns to defaults (repr for float, 8 for numpy). |
|
624 | 624 | |
|
625 | 625 | This parameter can be set via the '%precision' magic. |
|
626 | 626 | """ |
|
627 | 627 | |
|
628 | 628 | if '%' in new: |
|
629 | 629 | # got explicit format string |
|
630 | 630 | fmt = new |
|
631 | 631 | try: |
|
632 | 632 | fmt%3.14159 |
|
633 | 633 | except Exception: |
|
634 | 634 | raise ValueError("Precision must be int or format string, not %r"%new) |
|
635 | 635 | elif new: |
|
636 | 636 | # otherwise, should be an int |
|
637 | 637 | try: |
|
638 | 638 | i = int(new) |
|
639 | 639 | assert i >= 0 |
|
640 | 640 | except ValueError: |
|
641 | 641 | raise ValueError("Precision must be int or format string, not %r"%new) |
|
642 | 642 | except AssertionError: |
|
643 | 643 | raise ValueError("int precision must be non-negative, not %r"%i) |
|
644 | 644 | |
|
645 | 645 | fmt = '%%.%if'%i |
|
646 | 646 | if 'numpy' in sys.modules: |
|
647 | 647 | # set numpy precision if it has been imported |
|
648 | 648 | import numpy |
|
649 | 649 | numpy.set_printoptions(precision=i) |
|
650 | 650 | else: |
|
651 | 651 | # default back to repr |
|
652 | 652 | fmt = '%r' |
|
653 | 653 | if 'numpy' in sys.modules: |
|
654 | 654 | import numpy |
|
655 | 655 | # numpy default is 8 |
|
656 | 656 | numpy.set_printoptions(precision=8) |
|
657 | 657 | self.float_format = fmt |
|
658 | 658 | |
|
659 | 659 | # Use the default pretty printers from IPython.lib.pretty. |
|
660 | 660 | def _singleton_printers_default(self): |
|
661 | 661 | return pretty._singleton_pprinters.copy() |
|
662 | 662 | |
|
663 | 663 | def _type_printers_default(self): |
|
664 | 664 | d = pretty._type_pprinters.copy() |
|
665 | 665 | d[float] = lambda obj,p,cycle: p.text(self.float_format%obj) |
|
666 | 666 | return d |
|
667 | 667 | |
|
668 | 668 | def _deferred_printers_default(self): |
|
669 | 669 | return pretty._deferred_type_pprinters.copy() |
|
670 | 670 | |
|
671 | 671 | #### FormatterABC interface #### |
|
672 | 672 | |
|
673 | 673 | @catch_format_error |
|
674 | 674 | def __call__(self, obj): |
|
675 | 675 | """Compute the pretty representation of the object.""" |
|
676 | 676 | if not self.pprint: |
|
677 | 677 | return repr(obj) |
|
678 | 678 | else: |
|
679 | 679 | # handle str and unicode on Python 2 |
|
680 | 680 | # io.StringIO only accepts unicode, |
|
681 | 681 | # cStringIO doesn't handle unicode on py2, |
|
682 | 682 | # StringIO allows str, unicode but only ascii str |
|
683 | 683 | stream = pretty.CUnicodeIO() |
|
684 | 684 | printer = pretty.RepresentationPrinter(stream, self.verbose, |
|
685 | 685 | self.max_width, self.newline, |
|
686 | 686 | max_seq_length=self.max_seq_length, |
|
687 | 687 | singleton_pprinters=self.singleton_printers, |
|
688 | 688 | type_pprinters=self.type_printers, |
|
689 | 689 | deferred_pprinters=self.deferred_printers) |
|
690 | 690 | printer.pretty(obj) |
|
691 | 691 | printer.flush() |
|
692 | 692 | return stream.getvalue() |
|
693 | 693 | |
|
694 | 694 | |
|
695 | 695 | class HTMLFormatter(BaseFormatter): |
|
696 | 696 | """An HTML formatter. |
|
697 | 697 | |
|
698 | 698 | To define the callables that compute the HTML representation of your |
|
699 | 699 | objects, define a :meth:`_repr_html_` method or use the :meth:`for_type` |
|
700 | 700 | or :meth:`for_type_by_name` methods to register functions that handle |
|
701 | 701 | this. |
|
702 | 702 | |
|
703 | 703 | The return value of this formatter should be a valid HTML snippet that |
|
704 | 704 | could be injected into an existing DOM. It should *not* include the |
|
705 | 705 | ```<html>`` or ```<body>`` tags. |
|
706 | 706 | """ |
|
707 | 707 | format_type = Unicode('text/html') |
|
708 | 708 | |
|
709 | 709 | print_method = ObjectName('_repr_html_') |
|
710 | 710 | |
|
711 | 711 | |
|
712 | 712 | class MarkdownFormatter(BaseFormatter): |
|
713 | 713 | """A Markdown formatter. |
|
714 | 714 | |
|
715 | 715 | To define the callables that compute the Markdown representation of your |
|
716 | 716 | objects, define a :meth:`_repr_markdown_` method or use the :meth:`for_type` |
|
717 | 717 | or :meth:`for_type_by_name` methods to register functions that handle |
|
718 | 718 | this. |
|
719 | 719 | |
|
720 | 720 | The return value of this formatter should be a valid Markdown. |
|
721 | 721 | """ |
|
722 | 722 | format_type = Unicode('text/markdown') |
|
723 | 723 | |
|
724 | 724 | print_method = ObjectName('_repr_markdown_') |
|
725 | 725 | |
|
726 | 726 | class SVGFormatter(BaseFormatter): |
|
727 | 727 | """An SVG formatter. |
|
728 | 728 | |
|
729 | 729 | To define the callables that compute the SVG representation of your |
|
730 | 730 | objects, define a :meth:`_repr_svg_` method or use the :meth:`for_type` |
|
731 | 731 | or :meth:`for_type_by_name` methods to register functions that handle |
|
732 | 732 | this. |
|
733 | 733 | |
|
734 | 734 | The return value of this formatter should be valid SVG enclosed in |
|
735 | 735 | ```<svg>``` tags, that could be injected into an existing DOM. It should |
|
736 | 736 | *not* include the ```<html>`` or ```<body>`` tags. |
|
737 | 737 | """ |
|
738 | 738 | format_type = Unicode('image/svg+xml') |
|
739 | 739 | |
|
740 | 740 | print_method = ObjectName('_repr_svg_') |
|
741 | 741 | |
|
742 | 742 | |
|
743 | 743 | class PNGFormatter(BaseFormatter): |
|
744 | 744 | """A PNG formatter. |
|
745 | 745 | |
|
746 | 746 | To define the callables that compute the PNG representation of your |
|
747 | 747 | objects, define a :meth:`_repr_png_` method or use the :meth:`for_type` |
|
748 | 748 | or :meth:`for_type_by_name` methods to register functions that handle |
|
749 | 749 | this. |
|
750 | 750 | |
|
751 | 751 | The return value of this formatter should be raw PNG data, *not* |
|
752 | 752 | base64 encoded. |
|
753 | 753 | """ |
|
754 | 754 | format_type = Unicode('image/png') |
|
755 | 755 | |
|
756 | 756 | print_method = ObjectName('_repr_png_') |
|
757 | 757 | |
|
758 | 758 | _return_type = (bytes, unicode_type) |
|
759 | 759 | |
|
760 | 760 | |
|
761 | 761 | class JPEGFormatter(BaseFormatter): |
|
762 | 762 | """A JPEG formatter. |
|
763 | 763 | |
|
764 | 764 | To define the callables that compute the JPEG representation of your |
|
765 | 765 | objects, define a :meth:`_repr_jpeg_` method or use the :meth:`for_type` |
|
766 | 766 | or :meth:`for_type_by_name` methods to register functions that handle |
|
767 | 767 | this. |
|
768 | 768 | |
|
769 | 769 | The return value of this formatter should be raw JPEG data, *not* |
|
770 | 770 | base64 encoded. |
|
771 | 771 | """ |
|
772 | 772 | format_type = Unicode('image/jpeg') |
|
773 | 773 | |
|
774 | 774 | print_method = ObjectName('_repr_jpeg_') |
|
775 | 775 | |
|
776 | 776 | _return_type = (bytes, unicode_type) |
|
777 | 777 | |
|
778 | 778 | |
|
779 | 779 | class LatexFormatter(BaseFormatter): |
|
780 | 780 | """A LaTeX formatter. |
|
781 | 781 | |
|
782 | 782 | To define the callables that compute the LaTeX representation of your |
|
783 | 783 | objects, define a :meth:`_repr_latex_` method or use the :meth:`for_type` |
|
784 | 784 | or :meth:`for_type_by_name` methods to register functions that handle |
|
785 | 785 | this. |
|
786 | 786 | |
|
787 | 787 | The return value of this formatter should be a valid LaTeX equation, |
|
788 | 788 | enclosed in either ```$```, ```$$``` or another LaTeX equation |
|
789 | 789 | environment. |
|
790 | 790 | """ |
|
791 | 791 | format_type = Unicode('text/latex') |
|
792 | 792 | |
|
793 | 793 | print_method = ObjectName('_repr_latex_') |
|
794 | 794 | |
|
795 | 795 | |
|
796 | 796 | class JSONFormatter(BaseFormatter): |
|
797 | 797 | """A JSON string formatter. |
|
798 | 798 | |
|
799 | 799 | To define the callables that compute the JSONable representation of |
|
800 | 800 | your objects, define a :meth:`_repr_json_` method or use the :meth:`for_type` |
|
801 | 801 | or :meth:`for_type_by_name` methods to register functions that handle |
|
802 | 802 | this. |
|
803 | 803 | |
|
804 | 804 | The return value of this formatter should be a JSONable list or dict. |
|
805 | 805 | JSON scalars (None, number, string) are not allowed, only dict or list containers. |
|
806 | 806 | """ |
|
807 | 807 | format_type = Unicode('application/json') |
|
808 | 808 | _return_type = (list, dict) |
|
809 | 809 | |
|
810 | 810 | print_method = ObjectName('_repr_json_') |
|
811 | 811 | |
|
812 | 812 | def _check_return(self, r, obj): |
|
813 | 813 | """Check that a return value is appropriate |
|
814 | 814 | |
|
815 | 815 | Return the value if so, None otherwise, warning if invalid. |
|
816 | 816 | """ |
|
817 | 817 | if r is None: |
|
818 | 818 | return |
|
819 | 819 | md = None |
|
820 | 820 | if isinstance(r, tuple): |
|
821 | 821 | # unpack data, metadata tuple for type checking on first element |
|
822 | 822 | r, md = r |
|
823 | 823 | |
|
824 | 824 | # handle deprecated JSON-as-string form from IPython < 3 |
|
825 | 825 | if isinstance(r, string_types): |
|
826 | 826 | warnings.warn("JSON expects JSONable list/dict containers, not JSON strings", |
|
827 | 827 | FormatterWarning) |
|
828 | 828 | r = json.loads(r) |
|
829 | 829 | |
|
830 | 830 | if md is not None: |
|
831 | 831 | # put the tuple back together |
|
832 | 832 | r = (r, md) |
|
833 | 833 | return super(JSONFormatter, self)._check_return(r, obj) |
|
834 | 834 | |
|
835 | 835 | |
|
836 | 836 | class JavascriptFormatter(BaseFormatter): |
|
837 | 837 | """A Javascript formatter. |
|
838 | 838 | |
|
839 | 839 | To define the callables that compute the Javascript representation of |
|
840 | 840 | your objects, define a :meth:`_repr_javascript_` method or use the |
|
841 | 841 | :meth:`for_type` or :meth:`for_type_by_name` methods to register functions |
|
842 | 842 | that handle this. |
|
843 | 843 | |
|
844 | 844 | The return value of this formatter should be valid Javascript code and |
|
845 | 845 | should *not* be enclosed in ```<script>``` tags. |
|
846 | 846 | """ |
|
847 | 847 | format_type = Unicode('application/javascript') |
|
848 | 848 | |
|
849 | 849 | print_method = ObjectName('_repr_javascript_') |
|
850 | 850 | |
|
851 | 851 | |
|
852 | 852 | class PDFFormatter(BaseFormatter): |
|
853 | 853 | """A PDF formatter. |
|
854 | 854 | |
|
855 | 855 | To define the callables that compute the PDF representation of your |
|
856 | 856 | objects, define a :meth:`_repr_pdf_` method or use the :meth:`for_type` |
|
857 | 857 | or :meth:`for_type_by_name` methods to register functions that handle |
|
858 | 858 | this. |
|
859 | 859 | |
|
860 | 860 | The return value of this formatter should be raw PDF data, *not* |
|
861 | 861 | base64 encoded. |
|
862 | 862 | """ |
|
863 | 863 | format_type = Unicode('application/pdf') |
|
864 | 864 | |
|
865 | 865 | print_method = ObjectName('_repr_pdf_') |
|
866 | 866 | |
|
867 | 867 | _return_type = (bytes, unicode_type) |
|
868 | 868 | |
|
869 | 869 | class IPythonDisplayFormatter(BaseFormatter): |
|
870 | 870 | """A Formatter for objects that know how to display themselves. |
|
871 | 871 | |
|
872 | 872 | To define the callables that compute the representation of your |
|
873 | 873 | objects, define a :meth:`_ipython_display_` method or use the :meth:`for_type` |
|
874 | 874 | or :meth:`for_type_by_name` methods to register functions that handle |
|
875 | 875 | this. Unlike mime-type displays, this method should not return anything, |
|
876 | 876 | instead calling any appropriate display methods itself. |
|
877 | 877 | |
|
878 | 878 | This display formatter has highest priority. |
|
879 | 879 | If it fires, no other display formatter will be called. |
|
880 | 880 | """ |
|
881 | 881 | print_method = ObjectName('_ipython_display_') |
|
882 | 882 | _return_type = (type(None), bool) |
|
883 | 883 | |
|
884 | 884 | |
|
885 | 885 | @catch_format_error |
|
886 | 886 | def __call__(self, obj): |
|
887 | 887 | """Compute the format for an object.""" |
|
888 | 888 | if self.enabled: |
|
889 | 889 | # lookup registered printer |
|
890 | 890 | try: |
|
891 | 891 | printer = self.lookup(obj) |
|
892 | 892 | except KeyError: |
|
893 | 893 | pass |
|
894 | 894 | else: |
|
895 | 895 | printer(obj) |
|
896 | 896 | return True |
|
897 | 897 | # Finally look for special method names |
|
898 | 898 | method = _safe_get_formatter_method(obj, self.print_method) |
|
899 | 899 | if method is not None: |
|
900 | 900 | method() |
|
901 | 901 | return True |
|
902 | 902 | |
|
903 | 903 | |
|
904 | 904 | FormatterABC.register(BaseFormatter) |
|
905 | 905 | FormatterABC.register(PlainTextFormatter) |
|
906 | 906 | FormatterABC.register(HTMLFormatter) |
|
907 | 907 | FormatterABC.register(MarkdownFormatter) |
|
908 | 908 | FormatterABC.register(SVGFormatter) |
|
909 | 909 | FormatterABC.register(PNGFormatter) |
|
910 | 910 | FormatterABC.register(PDFFormatter) |
|
911 | 911 | FormatterABC.register(JPEGFormatter) |
|
912 | 912 | FormatterABC.register(LatexFormatter) |
|
913 | 913 | FormatterABC.register(JSONFormatter) |
|
914 | 914 | FormatterABC.register(JavascriptFormatter) |
|
915 | 915 | FormatterABC.register(IPythonDisplayFormatter) |
|
916 | 916 | |
|
917 | 917 | |
|
918 | 918 | def format_display_data(obj, include=None, exclude=None): |
|
919 | 919 | """Return a format data dict for an object. |
|
920 | 920 | |
|
921 | 921 | By default all format types will be computed. |
|
922 | 922 | |
|
923 | 923 | The following MIME types are currently implemented: |
|
924 | 924 | |
|
925 | 925 | * text/plain |
|
926 | 926 | * text/html |
|
927 | 927 | * text/markdown |
|
928 | 928 | * text/latex |
|
929 | 929 | * application/json |
|
930 | 930 | * application/javascript |
|
931 | 931 | * application/pdf |
|
932 | 932 | * image/png |
|
933 | 933 | * image/jpeg |
|
934 | 934 | * image/svg+xml |
|
935 | 935 | |
|
936 | 936 | Parameters |
|
937 | 937 | ---------- |
|
938 | 938 | obj : object |
|
939 | 939 | The Python object whose format data will be computed. |
|
940 | 940 | |
|
941 | 941 | Returns |
|
942 | 942 | ------- |
|
943 | 943 | format_dict : dict |
|
944 | 944 | A dictionary of key/value pairs, one or each format that was |
|
945 | 945 | generated for the object. The keys are the format types, which |
|
946 | 946 | will usually be MIME type strings and the values and JSON'able |
|
947 | 947 | data structure containing the raw data for the representation in |
|
948 | 948 | that format. |
|
949 | 949 | include : list or tuple, optional |
|
950 | 950 | A list of format type strings (MIME types) to include in the |
|
951 | 951 | format data dict. If this is set *only* the format types included |
|
952 | 952 | in this list will be computed. |
|
953 | 953 | exclude : list or tuple, optional |
|
954 | 954 | A list of format type string (MIME types) to exclue in the format |
|
955 | 955 | data dict. If this is set all format types will be computed, |
|
956 | 956 | except for those included in this argument. |
|
957 | 957 | """ |
|
958 | 958 | from IPython.core.interactiveshell import InteractiveShell |
|
959 | 959 | |
|
960 | 960 | InteractiveShell.instance().display_formatter.format( |
|
961 | 961 | obj, |
|
962 | 962 | include, |
|
963 | 963 | exclude |
|
964 | 964 | ) |
|
965 | 965 |
@@ -1,870 +1,870 b'' | |||
|
1 | 1 | """ History related magics and functionality """ |
|
2 | 2 | #----------------------------------------------------------------------------- |
|
3 | 3 | # Copyright (C) 2010-2011 The IPython Development Team. |
|
4 | 4 | # |
|
5 | 5 | # Distributed under the terms of the BSD License. |
|
6 | 6 | # |
|
7 | 7 | # The full license is in the file COPYING.txt, distributed with this software. |
|
8 | 8 | #----------------------------------------------------------------------------- |
|
9 | 9 | |
|
10 | 10 | #----------------------------------------------------------------------------- |
|
11 | 11 | # Imports |
|
12 | 12 | #----------------------------------------------------------------------------- |
|
13 | 13 | from __future__ import print_function |
|
14 | 14 | |
|
15 | 15 | # Stdlib imports |
|
16 | 16 | import atexit |
|
17 | 17 | import datetime |
|
18 | 18 | import os |
|
19 | 19 | import re |
|
20 | 20 | try: |
|
21 | 21 | import sqlite3 |
|
22 | 22 | except ImportError: |
|
23 | 23 | try: |
|
24 | 24 | from pysqlite2 import dbapi2 as sqlite3 |
|
25 | 25 | except ImportError: |
|
26 | 26 | sqlite3 = None |
|
27 | 27 | import threading |
|
28 | 28 | |
|
29 | 29 | # Our own packages |
|
30 | 30 | from IPython.config.configurable import Configurable |
|
31 |
from |
|
|
31 | from decorator import decorator | |
|
32 | 32 | from IPython.utils.decorators import undoc |
|
33 | 33 | from IPython.utils.path import locate_profile |
|
34 | 34 | from IPython.utils import py3compat |
|
35 | 35 | from IPython.utils.traitlets import ( |
|
36 | 36 | Any, Bool, Dict, Instance, Integer, List, Unicode, TraitError, |
|
37 | 37 | ) |
|
38 | 38 | from IPython.utils.warn import warn |
|
39 | 39 | |
|
40 | 40 | #----------------------------------------------------------------------------- |
|
41 | 41 | # Classes and functions |
|
42 | 42 | #----------------------------------------------------------------------------- |
|
43 | 43 | |
|
44 | 44 | @undoc |
|
45 | 45 | class DummyDB(object): |
|
46 | 46 | """Dummy DB that will act as a black hole for history. |
|
47 | 47 | |
|
48 | 48 | Only used in the absence of sqlite""" |
|
49 | 49 | def execute(*args, **kwargs): |
|
50 | 50 | return [] |
|
51 | 51 | |
|
52 | 52 | def commit(self, *args, **kwargs): |
|
53 | 53 | pass |
|
54 | 54 | |
|
55 | 55 | def __enter__(self, *args, **kwargs): |
|
56 | 56 | pass |
|
57 | 57 | |
|
58 | 58 | def __exit__(self, *args, **kwargs): |
|
59 | 59 | pass |
|
60 | 60 | |
|
61 | 61 | |
|
62 | 62 | @decorator |
|
63 | 63 | def needs_sqlite(f, self, *a, **kw): |
|
64 | 64 | """Decorator: return an empty list in the absence of sqlite.""" |
|
65 | 65 | if sqlite3 is None or not self.enabled: |
|
66 | 66 | return [] |
|
67 | 67 | else: |
|
68 | 68 | return f(self, *a, **kw) |
|
69 | 69 | |
|
70 | 70 | |
|
71 | 71 | if sqlite3 is not None: |
|
72 | 72 | DatabaseError = sqlite3.DatabaseError |
|
73 | 73 | else: |
|
74 | 74 | @undoc |
|
75 | 75 | class DatabaseError(Exception): |
|
76 | 76 | "Dummy exception when sqlite could not be imported. Should never occur." |
|
77 | 77 | |
|
78 | 78 | @decorator |
|
79 | 79 | def catch_corrupt_db(f, self, *a, **kw): |
|
80 | 80 | """A decorator which wraps HistoryAccessor method calls to catch errors from |
|
81 | 81 | a corrupt SQLite database, move the old database out of the way, and create |
|
82 | 82 | a new one. |
|
83 | 83 | """ |
|
84 | 84 | try: |
|
85 | 85 | return f(self, *a, **kw) |
|
86 | 86 | except DatabaseError: |
|
87 | 87 | if os.path.isfile(self.hist_file): |
|
88 | 88 | # Try to move the file out of the way |
|
89 | 89 | base,ext = os.path.splitext(self.hist_file) |
|
90 | 90 | newpath = base + '-corrupt' + ext |
|
91 | 91 | os.rename(self.hist_file, newpath) |
|
92 | 92 | self.init_db() |
|
93 | 93 | print("ERROR! History file wasn't a valid SQLite database.", |
|
94 | 94 | "It was moved to %s" % newpath, "and a new file created.") |
|
95 | 95 | return [] |
|
96 | 96 | |
|
97 | 97 | else: |
|
98 | 98 | # The hist_file is probably :memory: or something else. |
|
99 | 99 | raise |
|
100 | 100 | |
|
101 | 101 | class HistoryAccessorBase(Configurable): |
|
102 | 102 | """An abstract class for History Accessors """ |
|
103 | 103 | |
|
104 | 104 | def get_tail(self, n=10, raw=True, output=False, include_latest=False): |
|
105 | 105 | raise NotImplementedError |
|
106 | 106 | |
|
107 | 107 | def search(self, pattern="*", raw=True, search_raw=True, |
|
108 | 108 | output=False, n=None, unique=False): |
|
109 | 109 | raise NotImplementedError |
|
110 | 110 | |
|
111 | 111 | def get_range(self, session, start=1, stop=None, raw=True,output=False): |
|
112 | 112 | raise NotImplementedError |
|
113 | 113 | |
|
114 | 114 | def get_range_by_str(self, rangestr, raw=True, output=False): |
|
115 | 115 | raise NotImplementedError |
|
116 | 116 | |
|
117 | 117 | |
|
118 | 118 | class HistoryAccessor(HistoryAccessorBase): |
|
119 | 119 | """Access the history database without adding to it. |
|
120 | 120 | |
|
121 | 121 | This is intended for use by standalone history tools. IPython shells use |
|
122 | 122 | HistoryManager, below, which is a subclass of this.""" |
|
123 | 123 | |
|
124 | 124 | # String holding the path to the history file |
|
125 | 125 | hist_file = Unicode(config=True, |
|
126 | 126 | help="""Path to file to use for SQLite history database. |
|
127 | 127 | |
|
128 | 128 | By default, IPython will put the history database in the IPython |
|
129 | 129 | profile directory. If you would rather share one history among |
|
130 | 130 | profiles, you can set this value in each, so that they are consistent. |
|
131 | 131 | |
|
132 | 132 | Due to an issue with fcntl, SQLite is known to misbehave on some NFS |
|
133 | 133 | mounts. If you see IPython hanging, try setting this to something on a |
|
134 | 134 | local disk, e.g:: |
|
135 | 135 | |
|
136 | 136 | ipython --HistoryManager.hist_file=/tmp/ipython_hist.sqlite |
|
137 | 137 | |
|
138 | 138 | """) |
|
139 | 139 | |
|
140 | 140 | enabled = Bool(True, config=True, |
|
141 | 141 | help="""enable the SQLite history |
|
142 | 142 | |
|
143 | 143 | set enabled=False to disable the SQLite history, |
|
144 | 144 | in which case there will be no stored history, no SQLite connection, |
|
145 | 145 | and no background saving thread. This may be necessary in some |
|
146 | 146 | threaded environments where IPython is embedded. |
|
147 | 147 | """ |
|
148 | 148 | ) |
|
149 | 149 | |
|
150 | 150 | connection_options = Dict(config=True, |
|
151 | 151 | help="""Options for configuring the SQLite connection |
|
152 | 152 | |
|
153 | 153 | These options are passed as keyword args to sqlite3.connect |
|
154 | 154 | when establishing database conenctions. |
|
155 | 155 | """ |
|
156 | 156 | ) |
|
157 | 157 | |
|
158 | 158 | # The SQLite database |
|
159 | 159 | db = Any() |
|
160 | 160 | def _db_changed(self, name, old, new): |
|
161 | 161 | """validate the db, since it can be an Instance of two different types""" |
|
162 | 162 | connection_types = (DummyDB,) |
|
163 | 163 | if sqlite3 is not None: |
|
164 | 164 | connection_types = (DummyDB, sqlite3.Connection) |
|
165 | 165 | if not isinstance(new, connection_types): |
|
166 | 166 | msg = "%s.db must be sqlite3 Connection or DummyDB, not %r" % \ |
|
167 | 167 | (self.__class__.__name__, new) |
|
168 | 168 | raise TraitError(msg) |
|
169 | 169 | |
|
170 | 170 | def __init__(self, profile='default', hist_file=u'', **traits): |
|
171 | 171 | """Create a new history accessor. |
|
172 | 172 | |
|
173 | 173 | Parameters |
|
174 | 174 | ---------- |
|
175 | 175 | profile : str |
|
176 | 176 | The name of the profile from which to open history. |
|
177 | 177 | hist_file : str |
|
178 | 178 | Path to an SQLite history database stored by IPython. If specified, |
|
179 | 179 | hist_file overrides profile. |
|
180 | 180 | config : :class:`~IPython.config.loader.Config` |
|
181 | 181 | Config object. hist_file can also be set through this. |
|
182 | 182 | """ |
|
183 | 183 | # We need a pointer back to the shell for various tasks. |
|
184 | 184 | super(HistoryAccessor, self).__init__(**traits) |
|
185 | 185 | # defer setting hist_file from kwarg until after init, |
|
186 | 186 | # otherwise the default kwarg value would clobber any value |
|
187 | 187 | # set by config |
|
188 | 188 | if hist_file: |
|
189 | 189 | self.hist_file = hist_file |
|
190 | 190 | |
|
191 | 191 | if self.hist_file == u'': |
|
192 | 192 | # No one has set the hist_file, yet. |
|
193 | 193 | self.hist_file = self._get_hist_file_name(profile) |
|
194 | 194 | |
|
195 | 195 | if sqlite3 is None and self.enabled: |
|
196 | 196 | warn("IPython History requires SQLite, your history will not be saved") |
|
197 | 197 | self.enabled = False |
|
198 | 198 | |
|
199 | 199 | self.init_db() |
|
200 | 200 | |
|
201 | 201 | def _get_hist_file_name(self, profile='default'): |
|
202 | 202 | """Find the history file for the given profile name. |
|
203 | 203 | |
|
204 | 204 | This is overridden by the HistoryManager subclass, to use the shell's |
|
205 | 205 | active profile. |
|
206 | 206 | |
|
207 | 207 | Parameters |
|
208 | 208 | ---------- |
|
209 | 209 | profile : str |
|
210 | 210 | The name of a profile which has a history file. |
|
211 | 211 | """ |
|
212 | 212 | return os.path.join(locate_profile(profile), 'history.sqlite') |
|
213 | 213 | |
|
214 | 214 | @catch_corrupt_db |
|
215 | 215 | def init_db(self): |
|
216 | 216 | """Connect to the database, and create tables if necessary.""" |
|
217 | 217 | if not self.enabled: |
|
218 | 218 | self.db = DummyDB() |
|
219 | 219 | return |
|
220 | 220 | |
|
221 | 221 | # use detect_types so that timestamps return datetime objects |
|
222 | 222 | kwargs = dict(detect_types=sqlite3.PARSE_DECLTYPES|sqlite3.PARSE_COLNAMES) |
|
223 | 223 | kwargs.update(self.connection_options) |
|
224 | 224 | self.db = sqlite3.connect(self.hist_file, **kwargs) |
|
225 | 225 | self.db.execute("""CREATE TABLE IF NOT EXISTS sessions (session integer |
|
226 | 226 | primary key autoincrement, start timestamp, |
|
227 | 227 | end timestamp, num_cmds integer, remark text)""") |
|
228 | 228 | self.db.execute("""CREATE TABLE IF NOT EXISTS history |
|
229 | 229 | (session integer, line integer, source text, source_raw text, |
|
230 | 230 | PRIMARY KEY (session, line))""") |
|
231 | 231 | # Output history is optional, but ensure the table's there so it can be |
|
232 | 232 | # enabled later. |
|
233 | 233 | self.db.execute("""CREATE TABLE IF NOT EXISTS output_history |
|
234 | 234 | (session integer, line integer, output text, |
|
235 | 235 | PRIMARY KEY (session, line))""") |
|
236 | 236 | self.db.commit() |
|
237 | 237 | |
|
238 | 238 | def writeout_cache(self): |
|
239 | 239 | """Overridden by HistoryManager to dump the cache before certain |
|
240 | 240 | database lookups.""" |
|
241 | 241 | pass |
|
242 | 242 | |
|
243 | 243 | ## ------------------------------- |
|
244 | 244 | ## Methods for retrieving history: |
|
245 | 245 | ## ------------------------------- |
|
246 | 246 | def _run_sql(self, sql, params, raw=True, output=False): |
|
247 | 247 | """Prepares and runs an SQL query for the history database. |
|
248 | 248 | |
|
249 | 249 | Parameters |
|
250 | 250 | ---------- |
|
251 | 251 | sql : str |
|
252 | 252 | Any filtering expressions to go after SELECT ... FROM ... |
|
253 | 253 | params : tuple |
|
254 | 254 | Parameters passed to the SQL query (to replace "?") |
|
255 | 255 | raw, output : bool |
|
256 | 256 | See :meth:`get_range` |
|
257 | 257 | |
|
258 | 258 | Returns |
|
259 | 259 | ------- |
|
260 | 260 | Tuples as :meth:`get_range` |
|
261 | 261 | """ |
|
262 | 262 | toget = 'source_raw' if raw else 'source' |
|
263 | 263 | sqlfrom = "history" |
|
264 | 264 | if output: |
|
265 | 265 | sqlfrom = "history LEFT JOIN output_history USING (session, line)" |
|
266 | 266 | toget = "history.%s, output_history.output" % toget |
|
267 | 267 | cur = self.db.execute("SELECT session, line, %s FROM %s " %\ |
|
268 | 268 | (toget, sqlfrom) + sql, params) |
|
269 | 269 | if output: # Regroup into 3-tuples, and parse JSON |
|
270 | 270 | return ((ses, lin, (inp, out)) for ses, lin, inp, out in cur) |
|
271 | 271 | return cur |
|
272 | 272 | |
|
273 | 273 | @needs_sqlite |
|
274 | 274 | @catch_corrupt_db |
|
275 | 275 | def get_session_info(self, session): |
|
276 | 276 | """Get info about a session. |
|
277 | 277 | |
|
278 | 278 | Parameters |
|
279 | 279 | ---------- |
|
280 | 280 | |
|
281 | 281 | session : int |
|
282 | 282 | Session number to retrieve. |
|
283 | 283 | |
|
284 | 284 | Returns |
|
285 | 285 | ------- |
|
286 | 286 | |
|
287 | 287 | session_id : int |
|
288 | 288 | Session ID number |
|
289 | 289 | start : datetime |
|
290 | 290 | Timestamp for the start of the session. |
|
291 | 291 | end : datetime |
|
292 | 292 | Timestamp for the end of the session, or None if IPython crashed. |
|
293 | 293 | num_cmds : int |
|
294 | 294 | Number of commands run, or None if IPython crashed. |
|
295 | 295 | remark : unicode |
|
296 | 296 | A manually set description. |
|
297 | 297 | """ |
|
298 | 298 | query = "SELECT * from sessions where session == ?" |
|
299 | 299 | return self.db.execute(query, (session,)).fetchone() |
|
300 | 300 | |
|
301 | 301 | @catch_corrupt_db |
|
302 | 302 | def get_last_session_id(self): |
|
303 | 303 | """Get the last session ID currently in the database. |
|
304 | 304 | |
|
305 | 305 | Within IPython, this should be the same as the value stored in |
|
306 | 306 | :attr:`HistoryManager.session_number`. |
|
307 | 307 | """ |
|
308 | 308 | for record in self.get_tail(n=1, include_latest=True): |
|
309 | 309 | return record[0] |
|
310 | 310 | |
|
311 | 311 | @catch_corrupt_db |
|
312 | 312 | def get_tail(self, n=10, raw=True, output=False, include_latest=False): |
|
313 | 313 | """Get the last n lines from the history database. |
|
314 | 314 | |
|
315 | 315 | Parameters |
|
316 | 316 | ---------- |
|
317 | 317 | n : int |
|
318 | 318 | The number of lines to get |
|
319 | 319 | raw, output : bool |
|
320 | 320 | See :meth:`get_range` |
|
321 | 321 | include_latest : bool |
|
322 | 322 | If False (default), n+1 lines are fetched, and the latest one |
|
323 | 323 | is discarded. This is intended to be used where the function |
|
324 | 324 | is called by a user command, which it should not return. |
|
325 | 325 | |
|
326 | 326 | Returns |
|
327 | 327 | ------- |
|
328 | 328 | Tuples as :meth:`get_range` |
|
329 | 329 | """ |
|
330 | 330 | self.writeout_cache() |
|
331 | 331 | if not include_latest: |
|
332 | 332 | n += 1 |
|
333 | 333 | cur = self._run_sql("ORDER BY session DESC, line DESC LIMIT ?", |
|
334 | 334 | (n,), raw=raw, output=output) |
|
335 | 335 | if not include_latest: |
|
336 | 336 | return reversed(list(cur)[1:]) |
|
337 | 337 | return reversed(list(cur)) |
|
338 | 338 | |
|
339 | 339 | @catch_corrupt_db |
|
340 | 340 | def search(self, pattern="*", raw=True, search_raw=True, |
|
341 | 341 | output=False, n=None, unique=False): |
|
342 | 342 | """Search the database using unix glob-style matching (wildcards |
|
343 | 343 | * and ?). |
|
344 | 344 | |
|
345 | 345 | Parameters |
|
346 | 346 | ---------- |
|
347 | 347 | pattern : str |
|
348 | 348 | The wildcarded pattern to match when searching |
|
349 | 349 | search_raw : bool |
|
350 | 350 | If True, search the raw input, otherwise, the parsed input |
|
351 | 351 | raw, output : bool |
|
352 | 352 | See :meth:`get_range` |
|
353 | 353 | n : None or int |
|
354 | 354 | If an integer is given, it defines the limit of |
|
355 | 355 | returned entries. |
|
356 | 356 | unique : bool |
|
357 | 357 | When it is true, return only unique entries. |
|
358 | 358 | |
|
359 | 359 | Returns |
|
360 | 360 | ------- |
|
361 | 361 | Tuples as :meth:`get_range` |
|
362 | 362 | """ |
|
363 | 363 | tosearch = "source_raw" if search_raw else "source" |
|
364 | 364 | if output: |
|
365 | 365 | tosearch = "history." + tosearch |
|
366 | 366 | self.writeout_cache() |
|
367 | 367 | sqlform = "WHERE %s GLOB ?" % tosearch |
|
368 | 368 | params = (pattern,) |
|
369 | 369 | if unique: |
|
370 | 370 | sqlform += ' GROUP BY {0}'.format(tosearch) |
|
371 | 371 | if n is not None: |
|
372 | 372 | sqlform += " ORDER BY session DESC, line DESC LIMIT ?" |
|
373 | 373 | params += (n,) |
|
374 | 374 | elif unique: |
|
375 | 375 | sqlform += " ORDER BY session, line" |
|
376 | 376 | cur = self._run_sql(sqlform, params, raw=raw, output=output) |
|
377 | 377 | if n is not None: |
|
378 | 378 | return reversed(list(cur)) |
|
379 | 379 | return cur |
|
380 | 380 | |
|
381 | 381 | @catch_corrupt_db |
|
382 | 382 | def get_range(self, session, start=1, stop=None, raw=True,output=False): |
|
383 | 383 | """Retrieve input by session. |
|
384 | 384 | |
|
385 | 385 | Parameters |
|
386 | 386 | ---------- |
|
387 | 387 | session : int |
|
388 | 388 | Session number to retrieve. |
|
389 | 389 | start : int |
|
390 | 390 | First line to retrieve. |
|
391 | 391 | stop : int |
|
392 | 392 | End of line range (excluded from output itself). If None, retrieve |
|
393 | 393 | to the end of the session. |
|
394 | 394 | raw : bool |
|
395 | 395 | If True, return untranslated input |
|
396 | 396 | output : bool |
|
397 | 397 | If True, attempt to include output. This will be 'real' Python |
|
398 | 398 | objects for the current session, or text reprs from previous |
|
399 | 399 | sessions if db_log_output was enabled at the time. Where no output |
|
400 | 400 | is found, None is used. |
|
401 | 401 | |
|
402 | 402 | Returns |
|
403 | 403 | ------- |
|
404 | 404 | entries |
|
405 | 405 | An iterator over the desired lines. Each line is a 3-tuple, either |
|
406 | 406 | (session, line, input) if output is False, or |
|
407 | 407 | (session, line, (input, output)) if output is True. |
|
408 | 408 | """ |
|
409 | 409 | if stop: |
|
410 | 410 | lineclause = "line >= ? AND line < ?" |
|
411 | 411 | params = (session, start, stop) |
|
412 | 412 | else: |
|
413 | 413 | lineclause = "line>=?" |
|
414 | 414 | params = (session, start) |
|
415 | 415 | |
|
416 | 416 | return self._run_sql("WHERE session==? AND %s" % lineclause, |
|
417 | 417 | params, raw=raw, output=output) |
|
418 | 418 | |
|
419 | 419 | def get_range_by_str(self, rangestr, raw=True, output=False): |
|
420 | 420 | """Get lines of history from a string of ranges, as used by magic |
|
421 | 421 | commands %hist, %save, %macro, etc. |
|
422 | 422 | |
|
423 | 423 | Parameters |
|
424 | 424 | ---------- |
|
425 | 425 | rangestr : str |
|
426 | 426 | A string specifying ranges, e.g. "5 ~2/1-4". See |
|
427 | 427 | :func:`magic_history` for full details. |
|
428 | 428 | raw, output : bool |
|
429 | 429 | As :meth:`get_range` |
|
430 | 430 | |
|
431 | 431 | Returns |
|
432 | 432 | ------- |
|
433 | 433 | Tuples as :meth:`get_range` |
|
434 | 434 | """ |
|
435 | 435 | for sess, s, e in extract_hist_ranges(rangestr): |
|
436 | 436 | for line in self.get_range(sess, s, e, raw=raw, output=output): |
|
437 | 437 | yield line |
|
438 | 438 | |
|
439 | 439 | |
|
440 | 440 | class HistoryManager(HistoryAccessor): |
|
441 | 441 | """A class to organize all history-related functionality in one place. |
|
442 | 442 | """ |
|
443 | 443 | # Public interface |
|
444 | 444 | |
|
445 | 445 | # An instance of the IPython shell we are attached to |
|
446 | 446 | shell = Instance('IPython.core.interactiveshell.InteractiveShellABC') |
|
447 | 447 | # Lists to hold processed and raw history. These start with a blank entry |
|
448 | 448 | # so that we can index them starting from 1 |
|
449 | 449 | input_hist_parsed = List([""]) |
|
450 | 450 | input_hist_raw = List([""]) |
|
451 | 451 | # A list of directories visited during session |
|
452 | 452 | dir_hist = List() |
|
453 | 453 | def _dir_hist_default(self): |
|
454 | 454 | try: |
|
455 | 455 | return [py3compat.getcwd()] |
|
456 | 456 | except OSError: |
|
457 | 457 | return [] |
|
458 | 458 | |
|
459 | 459 | # A dict of output history, keyed with ints from the shell's |
|
460 | 460 | # execution count. |
|
461 | 461 | output_hist = Dict() |
|
462 | 462 | # The text/plain repr of outputs. |
|
463 | 463 | output_hist_reprs = Dict() |
|
464 | 464 | |
|
465 | 465 | # The number of the current session in the history database |
|
466 | 466 | session_number = Integer() |
|
467 | 467 | |
|
468 | 468 | db_log_output = Bool(False, config=True, |
|
469 | 469 | help="Should the history database include output? (default: no)" |
|
470 | 470 | ) |
|
471 | 471 | db_cache_size = Integer(0, config=True, |
|
472 | 472 | help="Write to database every x commands (higher values save disk access & power).\n" |
|
473 | 473 | "Values of 1 or less effectively disable caching." |
|
474 | 474 | ) |
|
475 | 475 | # The input and output caches |
|
476 | 476 | db_input_cache = List() |
|
477 | 477 | db_output_cache = List() |
|
478 | 478 | |
|
479 | 479 | # History saving in separate thread |
|
480 | 480 | save_thread = Instance('IPython.core.history.HistorySavingThread') |
|
481 | 481 | try: # Event is a function returning an instance of _Event... |
|
482 | 482 | save_flag = Instance(threading._Event) |
|
483 | 483 | except AttributeError: # ...until Python 3.3, when it's a class. |
|
484 | 484 | save_flag = Instance(threading.Event) |
|
485 | 485 | |
|
486 | 486 | # Private interface |
|
487 | 487 | # Variables used to store the three last inputs from the user. On each new |
|
488 | 488 | # history update, we populate the user's namespace with these, shifted as |
|
489 | 489 | # necessary. |
|
490 | 490 | _i00 = Unicode(u'') |
|
491 | 491 | _i = Unicode(u'') |
|
492 | 492 | _ii = Unicode(u'') |
|
493 | 493 | _iii = Unicode(u'') |
|
494 | 494 | |
|
495 | 495 | # A regex matching all forms of the exit command, so that we don't store |
|
496 | 496 | # them in the history (it's annoying to rewind the first entry and land on |
|
497 | 497 | # an exit call). |
|
498 | 498 | _exit_re = re.compile(r"(exit|quit)(\s*\(.*\))?$") |
|
499 | 499 | |
|
500 | 500 | def __init__(self, shell=None, config=None, **traits): |
|
501 | 501 | """Create a new history manager associated with a shell instance. |
|
502 | 502 | """ |
|
503 | 503 | # We need a pointer back to the shell for various tasks. |
|
504 | 504 | super(HistoryManager, self).__init__(shell=shell, config=config, |
|
505 | 505 | **traits) |
|
506 | 506 | self.save_flag = threading.Event() |
|
507 | 507 | self.db_input_cache_lock = threading.Lock() |
|
508 | 508 | self.db_output_cache_lock = threading.Lock() |
|
509 | 509 | if self.enabled and self.hist_file != ':memory:': |
|
510 | 510 | self.save_thread = HistorySavingThread(self) |
|
511 | 511 | self.save_thread.start() |
|
512 | 512 | |
|
513 | 513 | self.new_session() |
|
514 | 514 | |
|
515 | 515 | def _get_hist_file_name(self, profile=None): |
|
516 | 516 | """Get default history file name based on the Shell's profile. |
|
517 | 517 | |
|
518 | 518 | The profile parameter is ignored, but must exist for compatibility with |
|
519 | 519 | the parent class.""" |
|
520 | 520 | profile_dir = self.shell.profile_dir.location |
|
521 | 521 | return os.path.join(profile_dir, 'history.sqlite') |
|
522 | 522 | |
|
523 | 523 | @needs_sqlite |
|
524 | 524 | def new_session(self, conn=None): |
|
525 | 525 | """Get a new session number.""" |
|
526 | 526 | if conn is None: |
|
527 | 527 | conn = self.db |
|
528 | 528 | |
|
529 | 529 | with conn: |
|
530 | 530 | cur = conn.execute("""INSERT INTO sessions VALUES (NULL, ?, NULL, |
|
531 | 531 | NULL, "") """, (datetime.datetime.now(),)) |
|
532 | 532 | self.session_number = cur.lastrowid |
|
533 | 533 | |
|
534 | 534 | def end_session(self): |
|
535 | 535 | """Close the database session, filling in the end time and line count.""" |
|
536 | 536 | self.writeout_cache() |
|
537 | 537 | with self.db: |
|
538 | 538 | self.db.execute("""UPDATE sessions SET end=?, num_cmds=? WHERE |
|
539 | 539 | session==?""", (datetime.datetime.now(), |
|
540 | 540 | len(self.input_hist_parsed)-1, self.session_number)) |
|
541 | 541 | self.session_number = 0 |
|
542 | 542 | |
|
543 | 543 | def name_session(self, name): |
|
544 | 544 | """Give the current session a name in the history database.""" |
|
545 | 545 | with self.db: |
|
546 | 546 | self.db.execute("UPDATE sessions SET remark=? WHERE session==?", |
|
547 | 547 | (name, self.session_number)) |
|
548 | 548 | |
|
549 | 549 | def reset(self, new_session=True): |
|
550 | 550 | """Clear the session history, releasing all object references, and |
|
551 | 551 | optionally open a new session.""" |
|
552 | 552 | self.output_hist.clear() |
|
553 | 553 | # The directory history can't be completely empty |
|
554 | 554 | self.dir_hist[:] = [py3compat.getcwd()] |
|
555 | 555 | |
|
556 | 556 | if new_session: |
|
557 | 557 | if self.session_number: |
|
558 | 558 | self.end_session() |
|
559 | 559 | self.input_hist_parsed[:] = [""] |
|
560 | 560 | self.input_hist_raw[:] = [""] |
|
561 | 561 | self.new_session() |
|
562 | 562 | |
|
563 | 563 | # ------------------------------ |
|
564 | 564 | # Methods for retrieving history |
|
565 | 565 | # ------------------------------ |
|
566 | 566 | def get_session_info(self, session=0): |
|
567 | 567 | """Get info about a session. |
|
568 | 568 | |
|
569 | 569 | Parameters |
|
570 | 570 | ---------- |
|
571 | 571 | |
|
572 | 572 | session : int |
|
573 | 573 | Session number to retrieve. The current session is 0, and negative |
|
574 | 574 | numbers count back from current session, so -1 is the previous session. |
|
575 | 575 | |
|
576 | 576 | Returns |
|
577 | 577 | ------- |
|
578 | 578 | |
|
579 | 579 | session_id : int |
|
580 | 580 | Session ID number |
|
581 | 581 | start : datetime |
|
582 | 582 | Timestamp for the start of the session. |
|
583 | 583 | end : datetime |
|
584 | 584 | Timestamp for the end of the session, or None if IPython crashed. |
|
585 | 585 | num_cmds : int |
|
586 | 586 | Number of commands run, or None if IPython crashed. |
|
587 | 587 | remark : unicode |
|
588 | 588 | A manually set description. |
|
589 | 589 | """ |
|
590 | 590 | if session <= 0: |
|
591 | 591 | session += self.session_number |
|
592 | 592 | |
|
593 | 593 | return super(HistoryManager, self).get_session_info(session=session) |
|
594 | 594 | |
|
595 | 595 | def _get_range_session(self, start=1, stop=None, raw=True, output=False): |
|
596 | 596 | """Get input and output history from the current session. Called by |
|
597 | 597 | get_range, and takes similar parameters.""" |
|
598 | 598 | input_hist = self.input_hist_raw if raw else self.input_hist_parsed |
|
599 | 599 | |
|
600 | 600 | n = len(input_hist) |
|
601 | 601 | if start < 0: |
|
602 | 602 | start += n |
|
603 | 603 | if not stop or (stop > n): |
|
604 | 604 | stop = n |
|
605 | 605 | elif stop < 0: |
|
606 | 606 | stop += n |
|
607 | 607 | |
|
608 | 608 | for i in range(start, stop): |
|
609 | 609 | if output: |
|
610 | 610 | line = (input_hist[i], self.output_hist_reprs.get(i)) |
|
611 | 611 | else: |
|
612 | 612 | line = input_hist[i] |
|
613 | 613 | yield (0, i, line) |
|
614 | 614 | |
|
615 | 615 | def get_range(self, session=0, start=1, stop=None, raw=True,output=False): |
|
616 | 616 | """Retrieve input by session. |
|
617 | 617 | |
|
618 | 618 | Parameters |
|
619 | 619 | ---------- |
|
620 | 620 | session : int |
|
621 | 621 | Session number to retrieve. The current session is 0, and negative |
|
622 | 622 | numbers count back from current session, so -1 is previous session. |
|
623 | 623 | start : int |
|
624 | 624 | First line to retrieve. |
|
625 | 625 | stop : int |
|
626 | 626 | End of line range (excluded from output itself). If None, retrieve |
|
627 | 627 | to the end of the session. |
|
628 | 628 | raw : bool |
|
629 | 629 | If True, return untranslated input |
|
630 | 630 | output : bool |
|
631 | 631 | If True, attempt to include output. This will be 'real' Python |
|
632 | 632 | objects for the current session, or text reprs from previous |
|
633 | 633 | sessions if db_log_output was enabled at the time. Where no output |
|
634 | 634 | is found, None is used. |
|
635 | 635 | |
|
636 | 636 | Returns |
|
637 | 637 | ------- |
|
638 | 638 | entries |
|
639 | 639 | An iterator over the desired lines. Each line is a 3-tuple, either |
|
640 | 640 | (session, line, input) if output is False, or |
|
641 | 641 | (session, line, (input, output)) if output is True. |
|
642 | 642 | """ |
|
643 | 643 | if session <= 0: |
|
644 | 644 | session += self.session_number |
|
645 | 645 | if session==self.session_number: # Current session |
|
646 | 646 | return self._get_range_session(start, stop, raw, output) |
|
647 | 647 | return super(HistoryManager, self).get_range(session, start, stop, raw, |
|
648 | 648 | output) |
|
649 | 649 | |
|
650 | 650 | ## ---------------------------- |
|
651 | 651 | ## Methods for storing history: |
|
652 | 652 | ## ---------------------------- |
|
653 | 653 | def store_inputs(self, line_num, source, source_raw=None): |
|
654 | 654 | """Store source and raw input in history and create input cache |
|
655 | 655 | variables ``_i*``. |
|
656 | 656 | |
|
657 | 657 | Parameters |
|
658 | 658 | ---------- |
|
659 | 659 | line_num : int |
|
660 | 660 | The prompt number of this input. |
|
661 | 661 | |
|
662 | 662 | source : str |
|
663 | 663 | Python input. |
|
664 | 664 | |
|
665 | 665 | source_raw : str, optional |
|
666 | 666 | If given, this is the raw input without any IPython transformations |
|
667 | 667 | applied to it. If not given, ``source`` is used. |
|
668 | 668 | """ |
|
669 | 669 | if source_raw is None: |
|
670 | 670 | source_raw = source |
|
671 | 671 | source = source.rstrip('\n') |
|
672 | 672 | source_raw = source_raw.rstrip('\n') |
|
673 | 673 | |
|
674 | 674 | # do not store exit/quit commands |
|
675 | 675 | if self._exit_re.match(source_raw.strip()): |
|
676 | 676 | return |
|
677 | 677 | |
|
678 | 678 | self.input_hist_parsed.append(source) |
|
679 | 679 | self.input_hist_raw.append(source_raw) |
|
680 | 680 | |
|
681 | 681 | with self.db_input_cache_lock: |
|
682 | 682 | self.db_input_cache.append((line_num, source, source_raw)) |
|
683 | 683 | # Trigger to flush cache and write to DB. |
|
684 | 684 | if len(self.db_input_cache) >= self.db_cache_size: |
|
685 | 685 | self.save_flag.set() |
|
686 | 686 | |
|
687 | 687 | # update the auto _i variables |
|
688 | 688 | self._iii = self._ii |
|
689 | 689 | self._ii = self._i |
|
690 | 690 | self._i = self._i00 |
|
691 | 691 | self._i00 = source_raw |
|
692 | 692 | |
|
693 | 693 | # hackish access to user namespace to create _i1,_i2... dynamically |
|
694 | 694 | new_i = '_i%s' % line_num |
|
695 | 695 | to_main = {'_i': self._i, |
|
696 | 696 | '_ii': self._ii, |
|
697 | 697 | '_iii': self._iii, |
|
698 | 698 | new_i : self._i00 } |
|
699 | 699 | |
|
700 | 700 | if self.shell is not None: |
|
701 | 701 | self.shell.push(to_main, interactive=False) |
|
702 | 702 | |
|
703 | 703 | def store_output(self, line_num): |
|
704 | 704 | """If database output logging is enabled, this saves all the |
|
705 | 705 | outputs from the indicated prompt number to the database. It's |
|
706 | 706 | called by run_cell after code has been executed. |
|
707 | 707 | |
|
708 | 708 | Parameters |
|
709 | 709 | ---------- |
|
710 | 710 | line_num : int |
|
711 | 711 | The line number from which to save outputs |
|
712 | 712 | """ |
|
713 | 713 | if (not self.db_log_output) or (line_num not in self.output_hist_reprs): |
|
714 | 714 | return |
|
715 | 715 | output = self.output_hist_reprs[line_num] |
|
716 | 716 | |
|
717 | 717 | with self.db_output_cache_lock: |
|
718 | 718 | self.db_output_cache.append((line_num, output)) |
|
719 | 719 | if self.db_cache_size <= 1: |
|
720 | 720 | self.save_flag.set() |
|
721 | 721 | |
|
722 | 722 | def _writeout_input_cache(self, conn): |
|
723 | 723 | with conn: |
|
724 | 724 | for line in self.db_input_cache: |
|
725 | 725 | conn.execute("INSERT INTO history VALUES (?, ?, ?, ?)", |
|
726 | 726 | (self.session_number,)+line) |
|
727 | 727 | |
|
728 | 728 | def _writeout_output_cache(self, conn): |
|
729 | 729 | with conn: |
|
730 | 730 | for line in self.db_output_cache: |
|
731 | 731 | conn.execute("INSERT INTO output_history VALUES (?, ?, ?)", |
|
732 | 732 | (self.session_number,)+line) |
|
733 | 733 | |
|
734 | 734 | @needs_sqlite |
|
735 | 735 | def writeout_cache(self, conn=None): |
|
736 | 736 | """Write any entries in the cache to the database.""" |
|
737 | 737 | if conn is None: |
|
738 | 738 | conn = self.db |
|
739 | 739 | |
|
740 | 740 | with self.db_input_cache_lock: |
|
741 | 741 | try: |
|
742 | 742 | self._writeout_input_cache(conn) |
|
743 | 743 | except sqlite3.IntegrityError: |
|
744 | 744 | self.new_session(conn) |
|
745 | 745 | print("ERROR! Session/line number was not unique in", |
|
746 | 746 | "database. History logging moved to new session", |
|
747 | 747 | self.session_number) |
|
748 | 748 | try: |
|
749 | 749 | # Try writing to the new session. If this fails, don't |
|
750 | 750 | # recurse |
|
751 | 751 | self._writeout_input_cache(conn) |
|
752 | 752 | except sqlite3.IntegrityError: |
|
753 | 753 | pass |
|
754 | 754 | finally: |
|
755 | 755 | self.db_input_cache = [] |
|
756 | 756 | |
|
757 | 757 | with self.db_output_cache_lock: |
|
758 | 758 | try: |
|
759 | 759 | self._writeout_output_cache(conn) |
|
760 | 760 | except sqlite3.IntegrityError: |
|
761 | 761 | print("!! Session/line number for output was not unique", |
|
762 | 762 | "in database. Output will not be stored.") |
|
763 | 763 | finally: |
|
764 | 764 | self.db_output_cache = [] |
|
765 | 765 | |
|
766 | 766 | |
|
767 | 767 | class HistorySavingThread(threading.Thread): |
|
768 | 768 | """This thread takes care of writing history to the database, so that |
|
769 | 769 | the UI isn't held up while that happens. |
|
770 | 770 | |
|
771 | 771 | It waits for the HistoryManager's save_flag to be set, then writes out |
|
772 | 772 | the history cache. The main thread is responsible for setting the flag when |
|
773 | 773 | the cache size reaches a defined threshold.""" |
|
774 | 774 | daemon = True |
|
775 | 775 | stop_now = False |
|
776 | 776 | enabled = True |
|
777 | 777 | def __init__(self, history_manager): |
|
778 | 778 | super(HistorySavingThread, self).__init__(name="IPythonHistorySavingThread") |
|
779 | 779 | self.history_manager = history_manager |
|
780 | 780 | self.enabled = history_manager.enabled |
|
781 | 781 | atexit.register(self.stop) |
|
782 | 782 | |
|
783 | 783 | @needs_sqlite |
|
784 | 784 | def run(self): |
|
785 | 785 | # We need a separate db connection per thread: |
|
786 | 786 | try: |
|
787 | 787 | self.db = sqlite3.connect(self.history_manager.hist_file, |
|
788 | 788 | **self.history_manager.connection_options |
|
789 | 789 | ) |
|
790 | 790 | while True: |
|
791 | 791 | self.history_manager.save_flag.wait() |
|
792 | 792 | if self.stop_now: |
|
793 | 793 | self.db.close() |
|
794 | 794 | return |
|
795 | 795 | self.history_manager.save_flag.clear() |
|
796 | 796 | self.history_manager.writeout_cache(self.db) |
|
797 | 797 | except Exception as e: |
|
798 | 798 | print(("The history saving thread hit an unexpected error (%s)." |
|
799 | 799 | "History will not be written to the database.") % repr(e)) |
|
800 | 800 | |
|
801 | 801 | def stop(self): |
|
802 | 802 | """This can be called from the main thread to safely stop this thread. |
|
803 | 803 | |
|
804 | 804 | Note that it does not attempt to write out remaining history before |
|
805 | 805 | exiting. That should be done by calling the HistoryManager's |
|
806 | 806 | end_session method.""" |
|
807 | 807 | self.stop_now = True |
|
808 | 808 | self.history_manager.save_flag.set() |
|
809 | 809 | self.join() |
|
810 | 810 | |
|
811 | 811 | |
|
812 | 812 | # To match, e.g. ~5/8-~2/3 |
|
813 | 813 | range_re = re.compile(r""" |
|
814 | 814 | ((?P<startsess>~?\d+)/)? |
|
815 | 815 | (?P<start>\d+)? |
|
816 | 816 | ((?P<sep>[\-:]) |
|
817 | 817 | ((?P<endsess>~?\d+)/)? |
|
818 | 818 | (?P<end>\d+))? |
|
819 | 819 | $""", re.VERBOSE) |
|
820 | 820 | |
|
821 | 821 | |
|
822 | 822 | def extract_hist_ranges(ranges_str): |
|
823 | 823 | """Turn a string of history ranges into 3-tuples of (session, start, stop). |
|
824 | 824 | |
|
825 | 825 | Examples |
|
826 | 826 | -------- |
|
827 | 827 | >>> list(extract_hist_ranges("~8/5-~7/4 2")) |
|
828 | 828 | [(-8, 5, None), (-7, 1, 5), (0, 2, 3)] |
|
829 | 829 | """ |
|
830 | 830 | for range_str in ranges_str.split(): |
|
831 | 831 | rmatch = range_re.match(range_str) |
|
832 | 832 | if not rmatch: |
|
833 | 833 | continue |
|
834 | 834 | start = rmatch.group("start") |
|
835 | 835 | if start: |
|
836 | 836 | start = int(start) |
|
837 | 837 | end = rmatch.group("end") |
|
838 | 838 | # If no end specified, get (a, a + 1) |
|
839 | 839 | end = int(end) if end else start + 1 |
|
840 | 840 | else: # start not specified |
|
841 | 841 | if not rmatch.group('startsess'): # no startsess |
|
842 | 842 | continue |
|
843 | 843 | start = 1 |
|
844 | 844 | end = None # provide the entire session hist |
|
845 | 845 | |
|
846 | 846 | if rmatch.group("sep") == "-": # 1-3 == 1:4 --> [1, 2, 3] |
|
847 | 847 | end += 1 |
|
848 | 848 | startsess = rmatch.group("startsess") or "0" |
|
849 | 849 | endsess = rmatch.group("endsess") or startsess |
|
850 | 850 | startsess = int(startsess.replace("~","-")) |
|
851 | 851 | endsess = int(endsess.replace("~","-")) |
|
852 | 852 | assert endsess >= startsess, "start session must be earlier than end session" |
|
853 | 853 | |
|
854 | 854 | if endsess == startsess: |
|
855 | 855 | yield (startsess, start, end) |
|
856 | 856 | continue |
|
857 | 857 | # Multiple sessions in one range: |
|
858 | 858 | yield (startsess, start, None) |
|
859 | 859 | for sess in range(startsess+1, endsess): |
|
860 | 860 | yield (sess, 1, None) |
|
861 | 861 | yield (endsess, 1, end) |
|
862 | 862 | |
|
863 | 863 | |
|
864 | 864 | def _format_lineno(session, line): |
|
865 | 865 | """Helper function to format line numbers properly.""" |
|
866 | 866 | if session == 0: |
|
867 | 867 | return str(line) |
|
868 | 868 | return "%s#%s" % (session, line) |
|
869 | 869 | |
|
870 | 870 |
@@ -1,702 +1,702 b'' | |||
|
1 | 1 | # encoding: utf-8 |
|
2 | 2 | """Magic functions for InteractiveShell. |
|
3 | 3 | """ |
|
4 | 4 | from __future__ import print_function |
|
5 | 5 | |
|
6 | 6 | #----------------------------------------------------------------------------- |
|
7 | 7 | # Copyright (C) 2001 Janko Hauser <jhauser@zscout.de> and |
|
8 | 8 | # Copyright (C) 2001 Fernando Perez <fperez@colorado.edu> |
|
9 | 9 | # Copyright (C) 2008 The IPython Development Team |
|
10 | 10 | |
|
11 | 11 | # Distributed under the terms of the BSD License. The full license is in |
|
12 | 12 | # the file COPYING, distributed as part of this software. |
|
13 | 13 | #----------------------------------------------------------------------------- |
|
14 | 14 | |
|
15 | 15 | #----------------------------------------------------------------------------- |
|
16 | 16 | # Imports |
|
17 | 17 | #----------------------------------------------------------------------------- |
|
18 | 18 | # Stdlib |
|
19 | 19 | import os |
|
20 | 20 | import re |
|
21 | 21 | import sys |
|
22 | 22 | import types |
|
23 | 23 | from getopt import getopt, GetoptError |
|
24 | 24 | |
|
25 | 25 | # Our own |
|
26 | 26 | from IPython.config.configurable import Configurable |
|
27 | 27 | from IPython.core import oinspect |
|
28 | 28 | from IPython.core.error import UsageError |
|
29 | 29 | from IPython.core.inputsplitter import ESC_MAGIC, ESC_MAGIC2 |
|
30 |
from |
|
|
30 | from decorator import decorator | |
|
31 | 31 | from IPython.utils.ipstruct import Struct |
|
32 | 32 | from IPython.utils.process import arg_split |
|
33 | 33 | from IPython.utils.py3compat import string_types, iteritems |
|
34 | 34 | from IPython.utils.text import dedent |
|
35 | 35 | from IPython.utils.traitlets import Bool, Dict, Instance, MetaHasTraits |
|
36 | 36 | from IPython.utils.warn import error |
|
37 | 37 | |
|
38 | 38 | #----------------------------------------------------------------------------- |
|
39 | 39 | # Globals |
|
40 | 40 | #----------------------------------------------------------------------------- |
|
41 | 41 | |
|
42 | 42 | # A dict we'll use for each class that has magics, used as temporary storage to |
|
43 | 43 | # pass information between the @line/cell_magic method decorators and the |
|
44 | 44 | # @magics_class class decorator, because the method decorators have no |
|
45 | 45 | # access to the class when they run. See for more details: |
|
46 | 46 | # http://stackoverflow.com/questions/2366713/can-a-python-decorator-of-an-instance-method-access-the-class |
|
47 | 47 | |
|
48 | 48 | magics = dict(line={}, cell={}) |
|
49 | 49 | |
|
50 | 50 | magic_kinds = ('line', 'cell') |
|
51 | 51 | magic_spec = ('line', 'cell', 'line_cell') |
|
52 | 52 | magic_escapes = dict(line=ESC_MAGIC, cell=ESC_MAGIC2) |
|
53 | 53 | |
|
54 | 54 | #----------------------------------------------------------------------------- |
|
55 | 55 | # Utility classes and functions |
|
56 | 56 | #----------------------------------------------------------------------------- |
|
57 | 57 | |
|
58 | 58 | class Bunch: pass |
|
59 | 59 | |
|
60 | 60 | |
|
61 | 61 | def on_off(tag): |
|
62 | 62 | """Return an ON/OFF string for a 1/0 input. Simple utility function.""" |
|
63 | 63 | return ['OFF','ON'][tag] |
|
64 | 64 | |
|
65 | 65 | |
|
66 | 66 | def compress_dhist(dh): |
|
67 | 67 | """Compress a directory history into a new one with at most 20 entries. |
|
68 | 68 | |
|
69 | 69 | Return a new list made from the first and last 10 elements of dhist after |
|
70 | 70 | removal of duplicates. |
|
71 | 71 | """ |
|
72 | 72 | head, tail = dh[:-10], dh[-10:] |
|
73 | 73 | |
|
74 | 74 | newhead = [] |
|
75 | 75 | done = set() |
|
76 | 76 | for h in head: |
|
77 | 77 | if h in done: |
|
78 | 78 | continue |
|
79 | 79 | newhead.append(h) |
|
80 | 80 | done.add(h) |
|
81 | 81 | |
|
82 | 82 | return newhead + tail |
|
83 | 83 | |
|
84 | 84 | |
|
85 | 85 | def needs_local_scope(func): |
|
86 | 86 | """Decorator to mark magic functions which need to local scope to run.""" |
|
87 | 87 | func.needs_local_scope = True |
|
88 | 88 | return func |
|
89 | 89 | |
|
90 | 90 | #----------------------------------------------------------------------------- |
|
91 | 91 | # Class and method decorators for registering magics |
|
92 | 92 | #----------------------------------------------------------------------------- |
|
93 | 93 | |
|
94 | 94 | def magics_class(cls): |
|
95 | 95 | """Class decorator for all subclasses of the main Magics class. |
|
96 | 96 | |
|
97 | 97 | Any class that subclasses Magics *must* also apply this decorator, to |
|
98 | 98 | ensure that all the methods that have been decorated as line/cell magics |
|
99 | 99 | get correctly registered in the class instance. This is necessary because |
|
100 | 100 | when method decorators run, the class does not exist yet, so they |
|
101 | 101 | temporarily store their information into a module global. Application of |
|
102 | 102 | this class decorator copies that global data to the class instance and |
|
103 | 103 | clears the global. |
|
104 | 104 | |
|
105 | 105 | Obviously, this mechanism is not thread-safe, which means that the |
|
106 | 106 | *creation* of subclasses of Magic should only be done in a single-thread |
|
107 | 107 | context. Instantiation of the classes has no restrictions. Given that |
|
108 | 108 | these classes are typically created at IPython startup time and before user |
|
109 | 109 | application code becomes active, in practice this should not pose any |
|
110 | 110 | problems. |
|
111 | 111 | """ |
|
112 | 112 | cls.registered = True |
|
113 | 113 | cls.magics = dict(line = magics['line'], |
|
114 | 114 | cell = magics['cell']) |
|
115 | 115 | magics['line'] = {} |
|
116 | 116 | magics['cell'] = {} |
|
117 | 117 | return cls |
|
118 | 118 | |
|
119 | 119 | |
|
120 | 120 | def record_magic(dct, magic_kind, magic_name, func): |
|
121 | 121 | """Utility function to store a function as a magic of a specific kind. |
|
122 | 122 | |
|
123 | 123 | Parameters |
|
124 | 124 | ---------- |
|
125 | 125 | dct : dict |
|
126 | 126 | A dictionary with 'line' and 'cell' subdicts. |
|
127 | 127 | |
|
128 | 128 | magic_kind : str |
|
129 | 129 | Kind of magic to be stored. |
|
130 | 130 | |
|
131 | 131 | magic_name : str |
|
132 | 132 | Key to store the magic as. |
|
133 | 133 | |
|
134 | 134 | func : function |
|
135 | 135 | Callable object to store. |
|
136 | 136 | """ |
|
137 | 137 | if magic_kind == 'line_cell': |
|
138 | 138 | dct['line'][magic_name] = dct['cell'][magic_name] = func |
|
139 | 139 | else: |
|
140 | 140 | dct[magic_kind][magic_name] = func |
|
141 | 141 | |
|
142 | 142 | |
|
143 | 143 | def validate_type(magic_kind): |
|
144 | 144 | """Ensure that the given magic_kind is valid. |
|
145 | 145 | |
|
146 | 146 | Check that the given magic_kind is one of the accepted spec types (stored |
|
147 | 147 | in the global `magic_spec`), raise ValueError otherwise. |
|
148 | 148 | """ |
|
149 | 149 | if magic_kind not in magic_spec: |
|
150 | 150 | raise ValueError('magic_kind must be one of %s, %s given' % |
|
151 | 151 | magic_kinds, magic_kind) |
|
152 | 152 | |
|
153 | 153 | |
|
154 | 154 | # The docstrings for the decorator below will be fairly similar for the two |
|
155 | 155 | # types (method and function), so we generate them here once and reuse the |
|
156 | 156 | # templates below. |
|
157 | 157 | _docstring_template = \ |
|
158 | 158 | """Decorate the given {0} as {1} magic. |
|
159 | 159 | |
|
160 | 160 | The decorator can be used with or without arguments, as follows. |
|
161 | 161 | |
|
162 | 162 | i) without arguments: it will create a {1} magic named as the {0} being |
|
163 | 163 | decorated:: |
|
164 | 164 | |
|
165 | 165 | @deco |
|
166 | 166 | def foo(...) |
|
167 | 167 | |
|
168 | 168 | will create a {1} magic named `foo`. |
|
169 | 169 | |
|
170 | 170 | ii) with one string argument: which will be used as the actual name of the |
|
171 | 171 | resulting magic:: |
|
172 | 172 | |
|
173 | 173 | @deco('bar') |
|
174 | 174 | def foo(...) |
|
175 | 175 | |
|
176 | 176 | will create a {1} magic named `bar`. |
|
177 | 177 | """ |
|
178 | 178 | |
|
179 | 179 | # These two are decorator factories. While they are conceptually very similar, |
|
180 | 180 | # there are enough differences in the details that it's simpler to have them |
|
181 | 181 | # written as completely standalone functions rather than trying to share code |
|
182 | 182 | # and make a single one with convoluted logic. |
|
183 | 183 | |
|
184 | 184 | def _method_magic_marker(magic_kind): |
|
185 | 185 | """Decorator factory for methods in Magics subclasses. |
|
186 | 186 | """ |
|
187 | 187 | |
|
188 | 188 | validate_type(magic_kind) |
|
189 | 189 | |
|
190 | 190 | # This is a closure to capture the magic_kind. We could also use a class, |
|
191 | 191 | # but it's overkill for just that one bit of state. |
|
192 | 192 | def magic_deco(arg): |
|
193 | 193 | call = lambda f, *a, **k: f(*a, **k) |
|
194 | 194 | |
|
195 | 195 | if callable(arg): |
|
196 | 196 | # "Naked" decorator call (just @foo, no args) |
|
197 | 197 | func = arg |
|
198 | 198 | name = func.__name__ |
|
199 | 199 | retval = decorator(call, func) |
|
200 | 200 | record_magic(magics, magic_kind, name, name) |
|
201 | 201 | elif isinstance(arg, string_types): |
|
202 | 202 | # Decorator called with arguments (@foo('bar')) |
|
203 | 203 | name = arg |
|
204 | 204 | def mark(func, *a, **kw): |
|
205 | 205 | record_magic(magics, magic_kind, name, func.__name__) |
|
206 | 206 | return decorator(call, func) |
|
207 | 207 | retval = mark |
|
208 | 208 | else: |
|
209 | 209 | raise TypeError("Decorator can only be called with " |
|
210 | 210 | "string or function") |
|
211 | 211 | return retval |
|
212 | 212 | |
|
213 | 213 | # Ensure the resulting decorator has a usable docstring |
|
214 | 214 | magic_deco.__doc__ = _docstring_template.format('method', magic_kind) |
|
215 | 215 | return magic_deco |
|
216 | 216 | |
|
217 | 217 | |
|
218 | 218 | def _function_magic_marker(magic_kind): |
|
219 | 219 | """Decorator factory for standalone functions. |
|
220 | 220 | """ |
|
221 | 221 | validate_type(magic_kind) |
|
222 | 222 | |
|
223 | 223 | # This is a closure to capture the magic_kind. We could also use a class, |
|
224 | 224 | # but it's overkill for just that one bit of state. |
|
225 | 225 | def magic_deco(arg): |
|
226 | 226 | call = lambda f, *a, **k: f(*a, **k) |
|
227 | 227 | |
|
228 | 228 | # Find get_ipython() in the caller's namespace |
|
229 | 229 | caller = sys._getframe(1) |
|
230 | 230 | for ns in ['f_locals', 'f_globals', 'f_builtins']: |
|
231 | 231 | get_ipython = getattr(caller, ns).get('get_ipython') |
|
232 | 232 | if get_ipython is not None: |
|
233 | 233 | break |
|
234 | 234 | else: |
|
235 | 235 | raise NameError('Decorator can only run in context where ' |
|
236 | 236 | '`get_ipython` exists') |
|
237 | 237 | |
|
238 | 238 | ip = get_ipython() |
|
239 | 239 | |
|
240 | 240 | if callable(arg): |
|
241 | 241 | # "Naked" decorator call (just @foo, no args) |
|
242 | 242 | func = arg |
|
243 | 243 | name = func.__name__ |
|
244 | 244 | ip.register_magic_function(func, magic_kind, name) |
|
245 | 245 | retval = decorator(call, func) |
|
246 | 246 | elif isinstance(arg, string_types): |
|
247 | 247 | # Decorator called with arguments (@foo('bar')) |
|
248 | 248 | name = arg |
|
249 | 249 | def mark(func, *a, **kw): |
|
250 | 250 | ip.register_magic_function(func, magic_kind, name) |
|
251 | 251 | return decorator(call, func) |
|
252 | 252 | retval = mark |
|
253 | 253 | else: |
|
254 | 254 | raise TypeError("Decorator can only be called with " |
|
255 | 255 | "string or function") |
|
256 | 256 | return retval |
|
257 | 257 | |
|
258 | 258 | # Ensure the resulting decorator has a usable docstring |
|
259 | 259 | ds = _docstring_template.format('function', magic_kind) |
|
260 | 260 | |
|
261 | 261 | ds += dedent(""" |
|
262 | 262 | Note: this decorator can only be used in a context where IPython is already |
|
263 | 263 | active, so that the `get_ipython()` call succeeds. You can therefore use |
|
264 | 264 | it in your startup files loaded after IPython initializes, but *not* in the |
|
265 | 265 | IPython configuration file itself, which is executed before IPython is |
|
266 | 266 | fully up and running. Any file located in the `startup` subdirectory of |
|
267 | 267 | your configuration profile will be OK in this sense. |
|
268 | 268 | """) |
|
269 | 269 | |
|
270 | 270 | magic_deco.__doc__ = ds |
|
271 | 271 | return magic_deco |
|
272 | 272 | |
|
273 | 273 | |
|
274 | 274 | # Create the actual decorators for public use |
|
275 | 275 | |
|
276 | 276 | # These three are used to decorate methods in class definitions |
|
277 | 277 | line_magic = _method_magic_marker('line') |
|
278 | 278 | cell_magic = _method_magic_marker('cell') |
|
279 | 279 | line_cell_magic = _method_magic_marker('line_cell') |
|
280 | 280 | |
|
281 | 281 | # These three decorate standalone functions and perform the decoration |
|
282 | 282 | # immediately. They can only run where get_ipython() works |
|
283 | 283 | register_line_magic = _function_magic_marker('line') |
|
284 | 284 | register_cell_magic = _function_magic_marker('cell') |
|
285 | 285 | register_line_cell_magic = _function_magic_marker('line_cell') |
|
286 | 286 | |
|
287 | 287 | #----------------------------------------------------------------------------- |
|
288 | 288 | # Core Magic classes |
|
289 | 289 | #----------------------------------------------------------------------------- |
|
290 | 290 | |
|
291 | 291 | class MagicsManager(Configurable): |
|
292 | 292 | """Object that handles all magic-related functionality for IPython. |
|
293 | 293 | """ |
|
294 | 294 | # Non-configurable class attributes |
|
295 | 295 | |
|
296 | 296 | # A two-level dict, first keyed by magic type, then by magic function, and |
|
297 | 297 | # holding the actual callable object as value. This is the dict used for |
|
298 | 298 | # magic function dispatch |
|
299 | 299 | magics = Dict |
|
300 | 300 | |
|
301 | 301 | # A registry of the original objects that we've been given holding magics. |
|
302 | 302 | registry = Dict |
|
303 | 303 | |
|
304 | 304 | shell = Instance('IPython.core.interactiveshell.InteractiveShellABC') |
|
305 | 305 | |
|
306 | 306 | auto_magic = Bool(True, config=True, help= |
|
307 | 307 | "Automatically call line magics without requiring explicit % prefix") |
|
308 | 308 | |
|
309 | 309 | def _auto_magic_changed(self, name, value): |
|
310 | 310 | self.shell.automagic = value |
|
311 | 311 | |
|
312 | 312 | _auto_status = [ |
|
313 | 313 | 'Automagic is OFF, % prefix IS needed for line magics.', |
|
314 | 314 | 'Automagic is ON, % prefix IS NOT needed for line magics.'] |
|
315 | 315 | |
|
316 | 316 | user_magics = Instance('IPython.core.magics.UserMagics') |
|
317 | 317 | |
|
318 | 318 | def __init__(self, shell=None, config=None, user_magics=None, **traits): |
|
319 | 319 | |
|
320 | 320 | super(MagicsManager, self).__init__(shell=shell, config=config, |
|
321 | 321 | user_magics=user_magics, **traits) |
|
322 | 322 | self.magics = dict(line={}, cell={}) |
|
323 | 323 | # Let's add the user_magics to the registry for uniformity, so *all* |
|
324 | 324 | # registered magic containers can be found there. |
|
325 | 325 | self.registry[user_magics.__class__.__name__] = user_magics |
|
326 | 326 | |
|
327 | 327 | def auto_status(self): |
|
328 | 328 | """Return descriptive string with automagic status.""" |
|
329 | 329 | return self._auto_status[self.auto_magic] |
|
330 | 330 | |
|
331 | 331 | def lsmagic(self): |
|
332 | 332 | """Return a dict of currently available magic functions. |
|
333 | 333 | |
|
334 | 334 | The return dict has the keys 'line' and 'cell', corresponding to the |
|
335 | 335 | two types of magics we support. Each value is a list of names. |
|
336 | 336 | """ |
|
337 | 337 | return self.magics |
|
338 | 338 | |
|
339 | 339 | def lsmagic_docs(self, brief=False, missing=''): |
|
340 | 340 | """Return dict of documentation of magic functions. |
|
341 | 341 | |
|
342 | 342 | The return dict has the keys 'line' and 'cell', corresponding to the |
|
343 | 343 | two types of magics we support. Each value is a dict keyed by magic |
|
344 | 344 | name whose value is the function docstring. If a docstring is |
|
345 | 345 | unavailable, the value of `missing` is used instead. |
|
346 | 346 | |
|
347 | 347 | If brief is True, only the first line of each docstring will be returned. |
|
348 | 348 | """ |
|
349 | 349 | docs = {} |
|
350 | 350 | for m_type in self.magics: |
|
351 | 351 | m_docs = {} |
|
352 | 352 | for m_name, m_func in iteritems(self.magics[m_type]): |
|
353 | 353 | if m_func.__doc__: |
|
354 | 354 | if brief: |
|
355 | 355 | m_docs[m_name] = m_func.__doc__.split('\n', 1)[0] |
|
356 | 356 | else: |
|
357 | 357 | m_docs[m_name] = m_func.__doc__.rstrip() |
|
358 | 358 | else: |
|
359 | 359 | m_docs[m_name] = missing |
|
360 | 360 | docs[m_type] = m_docs |
|
361 | 361 | return docs |
|
362 | 362 | |
|
363 | 363 | def register(self, *magic_objects): |
|
364 | 364 | """Register one or more instances of Magics. |
|
365 | 365 | |
|
366 | 366 | Take one or more classes or instances of classes that subclass the main |
|
367 | 367 | `core.Magic` class, and register them with IPython to use the magic |
|
368 | 368 | functions they provide. The registration process will then ensure that |
|
369 | 369 | any methods that have decorated to provide line and/or cell magics will |
|
370 | 370 | be recognized with the `%x`/`%%x` syntax as a line/cell magic |
|
371 | 371 | respectively. |
|
372 | 372 | |
|
373 | 373 | If classes are given, they will be instantiated with the default |
|
374 | 374 | constructor. If your classes need a custom constructor, you should |
|
375 | 375 | instanitate them first and pass the instance. |
|
376 | 376 | |
|
377 | 377 | The provided arguments can be an arbitrary mix of classes and instances. |
|
378 | 378 | |
|
379 | 379 | Parameters |
|
380 | 380 | ---------- |
|
381 | 381 | magic_objects : one or more classes or instances |
|
382 | 382 | """ |
|
383 | 383 | # Start by validating them to ensure they have all had their magic |
|
384 | 384 | # methods registered at the instance level |
|
385 | 385 | for m in magic_objects: |
|
386 | 386 | if not m.registered: |
|
387 | 387 | raise ValueError("Class of magics %r was constructed without " |
|
388 | 388 | "the @register_magics class decorator") |
|
389 | 389 | if type(m) in (type, MetaHasTraits): |
|
390 | 390 | # If we're given an uninstantiated class |
|
391 | 391 | m = m(shell=self.shell) |
|
392 | 392 | |
|
393 | 393 | # Now that we have an instance, we can register it and update the |
|
394 | 394 | # table of callables |
|
395 | 395 | self.registry[m.__class__.__name__] = m |
|
396 | 396 | for mtype in magic_kinds: |
|
397 | 397 | self.magics[mtype].update(m.magics[mtype]) |
|
398 | 398 | |
|
399 | 399 | def register_function(self, func, magic_kind='line', magic_name=None): |
|
400 | 400 | """Expose a standalone function as magic function for IPython. |
|
401 | 401 | |
|
402 | 402 | This will create an IPython magic (line, cell or both) from a |
|
403 | 403 | standalone function. The functions should have the following |
|
404 | 404 | signatures: |
|
405 | 405 | |
|
406 | 406 | * For line magics: `def f(line)` |
|
407 | 407 | * For cell magics: `def f(line, cell)` |
|
408 | 408 | * For a function that does both: `def f(line, cell=None)` |
|
409 | 409 | |
|
410 | 410 | In the latter case, the function will be called with `cell==None` when |
|
411 | 411 | invoked as `%f`, and with cell as a string when invoked as `%%f`. |
|
412 | 412 | |
|
413 | 413 | Parameters |
|
414 | 414 | ---------- |
|
415 | 415 | func : callable |
|
416 | 416 | Function to be registered as a magic. |
|
417 | 417 | |
|
418 | 418 | magic_kind : str |
|
419 | 419 | Kind of magic, one of 'line', 'cell' or 'line_cell' |
|
420 | 420 | |
|
421 | 421 | magic_name : optional str |
|
422 | 422 | If given, the name the magic will have in the IPython namespace. By |
|
423 | 423 | default, the name of the function itself is used. |
|
424 | 424 | """ |
|
425 | 425 | |
|
426 | 426 | # Create the new method in the user_magics and register it in the |
|
427 | 427 | # global table |
|
428 | 428 | validate_type(magic_kind) |
|
429 | 429 | magic_name = func.__name__ if magic_name is None else magic_name |
|
430 | 430 | setattr(self.user_magics, magic_name, func) |
|
431 | 431 | record_magic(self.magics, magic_kind, magic_name, func) |
|
432 | 432 | |
|
433 | 433 | def define_magic(self, name, func): |
|
434 | 434 | """[Deprecated] Expose own function as magic function for IPython. |
|
435 | 435 | |
|
436 | 436 | Example:: |
|
437 | 437 | |
|
438 | 438 | def foo_impl(self, parameter_s=''): |
|
439 | 439 | 'My very own magic!. (Use docstrings, IPython reads them).' |
|
440 | 440 | print 'Magic function. Passed parameter is between < >:' |
|
441 | 441 | print '<%s>' % parameter_s |
|
442 | 442 | print 'The self object is:', self |
|
443 | 443 | |
|
444 | 444 | ip.define_magic('foo',foo_impl) |
|
445 | 445 | """ |
|
446 | 446 | meth = types.MethodType(func, self.user_magics) |
|
447 | 447 | setattr(self.user_magics, name, meth) |
|
448 | 448 | record_magic(self.magics, 'line', name, meth) |
|
449 | 449 | |
|
450 | 450 | def register_alias(self, alias_name, magic_name, magic_kind='line'): |
|
451 | 451 | """Register an alias to a magic function. |
|
452 | 452 | |
|
453 | 453 | The alias is an instance of :class:`MagicAlias`, which holds the |
|
454 | 454 | name and kind of the magic it should call. Binding is done at |
|
455 | 455 | call time, so if the underlying magic function is changed the alias |
|
456 | 456 | will call the new function. |
|
457 | 457 | |
|
458 | 458 | Parameters |
|
459 | 459 | ---------- |
|
460 | 460 | alias_name : str |
|
461 | 461 | The name of the magic to be registered. |
|
462 | 462 | |
|
463 | 463 | magic_name : str |
|
464 | 464 | The name of an existing magic. |
|
465 | 465 | |
|
466 | 466 | magic_kind : str |
|
467 | 467 | Kind of magic, one of 'line' or 'cell' |
|
468 | 468 | """ |
|
469 | 469 | |
|
470 | 470 | # `validate_type` is too permissive, as it allows 'line_cell' |
|
471 | 471 | # which we do not handle. |
|
472 | 472 | if magic_kind not in magic_kinds: |
|
473 | 473 | raise ValueError('magic_kind must be one of %s, %s given' % |
|
474 | 474 | magic_kinds, magic_kind) |
|
475 | 475 | |
|
476 | 476 | alias = MagicAlias(self.shell, magic_name, magic_kind) |
|
477 | 477 | setattr(self.user_magics, alias_name, alias) |
|
478 | 478 | record_magic(self.magics, magic_kind, alias_name, alias) |
|
479 | 479 | |
|
480 | 480 | # Key base class that provides the central functionality for magics. |
|
481 | 481 | |
|
482 | 482 | |
|
483 | 483 | class Magics(Configurable): |
|
484 | 484 | """Base class for implementing magic functions. |
|
485 | 485 | |
|
486 | 486 | Shell functions which can be reached as %function_name. All magic |
|
487 | 487 | functions should accept a string, which they can parse for their own |
|
488 | 488 | needs. This can make some functions easier to type, eg `%cd ../` |
|
489 | 489 | vs. `%cd("../")` |
|
490 | 490 | |
|
491 | 491 | Classes providing magic functions need to subclass this class, and they |
|
492 | 492 | MUST: |
|
493 | 493 | |
|
494 | 494 | - Use the method decorators `@line_magic` and `@cell_magic` to decorate |
|
495 | 495 | individual methods as magic functions, AND |
|
496 | 496 | |
|
497 | 497 | - Use the class decorator `@magics_class` to ensure that the magic |
|
498 | 498 | methods are properly registered at the instance level upon instance |
|
499 | 499 | initialization. |
|
500 | 500 | |
|
501 | 501 | See :mod:`magic_functions` for examples of actual implementation classes. |
|
502 | 502 | """ |
|
503 | 503 | # Dict holding all command-line options for each magic. |
|
504 | 504 | options_table = None |
|
505 | 505 | # Dict for the mapping of magic names to methods, set by class decorator |
|
506 | 506 | magics = None |
|
507 | 507 | # Flag to check that the class decorator was properly applied |
|
508 | 508 | registered = False |
|
509 | 509 | # Instance of IPython shell |
|
510 | 510 | shell = None |
|
511 | 511 | |
|
512 | 512 | def __init__(self, shell=None, **kwargs): |
|
513 | 513 | if not(self.__class__.registered): |
|
514 | 514 | raise ValueError('Magics subclass without registration - ' |
|
515 | 515 | 'did you forget to apply @magics_class?') |
|
516 | 516 | if shell is not None: |
|
517 | 517 | if hasattr(shell, 'configurables'): |
|
518 | 518 | shell.configurables.append(self) |
|
519 | 519 | if hasattr(shell, 'config'): |
|
520 | 520 | kwargs.setdefault('parent', shell) |
|
521 | 521 | kwargs['shell'] = shell |
|
522 | 522 | |
|
523 | 523 | self.shell = shell |
|
524 | 524 | self.options_table = {} |
|
525 | 525 | # The method decorators are run when the instance doesn't exist yet, so |
|
526 | 526 | # they can only record the names of the methods they are supposed to |
|
527 | 527 | # grab. Only now, that the instance exists, can we create the proper |
|
528 | 528 | # mapping to bound methods. So we read the info off the original names |
|
529 | 529 | # table and replace each method name by the actual bound method. |
|
530 | 530 | # But we mustn't clobber the *class* mapping, in case of multiple instances. |
|
531 | 531 | class_magics = self.magics |
|
532 | 532 | self.magics = {} |
|
533 | 533 | for mtype in magic_kinds: |
|
534 | 534 | tab = self.magics[mtype] = {} |
|
535 | 535 | cls_tab = class_magics[mtype] |
|
536 | 536 | for magic_name, meth_name in iteritems(cls_tab): |
|
537 | 537 | if isinstance(meth_name, string_types): |
|
538 | 538 | # it's a method name, grab it |
|
539 | 539 | tab[magic_name] = getattr(self, meth_name) |
|
540 | 540 | else: |
|
541 | 541 | # it's the real thing |
|
542 | 542 | tab[magic_name] = meth_name |
|
543 | 543 | # Configurable **needs** to be initiated at the end or the config |
|
544 | 544 | # magics get screwed up. |
|
545 | 545 | super(Magics, self).__init__(**kwargs) |
|
546 | 546 | |
|
547 | 547 | def arg_err(self,func): |
|
548 | 548 | """Print docstring if incorrect arguments were passed""" |
|
549 | 549 | print('Error in arguments:') |
|
550 | 550 | print(oinspect.getdoc(func)) |
|
551 | 551 | |
|
552 | 552 | def format_latex(self, strng): |
|
553 | 553 | """Format a string for latex inclusion.""" |
|
554 | 554 | |
|
555 | 555 | # Characters that need to be escaped for latex: |
|
556 | 556 | escape_re = re.compile(r'(%|_|\$|#|&)',re.MULTILINE) |
|
557 | 557 | # Magic command names as headers: |
|
558 | 558 | cmd_name_re = re.compile(r'^(%s.*?):' % ESC_MAGIC, |
|
559 | 559 | re.MULTILINE) |
|
560 | 560 | # Magic commands |
|
561 | 561 | cmd_re = re.compile(r'(?P<cmd>%s.+?\b)(?!\}\}:)' % ESC_MAGIC, |
|
562 | 562 | re.MULTILINE) |
|
563 | 563 | # Paragraph continue |
|
564 | 564 | par_re = re.compile(r'\\$',re.MULTILINE) |
|
565 | 565 | |
|
566 | 566 | # The "\n" symbol |
|
567 | 567 | newline_re = re.compile(r'\\n') |
|
568 | 568 | |
|
569 | 569 | # Now build the string for output: |
|
570 | 570 | #strng = cmd_name_re.sub(r'\n\\texttt{\\textsl{\\large \1}}:',strng) |
|
571 | 571 | strng = cmd_name_re.sub(r'\n\\bigskip\n\\texttt{\\textbf{ \1}}:', |
|
572 | 572 | strng) |
|
573 | 573 | strng = cmd_re.sub(r'\\texttt{\g<cmd>}',strng) |
|
574 | 574 | strng = par_re.sub(r'\\\\',strng) |
|
575 | 575 | strng = escape_re.sub(r'\\\1',strng) |
|
576 | 576 | strng = newline_re.sub(r'\\textbackslash{}n',strng) |
|
577 | 577 | return strng |
|
578 | 578 | |
|
579 | 579 | def parse_options(self, arg_str, opt_str, *long_opts, **kw): |
|
580 | 580 | """Parse options passed to an argument string. |
|
581 | 581 | |
|
582 | 582 | The interface is similar to that of :func:`getopt.getopt`, but it |
|
583 | 583 | returns a :class:`~IPython.utils.struct.Struct` with the options as keys |
|
584 | 584 | and the stripped argument string still as a string. |
|
585 | 585 | |
|
586 | 586 | arg_str is quoted as a true sys.argv vector by using shlex.split. |
|
587 | 587 | This allows us to easily expand variables, glob files, quote |
|
588 | 588 | arguments, etc. |
|
589 | 589 | |
|
590 | 590 | Parameters |
|
591 | 591 | ---------- |
|
592 | 592 | |
|
593 | 593 | arg_str : str |
|
594 | 594 | The arguments to parse. |
|
595 | 595 | |
|
596 | 596 | opt_str : str |
|
597 | 597 | The options specification. |
|
598 | 598 | |
|
599 | 599 | mode : str, default 'string' |
|
600 | 600 | If given as 'list', the argument string is returned as a list (split |
|
601 | 601 | on whitespace) instead of a string. |
|
602 | 602 | |
|
603 | 603 | list_all : bool, default False |
|
604 | 604 | Put all option values in lists. Normally only options |
|
605 | 605 | appearing more than once are put in a list. |
|
606 | 606 | |
|
607 | 607 | posix : bool, default True |
|
608 | 608 | Whether to split the input line in POSIX mode or not, as per the |
|
609 | 609 | conventions outlined in the :mod:`shlex` module from the standard |
|
610 | 610 | library. |
|
611 | 611 | """ |
|
612 | 612 | |
|
613 | 613 | # inject default options at the beginning of the input line |
|
614 | 614 | caller = sys._getframe(1).f_code.co_name |
|
615 | 615 | arg_str = '%s %s' % (self.options_table.get(caller,''),arg_str) |
|
616 | 616 | |
|
617 | 617 | mode = kw.get('mode','string') |
|
618 | 618 | if mode not in ['string','list']: |
|
619 | 619 | raise ValueError('incorrect mode given: %s' % mode) |
|
620 | 620 | # Get options |
|
621 | 621 | list_all = kw.get('list_all',0) |
|
622 | 622 | posix = kw.get('posix', os.name == 'posix') |
|
623 | 623 | strict = kw.get('strict', True) |
|
624 | 624 | |
|
625 | 625 | # Check if we have more than one argument to warrant extra processing: |
|
626 | 626 | odict = {} # Dictionary with options |
|
627 | 627 | args = arg_str.split() |
|
628 | 628 | if len(args) >= 1: |
|
629 | 629 | # If the list of inputs only has 0 or 1 thing in it, there's no |
|
630 | 630 | # need to look for options |
|
631 | 631 | argv = arg_split(arg_str, posix, strict) |
|
632 | 632 | # Do regular option processing |
|
633 | 633 | try: |
|
634 | 634 | opts,args = getopt(argv, opt_str, long_opts) |
|
635 | 635 | except GetoptError as e: |
|
636 | 636 | raise UsageError('%s ( allowed: "%s" %s)' % (e.msg,opt_str, |
|
637 | 637 | " ".join(long_opts))) |
|
638 | 638 | for o,a in opts: |
|
639 | 639 | if o.startswith('--'): |
|
640 | 640 | o = o[2:] |
|
641 | 641 | else: |
|
642 | 642 | o = o[1:] |
|
643 | 643 | try: |
|
644 | 644 | odict[o].append(a) |
|
645 | 645 | except AttributeError: |
|
646 | 646 | odict[o] = [odict[o],a] |
|
647 | 647 | except KeyError: |
|
648 | 648 | if list_all: |
|
649 | 649 | odict[o] = [a] |
|
650 | 650 | else: |
|
651 | 651 | odict[o] = a |
|
652 | 652 | |
|
653 | 653 | # Prepare opts,args for return |
|
654 | 654 | opts = Struct(odict) |
|
655 | 655 | if mode == 'string': |
|
656 | 656 | args = ' '.join(args) |
|
657 | 657 | |
|
658 | 658 | return opts,args |
|
659 | 659 | |
|
660 | 660 | def default_option(self, fn, optstr): |
|
661 | 661 | """Make an entry in the options_table for fn, with value optstr""" |
|
662 | 662 | |
|
663 | 663 | if fn not in self.lsmagic(): |
|
664 | 664 | error("%s is not a magic function" % fn) |
|
665 | 665 | self.options_table[fn] = optstr |
|
666 | 666 | |
|
667 | 667 | |
|
668 | 668 | class MagicAlias(object): |
|
669 | 669 | """An alias to another magic function. |
|
670 | 670 | |
|
671 | 671 | An alias is determined by its magic name and magic kind. Lookup |
|
672 | 672 | is done at call time, so if the underlying magic changes the alias |
|
673 | 673 | will call the new function. |
|
674 | 674 | |
|
675 | 675 | Use the :meth:`MagicsManager.register_alias` method or the |
|
676 | 676 | `%alias_magic` magic function to create and register a new alias. |
|
677 | 677 | """ |
|
678 | 678 | def __init__(self, shell, magic_name, magic_kind): |
|
679 | 679 | self.shell = shell |
|
680 | 680 | self.magic_name = magic_name |
|
681 | 681 | self.magic_kind = magic_kind |
|
682 | 682 | |
|
683 | 683 | self.pretty_target = '%s%s' % (magic_escapes[self.magic_kind], self.magic_name) |
|
684 | 684 | self.__doc__ = "Alias for `%s`." % self.pretty_target |
|
685 | 685 | |
|
686 | 686 | self._in_call = False |
|
687 | 687 | |
|
688 | 688 | def __call__(self, *args, **kwargs): |
|
689 | 689 | """Call the magic alias.""" |
|
690 | 690 | fn = self.shell.find_magic(self.magic_name, self.magic_kind) |
|
691 | 691 | if fn is None: |
|
692 | 692 | raise UsageError("Magic `%s` not found." % self.pretty_target) |
|
693 | 693 | |
|
694 | 694 | # Protect against infinite recursion. |
|
695 | 695 | if self._in_call: |
|
696 | 696 | raise UsageError("Infinite recursion detected; " |
|
697 | 697 | "magic aliases cannot call themselves.") |
|
698 | 698 | self._in_call = True |
|
699 | 699 | try: |
|
700 | 700 | return fn(*args, **kwargs) |
|
701 | 701 | finally: |
|
702 | 702 | self._in_call = False |
@@ -1,382 +1,382 b'' | |||
|
1 | 1 | """Tests for the object inspection functionality. |
|
2 | 2 | """ |
|
3 | 3 | #----------------------------------------------------------------------------- |
|
4 | 4 | # Copyright (C) 2010-2011 The IPython Development Team. |
|
5 | 5 | # |
|
6 | 6 | # Distributed under the terms of the BSD License. |
|
7 | 7 | # |
|
8 | 8 | # The full license is in the file COPYING.txt, distributed with this software. |
|
9 | 9 | #----------------------------------------------------------------------------- |
|
10 | 10 | |
|
11 | 11 | #----------------------------------------------------------------------------- |
|
12 | 12 | # Imports |
|
13 | 13 | #----------------------------------------------------------------------------- |
|
14 | 14 | from __future__ import print_function |
|
15 | 15 | |
|
16 | 16 | # Stdlib imports |
|
17 | 17 | import os |
|
18 | 18 | import re |
|
19 | 19 | |
|
20 | 20 | # Third-party imports |
|
21 | 21 | import nose.tools as nt |
|
22 | 22 | |
|
23 | 23 | # Our own imports |
|
24 | 24 | from .. import oinspect |
|
25 | 25 | from IPython.core.magic import (Magics, magics_class, line_magic, |
|
26 | 26 | cell_magic, line_cell_magic, |
|
27 | 27 | register_line_magic, register_cell_magic, |
|
28 | 28 | register_line_cell_magic) |
|
29 |
from |
|
|
29 | from decorator import decorator | |
|
30 | 30 | from IPython.testing.decorators import skipif |
|
31 | 31 | from IPython.testing.tools import AssertPrints |
|
32 | 32 | from IPython.utils.path import compress_user |
|
33 | 33 | from IPython.utils import py3compat |
|
34 | 34 | |
|
35 | 35 | |
|
36 | 36 | #----------------------------------------------------------------------------- |
|
37 | 37 | # Globals and constants |
|
38 | 38 | #----------------------------------------------------------------------------- |
|
39 | 39 | |
|
40 | 40 | inspector = oinspect.Inspector() |
|
41 | 41 | ip = get_ipython() |
|
42 | 42 | |
|
43 | 43 | #----------------------------------------------------------------------------- |
|
44 | 44 | # Local utilities |
|
45 | 45 | #----------------------------------------------------------------------------- |
|
46 | 46 | |
|
47 | 47 | # WARNING: since this test checks the line number where a function is |
|
48 | 48 | # defined, if any code is inserted above, the following line will need to be |
|
49 | 49 | # updated. Do NOT insert any whitespace between the next line and the function |
|
50 | 50 | # definition below. |
|
51 | 51 | THIS_LINE_NUMBER = 51 # Put here the actual number of this line |
|
52 | 52 | def test_find_source_lines(): |
|
53 | 53 | nt.assert_equal(oinspect.find_source_lines(test_find_source_lines), |
|
54 | 54 | THIS_LINE_NUMBER+1) |
|
55 | 55 | |
|
56 | 56 | |
|
57 | 57 | # A couple of utilities to ensure these tests work the same from a source or a |
|
58 | 58 | # binary install |
|
59 | 59 | def pyfile(fname): |
|
60 | 60 | return os.path.normcase(re.sub('.py[co]$', '.py', fname)) |
|
61 | 61 | |
|
62 | 62 | |
|
63 | 63 | def match_pyfiles(f1, f2): |
|
64 | 64 | nt.assert_equal(pyfile(f1), pyfile(f2)) |
|
65 | 65 | |
|
66 | 66 | |
|
67 | 67 | def test_find_file(): |
|
68 | 68 | match_pyfiles(oinspect.find_file(test_find_file), os.path.abspath(__file__)) |
|
69 | 69 | |
|
70 | 70 | |
|
71 | 71 | def test_find_file_decorated1(): |
|
72 | 72 | |
|
73 | 73 | @decorator |
|
74 | 74 | def noop1(f): |
|
75 | 75 | def wrapper(): |
|
76 | 76 | return f(*a, **kw) |
|
77 | 77 | return wrapper |
|
78 | 78 | |
|
79 | 79 | @noop1 |
|
80 | 80 | def f(x): |
|
81 | 81 | "My docstring" |
|
82 | 82 | |
|
83 | 83 | match_pyfiles(oinspect.find_file(f), os.path.abspath(__file__)) |
|
84 | 84 | nt.assert_equal(f.__doc__, "My docstring") |
|
85 | 85 | |
|
86 | 86 | |
|
87 | 87 | def test_find_file_decorated2(): |
|
88 | 88 | |
|
89 | 89 | @decorator |
|
90 | 90 | def noop2(f, *a, **kw): |
|
91 | 91 | return f(*a, **kw) |
|
92 | 92 | |
|
93 | 93 | @noop2 |
|
94 | 94 | def f(x): |
|
95 | 95 | "My docstring 2" |
|
96 | 96 | |
|
97 | 97 | match_pyfiles(oinspect.find_file(f), os.path.abspath(__file__)) |
|
98 | 98 | nt.assert_equal(f.__doc__, "My docstring 2") |
|
99 | 99 | |
|
100 | 100 | |
|
101 | 101 | def test_find_file_magic(): |
|
102 | 102 | run = ip.find_line_magic('run') |
|
103 | 103 | nt.assert_not_equal(oinspect.find_file(run), None) |
|
104 | 104 | |
|
105 | 105 | |
|
106 | 106 | # A few generic objects we can then inspect in the tests below |
|
107 | 107 | |
|
108 | 108 | class Call(object): |
|
109 | 109 | """This is the class docstring.""" |
|
110 | 110 | |
|
111 | 111 | def __init__(self, x, y=1): |
|
112 | 112 | """This is the constructor docstring.""" |
|
113 | 113 | |
|
114 | 114 | def __call__(self, *a, **kw): |
|
115 | 115 | """This is the call docstring.""" |
|
116 | 116 | |
|
117 | 117 | def method(self, x, z=2): |
|
118 | 118 | """Some method's docstring""" |
|
119 | 119 | |
|
120 | 120 | class SimpleClass(object): |
|
121 | 121 | def method(self, x, z=2): |
|
122 | 122 | """Some method's docstring""" |
|
123 | 123 | |
|
124 | 124 | |
|
125 | 125 | class OldStyle: |
|
126 | 126 | """An old-style class for testing.""" |
|
127 | 127 | pass |
|
128 | 128 | |
|
129 | 129 | |
|
130 | 130 | def f(x, y=2, *a, **kw): |
|
131 | 131 | """A simple function.""" |
|
132 | 132 | |
|
133 | 133 | |
|
134 | 134 | def g(y, z=3, *a, **kw): |
|
135 | 135 | pass # no docstring |
|
136 | 136 | |
|
137 | 137 | |
|
138 | 138 | @register_line_magic |
|
139 | 139 | def lmagic(line): |
|
140 | 140 | "A line magic" |
|
141 | 141 | |
|
142 | 142 | |
|
143 | 143 | @register_cell_magic |
|
144 | 144 | def cmagic(line, cell): |
|
145 | 145 | "A cell magic" |
|
146 | 146 | |
|
147 | 147 | |
|
148 | 148 | @register_line_cell_magic |
|
149 | 149 | def lcmagic(line, cell=None): |
|
150 | 150 | "A line/cell magic" |
|
151 | 151 | |
|
152 | 152 | |
|
153 | 153 | @magics_class |
|
154 | 154 | class SimpleMagics(Magics): |
|
155 | 155 | @line_magic |
|
156 | 156 | def Clmagic(self, cline): |
|
157 | 157 | "A class-based line magic" |
|
158 | 158 | |
|
159 | 159 | @cell_magic |
|
160 | 160 | def Ccmagic(self, cline, ccell): |
|
161 | 161 | "A class-based cell magic" |
|
162 | 162 | |
|
163 | 163 | @line_cell_magic |
|
164 | 164 | def Clcmagic(self, cline, ccell=None): |
|
165 | 165 | "A class-based line/cell magic" |
|
166 | 166 | |
|
167 | 167 | |
|
168 | 168 | class Awkward(object): |
|
169 | 169 | def __getattr__(self, name): |
|
170 | 170 | raise Exception(name) |
|
171 | 171 | |
|
172 | 172 | |
|
173 | 173 | def check_calltip(obj, name, call, docstring): |
|
174 | 174 | """Generic check pattern all calltip tests will use""" |
|
175 | 175 | info = inspector.info(obj, name) |
|
176 | 176 | call_line, ds = oinspect.call_tip(info) |
|
177 | 177 | nt.assert_equal(call_line, call) |
|
178 | 178 | nt.assert_equal(ds, docstring) |
|
179 | 179 | |
|
180 | 180 | #----------------------------------------------------------------------------- |
|
181 | 181 | # Tests |
|
182 | 182 | #----------------------------------------------------------------------------- |
|
183 | 183 | |
|
184 | 184 | def test_calltip_class(): |
|
185 | 185 | check_calltip(Call, 'Call', 'Call(x, y=1)', Call.__init__.__doc__) |
|
186 | 186 | |
|
187 | 187 | |
|
188 | 188 | def test_calltip_instance(): |
|
189 | 189 | c = Call(1) |
|
190 | 190 | check_calltip(c, 'c', 'c(*a, **kw)', c.__call__.__doc__) |
|
191 | 191 | |
|
192 | 192 | |
|
193 | 193 | def test_calltip_method(): |
|
194 | 194 | c = Call(1) |
|
195 | 195 | check_calltip(c.method, 'c.method', 'c.method(x, z=2)', c.method.__doc__) |
|
196 | 196 | |
|
197 | 197 | |
|
198 | 198 | def test_calltip_function(): |
|
199 | 199 | check_calltip(f, 'f', 'f(x, y=2, *a, **kw)', f.__doc__) |
|
200 | 200 | |
|
201 | 201 | |
|
202 | 202 | def test_calltip_function2(): |
|
203 | 203 | check_calltip(g, 'g', 'g(y, z=3, *a, **kw)', '<no docstring>') |
|
204 | 204 | |
|
205 | 205 | |
|
206 | 206 | def test_calltip_builtin(): |
|
207 | 207 | check_calltip(sum, 'sum', None, sum.__doc__) |
|
208 | 208 | |
|
209 | 209 | |
|
210 | 210 | def test_calltip_line_magic(): |
|
211 | 211 | check_calltip(lmagic, 'lmagic', 'lmagic(line)', "A line magic") |
|
212 | 212 | |
|
213 | 213 | |
|
214 | 214 | def test_calltip_cell_magic(): |
|
215 | 215 | check_calltip(cmagic, 'cmagic', 'cmagic(line, cell)', "A cell magic") |
|
216 | 216 | |
|
217 | 217 | |
|
218 | 218 | def test_calltip_line_cell_magic(): |
|
219 | 219 | check_calltip(lcmagic, 'lcmagic', 'lcmagic(line, cell=None)', |
|
220 | 220 | "A line/cell magic") |
|
221 | 221 | |
|
222 | 222 | |
|
223 | 223 | def test_class_magics(): |
|
224 | 224 | cm = SimpleMagics(ip) |
|
225 | 225 | ip.register_magics(cm) |
|
226 | 226 | check_calltip(cm.Clmagic, 'Clmagic', 'Clmagic(cline)', |
|
227 | 227 | "A class-based line magic") |
|
228 | 228 | check_calltip(cm.Ccmagic, 'Ccmagic', 'Ccmagic(cline, ccell)', |
|
229 | 229 | "A class-based cell magic") |
|
230 | 230 | check_calltip(cm.Clcmagic, 'Clcmagic', 'Clcmagic(cline, ccell=None)', |
|
231 | 231 | "A class-based line/cell magic") |
|
232 | 232 | |
|
233 | 233 | |
|
234 | 234 | def test_info(): |
|
235 | 235 | "Check that Inspector.info fills out various fields as expected." |
|
236 | 236 | i = inspector.info(Call, oname='Call') |
|
237 | 237 | nt.assert_equal(i['type_name'], 'type') |
|
238 | 238 | expted_class = str(type(type)) # <class 'type'> (Python 3) or <type 'type'> |
|
239 | 239 | nt.assert_equal(i['base_class'], expted_class) |
|
240 | 240 | nt.assert_equal(i['string_form'], "<class 'IPython.core.tests.test_oinspect.Call'>") |
|
241 | 241 | fname = __file__ |
|
242 | 242 | if fname.endswith(".pyc"): |
|
243 | 243 | fname = fname[:-1] |
|
244 | 244 | # case-insensitive comparison needed on some filesystems |
|
245 | 245 | # e.g. Windows: |
|
246 | 246 | nt.assert_equal(i['file'].lower(), compress_user(fname.lower())) |
|
247 | 247 | nt.assert_equal(i['definition'], None) |
|
248 | 248 | nt.assert_equal(i['docstring'], Call.__doc__) |
|
249 | 249 | nt.assert_equal(i['source'], None) |
|
250 | 250 | nt.assert_true(i['isclass']) |
|
251 | 251 | nt.assert_equal(i['init_definition'], "Call(self, x, y=1)\n") |
|
252 | 252 | nt.assert_equal(i['init_docstring'], Call.__init__.__doc__) |
|
253 | 253 | |
|
254 | 254 | i = inspector.info(Call, detail_level=1) |
|
255 | 255 | nt.assert_not_equal(i['source'], None) |
|
256 | 256 | nt.assert_equal(i['docstring'], None) |
|
257 | 257 | |
|
258 | 258 | c = Call(1) |
|
259 | 259 | c.__doc__ = "Modified instance docstring" |
|
260 | 260 | i = inspector.info(c) |
|
261 | 261 | nt.assert_equal(i['type_name'], 'Call') |
|
262 | 262 | nt.assert_equal(i['docstring'], "Modified instance docstring") |
|
263 | 263 | nt.assert_equal(i['class_docstring'], Call.__doc__) |
|
264 | 264 | nt.assert_equal(i['init_docstring'], Call.__init__.__doc__) |
|
265 | 265 | nt.assert_equal(i['call_docstring'], Call.__call__.__doc__) |
|
266 | 266 | |
|
267 | 267 | # Test old-style classes, which for example may not have an __init__ method. |
|
268 | 268 | if not py3compat.PY3: |
|
269 | 269 | i = inspector.info(OldStyle) |
|
270 | 270 | nt.assert_equal(i['type_name'], 'classobj') |
|
271 | 271 | |
|
272 | 272 | i = inspector.info(OldStyle()) |
|
273 | 273 | nt.assert_equal(i['type_name'], 'instance') |
|
274 | 274 | nt.assert_equal(i['docstring'], OldStyle.__doc__) |
|
275 | 275 | |
|
276 | 276 | def test_info_awkward(): |
|
277 | 277 | # Just test that this doesn't throw an error. |
|
278 | 278 | i = inspector.info(Awkward()) |
|
279 | 279 | |
|
280 | 280 | def test_calldef_none(): |
|
281 | 281 | # We should ignore __call__ for all of these. |
|
282 | 282 | for obj in [f, SimpleClass().method, any, str.upper]: |
|
283 | 283 | print(obj) |
|
284 | 284 | i = inspector.info(obj) |
|
285 | 285 | nt.assert_is(i['call_def'], None) |
|
286 | 286 | |
|
287 | 287 | if py3compat.PY3: |
|
288 | 288 | exec("def f_kwarg(pos, *, kwonly): pass") |
|
289 | 289 | |
|
290 | 290 | @skipif(not py3compat.PY3) |
|
291 | 291 | def test_definition_kwonlyargs(): |
|
292 | 292 | i = inspector.info(f_kwarg, oname='f_kwarg') # analysis:ignore |
|
293 | 293 | nt.assert_equal(i['definition'], "f_kwarg(pos, *, kwonly)\n") |
|
294 | 294 | |
|
295 | 295 | def test_getdoc(): |
|
296 | 296 | class A(object): |
|
297 | 297 | """standard docstring""" |
|
298 | 298 | pass |
|
299 | 299 | |
|
300 | 300 | class B(object): |
|
301 | 301 | """standard docstring""" |
|
302 | 302 | def getdoc(self): |
|
303 | 303 | return "custom docstring" |
|
304 | 304 | |
|
305 | 305 | class C(object): |
|
306 | 306 | """standard docstring""" |
|
307 | 307 | def getdoc(self): |
|
308 | 308 | return None |
|
309 | 309 | |
|
310 | 310 | a = A() |
|
311 | 311 | b = B() |
|
312 | 312 | c = C() |
|
313 | 313 | |
|
314 | 314 | nt.assert_equal(oinspect.getdoc(a), "standard docstring") |
|
315 | 315 | nt.assert_equal(oinspect.getdoc(b), "custom docstring") |
|
316 | 316 | nt.assert_equal(oinspect.getdoc(c), "standard docstring") |
|
317 | 317 | |
|
318 | 318 | |
|
319 | 319 | def test_empty_property_has_no_source(): |
|
320 | 320 | i = inspector.info(property(), detail_level=1) |
|
321 | 321 | nt.assert_is(i['source'], None) |
|
322 | 322 | |
|
323 | 323 | |
|
324 | 324 | def test_property_sources(): |
|
325 | 325 | import zlib |
|
326 | 326 | |
|
327 | 327 | class A(object): |
|
328 | 328 | @property |
|
329 | 329 | def foo(self): |
|
330 | 330 | return 'bar' |
|
331 | 331 | |
|
332 | 332 | foo = foo.setter(lambda self, v: setattr(self, 'bar', v)) |
|
333 | 333 | |
|
334 | 334 | id = property(id) |
|
335 | 335 | compress = property(zlib.compress) |
|
336 | 336 | |
|
337 | 337 | i = inspector.info(A.foo, detail_level=1) |
|
338 | 338 | nt.assert_in('def foo(self):', i['source']) |
|
339 | 339 | nt.assert_in('lambda self, v:', i['source']) |
|
340 | 340 | |
|
341 | 341 | i = inspector.info(A.id, detail_level=1) |
|
342 | 342 | nt.assert_in('fget = <function id>', i['source']) |
|
343 | 343 | |
|
344 | 344 | i = inspector.info(A.compress, detail_level=1) |
|
345 | 345 | nt.assert_in('fget = <function zlib.compress>', i['source']) |
|
346 | 346 | |
|
347 | 347 | |
|
348 | 348 | def test_property_docstring_is_in_info_for_detail_level_0(): |
|
349 | 349 | class A(object): |
|
350 | 350 | @property |
|
351 | 351 | def foobar(): |
|
352 | 352 | """This is `foobar` property.""" |
|
353 | 353 | pass |
|
354 | 354 | |
|
355 | 355 | ip.user_ns['a_obj'] = A() |
|
356 | 356 | nt.assert_equals( |
|
357 | 357 | 'This is `foobar` property.', |
|
358 | 358 | ip.object_inspect('a_obj.foobar', detail_level=0)['docstring']) |
|
359 | 359 | |
|
360 | 360 | ip.user_ns['a_cls'] = A |
|
361 | 361 | nt.assert_equals( |
|
362 | 362 | 'This is `foobar` property.', |
|
363 | 363 | ip.object_inspect('a_cls.foobar', detail_level=0)['docstring']) |
|
364 | 364 | |
|
365 | 365 | |
|
366 | 366 | def test_pdef(): |
|
367 | 367 | # See gh-1914 |
|
368 | 368 | def foo(): pass |
|
369 | 369 | inspector.pdef(foo, 'foo') |
|
370 | 370 | |
|
371 | 371 | def test_pinfo_nonascii(): |
|
372 | 372 | # See gh-1177 |
|
373 | 373 | from . import nonascii2 |
|
374 | 374 | ip.user_ns['nonascii2'] = nonascii2 |
|
375 | 375 | ip._inspect('pinfo', 'nonascii2', detail_level=1) |
|
376 | 376 | |
|
377 | 377 | def test_pinfo_magic(): |
|
378 | 378 | with AssertPrints('Docstring:'): |
|
379 | 379 | ip._inspect('pinfo', 'lsmagic', detail_level=0) |
|
380 | 380 | |
|
381 | 381 | with AssertPrints('Source:'): |
|
382 | 382 | ip._inspect('pinfo', 'lsmagic', detail_level=1) |
@@ -1,703 +1,703 b'' | |||
|
1 | 1 | """AsyncResult objects for the client""" |
|
2 | 2 | |
|
3 | 3 | # Copyright (c) IPython Development Team. |
|
4 | 4 | # Distributed under the terms of the Modified BSD License. |
|
5 | 5 | |
|
6 | 6 | from __future__ import print_function |
|
7 | 7 | |
|
8 | 8 | import sys |
|
9 | 9 | import time |
|
10 | 10 | from datetime import datetime |
|
11 | 11 | |
|
12 | 12 | from zmq import MessageTracker |
|
13 | 13 | |
|
14 | 14 | from IPython.core.display import clear_output, display, display_pretty |
|
15 |
from |
|
|
15 | from decorator import decorator | |
|
16 | 16 | from IPython.parallel import error |
|
17 | 17 | from IPython.utils.py3compat import string_types |
|
18 | 18 | |
|
19 | 19 | |
|
20 | 20 | def _raw_text(s): |
|
21 | 21 | display_pretty(s, raw=True) |
|
22 | 22 | |
|
23 | 23 | |
|
24 | 24 | # global empty tracker that's always done: |
|
25 | 25 | finished_tracker = MessageTracker() |
|
26 | 26 | |
|
27 | 27 | @decorator |
|
28 | 28 | def check_ready(f, self, *args, **kwargs): |
|
29 | 29 | """Call spin() to sync state prior to calling the method.""" |
|
30 | 30 | self.wait(0) |
|
31 | 31 | if not self._ready: |
|
32 | 32 | raise error.TimeoutError("result not ready") |
|
33 | 33 | return f(self, *args, **kwargs) |
|
34 | 34 | |
|
35 | 35 | class AsyncResult(object): |
|
36 | 36 | """Class for representing results of non-blocking calls. |
|
37 | 37 | |
|
38 | 38 | Provides the same interface as :py:class:`multiprocessing.pool.AsyncResult`. |
|
39 | 39 | """ |
|
40 | 40 | |
|
41 | 41 | msg_ids = None |
|
42 | 42 | _targets = None |
|
43 | 43 | _tracker = None |
|
44 | 44 | _single_result = False |
|
45 | 45 | owner = False, |
|
46 | 46 | |
|
47 | 47 | def __init__(self, client, msg_ids, fname='unknown', targets=None, tracker=None, |
|
48 | 48 | owner=False, |
|
49 | 49 | ): |
|
50 | 50 | if isinstance(msg_ids, string_types): |
|
51 | 51 | # always a list |
|
52 | 52 | msg_ids = [msg_ids] |
|
53 | 53 | self._single_result = True |
|
54 | 54 | else: |
|
55 | 55 | self._single_result = False |
|
56 | 56 | if tracker is None: |
|
57 | 57 | # default to always done |
|
58 | 58 | tracker = finished_tracker |
|
59 | 59 | self._client = client |
|
60 | 60 | self.msg_ids = msg_ids |
|
61 | 61 | self._fname=fname |
|
62 | 62 | self._targets = targets |
|
63 | 63 | self._tracker = tracker |
|
64 | 64 | self.owner = owner |
|
65 | 65 | |
|
66 | 66 | self._ready = False |
|
67 | 67 | self._outputs_ready = False |
|
68 | 68 | self._success = None |
|
69 | 69 | self._metadata = [self._client.metadata[id] for id in self.msg_ids] |
|
70 | 70 | |
|
71 | 71 | def __repr__(self): |
|
72 | 72 | if self._ready: |
|
73 | 73 | return "<%s: finished>"%(self.__class__.__name__) |
|
74 | 74 | else: |
|
75 | 75 | return "<%s: %s>"%(self.__class__.__name__,self._fname) |
|
76 | 76 | |
|
77 | 77 | |
|
78 | 78 | def _reconstruct_result(self, res): |
|
79 | 79 | """Reconstruct our result from actual result list (always a list) |
|
80 | 80 | |
|
81 | 81 | Override me in subclasses for turning a list of results |
|
82 | 82 | into the expected form. |
|
83 | 83 | """ |
|
84 | 84 | if self._single_result: |
|
85 | 85 | return res[0] |
|
86 | 86 | else: |
|
87 | 87 | return res |
|
88 | 88 | |
|
89 | 89 | def get(self, timeout=-1): |
|
90 | 90 | """Return the result when it arrives. |
|
91 | 91 | |
|
92 | 92 | If `timeout` is not ``None`` and the result does not arrive within |
|
93 | 93 | `timeout` seconds then ``TimeoutError`` is raised. If the |
|
94 | 94 | remote call raised an exception then that exception will be reraised |
|
95 | 95 | by get() inside a `RemoteError`. |
|
96 | 96 | """ |
|
97 | 97 | if not self.ready(): |
|
98 | 98 | self.wait(timeout) |
|
99 | 99 | |
|
100 | 100 | if self._ready: |
|
101 | 101 | if self._success: |
|
102 | 102 | return self._result |
|
103 | 103 | else: |
|
104 | 104 | raise self._exception |
|
105 | 105 | else: |
|
106 | 106 | raise error.TimeoutError("Result not ready.") |
|
107 | 107 | |
|
108 | 108 | def _check_ready(self): |
|
109 | 109 | if not self.ready(): |
|
110 | 110 | raise error.TimeoutError("Result not ready.") |
|
111 | 111 | |
|
112 | 112 | def ready(self): |
|
113 | 113 | """Return whether the call has completed.""" |
|
114 | 114 | if not self._ready: |
|
115 | 115 | self.wait(0) |
|
116 | 116 | elif not self._outputs_ready: |
|
117 | 117 | self._wait_for_outputs(0) |
|
118 | 118 | |
|
119 | 119 | return self._ready |
|
120 | 120 | |
|
121 | 121 | def wait(self, timeout=-1): |
|
122 | 122 | """Wait until the result is available or until `timeout` seconds pass. |
|
123 | 123 | |
|
124 | 124 | This method always returns None. |
|
125 | 125 | """ |
|
126 | 126 | if self._ready: |
|
127 | 127 | self._wait_for_outputs(timeout) |
|
128 | 128 | return |
|
129 | 129 | self._ready = self._client.wait(self.msg_ids, timeout) |
|
130 | 130 | if self._ready: |
|
131 | 131 | try: |
|
132 | 132 | results = list(map(self._client.results.get, self.msg_ids)) |
|
133 | 133 | self._result = results |
|
134 | 134 | if self._single_result: |
|
135 | 135 | r = results[0] |
|
136 | 136 | if isinstance(r, Exception): |
|
137 | 137 | raise r |
|
138 | 138 | else: |
|
139 | 139 | results = error.collect_exceptions(results, self._fname) |
|
140 | 140 | self._result = self._reconstruct_result(results) |
|
141 | 141 | except Exception as e: |
|
142 | 142 | self._exception = e |
|
143 | 143 | self._success = False |
|
144 | 144 | else: |
|
145 | 145 | self._success = True |
|
146 | 146 | finally: |
|
147 | 147 | if timeout is None or timeout < 0: |
|
148 | 148 | # cutoff infinite wait at 10s |
|
149 | 149 | timeout = 10 |
|
150 | 150 | self._wait_for_outputs(timeout) |
|
151 | 151 | |
|
152 | 152 | if self.owner: |
|
153 | 153 | |
|
154 | 154 | self._metadata = [self._client.metadata.pop(mid) for mid in self.msg_ids] |
|
155 | 155 | [self._client.results.pop(mid) for mid in self.msg_ids] |
|
156 | 156 | |
|
157 | 157 | |
|
158 | 158 | |
|
159 | 159 | def successful(self): |
|
160 | 160 | """Return whether the call completed without raising an exception. |
|
161 | 161 | |
|
162 | 162 | Will raise ``AssertionError`` if the result is not ready. |
|
163 | 163 | """ |
|
164 | 164 | assert self.ready() |
|
165 | 165 | return self._success |
|
166 | 166 | |
|
167 | 167 | #---------------------------------------------------------------- |
|
168 | 168 | # Extra methods not in mp.pool.AsyncResult |
|
169 | 169 | #---------------------------------------------------------------- |
|
170 | 170 | |
|
171 | 171 | def get_dict(self, timeout=-1): |
|
172 | 172 | """Get the results as a dict, keyed by engine_id. |
|
173 | 173 | |
|
174 | 174 | timeout behavior is described in `get()`. |
|
175 | 175 | """ |
|
176 | 176 | |
|
177 | 177 | results = self.get(timeout) |
|
178 | 178 | if self._single_result: |
|
179 | 179 | results = [results] |
|
180 | 180 | engine_ids = [ md['engine_id'] for md in self._metadata ] |
|
181 | 181 | |
|
182 | 182 | |
|
183 | 183 | rdict = {} |
|
184 | 184 | for engine_id, result in zip(engine_ids, results): |
|
185 | 185 | if engine_id in rdict: |
|
186 | 186 | raise ValueError("Cannot build dict, %i jobs ran on engine #%i" % ( |
|
187 | 187 | engine_ids.count(engine_id), engine_id) |
|
188 | 188 | ) |
|
189 | 189 | else: |
|
190 | 190 | rdict[engine_id] = result |
|
191 | 191 | |
|
192 | 192 | return rdict |
|
193 | 193 | |
|
194 | 194 | @property |
|
195 | 195 | def result(self): |
|
196 | 196 | """result property wrapper for `get(timeout=-1)`.""" |
|
197 | 197 | return self.get() |
|
198 | 198 | |
|
199 | 199 | # abbreviated alias: |
|
200 | 200 | r = result |
|
201 | 201 | |
|
202 | 202 | @property |
|
203 | 203 | def metadata(self): |
|
204 | 204 | """property for accessing execution metadata.""" |
|
205 | 205 | if self._single_result: |
|
206 | 206 | return self._metadata[0] |
|
207 | 207 | else: |
|
208 | 208 | return self._metadata |
|
209 | 209 | |
|
210 | 210 | @property |
|
211 | 211 | def result_dict(self): |
|
212 | 212 | """result property as a dict.""" |
|
213 | 213 | return self.get_dict() |
|
214 | 214 | |
|
215 | 215 | def __dict__(self): |
|
216 | 216 | return self.get_dict(0) |
|
217 | 217 | |
|
218 | 218 | def abort(self): |
|
219 | 219 | """abort my tasks.""" |
|
220 | 220 | assert not self.ready(), "Can't abort, I am already done!" |
|
221 | 221 | return self._client.abort(self.msg_ids, targets=self._targets, block=True) |
|
222 | 222 | |
|
223 | 223 | @property |
|
224 | 224 | def sent(self): |
|
225 | 225 | """check whether my messages have been sent.""" |
|
226 | 226 | return self._tracker.done |
|
227 | 227 | |
|
228 | 228 | def wait_for_send(self, timeout=-1): |
|
229 | 229 | """wait for pyzmq send to complete. |
|
230 | 230 | |
|
231 | 231 | This is necessary when sending arrays that you intend to edit in-place. |
|
232 | 232 | `timeout` is in seconds, and will raise TimeoutError if it is reached |
|
233 | 233 | before the send completes. |
|
234 | 234 | """ |
|
235 | 235 | return self._tracker.wait(timeout) |
|
236 | 236 | |
|
237 | 237 | #------------------------------------- |
|
238 | 238 | # dict-access |
|
239 | 239 | #------------------------------------- |
|
240 | 240 | |
|
241 | 241 | def __getitem__(self, key): |
|
242 | 242 | """getitem returns result value(s) if keyed by int/slice, or metadata if key is str. |
|
243 | 243 | """ |
|
244 | 244 | if isinstance(key, int): |
|
245 | 245 | self._check_ready() |
|
246 | 246 | return error.collect_exceptions([self._result[key]], self._fname)[0] |
|
247 | 247 | elif isinstance(key, slice): |
|
248 | 248 | self._check_ready() |
|
249 | 249 | return error.collect_exceptions(self._result[key], self._fname) |
|
250 | 250 | elif isinstance(key, string_types): |
|
251 | 251 | # metadata proxy *does not* require that results are done |
|
252 | 252 | self.wait(0) |
|
253 | 253 | values = [ md[key] for md in self._metadata ] |
|
254 | 254 | if self._single_result: |
|
255 | 255 | return values[0] |
|
256 | 256 | else: |
|
257 | 257 | return values |
|
258 | 258 | else: |
|
259 | 259 | raise TypeError("Invalid key type %r, must be 'int','slice', or 'str'"%type(key)) |
|
260 | 260 | |
|
261 | 261 | def __getattr__(self, key): |
|
262 | 262 | """getattr maps to getitem for convenient attr access to metadata.""" |
|
263 | 263 | try: |
|
264 | 264 | return self.__getitem__(key) |
|
265 | 265 | except (error.TimeoutError, KeyError): |
|
266 | 266 | raise AttributeError("%r object has no attribute %r"%( |
|
267 | 267 | self.__class__.__name__, key)) |
|
268 | 268 | |
|
269 | 269 | # asynchronous iterator: |
|
270 | 270 | def __iter__(self): |
|
271 | 271 | if self._single_result: |
|
272 | 272 | raise TypeError("AsyncResults with a single result are not iterable.") |
|
273 | 273 | try: |
|
274 | 274 | rlist = self.get(0) |
|
275 | 275 | except error.TimeoutError: |
|
276 | 276 | # wait for each result individually |
|
277 | 277 | for msg_id in self.msg_ids: |
|
278 | 278 | ar = AsyncResult(self._client, msg_id, self._fname) |
|
279 | 279 | yield ar.get() |
|
280 | 280 | else: |
|
281 | 281 | # already done |
|
282 | 282 | for r in rlist: |
|
283 | 283 | yield r |
|
284 | 284 | |
|
285 | 285 | def __len__(self): |
|
286 | 286 | return len(self.msg_ids) |
|
287 | 287 | |
|
288 | 288 | #------------------------------------- |
|
289 | 289 | # Sugar methods and attributes |
|
290 | 290 | #------------------------------------- |
|
291 | 291 | |
|
292 | 292 | def timedelta(self, start, end, start_key=min, end_key=max): |
|
293 | 293 | """compute the difference between two sets of timestamps |
|
294 | 294 | |
|
295 | 295 | The default behavior is to use the earliest of the first |
|
296 | 296 | and the latest of the second list, but this can be changed |
|
297 | 297 | by passing a different |
|
298 | 298 | |
|
299 | 299 | Parameters |
|
300 | 300 | ---------- |
|
301 | 301 | |
|
302 | 302 | start : one or more datetime objects (e.g. ar.submitted) |
|
303 | 303 | end : one or more datetime objects (e.g. ar.received) |
|
304 | 304 | start_key : callable |
|
305 | 305 | Function to call on `start` to extract the relevant |
|
306 | 306 | entry [defalt: min] |
|
307 | 307 | end_key : callable |
|
308 | 308 | Function to call on `end` to extract the relevant |
|
309 | 309 | entry [default: max] |
|
310 | 310 | |
|
311 | 311 | Returns |
|
312 | 312 | ------- |
|
313 | 313 | |
|
314 | 314 | dt : float |
|
315 | 315 | The time elapsed (in seconds) between the two selected timestamps. |
|
316 | 316 | """ |
|
317 | 317 | if not isinstance(start, datetime): |
|
318 | 318 | # handle single_result AsyncResults, where ar.stamp is single object, |
|
319 | 319 | # not a list |
|
320 | 320 | start = start_key(start) |
|
321 | 321 | if not isinstance(end, datetime): |
|
322 | 322 | # handle single_result AsyncResults, where ar.stamp is single object, |
|
323 | 323 | # not a list |
|
324 | 324 | end = end_key(end) |
|
325 | 325 | return (end - start).total_seconds() |
|
326 | 326 | |
|
327 | 327 | @property |
|
328 | 328 | def progress(self): |
|
329 | 329 | """the number of tasks which have been completed at this point. |
|
330 | 330 | |
|
331 | 331 | Fractional progress would be given by 1.0 * ar.progress / len(ar) |
|
332 | 332 | """ |
|
333 | 333 | self.wait(0) |
|
334 | 334 | return len(self) - len(set(self.msg_ids).intersection(self._client.outstanding)) |
|
335 | 335 | |
|
336 | 336 | @property |
|
337 | 337 | def elapsed(self): |
|
338 | 338 | """elapsed time since initial submission""" |
|
339 | 339 | if self.ready(): |
|
340 | 340 | return self.wall_time |
|
341 | 341 | |
|
342 | 342 | now = submitted = datetime.now() |
|
343 | 343 | for msg_id in self.msg_ids: |
|
344 | 344 | if msg_id in self._client.metadata: |
|
345 | 345 | stamp = self._client.metadata[msg_id]['submitted'] |
|
346 | 346 | if stamp and stamp < submitted: |
|
347 | 347 | submitted = stamp |
|
348 | 348 | return (now-submitted).total_seconds() |
|
349 | 349 | |
|
350 | 350 | @property |
|
351 | 351 | @check_ready |
|
352 | 352 | def serial_time(self): |
|
353 | 353 | """serial computation time of a parallel calculation |
|
354 | 354 | |
|
355 | 355 | Computed as the sum of (completed-started) of each task |
|
356 | 356 | """ |
|
357 | 357 | t = 0 |
|
358 | 358 | for md in self._metadata: |
|
359 | 359 | t += (md['completed'] - md['started']).total_seconds() |
|
360 | 360 | return t |
|
361 | 361 | |
|
362 | 362 | @property |
|
363 | 363 | @check_ready |
|
364 | 364 | def wall_time(self): |
|
365 | 365 | """actual computation time of a parallel calculation |
|
366 | 366 | |
|
367 | 367 | Computed as the time between the latest `received` stamp |
|
368 | 368 | and the earliest `submitted`. |
|
369 | 369 | |
|
370 | 370 | Only reliable if Client was spinning/waiting when the task finished, because |
|
371 | 371 | the `received` timestamp is created when a result is pulled off of the zmq queue, |
|
372 | 372 | which happens as a result of `client.spin()`. |
|
373 | 373 | |
|
374 | 374 | For similar comparison of other timestamp pairs, check out AsyncResult.timedelta. |
|
375 | 375 | |
|
376 | 376 | """ |
|
377 | 377 | return self.timedelta(self.submitted, self.received) |
|
378 | 378 | |
|
379 | 379 | def wait_interactive(self, interval=1., timeout=-1): |
|
380 | 380 | """interactive wait, printing progress at regular intervals""" |
|
381 | 381 | if timeout is None: |
|
382 | 382 | timeout = -1 |
|
383 | 383 | N = len(self) |
|
384 | 384 | tic = time.time() |
|
385 | 385 | while not self.ready() and (timeout < 0 or time.time() - tic <= timeout): |
|
386 | 386 | self.wait(interval) |
|
387 | 387 | clear_output(wait=True) |
|
388 | 388 | print("%4i/%i tasks finished after %4i s" % (self.progress, N, self.elapsed), end="") |
|
389 | 389 | sys.stdout.flush() |
|
390 | 390 | print() |
|
391 | 391 | print("done") |
|
392 | 392 | |
|
393 | 393 | def _republish_displaypub(self, content, eid): |
|
394 | 394 | """republish individual displaypub content dicts""" |
|
395 | 395 | try: |
|
396 | 396 | ip = get_ipython() |
|
397 | 397 | except NameError: |
|
398 | 398 | # displaypub is meaningless outside IPython |
|
399 | 399 | return |
|
400 | 400 | md = content['metadata'] or {} |
|
401 | 401 | md['engine'] = eid |
|
402 | 402 | ip.display_pub.publish(data=content['data'], metadata=md) |
|
403 | 403 | |
|
404 | 404 | def _display_stream(self, text, prefix='', file=None): |
|
405 | 405 | if not text: |
|
406 | 406 | # nothing to display |
|
407 | 407 | return |
|
408 | 408 | if file is None: |
|
409 | 409 | file = sys.stdout |
|
410 | 410 | end = '' if text.endswith('\n') else '\n' |
|
411 | 411 | |
|
412 | 412 | multiline = text.count('\n') > int(text.endswith('\n')) |
|
413 | 413 | if prefix and multiline and not text.startswith('\n'): |
|
414 | 414 | prefix = prefix + '\n' |
|
415 | 415 | print("%s%s" % (prefix, text), file=file, end=end) |
|
416 | 416 | |
|
417 | 417 | |
|
418 | 418 | def _display_single_result(self): |
|
419 | 419 | self._display_stream(self.stdout) |
|
420 | 420 | self._display_stream(self.stderr, file=sys.stderr) |
|
421 | 421 | |
|
422 | 422 | try: |
|
423 | 423 | get_ipython() |
|
424 | 424 | except NameError: |
|
425 | 425 | # displaypub is meaningless outside IPython |
|
426 | 426 | return |
|
427 | 427 | |
|
428 | 428 | for output in self.outputs: |
|
429 | 429 | self._republish_displaypub(output, self.engine_id) |
|
430 | 430 | |
|
431 | 431 | if self.execute_result is not None: |
|
432 | 432 | display(self.get()) |
|
433 | 433 | |
|
434 | 434 | def _wait_for_outputs(self, timeout=-1): |
|
435 | 435 | """wait for the 'status=idle' message that indicates we have all outputs |
|
436 | 436 | """ |
|
437 | 437 | if self._outputs_ready or not self._success: |
|
438 | 438 | # don't wait on errors |
|
439 | 439 | return |
|
440 | 440 | |
|
441 | 441 | # cast None to -1 for infinite timeout |
|
442 | 442 | if timeout is None: |
|
443 | 443 | timeout = -1 |
|
444 | 444 | |
|
445 | 445 | tic = time.time() |
|
446 | 446 | while True: |
|
447 | 447 | self._client._flush_iopub(self._client._iopub_socket) |
|
448 | 448 | self._outputs_ready = all(md['outputs_ready'] |
|
449 | 449 | for md in self._metadata) |
|
450 | 450 | if self._outputs_ready or \ |
|
451 | 451 | (timeout >= 0 and time.time() > tic + timeout): |
|
452 | 452 | break |
|
453 | 453 | time.sleep(0.01) |
|
454 | 454 | |
|
455 | 455 | @check_ready |
|
456 | 456 | def display_outputs(self, groupby="type"): |
|
457 | 457 | """republish the outputs of the computation |
|
458 | 458 | |
|
459 | 459 | Parameters |
|
460 | 460 | ---------- |
|
461 | 461 | |
|
462 | 462 | groupby : str [default: type] |
|
463 | 463 | if 'type': |
|
464 | 464 | Group outputs by type (show all stdout, then all stderr, etc.): |
|
465 | 465 | |
|
466 | 466 | [stdout:1] foo |
|
467 | 467 | [stdout:2] foo |
|
468 | 468 | [stderr:1] bar |
|
469 | 469 | [stderr:2] bar |
|
470 | 470 | if 'engine': |
|
471 | 471 | Display outputs for each engine before moving on to the next: |
|
472 | 472 | |
|
473 | 473 | [stdout:1] foo |
|
474 | 474 | [stderr:1] bar |
|
475 | 475 | [stdout:2] foo |
|
476 | 476 | [stderr:2] bar |
|
477 | 477 | |
|
478 | 478 | if 'order': |
|
479 | 479 | Like 'type', but further collate individual displaypub |
|
480 | 480 | outputs. This is meant for cases of each command producing |
|
481 | 481 | several plots, and you would like to see all of the first |
|
482 | 482 | plots together, then all of the second plots, and so on. |
|
483 | 483 | """ |
|
484 | 484 | if self._single_result: |
|
485 | 485 | self._display_single_result() |
|
486 | 486 | return |
|
487 | 487 | |
|
488 | 488 | stdouts = self.stdout |
|
489 | 489 | stderrs = self.stderr |
|
490 | 490 | execute_results = self.execute_result |
|
491 | 491 | output_lists = self.outputs |
|
492 | 492 | results = self.get() |
|
493 | 493 | |
|
494 | 494 | targets = self.engine_id |
|
495 | 495 | |
|
496 | 496 | if groupby == "engine": |
|
497 | 497 | for eid,stdout,stderr,outputs,r,execute_result in zip( |
|
498 | 498 | targets, stdouts, stderrs, output_lists, results, execute_results |
|
499 | 499 | ): |
|
500 | 500 | self._display_stream(stdout, '[stdout:%i] ' % eid) |
|
501 | 501 | self._display_stream(stderr, '[stderr:%i] ' % eid, file=sys.stderr) |
|
502 | 502 | |
|
503 | 503 | try: |
|
504 | 504 | get_ipython() |
|
505 | 505 | except NameError: |
|
506 | 506 | # displaypub is meaningless outside IPython |
|
507 | 507 | return |
|
508 | 508 | |
|
509 | 509 | if outputs or execute_result is not None: |
|
510 | 510 | _raw_text('[output:%i]' % eid) |
|
511 | 511 | |
|
512 | 512 | for output in outputs: |
|
513 | 513 | self._republish_displaypub(output, eid) |
|
514 | 514 | |
|
515 | 515 | if execute_result is not None: |
|
516 | 516 | display(r) |
|
517 | 517 | |
|
518 | 518 | elif groupby in ('type', 'order'): |
|
519 | 519 | # republish stdout: |
|
520 | 520 | for eid,stdout in zip(targets, stdouts): |
|
521 | 521 | self._display_stream(stdout, '[stdout:%i] ' % eid) |
|
522 | 522 | |
|
523 | 523 | # republish stderr: |
|
524 | 524 | for eid,stderr in zip(targets, stderrs): |
|
525 | 525 | self._display_stream(stderr, '[stderr:%i] ' % eid, file=sys.stderr) |
|
526 | 526 | |
|
527 | 527 | try: |
|
528 | 528 | get_ipython() |
|
529 | 529 | except NameError: |
|
530 | 530 | # displaypub is meaningless outside IPython |
|
531 | 531 | return |
|
532 | 532 | |
|
533 | 533 | if groupby == 'order': |
|
534 | 534 | output_dict = dict((eid, outputs) for eid,outputs in zip(targets, output_lists)) |
|
535 | 535 | N = max(len(outputs) for outputs in output_lists) |
|
536 | 536 | for i in range(N): |
|
537 | 537 | for eid in targets: |
|
538 | 538 | outputs = output_dict[eid] |
|
539 | 539 | if len(outputs) >= N: |
|
540 | 540 | _raw_text('[output:%i]' % eid) |
|
541 | 541 | self._republish_displaypub(outputs[i], eid) |
|
542 | 542 | else: |
|
543 | 543 | # republish displaypub output |
|
544 | 544 | for eid,outputs in zip(targets, output_lists): |
|
545 | 545 | if outputs: |
|
546 | 546 | _raw_text('[output:%i]' % eid) |
|
547 | 547 | for output in outputs: |
|
548 | 548 | self._republish_displaypub(output, eid) |
|
549 | 549 | |
|
550 | 550 | # finally, add execute_result: |
|
551 | 551 | for eid,r,execute_result in zip(targets, results, execute_results): |
|
552 | 552 | if execute_result is not None: |
|
553 | 553 | display(r) |
|
554 | 554 | |
|
555 | 555 | else: |
|
556 | 556 | raise ValueError("groupby must be one of 'type', 'engine', 'collate', not %r" % groupby) |
|
557 | 557 | |
|
558 | 558 | |
|
559 | 559 | |
|
560 | 560 | |
|
561 | 561 | class AsyncMapResult(AsyncResult): |
|
562 | 562 | """Class for representing results of non-blocking gathers. |
|
563 | 563 | |
|
564 | 564 | This will properly reconstruct the gather. |
|
565 | 565 | |
|
566 | 566 | This class is iterable at any time, and will wait on results as they come. |
|
567 | 567 | |
|
568 | 568 | If ordered=False, then the first results to arrive will come first, otherwise |
|
569 | 569 | results will be yielded in the order they were submitted. |
|
570 | 570 | |
|
571 | 571 | """ |
|
572 | 572 | |
|
573 | 573 | def __init__(self, client, msg_ids, mapObject, fname='', ordered=True): |
|
574 | 574 | AsyncResult.__init__(self, client, msg_ids, fname=fname) |
|
575 | 575 | self._mapObject = mapObject |
|
576 | 576 | self._single_result = False |
|
577 | 577 | self.ordered = ordered |
|
578 | 578 | |
|
579 | 579 | def _reconstruct_result(self, res): |
|
580 | 580 | """Perform the gather on the actual results.""" |
|
581 | 581 | return self._mapObject.joinPartitions(res) |
|
582 | 582 | |
|
583 | 583 | # asynchronous iterator: |
|
584 | 584 | def __iter__(self): |
|
585 | 585 | it = self._ordered_iter if self.ordered else self._unordered_iter |
|
586 | 586 | for r in it(): |
|
587 | 587 | yield r |
|
588 | 588 | |
|
589 | 589 | # asynchronous ordered iterator: |
|
590 | 590 | def _ordered_iter(self): |
|
591 | 591 | """iterator for results *as they arrive*, preserving submission order.""" |
|
592 | 592 | try: |
|
593 | 593 | rlist = self.get(0) |
|
594 | 594 | except error.TimeoutError: |
|
595 | 595 | # wait for each result individually |
|
596 | 596 | for msg_id in self.msg_ids: |
|
597 | 597 | ar = AsyncResult(self._client, msg_id, self._fname) |
|
598 | 598 | rlist = ar.get() |
|
599 | 599 | try: |
|
600 | 600 | for r in rlist: |
|
601 | 601 | yield r |
|
602 | 602 | except TypeError: |
|
603 | 603 | # flattened, not a list |
|
604 | 604 | # this could get broken by flattened data that returns iterables |
|
605 | 605 | # but most calls to map do not expose the `flatten` argument |
|
606 | 606 | yield rlist |
|
607 | 607 | else: |
|
608 | 608 | # already done |
|
609 | 609 | for r in rlist: |
|
610 | 610 | yield r |
|
611 | 611 | |
|
612 | 612 | # asynchronous unordered iterator: |
|
613 | 613 | def _unordered_iter(self): |
|
614 | 614 | """iterator for results *as they arrive*, on FCFS basis, ignoring submission order.""" |
|
615 | 615 | try: |
|
616 | 616 | rlist = self.get(0) |
|
617 | 617 | except error.TimeoutError: |
|
618 | 618 | pending = set(self.msg_ids) |
|
619 | 619 | while pending: |
|
620 | 620 | try: |
|
621 | 621 | self._client.wait(pending, 1e-3) |
|
622 | 622 | except error.TimeoutError: |
|
623 | 623 | # ignore timeout error, because that only means |
|
624 | 624 | # *some* jobs are outstanding |
|
625 | 625 | pass |
|
626 | 626 | # update ready set with those no longer outstanding: |
|
627 | 627 | ready = pending.difference(self._client.outstanding) |
|
628 | 628 | # update pending to exclude those that are finished |
|
629 | 629 | pending = pending.difference(ready) |
|
630 | 630 | while ready: |
|
631 | 631 | msg_id = ready.pop() |
|
632 | 632 | ar = AsyncResult(self._client, msg_id, self._fname) |
|
633 | 633 | rlist = ar.get() |
|
634 | 634 | try: |
|
635 | 635 | for r in rlist: |
|
636 | 636 | yield r |
|
637 | 637 | except TypeError: |
|
638 | 638 | # flattened, not a list |
|
639 | 639 | # this could get broken by flattened data that returns iterables |
|
640 | 640 | # but most calls to map do not expose the `flatten` argument |
|
641 | 641 | yield rlist |
|
642 | 642 | else: |
|
643 | 643 | # already done |
|
644 | 644 | for r in rlist: |
|
645 | 645 | yield r |
|
646 | 646 | |
|
647 | 647 | |
|
648 | 648 | class AsyncHubResult(AsyncResult): |
|
649 | 649 | """Class to wrap pending results that must be requested from the Hub. |
|
650 | 650 | |
|
651 | 651 | Note that waiting/polling on these objects requires polling the Hubover the network, |
|
652 | 652 | so use `AsyncHubResult.wait()` sparingly. |
|
653 | 653 | """ |
|
654 | 654 | |
|
655 | 655 | def _wait_for_outputs(self, timeout=-1): |
|
656 | 656 | """no-op, because HubResults are never incomplete""" |
|
657 | 657 | self._outputs_ready = True |
|
658 | 658 | |
|
659 | 659 | def wait(self, timeout=-1): |
|
660 | 660 | """wait for result to complete.""" |
|
661 | 661 | start = time.time() |
|
662 | 662 | if self._ready: |
|
663 | 663 | return |
|
664 | 664 | local_ids = [m for m in self.msg_ids if m in self._client.outstanding] |
|
665 | 665 | local_ready = self._client.wait(local_ids, timeout) |
|
666 | 666 | if local_ready: |
|
667 | 667 | remote_ids = [m for m in self.msg_ids if m not in self._client.results] |
|
668 | 668 | if not remote_ids: |
|
669 | 669 | self._ready = True |
|
670 | 670 | else: |
|
671 | 671 | rdict = self._client.result_status(remote_ids, status_only=False) |
|
672 | 672 | pending = rdict['pending'] |
|
673 | 673 | while pending and (timeout < 0 or time.time() < start+timeout): |
|
674 | 674 | rdict = self._client.result_status(remote_ids, status_only=False) |
|
675 | 675 | pending = rdict['pending'] |
|
676 | 676 | if pending: |
|
677 | 677 | time.sleep(0.1) |
|
678 | 678 | if not pending: |
|
679 | 679 | self._ready = True |
|
680 | 680 | if self._ready: |
|
681 | 681 | try: |
|
682 | 682 | results = list(map(self._client.results.get, self.msg_ids)) |
|
683 | 683 | self._result = results |
|
684 | 684 | if self._single_result: |
|
685 | 685 | r = results[0] |
|
686 | 686 | if isinstance(r, Exception): |
|
687 | 687 | raise r |
|
688 | 688 | else: |
|
689 | 689 | results = error.collect_exceptions(results, self._fname) |
|
690 | 690 | self._result = self._reconstruct_result(results) |
|
691 | 691 | except Exception as e: |
|
692 | 692 | self._exception = e |
|
693 | 693 | self._success = False |
|
694 | 694 | else: |
|
695 | 695 | self._success = True |
|
696 | 696 | finally: |
|
697 | 697 | self._metadata = [self._client.metadata[mid] for mid in self.msg_ids] |
|
698 | 698 | if self.owner: |
|
699 | 699 | [self._client.metadata.pop(mid) for mid in self.msg_ids] |
|
700 | 700 | [self._client.results.pop(mid) for mid in self.msg_ids] |
|
701 | 701 | |
|
702 | 702 | |
|
703 | 703 | __all__ = ['AsyncResult', 'AsyncMapResult', 'AsyncHubResult'] |
@@ -1,1893 +1,1893 b'' | |||
|
1 | 1 | """A semi-synchronous Client for IPython parallel""" |
|
2 | 2 | |
|
3 | 3 | # Copyright (c) IPython Development Team. |
|
4 | 4 | # Distributed under the terms of the Modified BSD License. |
|
5 | 5 | |
|
6 | 6 | from __future__ import print_function |
|
7 | 7 | |
|
8 | 8 | import os |
|
9 | 9 | import json |
|
10 | 10 | import sys |
|
11 | 11 | from threading import Thread, Event |
|
12 | 12 | import time |
|
13 | 13 | import warnings |
|
14 | 14 | from datetime import datetime |
|
15 | 15 | from getpass import getpass |
|
16 | 16 | from pprint import pprint |
|
17 | 17 | |
|
18 | 18 | pjoin = os.path.join |
|
19 | 19 | |
|
20 | 20 | import zmq |
|
21 | 21 | |
|
22 | 22 | from IPython.config.configurable import MultipleInstanceError |
|
23 | 23 | from IPython.core.application import BaseIPythonApplication |
|
24 | 24 | from IPython.core.profiledir import ProfileDir, ProfileDirError |
|
25 | 25 | |
|
26 | 26 | from IPython.utils.capture import RichOutput |
|
27 | 27 | from IPython.utils.coloransi import TermColors |
|
28 | 28 | from IPython.utils.jsonutil import rekey, extract_dates, parse_date |
|
29 | 29 | from IPython.utils.localinterfaces import localhost, is_local_ip |
|
30 | 30 | from IPython.utils.path import get_ipython_dir, compress_user |
|
31 | 31 | from IPython.utils.py3compat import cast_bytes, string_types, xrange, iteritems |
|
32 | 32 | from IPython.utils.traitlets import (HasTraits, Integer, Instance, Unicode, |
|
33 | 33 | Dict, List, Bool, Set, Any) |
|
34 |
from |
|
|
34 | from decorator import decorator | |
|
35 | 35 | |
|
36 | 36 | from IPython.parallel import Reference |
|
37 | 37 | from IPython.parallel import error |
|
38 | 38 | from IPython.parallel import util |
|
39 | 39 | |
|
40 | 40 | from IPython.kernel.zmq.session import Session, Message |
|
41 | 41 | from IPython.kernel.zmq import serialize |
|
42 | 42 | |
|
43 | 43 | from .asyncresult import AsyncResult, AsyncHubResult |
|
44 | 44 | from .view import DirectView, LoadBalancedView |
|
45 | 45 | |
|
46 | 46 | #-------------------------------------------------------------------------- |
|
47 | 47 | # Decorators for Client methods |
|
48 | 48 | #-------------------------------------------------------------------------- |
|
49 | 49 | |
|
50 | 50 | |
|
51 | 51 | @decorator |
|
52 | 52 | def spin_first(f, self, *args, **kwargs): |
|
53 | 53 | """Call spin() to sync state prior to calling the method.""" |
|
54 | 54 | self.spin() |
|
55 | 55 | return f(self, *args, **kwargs) |
|
56 | 56 | |
|
57 | 57 | |
|
58 | 58 | #-------------------------------------------------------------------------- |
|
59 | 59 | # Classes |
|
60 | 60 | #-------------------------------------------------------------------------- |
|
61 | 61 | |
|
62 | 62 | _no_connection_file_msg = """ |
|
63 | 63 | Failed to connect because no Controller could be found. |
|
64 | 64 | Please double-check your profile and ensure that a cluster is running. |
|
65 | 65 | """ |
|
66 | 66 | |
|
67 | 67 | class ExecuteReply(RichOutput): |
|
68 | 68 | """wrapper for finished Execute results""" |
|
69 | 69 | def __init__(self, msg_id, content, metadata): |
|
70 | 70 | self.msg_id = msg_id |
|
71 | 71 | self._content = content |
|
72 | 72 | self.execution_count = content['execution_count'] |
|
73 | 73 | self.metadata = metadata |
|
74 | 74 | |
|
75 | 75 | # RichOutput overrides |
|
76 | 76 | |
|
77 | 77 | @property |
|
78 | 78 | def source(self): |
|
79 | 79 | execute_result = self.metadata['execute_result'] |
|
80 | 80 | if execute_result: |
|
81 | 81 | return execute_result.get('source', '') |
|
82 | 82 | |
|
83 | 83 | @property |
|
84 | 84 | def data(self): |
|
85 | 85 | execute_result = self.metadata['execute_result'] |
|
86 | 86 | if execute_result: |
|
87 | 87 | return execute_result.get('data', {}) |
|
88 | 88 | |
|
89 | 89 | @property |
|
90 | 90 | def _metadata(self): |
|
91 | 91 | execute_result = self.metadata['execute_result'] |
|
92 | 92 | if execute_result: |
|
93 | 93 | return execute_result.get('metadata', {}) |
|
94 | 94 | |
|
95 | 95 | def display(self): |
|
96 | 96 | from IPython.display import publish_display_data |
|
97 | 97 | publish_display_data(self.data, self.metadata) |
|
98 | 98 | |
|
99 | 99 | def _repr_mime_(self, mime): |
|
100 | 100 | if mime not in self.data: |
|
101 | 101 | return |
|
102 | 102 | data = self.data[mime] |
|
103 | 103 | if mime in self._metadata: |
|
104 | 104 | return data, self._metadata[mime] |
|
105 | 105 | else: |
|
106 | 106 | return data |
|
107 | 107 | |
|
108 | 108 | def __getitem__(self, key): |
|
109 | 109 | return self.metadata[key] |
|
110 | 110 | |
|
111 | 111 | def __getattr__(self, key): |
|
112 | 112 | if key not in self.metadata: |
|
113 | 113 | raise AttributeError(key) |
|
114 | 114 | return self.metadata[key] |
|
115 | 115 | |
|
116 | 116 | def __repr__(self): |
|
117 | 117 | execute_result = self.metadata['execute_result'] or {'data':{}} |
|
118 | 118 | text_out = execute_result['data'].get('text/plain', '') |
|
119 | 119 | if len(text_out) > 32: |
|
120 | 120 | text_out = text_out[:29] + '...' |
|
121 | 121 | |
|
122 | 122 | return "<ExecuteReply[%i]: %s>" % (self.execution_count, text_out) |
|
123 | 123 | |
|
124 | 124 | def _repr_pretty_(self, p, cycle): |
|
125 | 125 | execute_result = self.metadata['execute_result'] or {'data':{}} |
|
126 | 126 | text_out = execute_result['data'].get('text/plain', '') |
|
127 | 127 | |
|
128 | 128 | if not text_out: |
|
129 | 129 | return |
|
130 | 130 | |
|
131 | 131 | try: |
|
132 | 132 | ip = get_ipython() |
|
133 | 133 | except NameError: |
|
134 | 134 | colors = "NoColor" |
|
135 | 135 | else: |
|
136 | 136 | colors = ip.colors |
|
137 | 137 | |
|
138 | 138 | if colors == "NoColor": |
|
139 | 139 | out = normal = "" |
|
140 | 140 | else: |
|
141 | 141 | out = TermColors.Red |
|
142 | 142 | normal = TermColors.Normal |
|
143 | 143 | |
|
144 | 144 | if '\n' in text_out and not text_out.startswith('\n'): |
|
145 | 145 | # add newline for multiline reprs |
|
146 | 146 | text_out = '\n' + text_out |
|
147 | 147 | |
|
148 | 148 | p.text( |
|
149 | 149 | out + u'Out[%i:%i]: ' % ( |
|
150 | 150 | self.metadata['engine_id'], self.execution_count |
|
151 | 151 | ) + normal + text_out |
|
152 | 152 | ) |
|
153 | 153 | |
|
154 | 154 | |
|
155 | 155 | class Metadata(dict): |
|
156 | 156 | """Subclass of dict for initializing metadata values. |
|
157 | 157 | |
|
158 | 158 | Attribute access works on keys. |
|
159 | 159 | |
|
160 | 160 | These objects have a strict set of keys - errors will raise if you try |
|
161 | 161 | to add new keys. |
|
162 | 162 | """ |
|
163 | 163 | def __init__(self, *args, **kwargs): |
|
164 | 164 | dict.__init__(self) |
|
165 | 165 | md = {'msg_id' : None, |
|
166 | 166 | 'submitted' : None, |
|
167 | 167 | 'started' : None, |
|
168 | 168 | 'completed' : None, |
|
169 | 169 | 'received' : None, |
|
170 | 170 | 'engine_uuid' : None, |
|
171 | 171 | 'engine_id' : None, |
|
172 | 172 | 'follow' : None, |
|
173 | 173 | 'after' : None, |
|
174 | 174 | 'status' : None, |
|
175 | 175 | |
|
176 | 176 | 'execute_input' : None, |
|
177 | 177 | 'execute_result' : None, |
|
178 | 178 | 'error' : None, |
|
179 | 179 | 'stdout' : '', |
|
180 | 180 | 'stderr' : '', |
|
181 | 181 | 'outputs' : [], |
|
182 | 182 | 'data': {}, |
|
183 | 183 | 'outputs_ready' : False, |
|
184 | 184 | } |
|
185 | 185 | self.update(md) |
|
186 | 186 | self.update(dict(*args, **kwargs)) |
|
187 | 187 | |
|
188 | 188 | def __getattr__(self, key): |
|
189 | 189 | """getattr aliased to getitem""" |
|
190 | 190 | if key in self: |
|
191 | 191 | return self[key] |
|
192 | 192 | else: |
|
193 | 193 | raise AttributeError(key) |
|
194 | 194 | |
|
195 | 195 | def __setattr__(self, key, value): |
|
196 | 196 | """setattr aliased to setitem, with strict""" |
|
197 | 197 | if key in self: |
|
198 | 198 | self[key] = value |
|
199 | 199 | else: |
|
200 | 200 | raise AttributeError(key) |
|
201 | 201 | |
|
202 | 202 | def __setitem__(self, key, value): |
|
203 | 203 | """strict static key enforcement""" |
|
204 | 204 | if key in self: |
|
205 | 205 | dict.__setitem__(self, key, value) |
|
206 | 206 | else: |
|
207 | 207 | raise KeyError(key) |
|
208 | 208 | |
|
209 | 209 | |
|
210 | 210 | class Client(HasTraits): |
|
211 | 211 | """A semi-synchronous client to the IPython ZMQ cluster |
|
212 | 212 | |
|
213 | 213 | Parameters |
|
214 | 214 | ---------- |
|
215 | 215 | |
|
216 | 216 | url_file : str/unicode; path to ipcontroller-client.json |
|
217 | 217 | This JSON file should contain all the information needed to connect to a cluster, |
|
218 | 218 | and is likely the only argument needed. |
|
219 | 219 | Connection information for the Hub's registration. If a json connector |
|
220 | 220 | file is given, then likely no further configuration is necessary. |
|
221 | 221 | [Default: use profile] |
|
222 | 222 | profile : bytes |
|
223 | 223 | The name of the Cluster profile to be used to find connector information. |
|
224 | 224 | If run from an IPython application, the default profile will be the same |
|
225 | 225 | as the running application, otherwise it will be 'default'. |
|
226 | 226 | cluster_id : str |
|
227 | 227 | String id to added to runtime files, to prevent name collisions when using |
|
228 | 228 | multiple clusters with a single profile simultaneously. |
|
229 | 229 | When set, will look for files named like: 'ipcontroller-<cluster_id>-client.json' |
|
230 | 230 | Since this is text inserted into filenames, typical recommendations apply: |
|
231 | 231 | Simple character strings are ideal, and spaces are not recommended (but |
|
232 | 232 | should generally work) |
|
233 | 233 | context : zmq.Context |
|
234 | 234 | Pass an existing zmq.Context instance, otherwise the client will create its own. |
|
235 | 235 | debug : bool |
|
236 | 236 | flag for lots of message printing for debug purposes |
|
237 | 237 | timeout : int/float |
|
238 | 238 | time (in seconds) to wait for connection replies from the Hub |
|
239 | 239 | [Default: 10] |
|
240 | 240 | |
|
241 | 241 | #-------------- session related args ---------------- |
|
242 | 242 | |
|
243 | 243 | config : Config object |
|
244 | 244 | If specified, this will be relayed to the Session for configuration |
|
245 | 245 | username : str |
|
246 | 246 | set username for the session object |
|
247 | 247 | |
|
248 | 248 | #-------------- ssh related args ---------------- |
|
249 | 249 | # These are args for configuring the ssh tunnel to be used |
|
250 | 250 | # credentials are used to forward connections over ssh to the Controller |
|
251 | 251 | # Note that the ip given in `addr` needs to be relative to sshserver |
|
252 | 252 | # The most basic case is to leave addr as pointing to localhost (127.0.0.1), |
|
253 | 253 | # and set sshserver as the same machine the Controller is on. However, |
|
254 | 254 | # the only requirement is that sshserver is able to see the Controller |
|
255 | 255 | # (i.e. is within the same trusted network). |
|
256 | 256 | |
|
257 | 257 | sshserver : str |
|
258 | 258 | A string of the form passed to ssh, i.e. 'server.tld' or 'user@server.tld:port' |
|
259 | 259 | If keyfile or password is specified, and this is not, it will default to |
|
260 | 260 | the ip given in addr. |
|
261 | 261 | sshkey : str; path to ssh private key file |
|
262 | 262 | This specifies a key to be used in ssh login, default None. |
|
263 | 263 | Regular default ssh keys will be used without specifying this argument. |
|
264 | 264 | password : str |
|
265 | 265 | Your ssh password to sshserver. Note that if this is left None, |
|
266 | 266 | you will be prompted for it if passwordless key based login is unavailable. |
|
267 | 267 | paramiko : bool |
|
268 | 268 | flag for whether to use paramiko instead of shell ssh for tunneling. |
|
269 | 269 | [default: True on win32, False else] |
|
270 | 270 | |
|
271 | 271 | |
|
272 | 272 | Attributes |
|
273 | 273 | ---------- |
|
274 | 274 | |
|
275 | 275 | ids : list of int engine IDs |
|
276 | 276 | requesting the ids attribute always synchronizes |
|
277 | 277 | the registration state. To request ids without synchronization, |
|
278 | 278 | use semi-private _ids attributes. |
|
279 | 279 | |
|
280 | 280 | history : list of msg_ids |
|
281 | 281 | a list of msg_ids, keeping track of all the execution |
|
282 | 282 | messages you have submitted in order. |
|
283 | 283 | |
|
284 | 284 | outstanding : set of msg_ids |
|
285 | 285 | a set of msg_ids that have been submitted, but whose |
|
286 | 286 | results have not yet been received. |
|
287 | 287 | |
|
288 | 288 | results : dict |
|
289 | 289 | a dict of all our results, keyed by msg_id |
|
290 | 290 | |
|
291 | 291 | block : bool |
|
292 | 292 | determines default behavior when block not specified |
|
293 | 293 | in execution methods |
|
294 | 294 | |
|
295 | 295 | Methods |
|
296 | 296 | ------- |
|
297 | 297 | |
|
298 | 298 | spin |
|
299 | 299 | flushes incoming results and registration state changes |
|
300 | 300 | control methods spin, and requesting `ids` also ensures up to date |
|
301 | 301 | |
|
302 | 302 | wait |
|
303 | 303 | wait on one or more msg_ids |
|
304 | 304 | |
|
305 | 305 | execution methods |
|
306 | 306 | apply |
|
307 | 307 | legacy: execute, run |
|
308 | 308 | |
|
309 | 309 | data movement |
|
310 | 310 | push, pull, scatter, gather |
|
311 | 311 | |
|
312 | 312 | query methods |
|
313 | 313 | queue_status, get_result, purge, result_status |
|
314 | 314 | |
|
315 | 315 | control methods |
|
316 | 316 | abort, shutdown |
|
317 | 317 | |
|
318 | 318 | """ |
|
319 | 319 | |
|
320 | 320 | |
|
321 | 321 | block = Bool(False) |
|
322 | 322 | outstanding = Set() |
|
323 | 323 | results = Instance('collections.defaultdict', (dict,)) |
|
324 | 324 | metadata = Instance('collections.defaultdict', (Metadata,)) |
|
325 | 325 | history = List() |
|
326 | 326 | debug = Bool(False) |
|
327 | 327 | _spin_thread = Any() |
|
328 | 328 | _stop_spinning = Any() |
|
329 | 329 | |
|
330 | 330 | profile=Unicode() |
|
331 | 331 | def _profile_default(self): |
|
332 | 332 | if BaseIPythonApplication.initialized(): |
|
333 | 333 | # an IPython app *might* be running, try to get its profile |
|
334 | 334 | try: |
|
335 | 335 | return BaseIPythonApplication.instance().profile |
|
336 | 336 | except (AttributeError, MultipleInstanceError): |
|
337 | 337 | # could be a *different* subclass of config.Application, |
|
338 | 338 | # which would raise one of these two errors. |
|
339 | 339 | return u'default' |
|
340 | 340 | else: |
|
341 | 341 | return u'default' |
|
342 | 342 | |
|
343 | 343 | |
|
344 | 344 | _outstanding_dict = Instance('collections.defaultdict', (set,)) |
|
345 | 345 | _ids = List() |
|
346 | 346 | _connected=Bool(False) |
|
347 | 347 | _ssh=Bool(False) |
|
348 | 348 | _context = Instance('zmq.Context') |
|
349 | 349 | _config = Dict() |
|
350 | 350 | _engines=Instance(util.ReverseDict, (), {}) |
|
351 | 351 | # _hub_socket=Instance('zmq.Socket') |
|
352 | 352 | _query_socket=Instance('zmq.Socket') |
|
353 | 353 | _control_socket=Instance('zmq.Socket') |
|
354 | 354 | _iopub_socket=Instance('zmq.Socket') |
|
355 | 355 | _notification_socket=Instance('zmq.Socket') |
|
356 | 356 | _mux_socket=Instance('zmq.Socket') |
|
357 | 357 | _task_socket=Instance('zmq.Socket') |
|
358 | 358 | _task_scheme=Unicode() |
|
359 | 359 | _closed = False |
|
360 | 360 | _ignored_control_replies=Integer(0) |
|
361 | 361 | _ignored_hub_replies=Integer(0) |
|
362 | 362 | |
|
363 | 363 | def __new__(self, *args, **kw): |
|
364 | 364 | # don't raise on positional args |
|
365 | 365 | return HasTraits.__new__(self, **kw) |
|
366 | 366 | |
|
367 | 367 | def __init__(self, url_file=None, profile=None, profile_dir=None, ipython_dir=None, |
|
368 | 368 | context=None, debug=False, |
|
369 | 369 | sshserver=None, sshkey=None, password=None, paramiko=None, |
|
370 | 370 | timeout=10, cluster_id=None, **extra_args |
|
371 | 371 | ): |
|
372 | 372 | if profile: |
|
373 | 373 | super(Client, self).__init__(debug=debug, profile=profile) |
|
374 | 374 | else: |
|
375 | 375 | super(Client, self).__init__(debug=debug) |
|
376 | 376 | if context is None: |
|
377 | 377 | context = zmq.Context.instance() |
|
378 | 378 | self._context = context |
|
379 | 379 | self._stop_spinning = Event() |
|
380 | 380 | |
|
381 | 381 | if 'url_or_file' in extra_args: |
|
382 | 382 | url_file = extra_args['url_or_file'] |
|
383 | 383 | warnings.warn("url_or_file arg no longer supported, use url_file", DeprecationWarning) |
|
384 | 384 | |
|
385 | 385 | if url_file and util.is_url(url_file): |
|
386 | 386 | raise ValueError("single urls cannot be specified, url-files must be used.") |
|
387 | 387 | |
|
388 | 388 | self._setup_profile_dir(self.profile, profile_dir, ipython_dir) |
|
389 | 389 | |
|
390 | 390 | no_file_msg = '\n'.join([ |
|
391 | 391 | "You have attempted to connect to an IPython Cluster but no Controller could be found.", |
|
392 | 392 | "Please double-check your configuration and ensure that a cluster is running.", |
|
393 | 393 | ]) |
|
394 | 394 | |
|
395 | 395 | if self._cd is not None: |
|
396 | 396 | if url_file is None: |
|
397 | 397 | if not cluster_id: |
|
398 | 398 | client_json = 'ipcontroller-client.json' |
|
399 | 399 | else: |
|
400 | 400 | client_json = 'ipcontroller-%s-client.json' % cluster_id |
|
401 | 401 | url_file = pjoin(self._cd.security_dir, client_json) |
|
402 | 402 | if not os.path.exists(url_file): |
|
403 | 403 | msg = '\n'.join([ |
|
404 | 404 | "Connection file %r not found." % compress_user(url_file), |
|
405 | 405 | no_file_msg, |
|
406 | 406 | ]) |
|
407 | 407 | raise IOError(msg) |
|
408 | 408 | if url_file is None: |
|
409 | 409 | raise IOError(no_file_msg) |
|
410 | 410 | |
|
411 | 411 | if not os.path.exists(url_file): |
|
412 | 412 | # Connection file explicitly specified, but not found |
|
413 | 413 | raise IOError("Connection file %r not found. Is a controller running?" % \ |
|
414 | 414 | compress_user(url_file) |
|
415 | 415 | ) |
|
416 | 416 | |
|
417 | 417 | with open(url_file) as f: |
|
418 | 418 | cfg = json.load(f) |
|
419 | 419 | |
|
420 | 420 | self._task_scheme = cfg['task_scheme'] |
|
421 | 421 | |
|
422 | 422 | # sync defaults from args, json: |
|
423 | 423 | if sshserver: |
|
424 | 424 | cfg['ssh'] = sshserver |
|
425 | 425 | |
|
426 | 426 | location = cfg.setdefault('location', None) |
|
427 | 427 | |
|
428 | 428 | proto,addr = cfg['interface'].split('://') |
|
429 | 429 | addr = util.disambiguate_ip_address(addr, location) |
|
430 | 430 | cfg['interface'] = "%s://%s" % (proto, addr) |
|
431 | 431 | |
|
432 | 432 | # turn interface,port into full urls: |
|
433 | 433 | for key in ('control', 'task', 'mux', 'iopub', 'notification', 'registration'): |
|
434 | 434 | cfg[key] = cfg['interface'] + ':%i' % cfg[key] |
|
435 | 435 | |
|
436 | 436 | url = cfg['registration'] |
|
437 | 437 | |
|
438 | 438 | if location is not None and addr == localhost(): |
|
439 | 439 | # location specified, and connection is expected to be local |
|
440 | 440 | if not is_local_ip(location) and not sshserver: |
|
441 | 441 | # load ssh from JSON *only* if the controller is not on |
|
442 | 442 | # this machine |
|
443 | 443 | sshserver=cfg['ssh'] |
|
444 | 444 | if not is_local_ip(location) and not sshserver: |
|
445 | 445 | # warn if no ssh specified, but SSH is probably needed |
|
446 | 446 | # This is only a warning, because the most likely cause |
|
447 | 447 | # is a local Controller on a laptop whose IP is dynamic |
|
448 | 448 | warnings.warn(""" |
|
449 | 449 | Controller appears to be listening on localhost, but not on this machine. |
|
450 | 450 | If this is true, you should specify Client(...,sshserver='you@%s') |
|
451 | 451 | or instruct your controller to listen on an external IP."""%location, |
|
452 | 452 | RuntimeWarning) |
|
453 | 453 | elif not sshserver: |
|
454 | 454 | # otherwise sync with cfg |
|
455 | 455 | sshserver = cfg['ssh'] |
|
456 | 456 | |
|
457 | 457 | self._config = cfg |
|
458 | 458 | |
|
459 | 459 | self._ssh = bool(sshserver or sshkey or password) |
|
460 | 460 | if self._ssh and sshserver is None: |
|
461 | 461 | # default to ssh via localhost |
|
462 | 462 | sshserver = addr |
|
463 | 463 | if self._ssh and password is None: |
|
464 | 464 | from zmq.ssh import tunnel |
|
465 | 465 | if tunnel.try_passwordless_ssh(sshserver, sshkey, paramiko): |
|
466 | 466 | password=False |
|
467 | 467 | else: |
|
468 | 468 | password = getpass("SSH Password for %s: "%sshserver) |
|
469 | 469 | ssh_kwargs = dict(keyfile=sshkey, password=password, paramiko=paramiko) |
|
470 | 470 | |
|
471 | 471 | # configure and construct the session |
|
472 | 472 | try: |
|
473 | 473 | extra_args['packer'] = cfg['pack'] |
|
474 | 474 | extra_args['unpacker'] = cfg['unpack'] |
|
475 | 475 | extra_args['key'] = cast_bytes(cfg['key']) |
|
476 | 476 | extra_args['signature_scheme'] = cfg['signature_scheme'] |
|
477 | 477 | except KeyError as exc: |
|
478 | 478 | msg = '\n'.join([ |
|
479 | 479 | "Connection file is invalid (missing '{}'), possibly from an old version of IPython.", |
|
480 | 480 | "If you are reusing connection files, remove them and start ipcontroller again." |
|
481 | 481 | ]) |
|
482 | 482 | raise ValueError(msg.format(exc.message)) |
|
483 | 483 | |
|
484 | 484 | self.session = Session(**extra_args) |
|
485 | 485 | |
|
486 | 486 | self._query_socket = self._context.socket(zmq.DEALER) |
|
487 | 487 | |
|
488 | 488 | if self._ssh: |
|
489 | 489 | from zmq.ssh import tunnel |
|
490 | 490 | tunnel.tunnel_connection(self._query_socket, cfg['registration'], sshserver, **ssh_kwargs) |
|
491 | 491 | else: |
|
492 | 492 | self._query_socket.connect(cfg['registration']) |
|
493 | 493 | |
|
494 | 494 | self.session.debug = self.debug |
|
495 | 495 | |
|
496 | 496 | self._notification_handlers = {'registration_notification' : self._register_engine, |
|
497 | 497 | 'unregistration_notification' : self._unregister_engine, |
|
498 | 498 | 'shutdown_notification' : lambda msg: self.close(), |
|
499 | 499 | } |
|
500 | 500 | self._queue_handlers = {'execute_reply' : self._handle_execute_reply, |
|
501 | 501 | 'apply_reply' : self._handle_apply_reply} |
|
502 | 502 | |
|
503 | 503 | try: |
|
504 | 504 | self._connect(sshserver, ssh_kwargs, timeout) |
|
505 | 505 | except: |
|
506 | 506 | self.close(linger=0) |
|
507 | 507 | raise |
|
508 | 508 | |
|
509 | 509 | # last step: setup magics, if we are in IPython: |
|
510 | 510 | |
|
511 | 511 | try: |
|
512 | 512 | ip = get_ipython() |
|
513 | 513 | except NameError: |
|
514 | 514 | return |
|
515 | 515 | else: |
|
516 | 516 | if 'px' not in ip.magics_manager.magics: |
|
517 | 517 | # in IPython but we are the first Client. |
|
518 | 518 | # activate a default view for parallel magics. |
|
519 | 519 | self.activate() |
|
520 | 520 | |
|
521 | 521 | def __del__(self): |
|
522 | 522 | """cleanup sockets, but _not_ context.""" |
|
523 | 523 | self.close() |
|
524 | 524 | |
|
525 | 525 | def _setup_profile_dir(self, profile, profile_dir, ipython_dir): |
|
526 | 526 | if ipython_dir is None: |
|
527 | 527 | ipython_dir = get_ipython_dir() |
|
528 | 528 | if profile_dir is not None: |
|
529 | 529 | try: |
|
530 | 530 | self._cd = ProfileDir.find_profile_dir(profile_dir) |
|
531 | 531 | return |
|
532 | 532 | except ProfileDirError: |
|
533 | 533 | pass |
|
534 | 534 | elif profile is not None: |
|
535 | 535 | try: |
|
536 | 536 | self._cd = ProfileDir.find_profile_dir_by_name( |
|
537 | 537 | ipython_dir, profile) |
|
538 | 538 | return |
|
539 | 539 | except ProfileDirError: |
|
540 | 540 | pass |
|
541 | 541 | self._cd = None |
|
542 | 542 | |
|
543 | 543 | def _update_engines(self, engines): |
|
544 | 544 | """Update our engines dict and _ids from a dict of the form: {id:uuid}.""" |
|
545 | 545 | for k,v in iteritems(engines): |
|
546 | 546 | eid = int(k) |
|
547 | 547 | if eid not in self._engines: |
|
548 | 548 | self._ids.append(eid) |
|
549 | 549 | self._engines[eid] = v |
|
550 | 550 | self._ids = sorted(self._ids) |
|
551 | 551 | if sorted(self._engines.keys()) != list(range(len(self._engines))) and \ |
|
552 | 552 | self._task_scheme == 'pure' and self._task_socket: |
|
553 | 553 | self._stop_scheduling_tasks() |
|
554 | 554 | |
|
555 | 555 | def _stop_scheduling_tasks(self): |
|
556 | 556 | """Stop scheduling tasks because an engine has been unregistered |
|
557 | 557 | from a pure ZMQ scheduler. |
|
558 | 558 | """ |
|
559 | 559 | self._task_socket.close() |
|
560 | 560 | self._task_socket = None |
|
561 | 561 | msg = "An engine has been unregistered, and we are using pure " +\ |
|
562 | 562 | "ZMQ task scheduling. Task farming will be disabled." |
|
563 | 563 | if self.outstanding: |
|
564 | 564 | msg += " If you were running tasks when this happened, " +\ |
|
565 | 565 | "some `outstanding` msg_ids may never resolve." |
|
566 | 566 | warnings.warn(msg, RuntimeWarning) |
|
567 | 567 | |
|
568 | 568 | def _build_targets(self, targets): |
|
569 | 569 | """Turn valid target IDs or 'all' into two lists: |
|
570 | 570 | (int_ids, uuids). |
|
571 | 571 | """ |
|
572 | 572 | if not self._ids: |
|
573 | 573 | # flush notification socket if no engines yet, just in case |
|
574 | 574 | if not self.ids: |
|
575 | 575 | raise error.NoEnginesRegistered("Can't build targets without any engines") |
|
576 | 576 | |
|
577 | 577 | if targets is None: |
|
578 | 578 | targets = self._ids |
|
579 | 579 | elif isinstance(targets, string_types): |
|
580 | 580 | if targets.lower() == 'all': |
|
581 | 581 | targets = self._ids |
|
582 | 582 | else: |
|
583 | 583 | raise TypeError("%r not valid str target, must be 'all'"%(targets)) |
|
584 | 584 | elif isinstance(targets, int): |
|
585 | 585 | if targets < 0: |
|
586 | 586 | targets = self.ids[targets] |
|
587 | 587 | if targets not in self._ids: |
|
588 | 588 | raise IndexError("No such engine: %i"%targets) |
|
589 | 589 | targets = [targets] |
|
590 | 590 | |
|
591 | 591 | if isinstance(targets, slice): |
|
592 | 592 | indices = list(range(len(self._ids))[targets]) |
|
593 | 593 | ids = self.ids |
|
594 | 594 | targets = [ ids[i] for i in indices ] |
|
595 | 595 | |
|
596 | 596 | if not isinstance(targets, (tuple, list, xrange)): |
|
597 | 597 | raise TypeError("targets by int/slice/collection of ints only, not %s"%(type(targets))) |
|
598 | 598 | |
|
599 | 599 | return [cast_bytes(self._engines[t]) for t in targets], list(targets) |
|
600 | 600 | |
|
601 | 601 | def _connect(self, sshserver, ssh_kwargs, timeout): |
|
602 | 602 | """setup all our socket connections to the cluster. This is called from |
|
603 | 603 | __init__.""" |
|
604 | 604 | |
|
605 | 605 | # Maybe allow reconnecting? |
|
606 | 606 | if self._connected: |
|
607 | 607 | return |
|
608 | 608 | self._connected=True |
|
609 | 609 | |
|
610 | 610 | def connect_socket(s, url): |
|
611 | 611 | if self._ssh: |
|
612 | 612 | from zmq.ssh import tunnel |
|
613 | 613 | return tunnel.tunnel_connection(s, url, sshserver, **ssh_kwargs) |
|
614 | 614 | else: |
|
615 | 615 | return s.connect(url) |
|
616 | 616 | |
|
617 | 617 | self.session.send(self._query_socket, 'connection_request') |
|
618 | 618 | # use Poller because zmq.select has wrong units in pyzmq 2.1.7 |
|
619 | 619 | poller = zmq.Poller() |
|
620 | 620 | poller.register(self._query_socket, zmq.POLLIN) |
|
621 | 621 | # poll expects milliseconds, timeout is seconds |
|
622 | 622 | evts = poller.poll(timeout*1000) |
|
623 | 623 | if not evts: |
|
624 | 624 | raise error.TimeoutError("Hub connection request timed out") |
|
625 | 625 | idents,msg = self.session.recv(self._query_socket,mode=0) |
|
626 | 626 | if self.debug: |
|
627 | 627 | pprint(msg) |
|
628 | 628 | content = msg['content'] |
|
629 | 629 | # self._config['registration'] = dict(content) |
|
630 | 630 | cfg = self._config |
|
631 | 631 | if content['status'] == 'ok': |
|
632 | 632 | self._mux_socket = self._context.socket(zmq.DEALER) |
|
633 | 633 | connect_socket(self._mux_socket, cfg['mux']) |
|
634 | 634 | |
|
635 | 635 | self._task_socket = self._context.socket(zmq.DEALER) |
|
636 | 636 | connect_socket(self._task_socket, cfg['task']) |
|
637 | 637 | |
|
638 | 638 | self._notification_socket = self._context.socket(zmq.SUB) |
|
639 | 639 | self._notification_socket.setsockopt(zmq.SUBSCRIBE, b'') |
|
640 | 640 | connect_socket(self._notification_socket, cfg['notification']) |
|
641 | 641 | |
|
642 | 642 | self._control_socket = self._context.socket(zmq.DEALER) |
|
643 | 643 | connect_socket(self._control_socket, cfg['control']) |
|
644 | 644 | |
|
645 | 645 | self._iopub_socket = self._context.socket(zmq.SUB) |
|
646 | 646 | self._iopub_socket.setsockopt(zmq.SUBSCRIBE, b'') |
|
647 | 647 | connect_socket(self._iopub_socket, cfg['iopub']) |
|
648 | 648 | |
|
649 | 649 | self._update_engines(dict(content['engines'])) |
|
650 | 650 | else: |
|
651 | 651 | self._connected = False |
|
652 | 652 | raise Exception("Failed to connect!") |
|
653 | 653 | |
|
654 | 654 | #-------------------------------------------------------------------------- |
|
655 | 655 | # handlers and callbacks for incoming messages |
|
656 | 656 | #-------------------------------------------------------------------------- |
|
657 | 657 | |
|
658 | 658 | def _unwrap_exception(self, content): |
|
659 | 659 | """unwrap exception, and remap engine_id to int.""" |
|
660 | 660 | e = error.unwrap_exception(content) |
|
661 | 661 | # print e.traceback |
|
662 | 662 | if e.engine_info: |
|
663 | 663 | e_uuid = e.engine_info['engine_uuid'] |
|
664 | 664 | eid = self._engines[e_uuid] |
|
665 | 665 | e.engine_info['engine_id'] = eid |
|
666 | 666 | return e |
|
667 | 667 | |
|
668 | 668 | def _extract_metadata(self, msg): |
|
669 | 669 | header = msg['header'] |
|
670 | 670 | parent = msg['parent_header'] |
|
671 | 671 | msg_meta = msg['metadata'] |
|
672 | 672 | content = msg['content'] |
|
673 | 673 | md = {'msg_id' : parent['msg_id'], |
|
674 | 674 | 'received' : datetime.now(), |
|
675 | 675 | 'engine_uuid' : msg_meta.get('engine', None), |
|
676 | 676 | 'follow' : msg_meta.get('follow', []), |
|
677 | 677 | 'after' : msg_meta.get('after', []), |
|
678 | 678 | 'status' : content['status'], |
|
679 | 679 | } |
|
680 | 680 | |
|
681 | 681 | if md['engine_uuid'] is not None: |
|
682 | 682 | md['engine_id'] = self._engines.get(md['engine_uuid'], None) |
|
683 | 683 | |
|
684 | 684 | if 'date' in parent: |
|
685 | 685 | md['submitted'] = parent['date'] |
|
686 | 686 | if 'started' in msg_meta: |
|
687 | 687 | md['started'] = parse_date(msg_meta['started']) |
|
688 | 688 | if 'date' in header: |
|
689 | 689 | md['completed'] = header['date'] |
|
690 | 690 | return md |
|
691 | 691 | |
|
692 | 692 | def _register_engine(self, msg): |
|
693 | 693 | """Register a new engine, and update our connection info.""" |
|
694 | 694 | content = msg['content'] |
|
695 | 695 | eid = content['id'] |
|
696 | 696 | d = {eid : content['uuid']} |
|
697 | 697 | self._update_engines(d) |
|
698 | 698 | |
|
699 | 699 | def _unregister_engine(self, msg): |
|
700 | 700 | """Unregister an engine that has died.""" |
|
701 | 701 | content = msg['content'] |
|
702 | 702 | eid = int(content['id']) |
|
703 | 703 | if eid in self._ids: |
|
704 | 704 | self._ids.remove(eid) |
|
705 | 705 | uuid = self._engines.pop(eid) |
|
706 | 706 | |
|
707 | 707 | self._handle_stranded_msgs(eid, uuid) |
|
708 | 708 | |
|
709 | 709 | if self._task_socket and self._task_scheme == 'pure': |
|
710 | 710 | self._stop_scheduling_tasks() |
|
711 | 711 | |
|
712 | 712 | def _handle_stranded_msgs(self, eid, uuid): |
|
713 | 713 | """Handle messages known to be on an engine when the engine unregisters. |
|
714 | 714 | |
|
715 | 715 | It is possible that this will fire prematurely - that is, an engine will |
|
716 | 716 | go down after completing a result, and the client will be notified |
|
717 | 717 | of the unregistration and later receive the successful result. |
|
718 | 718 | """ |
|
719 | 719 | |
|
720 | 720 | outstanding = self._outstanding_dict[uuid] |
|
721 | 721 | |
|
722 | 722 | for msg_id in list(outstanding): |
|
723 | 723 | if msg_id in self.results: |
|
724 | 724 | # we already |
|
725 | 725 | continue |
|
726 | 726 | try: |
|
727 | 727 | raise error.EngineError("Engine %r died while running task %r"%(eid, msg_id)) |
|
728 | 728 | except: |
|
729 | 729 | content = error.wrap_exception() |
|
730 | 730 | # build a fake message: |
|
731 | 731 | msg = self.session.msg('apply_reply', content=content) |
|
732 | 732 | msg['parent_header']['msg_id'] = msg_id |
|
733 | 733 | msg['metadata']['engine'] = uuid |
|
734 | 734 | self._handle_apply_reply(msg) |
|
735 | 735 | |
|
736 | 736 | def _handle_execute_reply(self, msg): |
|
737 | 737 | """Save the reply to an execute_request into our results. |
|
738 | 738 | |
|
739 | 739 | execute messages are never actually used. apply is used instead. |
|
740 | 740 | """ |
|
741 | 741 | |
|
742 | 742 | parent = msg['parent_header'] |
|
743 | 743 | msg_id = parent['msg_id'] |
|
744 | 744 | if msg_id not in self.outstanding: |
|
745 | 745 | if msg_id in self.history: |
|
746 | 746 | print("got stale result: %s"%msg_id) |
|
747 | 747 | else: |
|
748 | 748 | print("got unknown result: %s"%msg_id) |
|
749 | 749 | else: |
|
750 | 750 | self.outstanding.remove(msg_id) |
|
751 | 751 | |
|
752 | 752 | content = msg['content'] |
|
753 | 753 | header = msg['header'] |
|
754 | 754 | |
|
755 | 755 | # construct metadata: |
|
756 | 756 | md = self.metadata[msg_id] |
|
757 | 757 | md.update(self._extract_metadata(msg)) |
|
758 | 758 | # is this redundant? |
|
759 | 759 | self.metadata[msg_id] = md |
|
760 | 760 | |
|
761 | 761 | e_outstanding = self._outstanding_dict[md['engine_uuid']] |
|
762 | 762 | if msg_id in e_outstanding: |
|
763 | 763 | e_outstanding.remove(msg_id) |
|
764 | 764 | |
|
765 | 765 | # construct result: |
|
766 | 766 | if content['status'] == 'ok': |
|
767 | 767 | self.results[msg_id] = ExecuteReply(msg_id, content, md) |
|
768 | 768 | elif content['status'] == 'aborted': |
|
769 | 769 | self.results[msg_id] = error.TaskAborted(msg_id) |
|
770 | 770 | elif content['status'] == 'resubmitted': |
|
771 | 771 | # TODO: handle resubmission |
|
772 | 772 | pass |
|
773 | 773 | else: |
|
774 | 774 | self.results[msg_id] = self._unwrap_exception(content) |
|
775 | 775 | |
|
776 | 776 | def _handle_apply_reply(self, msg): |
|
777 | 777 | """Save the reply to an apply_request into our results.""" |
|
778 | 778 | parent = msg['parent_header'] |
|
779 | 779 | msg_id = parent['msg_id'] |
|
780 | 780 | if msg_id not in self.outstanding: |
|
781 | 781 | if msg_id in self.history: |
|
782 | 782 | print("got stale result: %s"%msg_id) |
|
783 | 783 | print(self.results[msg_id]) |
|
784 | 784 | print(msg) |
|
785 | 785 | else: |
|
786 | 786 | print("got unknown result: %s"%msg_id) |
|
787 | 787 | else: |
|
788 | 788 | self.outstanding.remove(msg_id) |
|
789 | 789 | content = msg['content'] |
|
790 | 790 | header = msg['header'] |
|
791 | 791 | |
|
792 | 792 | # construct metadata: |
|
793 | 793 | md = self.metadata[msg_id] |
|
794 | 794 | md.update(self._extract_metadata(msg)) |
|
795 | 795 | # is this redundant? |
|
796 | 796 | self.metadata[msg_id] = md |
|
797 | 797 | |
|
798 | 798 | e_outstanding = self._outstanding_dict[md['engine_uuid']] |
|
799 | 799 | if msg_id in e_outstanding: |
|
800 | 800 | e_outstanding.remove(msg_id) |
|
801 | 801 | |
|
802 | 802 | # construct result: |
|
803 | 803 | if content['status'] == 'ok': |
|
804 | 804 | self.results[msg_id] = serialize.deserialize_object(msg['buffers'])[0] |
|
805 | 805 | elif content['status'] == 'aborted': |
|
806 | 806 | self.results[msg_id] = error.TaskAborted(msg_id) |
|
807 | 807 | elif content['status'] == 'resubmitted': |
|
808 | 808 | # TODO: handle resubmission |
|
809 | 809 | pass |
|
810 | 810 | else: |
|
811 | 811 | self.results[msg_id] = self._unwrap_exception(content) |
|
812 | 812 | |
|
813 | 813 | def _flush_notifications(self): |
|
814 | 814 | """Flush notifications of engine registrations waiting |
|
815 | 815 | in ZMQ queue.""" |
|
816 | 816 | idents,msg = self.session.recv(self._notification_socket, mode=zmq.NOBLOCK) |
|
817 | 817 | while msg is not None: |
|
818 | 818 | if self.debug: |
|
819 | 819 | pprint(msg) |
|
820 | 820 | msg_type = msg['header']['msg_type'] |
|
821 | 821 | handler = self._notification_handlers.get(msg_type, None) |
|
822 | 822 | if handler is None: |
|
823 | 823 | raise Exception("Unhandled message type: %s" % msg_type) |
|
824 | 824 | else: |
|
825 | 825 | handler(msg) |
|
826 | 826 | idents,msg = self.session.recv(self._notification_socket, mode=zmq.NOBLOCK) |
|
827 | 827 | |
|
828 | 828 | def _flush_results(self, sock): |
|
829 | 829 | """Flush task or queue results waiting in ZMQ queue.""" |
|
830 | 830 | idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK) |
|
831 | 831 | while msg is not None: |
|
832 | 832 | if self.debug: |
|
833 | 833 | pprint(msg) |
|
834 | 834 | msg_type = msg['header']['msg_type'] |
|
835 | 835 | handler = self._queue_handlers.get(msg_type, None) |
|
836 | 836 | if handler is None: |
|
837 | 837 | raise Exception("Unhandled message type: %s" % msg_type) |
|
838 | 838 | else: |
|
839 | 839 | handler(msg) |
|
840 | 840 | idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK) |
|
841 | 841 | |
|
842 | 842 | def _flush_control(self, sock): |
|
843 | 843 | """Flush replies from the control channel waiting |
|
844 | 844 | in the ZMQ queue. |
|
845 | 845 | |
|
846 | 846 | Currently: ignore them.""" |
|
847 | 847 | if self._ignored_control_replies <= 0: |
|
848 | 848 | return |
|
849 | 849 | idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK) |
|
850 | 850 | while msg is not None: |
|
851 | 851 | self._ignored_control_replies -= 1 |
|
852 | 852 | if self.debug: |
|
853 | 853 | pprint(msg) |
|
854 | 854 | idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK) |
|
855 | 855 | |
|
856 | 856 | def _flush_ignored_control(self): |
|
857 | 857 | """flush ignored control replies""" |
|
858 | 858 | while self._ignored_control_replies > 0: |
|
859 | 859 | self.session.recv(self._control_socket) |
|
860 | 860 | self._ignored_control_replies -= 1 |
|
861 | 861 | |
|
862 | 862 | def _flush_ignored_hub_replies(self): |
|
863 | 863 | ident,msg = self.session.recv(self._query_socket, mode=zmq.NOBLOCK) |
|
864 | 864 | while msg is not None: |
|
865 | 865 | ident,msg = self.session.recv(self._query_socket, mode=zmq.NOBLOCK) |
|
866 | 866 | |
|
867 | 867 | def _flush_iopub(self, sock): |
|
868 | 868 | """Flush replies from the iopub channel waiting |
|
869 | 869 | in the ZMQ queue. |
|
870 | 870 | """ |
|
871 | 871 | idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK) |
|
872 | 872 | while msg is not None: |
|
873 | 873 | if self.debug: |
|
874 | 874 | pprint(msg) |
|
875 | 875 | parent = msg['parent_header'] |
|
876 | 876 | if not parent or parent['session'] != self.session.session: |
|
877 | 877 | # ignore IOPub messages not from here |
|
878 | 878 | idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK) |
|
879 | 879 | continue |
|
880 | 880 | msg_id = parent['msg_id'] |
|
881 | 881 | content = msg['content'] |
|
882 | 882 | header = msg['header'] |
|
883 | 883 | msg_type = msg['header']['msg_type'] |
|
884 | 884 | |
|
885 | 885 | if msg_type == 'status' and msg_id not in self.metadata: |
|
886 | 886 | # ignore status messages if they aren't mine |
|
887 | 887 | idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK) |
|
888 | 888 | continue |
|
889 | 889 | |
|
890 | 890 | # init metadata: |
|
891 | 891 | md = self.metadata[msg_id] |
|
892 | 892 | |
|
893 | 893 | if msg_type == 'stream': |
|
894 | 894 | name = content['name'] |
|
895 | 895 | s = md[name] or '' |
|
896 | 896 | md[name] = s + content['text'] |
|
897 | 897 | elif msg_type == 'error': |
|
898 | 898 | md.update({'error' : self._unwrap_exception(content)}) |
|
899 | 899 | elif msg_type == 'execute_input': |
|
900 | 900 | md.update({'execute_input' : content['code']}) |
|
901 | 901 | elif msg_type == 'display_data': |
|
902 | 902 | md['outputs'].append(content) |
|
903 | 903 | elif msg_type == 'execute_result': |
|
904 | 904 | md['execute_result'] = content |
|
905 | 905 | elif msg_type == 'data_message': |
|
906 | 906 | data, remainder = serialize.deserialize_object(msg['buffers']) |
|
907 | 907 | md['data'].update(data) |
|
908 | 908 | elif msg_type == 'status': |
|
909 | 909 | # idle message comes after all outputs |
|
910 | 910 | if content['execution_state'] == 'idle': |
|
911 | 911 | md['outputs_ready'] = True |
|
912 | 912 | else: |
|
913 | 913 | # unhandled msg_type (status, etc.) |
|
914 | 914 | pass |
|
915 | 915 | |
|
916 | 916 | # reduntant? |
|
917 | 917 | self.metadata[msg_id] = md |
|
918 | 918 | |
|
919 | 919 | idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK) |
|
920 | 920 | |
|
921 | 921 | #-------------------------------------------------------------------------- |
|
922 | 922 | # len, getitem |
|
923 | 923 | #-------------------------------------------------------------------------- |
|
924 | 924 | |
|
925 | 925 | def __len__(self): |
|
926 | 926 | """len(client) returns # of engines.""" |
|
927 | 927 | return len(self.ids) |
|
928 | 928 | |
|
929 | 929 | def __getitem__(self, key): |
|
930 | 930 | """index access returns DirectView multiplexer objects |
|
931 | 931 | |
|
932 | 932 | Must be int, slice, or list/tuple/xrange of ints""" |
|
933 | 933 | if not isinstance(key, (int, slice, tuple, list, xrange)): |
|
934 | 934 | raise TypeError("key by int/slice/iterable of ints only, not %s"%(type(key))) |
|
935 | 935 | else: |
|
936 | 936 | return self.direct_view(key) |
|
937 | 937 | |
|
938 | 938 | def __iter__(self): |
|
939 | 939 | """Since we define getitem, Client is iterable |
|
940 | 940 | |
|
941 | 941 | but unless we also define __iter__, it won't work correctly unless engine IDs |
|
942 | 942 | start at zero and are continuous. |
|
943 | 943 | """ |
|
944 | 944 | for eid in self.ids: |
|
945 | 945 | yield self.direct_view(eid) |
|
946 | 946 | |
|
947 | 947 | #-------------------------------------------------------------------------- |
|
948 | 948 | # Begin public methods |
|
949 | 949 | #-------------------------------------------------------------------------- |
|
950 | 950 | |
|
951 | 951 | @property |
|
952 | 952 | def ids(self): |
|
953 | 953 | """Always up-to-date ids property.""" |
|
954 | 954 | self._flush_notifications() |
|
955 | 955 | # always copy: |
|
956 | 956 | return list(self._ids) |
|
957 | 957 | |
|
958 | 958 | def activate(self, targets='all', suffix=''): |
|
959 | 959 | """Create a DirectView and register it with IPython magics |
|
960 | 960 | |
|
961 | 961 | Defines the magics `%px, %autopx, %pxresult, %%px` |
|
962 | 962 | |
|
963 | 963 | Parameters |
|
964 | 964 | ---------- |
|
965 | 965 | |
|
966 | 966 | targets: int, list of ints, or 'all' |
|
967 | 967 | The engines on which the view's magics will run |
|
968 | 968 | suffix: str [default: ''] |
|
969 | 969 | The suffix, if any, for the magics. This allows you to have |
|
970 | 970 | multiple views associated with parallel magics at the same time. |
|
971 | 971 | |
|
972 | 972 | e.g. ``rc.activate(targets=0, suffix='0')`` will give you |
|
973 | 973 | the magics ``%px0``, ``%pxresult0``, etc. for running magics just |
|
974 | 974 | on engine 0. |
|
975 | 975 | """ |
|
976 | 976 | view = self.direct_view(targets) |
|
977 | 977 | view.block = True |
|
978 | 978 | view.activate(suffix) |
|
979 | 979 | return view |
|
980 | 980 | |
|
981 | 981 | def close(self, linger=None): |
|
982 | 982 | """Close my zmq Sockets |
|
983 | 983 | |
|
984 | 984 | If `linger`, set the zmq LINGER socket option, |
|
985 | 985 | which allows discarding of messages. |
|
986 | 986 | """ |
|
987 | 987 | if self._closed: |
|
988 | 988 | return |
|
989 | 989 | self.stop_spin_thread() |
|
990 | 990 | snames = [ trait for trait in self.trait_names() if trait.endswith("socket") ] |
|
991 | 991 | for name in snames: |
|
992 | 992 | socket = getattr(self, name) |
|
993 | 993 | if socket is not None and not socket.closed: |
|
994 | 994 | if linger is not None: |
|
995 | 995 | socket.close(linger=linger) |
|
996 | 996 | else: |
|
997 | 997 | socket.close() |
|
998 | 998 | self._closed = True |
|
999 | 999 | |
|
1000 | 1000 | def _spin_every(self, interval=1): |
|
1001 | 1001 | """target func for use in spin_thread""" |
|
1002 | 1002 | while True: |
|
1003 | 1003 | if self._stop_spinning.is_set(): |
|
1004 | 1004 | return |
|
1005 | 1005 | time.sleep(interval) |
|
1006 | 1006 | self.spin() |
|
1007 | 1007 | |
|
1008 | 1008 | def spin_thread(self, interval=1): |
|
1009 | 1009 | """call Client.spin() in a background thread on some regular interval |
|
1010 | 1010 | |
|
1011 | 1011 | This helps ensure that messages don't pile up too much in the zmq queue |
|
1012 | 1012 | while you are working on other things, or just leaving an idle terminal. |
|
1013 | 1013 | |
|
1014 | 1014 | It also helps limit potential padding of the `received` timestamp |
|
1015 | 1015 | on AsyncResult objects, used for timings. |
|
1016 | 1016 | |
|
1017 | 1017 | Parameters |
|
1018 | 1018 | ---------- |
|
1019 | 1019 | |
|
1020 | 1020 | interval : float, optional |
|
1021 | 1021 | The interval on which to spin the client in the background thread |
|
1022 | 1022 | (simply passed to time.sleep). |
|
1023 | 1023 | |
|
1024 | 1024 | Notes |
|
1025 | 1025 | ----- |
|
1026 | 1026 | |
|
1027 | 1027 | For precision timing, you may want to use this method to put a bound |
|
1028 | 1028 | on the jitter (in seconds) in `received` timestamps used |
|
1029 | 1029 | in AsyncResult.wall_time. |
|
1030 | 1030 | |
|
1031 | 1031 | """ |
|
1032 | 1032 | if self._spin_thread is not None: |
|
1033 | 1033 | self.stop_spin_thread() |
|
1034 | 1034 | self._stop_spinning.clear() |
|
1035 | 1035 | self._spin_thread = Thread(target=self._spin_every, args=(interval,)) |
|
1036 | 1036 | self._spin_thread.daemon = True |
|
1037 | 1037 | self._spin_thread.start() |
|
1038 | 1038 | |
|
1039 | 1039 | def stop_spin_thread(self): |
|
1040 | 1040 | """stop background spin_thread, if any""" |
|
1041 | 1041 | if self._spin_thread is not None: |
|
1042 | 1042 | self._stop_spinning.set() |
|
1043 | 1043 | self._spin_thread.join() |
|
1044 | 1044 | self._spin_thread = None |
|
1045 | 1045 | |
|
1046 | 1046 | def spin(self): |
|
1047 | 1047 | """Flush any registration notifications and execution results |
|
1048 | 1048 | waiting in the ZMQ queue. |
|
1049 | 1049 | """ |
|
1050 | 1050 | if self._notification_socket: |
|
1051 | 1051 | self._flush_notifications() |
|
1052 | 1052 | if self._iopub_socket: |
|
1053 | 1053 | self._flush_iopub(self._iopub_socket) |
|
1054 | 1054 | if self._mux_socket: |
|
1055 | 1055 | self._flush_results(self._mux_socket) |
|
1056 | 1056 | if self._task_socket: |
|
1057 | 1057 | self._flush_results(self._task_socket) |
|
1058 | 1058 | if self._control_socket: |
|
1059 | 1059 | self._flush_control(self._control_socket) |
|
1060 | 1060 | if self._query_socket: |
|
1061 | 1061 | self._flush_ignored_hub_replies() |
|
1062 | 1062 | |
|
1063 | 1063 | def wait(self, jobs=None, timeout=-1): |
|
1064 | 1064 | """waits on one or more `jobs`, for up to `timeout` seconds. |
|
1065 | 1065 | |
|
1066 | 1066 | Parameters |
|
1067 | 1067 | ---------- |
|
1068 | 1068 | |
|
1069 | 1069 | jobs : int, str, or list of ints and/or strs, or one or more AsyncResult objects |
|
1070 | 1070 | ints are indices to self.history |
|
1071 | 1071 | strs are msg_ids |
|
1072 | 1072 | default: wait on all outstanding messages |
|
1073 | 1073 | timeout : float |
|
1074 | 1074 | a time in seconds, after which to give up. |
|
1075 | 1075 | default is -1, which means no timeout |
|
1076 | 1076 | |
|
1077 | 1077 | Returns |
|
1078 | 1078 | ------- |
|
1079 | 1079 | |
|
1080 | 1080 | True : when all msg_ids are done |
|
1081 | 1081 | False : timeout reached, some msg_ids still outstanding |
|
1082 | 1082 | """ |
|
1083 | 1083 | tic = time.time() |
|
1084 | 1084 | if jobs is None: |
|
1085 | 1085 | theids = self.outstanding |
|
1086 | 1086 | else: |
|
1087 | 1087 | if isinstance(jobs, string_types + (int, AsyncResult)): |
|
1088 | 1088 | jobs = [jobs] |
|
1089 | 1089 | theids = set() |
|
1090 | 1090 | for job in jobs: |
|
1091 | 1091 | if isinstance(job, int): |
|
1092 | 1092 | # index access |
|
1093 | 1093 | job = self.history[job] |
|
1094 | 1094 | elif isinstance(job, AsyncResult): |
|
1095 | 1095 | theids.update(job.msg_ids) |
|
1096 | 1096 | continue |
|
1097 | 1097 | theids.add(job) |
|
1098 | 1098 | if not theids.intersection(self.outstanding): |
|
1099 | 1099 | return True |
|
1100 | 1100 | self.spin() |
|
1101 | 1101 | while theids.intersection(self.outstanding): |
|
1102 | 1102 | if timeout >= 0 and ( time.time()-tic ) > timeout: |
|
1103 | 1103 | break |
|
1104 | 1104 | time.sleep(1e-3) |
|
1105 | 1105 | self.spin() |
|
1106 | 1106 | return len(theids.intersection(self.outstanding)) == 0 |
|
1107 | 1107 | |
|
1108 | 1108 | #-------------------------------------------------------------------------- |
|
1109 | 1109 | # Control methods |
|
1110 | 1110 | #-------------------------------------------------------------------------- |
|
1111 | 1111 | |
|
1112 | 1112 | @spin_first |
|
1113 | 1113 | def clear(self, targets=None, block=None): |
|
1114 | 1114 | """Clear the namespace in target(s).""" |
|
1115 | 1115 | block = self.block if block is None else block |
|
1116 | 1116 | targets = self._build_targets(targets)[0] |
|
1117 | 1117 | for t in targets: |
|
1118 | 1118 | self.session.send(self._control_socket, 'clear_request', content={}, ident=t) |
|
1119 | 1119 | error = False |
|
1120 | 1120 | if block: |
|
1121 | 1121 | self._flush_ignored_control() |
|
1122 | 1122 | for i in range(len(targets)): |
|
1123 | 1123 | idents,msg = self.session.recv(self._control_socket,0) |
|
1124 | 1124 | if self.debug: |
|
1125 | 1125 | pprint(msg) |
|
1126 | 1126 | if msg['content']['status'] != 'ok': |
|
1127 | 1127 | error = self._unwrap_exception(msg['content']) |
|
1128 | 1128 | else: |
|
1129 | 1129 | self._ignored_control_replies += len(targets) |
|
1130 | 1130 | if error: |
|
1131 | 1131 | raise error |
|
1132 | 1132 | |
|
1133 | 1133 | |
|
1134 | 1134 | @spin_first |
|
1135 | 1135 | def abort(self, jobs=None, targets=None, block=None): |
|
1136 | 1136 | """Abort specific jobs from the execution queues of target(s). |
|
1137 | 1137 | |
|
1138 | 1138 | This is a mechanism to prevent jobs that have already been submitted |
|
1139 | 1139 | from executing. |
|
1140 | 1140 | |
|
1141 | 1141 | Parameters |
|
1142 | 1142 | ---------- |
|
1143 | 1143 | |
|
1144 | 1144 | jobs : msg_id, list of msg_ids, or AsyncResult |
|
1145 | 1145 | The jobs to be aborted |
|
1146 | 1146 | |
|
1147 | 1147 | If unspecified/None: abort all outstanding jobs. |
|
1148 | 1148 | |
|
1149 | 1149 | """ |
|
1150 | 1150 | block = self.block if block is None else block |
|
1151 | 1151 | jobs = jobs if jobs is not None else list(self.outstanding) |
|
1152 | 1152 | targets = self._build_targets(targets)[0] |
|
1153 | 1153 | |
|
1154 | 1154 | msg_ids = [] |
|
1155 | 1155 | if isinstance(jobs, string_types + (AsyncResult,)): |
|
1156 | 1156 | jobs = [jobs] |
|
1157 | 1157 | bad_ids = [obj for obj in jobs if not isinstance(obj, string_types + (AsyncResult,))] |
|
1158 | 1158 | if bad_ids: |
|
1159 | 1159 | raise TypeError("Invalid msg_id type %r, expected str or AsyncResult"%bad_ids[0]) |
|
1160 | 1160 | for j in jobs: |
|
1161 | 1161 | if isinstance(j, AsyncResult): |
|
1162 | 1162 | msg_ids.extend(j.msg_ids) |
|
1163 | 1163 | else: |
|
1164 | 1164 | msg_ids.append(j) |
|
1165 | 1165 | content = dict(msg_ids=msg_ids) |
|
1166 | 1166 | for t in targets: |
|
1167 | 1167 | self.session.send(self._control_socket, 'abort_request', |
|
1168 | 1168 | content=content, ident=t) |
|
1169 | 1169 | error = False |
|
1170 | 1170 | if block: |
|
1171 | 1171 | self._flush_ignored_control() |
|
1172 | 1172 | for i in range(len(targets)): |
|
1173 | 1173 | idents,msg = self.session.recv(self._control_socket,0) |
|
1174 | 1174 | if self.debug: |
|
1175 | 1175 | pprint(msg) |
|
1176 | 1176 | if msg['content']['status'] != 'ok': |
|
1177 | 1177 | error = self._unwrap_exception(msg['content']) |
|
1178 | 1178 | else: |
|
1179 | 1179 | self._ignored_control_replies += len(targets) |
|
1180 | 1180 | if error: |
|
1181 | 1181 | raise error |
|
1182 | 1182 | |
|
1183 | 1183 | @spin_first |
|
1184 | 1184 | def shutdown(self, targets='all', restart=False, hub=False, block=None): |
|
1185 | 1185 | """Terminates one or more engine processes, optionally including the hub. |
|
1186 | 1186 | |
|
1187 | 1187 | Parameters |
|
1188 | 1188 | ---------- |
|
1189 | 1189 | |
|
1190 | 1190 | targets: list of ints or 'all' [default: all] |
|
1191 | 1191 | Which engines to shutdown. |
|
1192 | 1192 | hub: bool [default: False] |
|
1193 | 1193 | Whether to include the Hub. hub=True implies targets='all'. |
|
1194 | 1194 | block: bool [default: self.block] |
|
1195 | 1195 | Whether to wait for clean shutdown replies or not. |
|
1196 | 1196 | restart: bool [default: False] |
|
1197 | 1197 | NOT IMPLEMENTED |
|
1198 | 1198 | whether to restart engines after shutting them down. |
|
1199 | 1199 | """ |
|
1200 | 1200 | from IPython.parallel.error import NoEnginesRegistered |
|
1201 | 1201 | if restart: |
|
1202 | 1202 | raise NotImplementedError("Engine restart is not yet implemented") |
|
1203 | 1203 | |
|
1204 | 1204 | block = self.block if block is None else block |
|
1205 | 1205 | if hub: |
|
1206 | 1206 | targets = 'all' |
|
1207 | 1207 | try: |
|
1208 | 1208 | targets = self._build_targets(targets)[0] |
|
1209 | 1209 | except NoEnginesRegistered: |
|
1210 | 1210 | targets = [] |
|
1211 | 1211 | for t in targets: |
|
1212 | 1212 | self.session.send(self._control_socket, 'shutdown_request', |
|
1213 | 1213 | content={'restart':restart},ident=t) |
|
1214 | 1214 | error = False |
|
1215 | 1215 | if block or hub: |
|
1216 | 1216 | self._flush_ignored_control() |
|
1217 | 1217 | for i in range(len(targets)): |
|
1218 | 1218 | idents,msg = self.session.recv(self._control_socket, 0) |
|
1219 | 1219 | if self.debug: |
|
1220 | 1220 | pprint(msg) |
|
1221 | 1221 | if msg['content']['status'] != 'ok': |
|
1222 | 1222 | error = self._unwrap_exception(msg['content']) |
|
1223 | 1223 | else: |
|
1224 | 1224 | self._ignored_control_replies += len(targets) |
|
1225 | 1225 | |
|
1226 | 1226 | if hub: |
|
1227 | 1227 | time.sleep(0.25) |
|
1228 | 1228 | self.session.send(self._query_socket, 'shutdown_request') |
|
1229 | 1229 | idents,msg = self.session.recv(self._query_socket, 0) |
|
1230 | 1230 | if self.debug: |
|
1231 | 1231 | pprint(msg) |
|
1232 | 1232 | if msg['content']['status'] != 'ok': |
|
1233 | 1233 | error = self._unwrap_exception(msg['content']) |
|
1234 | 1234 | |
|
1235 | 1235 | if error: |
|
1236 | 1236 | raise error |
|
1237 | 1237 | |
|
1238 | 1238 | #-------------------------------------------------------------------------- |
|
1239 | 1239 | # Execution related methods |
|
1240 | 1240 | #-------------------------------------------------------------------------- |
|
1241 | 1241 | |
|
1242 | 1242 | def _maybe_raise(self, result): |
|
1243 | 1243 | """wrapper for maybe raising an exception if apply failed.""" |
|
1244 | 1244 | if isinstance(result, error.RemoteError): |
|
1245 | 1245 | raise result |
|
1246 | 1246 | |
|
1247 | 1247 | return result |
|
1248 | 1248 | |
|
1249 | 1249 | def send_apply_request(self, socket, f, args=None, kwargs=None, metadata=None, track=False, |
|
1250 | 1250 | ident=None): |
|
1251 | 1251 | """construct and send an apply message via a socket. |
|
1252 | 1252 | |
|
1253 | 1253 | This is the principal method with which all engine execution is performed by views. |
|
1254 | 1254 | """ |
|
1255 | 1255 | |
|
1256 | 1256 | if self._closed: |
|
1257 | 1257 | raise RuntimeError("Client cannot be used after its sockets have been closed") |
|
1258 | 1258 | |
|
1259 | 1259 | # defaults: |
|
1260 | 1260 | args = args if args is not None else [] |
|
1261 | 1261 | kwargs = kwargs if kwargs is not None else {} |
|
1262 | 1262 | metadata = metadata if metadata is not None else {} |
|
1263 | 1263 | |
|
1264 | 1264 | # validate arguments |
|
1265 | 1265 | if not callable(f) and not isinstance(f, Reference): |
|
1266 | 1266 | raise TypeError("f must be callable, not %s"%type(f)) |
|
1267 | 1267 | if not isinstance(args, (tuple, list)): |
|
1268 | 1268 | raise TypeError("args must be tuple or list, not %s"%type(args)) |
|
1269 | 1269 | if not isinstance(kwargs, dict): |
|
1270 | 1270 | raise TypeError("kwargs must be dict, not %s"%type(kwargs)) |
|
1271 | 1271 | if not isinstance(metadata, dict): |
|
1272 | 1272 | raise TypeError("metadata must be dict, not %s"%type(metadata)) |
|
1273 | 1273 | |
|
1274 | 1274 | bufs = serialize.pack_apply_message(f, args, kwargs, |
|
1275 | 1275 | buffer_threshold=self.session.buffer_threshold, |
|
1276 | 1276 | item_threshold=self.session.item_threshold, |
|
1277 | 1277 | ) |
|
1278 | 1278 | |
|
1279 | 1279 | msg = self.session.send(socket, "apply_request", buffers=bufs, ident=ident, |
|
1280 | 1280 | metadata=metadata, track=track) |
|
1281 | 1281 | |
|
1282 | 1282 | msg_id = msg['header']['msg_id'] |
|
1283 | 1283 | self.outstanding.add(msg_id) |
|
1284 | 1284 | if ident: |
|
1285 | 1285 | # possibly routed to a specific engine |
|
1286 | 1286 | if isinstance(ident, list): |
|
1287 | 1287 | ident = ident[-1] |
|
1288 | 1288 | if ident in self._engines.values(): |
|
1289 | 1289 | # save for later, in case of engine death |
|
1290 | 1290 | self._outstanding_dict[ident].add(msg_id) |
|
1291 | 1291 | self.history.append(msg_id) |
|
1292 | 1292 | self.metadata[msg_id]['submitted'] = datetime.now() |
|
1293 | 1293 | |
|
1294 | 1294 | return msg |
|
1295 | 1295 | |
|
1296 | 1296 | def send_execute_request(self, socket, code, silent=True, metadata=None, ident=None): |
|
1297 | 1297 | """construct and send an execute request via a socket. |
|
1298 | 1298 | |
|
1299 | 1299 | """ |
|
1300 | 1300 | |
|
1301 | 1301 | if self._closed: |
|
1302 | 1302 | raise RuntimeError("Client cannot be used after its sockets have been closed") |
|
1303 | 1303 | |
|
1304 | 1304 | # defaults: |
|
1305 | 1305 | metadata = metadata if metadata is not None else {} |
|
1306 | 1306 | |
|
1307 | 1307 | # validate arguments |
|
1308 | 1308 | if not isinstance(code, string_types): |
|
1309 | 1309 | raise TypeError("code must be text, not %s" % type(code)) |
|
1310 | 1310 | if not isinstance(metadata, dict): |
|
1311 | 1311 | raise TypeError("metadata must be dict, not %s" % type(metadata)) |
|
1312 | 1312 | |
|
1313 | 1313 | content = dict(code=code, silent=bool(silent), user_expressions={}) |
|
1314 | 1314 | |
|
1315 | 1315 | |
|
1316 | 1316 | msg = self.session.send(socket, "execute_request", content=content, ident=ident, |
|
1317 | 1317 | metadata=metadata) |
|
1318 | 1318 | |
|
1319 | 1319 | msg_id = msg['header']['msg_id'] |
|
1320 | 1320 | self.outstanding.add(msg_id) |
|
1321 | 1321 | if ident: |
|
1322 | 1322 | # possibly routed to a specific engine |
|
1323 | 1323 | if isinstance(ident, list): |
|
1324 | 1324 | ident = ident[-1] |
|
1325 | 1325 | if ident in self._engines.values(): |
|
1326 | 1326 | # save for later, in case of engine death |
|
1327 | 1327 | self._outstanding_dict[ident].add(msg_id) |
|
1328 | 1328 | self.history.append(msg_id) |
|
1329 | 1329 | self.metadata[msg_id]['submitted'] = datetime.now() |
|
1330 | 1330 | |
|
1331 | 1331 | return msg |
|
1332 | 1332 | |
|
1333 | 1333 | #-------------------------------------------------------------------------- |
|
1334 | 1334 | # construct a View object |
|
1335 | 1335 | #-------------------------------------------------------------------------- |
|
1336 | 1336 | |
|
1337 | 1337 | def load_balanced_view(self, targets=None): |
|
1338 | 1338 | """construct a DirectView object. |
|
1339 | 1339 | |
|
1340 | 1340 | If no arguments are specified, create a LoadBalancedView |
|
1341 | 1341 | using all engines. |
|
1342 | 1342 | |
|
1343 | 1343 | Parameters |
|
1344 | 1344 | ---------- |
|
1345 | 1345 | |
|
1346 | 1346 | targets: list,slice,int,etc. [default: use all engines] |
|
1347 | 1347 | The subset of engines across which to load-balance |
|
1348 | 1348 | """ |
|
1349 | 1349 | if targets == 'all': |
|
1350 | 1350 | targets = None |
|
1351 | 1351 | if targets is not None: |
|
1352 | 1352 | targets = self._build_targets(targets)[1] |
|
1353 | 1353 | return LoadBalancedView(client=self, socket=self._task_socket, targets=targets) |
|
1354 | 1354 | |
|
1355 | 1355 | def direct_view(self, targets='all'): |
|
1356 | 1356 | """construct a DirectView object. |
|
1357 | 1357 | |
|
1358 | 1358 | If no targets are specified, create a DirectView using all engines. |
|
1359 | 1359 | |
|
1360 | 1360 | rc.direct_view('all') is distinguished from rc[:] in that 'all' will |
|
1361 | 1361 | evaluate the target engines at each execution, whereas rc[:] will connect to |
|
1362 | 1362 | all *current* engines, and that list will not change. |
|
1363 | 1363 | |
|
1364 | 1364 | That is, 'all' will always use all engines, whereas rc[:] will not use |
|
1365 | 1365 | engines added after the DirectView is constructed. |
|
1366 | 1366 | |
|
1367 | 1367 | Parameters |
|
1368 | 1368 | ---------- |
|
1369 | 1369 | |
|
1370 | 1370 | targets: list,slice,int,etc. [default: use all engines] |
|
1371 | 1371 | The engines to use for the View |
|
1372 | 1372 | """ |
|
1373 | 1373 | single = isinstance(targets, int) |
|
1374 | 1374 | # allow 'all' to be lazily evaluated at each execution |
|
1375 | 1375 | if targets != 'all': |
|
1376 | 1376 | targets = self._build_targets(targets)[1] |
|
1377 | 1377 | if single: |
|
1378 | 1378 | targets = targets[0] |
|
1379 | 1379 | return DirectView(client=self, socket=self._mux_socket, targets=targets) |
|
1380 | 1380 | |
|
1381 | 1381 | #-------------------------------------------------------------------------- |
|
1382 | 1382 | # Query methods |
|
1383 | 1383 | #-------------------------------------------------------------------------- |
|
1384 | 1384 | |
|
1385 | 1385 | @spin_first |
|
1386 | 1386 | def get_result(self, indices_or_msg_ids=None, block=None, owner=True): |
|
1387 | 1387 | """Retrieve a result by msg_id or history index, wrapped in an AsyncResult object. |
|
1388 | 1388 | |
|
1389 | 1389 | If the client already has the results, no request to the Hub will be made. |
|
1390 | 1390 | |
|
1391 | 1391 | This is a convenient way to construct AsyncResult objects, which are wrappers |
|
1392 | 1392 | that include metadata about execution, and allow for awaiting results that |
|
1393 | 1393 | were not submitted by this Client. |
|
1394 | 1394 | |
|
1395 | 1395 | It can also be a convenient way to retrieve the metadata associated with |
|
1396 | 1396 | blocking execution, since it always retrieves |
|
1397 | 1397 | |
|
1398 | 1398 | Examples |
|
1399 | 1399 | -------- |
|
1400 | 1400 | :: |
|
1401 | 1401 | |
|
1402 | 1402 | In [10]: r = client.apply() |
|
1403 | 1403 | |
|
1404 | 1404 | Parameters |
|
1405 | 1405 | ---------- |
|
1406 | 1406 | |
|
1407 | 1407 | indices_or_msg_ids : integer history index, str msg_id, or list of either |
|
1408 | 1408 | The indices or msg_ids of indices to be retrieved |
|
1409 | 1409 | |
|
1410 | 1410 | block : bool |
|
1411 | 1411 | Whether to wait for the result to be done |
|
1412 | 1412 | owner : bool [default: True] |
|
1413 | 1413 | Whether this AsyncResult should own the result. |
|
1414 | 1414 | If so, calling `ar.get()` will remove data from the |
|
1415 | 1415 | client's result and metadata cache. |
|
1416 | 1416 | There should only be one owner of any given msg_id. |
|
1417 | 1417 | |
|
1418 | 1418 | Returns |
|
1419 | 1419 | ------- |
|
1420 | 1420 | |
|
1421 | 1421 | AsyncResult |
|
1422 | 1422 | A single AsyncResult object will always be returned. |
|
1423 | 1423 | |
|
1424 | 1424 | AsyncHubResult |
|
1425 | 1425 | A subclass of AsyncResult that retrieves results from the Hub |
|
1426 | 1426 | |
|
1427 | 1427 | """ |
|
1428 | 1428 | block = self.block if block is None else block |
|
1429 | 1429 | if indices_or_msg_ids is None: |
|
1430 | 1430 | indices_or_msg_ids = -1 |
|
1431 | 1431 | |
|
1432 | 1432 | single_result = False |
|
1433 | 1433 | if not isinstance(indices_or_msg_ids, (list,tuple)): |
|
1434 | 1434 | indices_or_msg_ids = [indices_or_msg_ids] |
|
1435 | 1435 | single_result = True |
|
1436 | 1436 | |
|
1437 | 1437 | theids = [] |
|
1438 | 1438 | for id in indices_or_msg_ids: |
|
1439 | 1439 | if isinstance(id, int): |
|
1440 | 1440 | id = self.history[id] |
|
1441 | 1441 | if not isinstance(id, string_types): |
|
1442 | 1442 | raise TypeError("indices must be str or int, not %r"%id) |
|
1443 | 1443 | theids.append(id) |
|
1444 | 1444 | |
|
1445 | 1445 | local_ids = [msg_id for msg_id in theids if (msg_id in self.outstanding or msg_id in self.results)] |
|
1446 | 1446 | remote_ids = [msg_id for msg_id in theids if msg_id not in local_ids] |
|
1447 | 1447 | |
|
1448 | 1448 | # given single msg_id initially, get_result shot get the result itself, |
|
1449 | 1449 | # not a length-one list |
|
1450 | 1450 | if single_result: |
|
1451 | 1451 | theids = theids[0] |
|
1452 | 1452 | |
|
1453 | 1453 | if remote_ids: |
|
1454 | 1454 | ar = AsyncHubResult(self, msg_ids=theids, owner=owner) |
|
1455 | 1455 | else: |
|
1456 | 1456 | ar = AsyncResult(self, msg_ids=theids, owner=owner) |
|
1457 | 1457 | |
|
1458 | 1458 | if block: |
|
1459 | 1459 | ar.wait() |
|
1460 | 1460 | |
|
1461 | 1461 | return ar |
|
1462 | 1462 | |
|
1463 | 1463 | @spin_first |
|
1464 | 1464 | def resubmit(self, indices_or_msg_ids=None, metadata=None, block=None): |
|
1465 | 1465 | """Resubmit one or more tasks. |
|
1466 | 1466 | |
|
1467 | 1467 | in-flight tasks may not be resubmitted. |
|
1468 | 1468 | |
|
1469 | 1469 | Parameters |
|
1470 | 1470 | ---------- |
|
1471 | 1471 | |
|
1472 | 1472 | indices_or_msg_ids : integer history index, str msg_id, or list of either |
|
1473 | 1473 | The indices or msg_ids of indices to be retrieved |
|
1474 | 1474 | |
|
1475 | 1475 | block : bool |
|
1476 | 1476 | Whether to wait for the result to be done |
|
1477 | 1477 | |
|
1478 | 1478 | Returns |
|
1479 | 1479 | ------- |
|
1480 | 1480 | |
|
1481 | 1481 | AsyncHubResult |
|
1482 | 1482 | A subclass of AsyncResult that retrieves results from the Hub |
|
1483 | 1483 | |
|
1484 | 1484 | """ |
|
1485 | 1485 | block = self.block if block is None else block |
|
1486 | 1486 | if indices_or_msg_ids is None: |
|
1487 | 1487 | indices_or_msg_ids = -1 |
|
1488 | 1488 | |
|
1489 | 1489 | if not isinstance(indices_or_msg_ids, (list,tuple)): |
|
1490 | 1490 | indices_or_msg_ids = [indices_or_msg_ids] |
|
1491 | 1491 | |
|
1492 | 1492 | theids = [] |
|
1493 | 1493 | for id in indices_or_msg_ids: |
|
1494 | 1494 | if isinstance(id, int): |
|
1495 | 1495 | id = self.history[id] |
|
1496 | 1496 | if not isinstance(id, string_types): |
|
1497 | 1497 | raise TypeError("indices must be str or int, not %r"%id) |
|
1498 | 1498 | theids.append(id) |
|
1499 | 1499 | |
|
1500 | 1500 | content = dict(msg_ids = theids) |
|
1501 | 1501 | |
|
1502 | 1502 | self.session.send(self._query_socket, 'resubmit_request', content) |
|
1503 | 1503 | |
|
1504 | 1504 | zmq.select([self._query_socket], [], []) |
|
1505 | 1505 | idents,msg = self.session.recv(self._query_socket, zmq.NOBLOCK) |
|
1506 | 1506 | if self.debug: |
|
1507 | 1507 | pprint(msg) |
|
1508 | 1508 | content = msg['content'] |
|
1509 | 1509 | if content['status'] != 'ok': |
|
1510 | 1510 | raise self._unwrap_exception(content) |
|
1511 | 1511 | mapping = content['resubmitted'] |
|
1512 | 1512 | new_ids = [ mapping[msg_id] for msg_id in theids ] |
|
1513 | 1513 | |
|
1514 | 1514 | ar = AsyncHubResult(self, msg_ids=new_ids) |
|
1515 | 1515 | |
|
1516 | 1516 | if block: |
|
1517 | 1517 | ar.wait() |
|
1518 | 1518 | |
|
1519 | 1519 | return ar |
|
1520 | 1520 | |
|
1521 | 1521 | @spin_first |
|
1522 | 1522 | def result_status(self, msg_ids, status_only=True): |
|
1523 | 1523 | """Check on the status of the result(s) of the apply request with `msg_ids`. |
|
1524 | 1524 | |
|
1525 | 1525 | If status_only is False, then the actual results will be retrieved, else |
|
1526 | 1526 | only the status of the results will be checked. |
|
1527 | 1527 | |
|
1528 | 1528 | Parameters |
|
1529 | 1529 | ---------- |
|
1530 | 1530 | |
|
1531 | 1531 | msg_ids : list of msg_ids |
|
1532 | 1532 | if int: |
|
1533 | 1533 | Passed as index to self.history for convenience. |
|
1534 | 1534 | status_only : bool (default: True) |
|
1535 | 1535 | if False: |
|
1536 | 1536 | Retrieve the actual results of completed tasks. |
|
1537 | 1537 | |
|
1538 | 1538 | Returns |
|
1539 | 1539 | ------- |
|
1540 | 1540 | |
|
1541 | 1541 | results : dict |
|
1542 | 1542 | There will always be the keys 'pending' and 'completed', which will |
|
1543 | 1543 | be lists of msg_ids that are incomplete or complete. If `status_only` |
|
1544 | 1544 | is False, then completed results will be keyed by their `msg_id`. |
|
1545 | 1545 | """ |
|
1546 | 1546 | if not isinstance(msg_ids, (list,tuple)): |
|
1547 | 1547 | msg_ids = [msg_ids] |
|
1548 | 1548 | |
|
1549 | 1549 | theids = [] |
|
1550 | 1550 | for msg_id in msg_ids: |
|
1551 | 1551 | if isinstance(msg_id, int): |
|
1552 | 1552 | msg_id = self.history[msg_id] |
|
1553 | 1553 | if not isinstance(msg_id, string_types): |
|
1554 | 1554 | raise TypeError("msg_ids must be str, not %r"%msg_id) |
|
1555 | 1555 | theids.append(msg_id) |
|
1556 | 1556 | |
|
1557 | 1557 | completed = [] |
|
1558 | 1558 | local_results = {} |
|
1559 | 1559 | |
|
1560 | 1560 | # comment this block out to temporarily disable local shortcut: |
|
1561 | 1561 | for msg_id in theids: |
|
1562 | 1562 | if msg_id in self.results: |
|
1563 | 1563 | completed.append(msg_id) |
|
1564 | 1564 | local_results[msg_id] = self.results[msg_id] |
|
1565 | 1565 | theids.remove(msg_id) |
|
1566 | 1566 | |
|
1567 | 1567 | if theids: # some not locally cached |
|
1568 | 1568 | content = dict(msg_ids=theids, status_only=status_only) |
|
1569 | 1569 | msg = self.session.send(self._query_socket, "result_request", content=content) |
|
1570 | 1570 | zmq.select([self._query_socket], [], []) |
|
1571 | 1571 | idents,msg = self.session.recv(self._query_socket, zmq.NOBLOCK) |
|
1572 | 1572 | if self.debug: |
|
1573 | 1573 | pprint(msg) |
|
1574 | 1574 | content = msg['content'] |
|
1575 | 1575 | if content['status'] != 'ok': |
|
1576 | 1576 | raise self._unwrap_exception(content) |
|
1577 | 1577 | buffers = msg['buffers'] |
|
1578 | 1578 | else: |
|
1579 | 1579 | content = dict(completed=[],pending=[]) |
|
1580 | 1580 | |
|
1581 | 1581 | content['completed'].extend(completed) |
|
1582 | 1582 | |
|
1583 | 1583 | if status_only: |
|
1584 | 1584 | return content |
|
1585 | 1585 | |
|
1586 | 1586 | failures = [] |
|
1587 | 1587 | # load cached results into result: |
|
1588 | 1588 | content.update(local_results) |
|
1589 | 1589 | |
|
1590 | 1590 | # update cache with results: |
|
1591 | 1591 | for msg_id in sorted(theids): |
|
1592 | 1592 | if msg_id in content['completed']: |
|
1593 | 1593 | rec = content[msg_id] |
|
1594 | 1594 | parent = extract_dates(rec['header']) |
|
1595 | 1595 | header = extract_dates(rec['result_header']) |
|
1596 | 1596 | rcontent = rec['result_content'] |
|
1597 | 1597 | iodict = rec['io'] |
|
1598 | 1598 | if isinstance(rcontent, str): |
|
1599 | 1599 | rcontent = self.session.unpack(rcontent) |
|
1600 | 1600 | |
|
1601 | 1601 | md = self.metadata[msg_id] |
|
1602 | 1602 | md_msg = dict( |
|
1603 | 1603 | content=rcontent, |
|
1604 | 1604 | parent_header=parent, |
|
1605 | 1605 | header=header, |
|
1606 | 1606 | metadata=rec['result_metadata'], |
|
1607 | 1607 | ) |
|
1608 | 1608 | md.update(self._extract_metadata(md_msg)) |
|
1609 | 1609 | if rec.get('received'): |
|
1610 | 1610 | md['received'] = parse_date(rec['received']) |
|
1611 | 1611 | md.update(iodict) |
|
1612 | 1612 | |
|
1613 | 1613 | if rcontent['status'] == 'ok': |
|
1614 | 1614 | if header['msg_type'] == 'apply_reply': |
|
1615 | 1615 | res,buffers = serialize.deserialize_object(buffers) |
|
1616 | 1616 | elif header['msg_type'] == 'execute_reply': |
|
1617 | 1617 | res = ExecuteReply(msg_id, rcontent, md) |
|
1618 | 1618 | else: |
|
1619 | 1619 | raise KeyError("unhandled msg type: %r" % header['msg_type']) |
|
1620 | 1620 | else: |
|
1621 | 1621 | res = self._unwrap_exception(rcontent) |
|
1622 | 1622 | failures.append(res) |
|
1623 | 1623 | |
|
1624 | 1624 | self.results[msg_id] = res |
|
1625 | 1625 | content[msg_id] = res |
|
1626 | 1626 | |
|
1627 | 1627 | if len(theids) == 1 and failures: |
|
1628 | 1628 | raise failures[0] |
|
1629 | 1629 | |
|
1630 | 1630 | error.collect_exceptions(failures, "result_status") |
|
1631 | 1631 | return content |
|
1632 | 1632 | |
|
1633 | 1633 | @spin_first |
|
1634 | 1634 | def queue_status(self, targets='all', verbose=False): |
|
1635 | 1635 | """Fetch the status of engine queues. |
|
1636 | 1636 | |
|
1637 | 1637 | Parameters |
|
1638 | 1638 | ---------- |
|
1639 | 1639 | |
|
1640 | 1640 | targets : int/str/list of ints/strs |
|
1641 | 1641 | the engines whose states are to be queried. |
|
1642 | 1642 | default : all |
|
1643 | 1643 | verbose : bool |
|
1644 | 1644 | Whether to return lengths only, or lists of ids for each element |
|
1645 | 1645 | """ |
|
1646 | 1646 | if targets == 'all': |
|
1647 | 1647 | # allow 'all' to be evaluated on the engine |
|
1648 | 1648 | engine_ids = None |
|
1649 | 1649 | else: |
|
1650 | 1650 | engine_ids = self._build_targets(targets)[1] |
|
1651 | 1651 | content = dict(targets=engine_ids, verbose=verbose) |
|
1652 | 1652 | self.session.send(self._query_socket, "queue_request", content=content) |
|
1653 | 1653 | idents,msg = self.session.recv(self._query_socket, 0) |
|
1654 | 1654 | if self.debug: |
|
1655 | 1655 | pprint(msg) |
|
1656 | 1656 | content = msg['content'] |
|
1657 | 1657 | status = content.pop('status') |
|
1658 | 1658 | if status != 'ok': |
|
1659 | 1659 | raise self._unwrap_exception(content) |
|
1660 | 1660 | content = rekey(content) |
|
1661 | 1661 | if isinstance(targets, int): |
|
1662 | 1662 | return content[targets] |
|
1663 | 1663 | else: |
|
1664 | 1664 | return content |
|
1665 | 1665 | |
|
1666 | 1666 | def _build_msgids_from_target(self, targets=None): |
|
1667 | 1667 | """Build a list of msg_ids from the list of engine targets""" |
|
1668 | 1668 | if not targets: # needed as _build_targets otherwise uses all engines |
|
1669 | 1669 | return [] |
|
1670 | 1670 | target_ids = self._build_targets(targets)[0] |
|
1671 | 1671 | return [md_id for md_id in self.metadata if self.metadata[md_id]["engine_uuid"] in target_ids] |
|
1672 | 1672 | |
|
1673 | 1673 | def _build_msgids_from_jobs(self, jobs=None): |
|
1674 | 1674 | """Build a list of msg_ids from "jobs" """ |
|
1675 | 1675 | if not jobs: |
|
1676 | 1676 | return [] |
|
1677 | 1677 | msg_ids = [] |
|
1678 | 1678 | if isinstance(jobs, string_types + (AsyncResult,)): |
|
1679 | 1679 | jobs = [jobs] |
|
1680 | 1680 | bad_ids = [obj for obj in jobs if not isinstance(obj, string_types + (AsyncResult,))] |
|
1681 | 1681 | if bad_ids: |
|
1682 | 1682 | raise TypeError("Invalid msg_id type %r, expected str or AsyncResult"%bad_ids[0]) |
|
1683 | 1683 | for j in jobs: |
|
1684 | 1684 | if isinstance(j, AsyncResult): |
|
1685 | 1685 | msg_ids.extend(j.msg_ids) |
|
1686 | 1686 | else: |
|
1687 | 1687 | msg_ids.append(j) |
|
1688 | 1688 | return msg_ids |
|
1689 | 1689 | |
|
1690 | 1690 | def purge_local_results(self, jobs=[], targets=[]): |
|
1691 | 1691 | """Clears the client caches of results and their metadata. |
|
1692 | 1692 | |
|
1693 | 1693 | Individual results can be purged by msg_id, or the entire |
|
1694 | 1694 | history of specific targets can be purged. |
|
1695 | 1695 | |
|
1696 | 1696 | Use `purge_local_results('all')` to scrub everything from the Clients's |
|
1697 | 1697 | results and metadata caches. |
|
1698 | 1698 | |
|
1699 | 1699 | After this call all `AsyncResults` are invalid and should be discarded. |
|
1700 | 1700 | |
|
1701 | 1701 | If you must "reget" the results, you can still do so by using |
|
1702 | 1702 | `client.get_result(msg_id)` or `client.get_result(asyncresult)`. This will |
|
1703 | 1703 | redownload the results from the hub if they are still available |
|
1704 | 1704 | (i.e `client.purge_hub_results(...)` has not been called. |
|
1705 | 1705 | |
|
1706 | 1706 | Parameters |
|
1707 | 1707 | ---------- |
|
1708 | 1708 | |
|
1709 | 1709 | jobs : str or list of str or AsyncResult objects |
|
1710 | 1710 | the msg_ids whose results should be purged. |
|
1711 | 1711 | targets : int/list of ints |
|
1712 | 1712 | The engines, by integer ID, whose entire result histories are to be purged. |
|
1713 | 1713 | |
|
1714 | 1714 | Raises |
|
1715 | 1715 | ------ |
|
1716 | 1716 | |
|
1717 | 1717 | RuntimeError : if any of the tasks to be purged are still outstanding. |
|
1718 | 1718 | |
|
1719 | 1719 | """ |
|
1720 | 1720 | if not targets and not jobs: |
|
1721 | 1721 | raise ValueError("Must specify at least one of `targets` and `jobs`") |
|
1722 | 1722 | |
|
1723 | 1723 | if jobs == 'all': |
|
1724 | 1724 | if self.outstanding: |
|
1725 | 1725 | raise RuntimeError("Can't purge outstanding tasks: %s" % self.outstanding) |
|
1726 | 1726 | self.results.clear() |
|
1727 | 1727 | self.metadata.clear() |
|
1728 | 1728 | else: |
|
1729 | 1729 | msg_ids = set() |
|
1730 | 1730 | msg_ids.update(self._build_msgids_from_target(targets)) |
|
1731 | 1731 | msg_ids.update(self._build_msgids_from_jobs(jobs)) |
|
1732 | 1732 | still_outstanding = self.outstanding.intersection(msg_ids) |
|
1733 | 1733 | if still_outstanding: |
|
1734 | 1734 | raise RuntimeError("Can't purge outstanding tasks: %s" % still_outstanding) |
|
1735 | 1735 | for mid in msg_ids: |
|
1736 | 1736 | self.results.pop(mid, None) |
|
1737 | 1737 | self.metadata.pop(mid, None) |
|
1738 | 1738 | |
|
1739 | 1739 | |
|
1740 | 1740 | @spin_first |
|
1741 | 1741 | def purge_hub_results(self, jobs=[], targets=[]): |
|
1742 | 1742 | """Tell the Hub to forget results. |
|
1743 | 1743 | |
|
1744 | 1744 | Individual results can be purged by msg_id, or the entire |
|
1745 | 1745 | history of specific targets can be purged. |
|
1746 | 1746 | |
|
1747 | 1747 | Use `purge_results('all')` to scrub everything from the Hub's db. |
|
1748 | 1748 | |
|
1749 | 1749 | Parameters |
|
1750 | 1750 | ---------- |
|
1751 | 1751 | |
|
1752 | 1752 | jobs : str or list of str or AsyncResult objects |
|
1753 | 1753 | the msg_ids whose results should be forgotten. |
|
1754 | 1754 | targets : int/str/list of ints/strs |
|
1755 | 1755 | The targets, by int_id, whose entire history is to be purged. |
|
1756 | 1756 | |
|
1757 | 1757 | default : None |
|
1758 | 1758 | """ |
|
1759 | 1759 | if not targets and not jobs: |
|
1760 | 1760 | raise ValueError("Must specify at least one of `targets` and `jobs`") |
|
1761 | 1761 | if targets: |
|
1762 | 1762 | targets = self._build_targets(targets)[1] |
|
1763 | 1763 | |
|
1764 | 1764 | # construct msg_ids from jobs |
|
1765 | 1765 | if jobs == 'all': |
|
1766 | 1766 | msg_ids = jobs |
|
1767 | 1767 | else: |
|
1768 | 1768 | msg_ids = self._build_msgids_from_jobs(jobs) |
|
1769 | 1769 | |
|
1770 | 1770 | content = dict(engine_ids=targets, msg_ids=msg_ids) |
|
1771 | 1771 | self.session.send(self._query_socket, "purge_request", content=content) |
|
1772 | 1772 | idents, msg = self.session.recv(self._query_socket, 0) |
|
1773 | 1773 | if self.debug: |
|
1774 | 1774 | pprint(msg) |
|
1775 | 1775 | content = msg['content'] |
|
1776 | 1776 | if content['status'] != 'ok': |
|
1777 | 1777 | raise self._unwrap_exception(content) |
|
1778 | 1778 | |
|
1779 | 1779 | def purge_results(self, jobs=[], targets=[]): |
|
1780 | 1780 | """Clears the cached results from both the hub and the local client |
|
1781 | 1781 | |
|
1782 | 1782 | Individual results can be purged by msg_id, or the entire |
|
1783 | 1783 | history of specific targets can be purged. |
|
1784 | 1784 | |
|
1785 | 1785 | Use `purge_results('all')` to scrub every cached result from both the Hub's and |
|
1786 | 1786 | the Client's db. |
|
1787 | 1787 | |
|
1788 | 1788 | Equivalent to calling both `purge_hub_results()` and `purge_client_results()` with |
|
1789 | 1789 | the same arguments. |
|
1790 | 1790 | |
|
1791 | 1791 | Parameters |
|
1792 | 1792 | ---------- |
|
1793 | 1793 | |
|
1794 | 1794 | jobs : str or list of str or AsyncResult objects |
|
1795 | 1795 | the msg_ids whose results should be forgotten. |
|
1796 | 1796 | targets : int/str/list of ints/strs |
|
1797 | 1797 | The targets, by int_id, whose entire history is to be purged. |
|
1798 | 1798 | |
|
1799 | 1799 | default : None |
|
1800 | 1800 | """ |
|
1801 | 1801 | self.purge_local_results(jobs=jobs, targets=targets) |
|
1802 | 1802 | self.purge_hub_results(jobs=jobs, targets=targets) |
|
1803 | 1803 | |
|
1804 | 1804 | def purge_everything(self): |
|
1805 | 1805 | """Clears all content from previous Tasks from both the hub and the local client |
|
1806 | 1806 | |
|
1807 | 1807 | In addition to calling `purge_results("all")` it also deletes the history and |
|
1808 | 1808 | other bookkeeping lists. |
|
1809 | 1809 | """ |
|
1810 | 1810 | self.purge_results("all") |
|
1811 | 1811 | self.history = [] |
|
1812 | 1812 | self.session.digest_history.clear() |
|
1813 | 1813 | |
|
1814 | 1814 | @spin_first |
|
1815 | 1815 | def hub_history(self): |
|
1816 | 1816 | """Get the Hub's history |
|
1817 | 1817 | |
|
1818 | 1818 | Just like the Client, the Hub has a history, which is a list of msg_ids. |
|
1819 | 1819 | This will contain the history of all clients, and, depending on configuration, |
|
1820 | 1820 | may contain history across multiple cluster sessions. |
|
1821 | 1821 | |
|
1822 | 1822 | Any msg_id returned here is a valid argument to `get_result`. |
|
1823 | 1823 | |
|
1824 | 1824 | Returns |
|
1825 | 1825 | ------- |
|
1826 | 1826 | |
|
1827 | 1827 | msg_ids : list of strs |
|
1828 | 1828 | list of all msg_ids, ordered by task submission time. |
|
1829 | 1829 | """ |
|
1830 | 1830 | |
|
1831 | 1831 | self.session.send(self._query_socket, "history_request", content={}) |
|
1832 | 1832 | idents, msg = self.session.recv(self._query_socket, 0) |
|
1833 | 1833 | |
|
1834 | 1834 | if self.debug: |
|
1835 | 1835 | pprint(msg) |
|
1836 | 1836 | content = msg['content'] |
|
1837 | 1837 | if content['status'] != 'ok': |
|
1838 | 1838 | raise self._unwrap_exception(content) |
|
1839 | 1839 | else: |
|
1840 | 1840 | return content['history'] |
|
1841 | 1841 | |
|
1842 | 1842 | @spin_first |
|
1843 | 1843 | def db_query(self, query, keys=None): |
|
1844 | 1844 | """Query the Hub's TaskRecord database |
|
1845 | 1845 | |
|
1846 | 1846 | This will return a list of task record dicts that match `query` |
|
1847 | 1847 | |
|
1848 | 1848 | Parameters |
|
1849 | 1849 | ---------- |
|
1850 | 1850 | |
|
1851 | 1851 | query : mongodb query dict |
|
1852 | 1852 | The search dict. See mongodb query docs for details. |
|
1853 | 1853 | keys : list of strs [optional] |
|
1854 | 1854 | The subset of keys to be returned. The default is to fetch everything but buffers. |
|
1855 | 1855 | 'msg_id' will *always* be included. |
|
1856 | 1856 | """ |
|
1857 | 1857 | if isinstance(keys, string_types): |
|
1858 | 1858 | keys = [keys] |
|
1859 | 1859 | content = dict(query=query, keys=keys) |
|
1860 | 1860 | self.session.send(self._query_socket, "db_request", content=content) |
|
1861 | 1861 | idents, msg = self.session.recv(self._query_socket, 0) |
|
1862 | 1862 | if self.debug: |
|
1863 | 1863 | pprint(msg) |
|
1864 | 1864 | content = msg['content'] |
|
1865 | 1865 | if content['status'] != 'ok': |
|
1866 | 1866 | raise self._unwrap_exception(content) |
|
1867 | 1867 | |
|
1868 | 1868 | records = content['records'] |
|
1869 | 1869 | |
|
1870 | 1870 | buffer_lens = content['buffer_lens'] |
|
1871 | 1871 | result_buffer_lens = content['result_buffer_lens'] |
|
1872 | 1872 | buffers = msg['buffers'] |
|
1873 | 1873 | has_bufs = buffer_lens is not None |
|
1874 | 1874 | has_rbufs = result_buffer_lens is not None |
|
1875 | 1875 | for i,rec in enumerate(records): |
|
1876 | 1876 | # unpack datetime objects |
|
1877 | 1877 | for hkey in ('header', 'result_header'): |
|
1878 | 1878 | if hkey in rec: |
|
1879 | 1879 | rec[hkey] = extract_dates(rec[hkey]) |
|
1880 | 1880 | for dtkey in ('submitted', 'started', 'completed', 'received'): |
|
1881 | 1881 | if dtkey in rec: |
|
1882 | 1882 | rec[dtkey] = parse_date(rec[dtkey]) |
|
1883 | 1883 | # relink buffers |
|
1884 | 1884 | if has_bufs: |
|
1885 | 1885 | blen = buffer_lens[i] |
|
1886 | 1886 | rec['buffers'], buffers = buffers[:blen],buffers[blen:] |
|
1887 | 1887 | if has_rbufs: |
|
1888 | 1888 | blen = result_buffer_lens[i] |
|
1889 | 1889 | rec['result_buffers'], buffers = buffers[:blen],buffers[blen:] |
|
1890 | 1890 | |
|
1891 | 1891 | return records |
|
1892 | 1892 | |
|
1893 | 1893 | __all__ = [ 'Client' ] |
@@ -1,276 +1,276 b'' | |||
|
1 | 1 | """Remote Functions and decorators for Views.""" |
|
2 | 2 | |
|
3 | 3 | # Copyright (c) IPython Development Team. |
|
4 | 4 | # Distributed under the terms of the Modified BSD License. |
|
5 | 5 | |
|
6 | 6 | from __future__ import division |
|
7 | 7 | |
|
8 | 8 | import sys |
|
9 | 9 | import warnings |
|
10 | 10 | |
|
11 |
from |
|
|
11 | from decorator import decorator | |
|
12 | 12 | from IPython.testing.skipdoctest import skip_doctest |
|
13 | 13 | |
|
14 | 14 | from . import map as Map |
|
15 | 15 | from .asyncresult import AsyncMapResult |
|
16 | 16 | |
|
17 | 17 | #----------------------------------------------------------------------------- |
|
18 | 18 | # Functions and Decorators |
|
19 | 19 | #----------------------------------------------------------------------------- |
|
20 | 20 | |
|
21 | 21 | @skip_doctest |
|
22 | 22 | def remote(view, block=None, **flags): |
|
23 | 23 | """Turn a function into a remote function. |
|
24 | 24 | |
|
25 | 25 | This method can be used for map: |
|
26 | 26 | |
|
27 | 27 | In [1]: @remote(view,block=True) |
|
28 | 28 | ...: def func(a): |
|
29 | 29 | ...: pass |
|
30 | 30 | """ |
|
31 | 31 | |
|
32 | 32 | def remote_function(f): |
|
33 | 33 | return RemoteFunction(view, f, block=block, **flags) |
|
34 | 34 | return remote_function |
|
35 | 35 | |
|
36 | 36 | @skip_doctest |
|
37 | 37 | def parallel(view, dist='b', block=None, ordered=True, **flags): |
|
38 | 38 | """Turn a function into a parallel remote function. |
|
39 | 39 | |
|
40 | 40 | This method can be used for map: |
|
41 | 41 | |
|
42 | 42 | In [1]: @parallel(view, block=True) |
|
43 | 43 | ...: def func(a): |
|
44 | 44 | ...: pass |
|
45 | 45 | """ |
|
46 | 46 | |
|
47 | 47 | def parallel_function(f): |
|
48 | 48 | return ParallelFunction(view, f, dist=dist, block=block, ordered=ordered, **flags) |
|
49 | 49 | return parallel_function |
|
50 | 50 | |
|
51 | 51 | def getname(f): |
|
52 | 52 | """Get the name of an object. |
|
53 | 53 | |
|
54 | 54 | For use in case of callables that are not functions, and |
|
55 | 55 | thus may not have __name__ defined. |
|
56 | 56 | |
|
57 | 57 | Order: f.__name__ > f.name > str(f) |
|
58 | 58 | """ |
|
59 | 59 | try: |
|
60 | 60 | return f.__name__ |
|
61 | 61 | except: |
|
62 | 62 | pass |
|
63 | 63 | try: |
|
64 | 64 | return f.name |
|
65 | 65 | except: |
|
66 | 66 | pass |
|
67 | 67 | |
|
68 | 68 | return str(f) |
|
69 | 69 | |
|
70 | 70 | @decorator |
|
71 | 71 | def sync_view_results(f, self, *args, **kwargs): |
|
72 | 72 | """sync relevant results from self.client to our results attribute. |
|
73 | 73 | |
|
74 | 74 | This is a clone of view.sync_results, but for remote functions |
|
75 | 75 | """ |
|
76 | 76 | view = self.view |
|
77 | 77 | if view._in_sync_results: |
|
78 | 78 | return f(self, *args, **kwargs) |
|
79 | 79 | view._in_sync_results = True |
|
80 | 80 | try: |
|
81 | 81 | ret = f(self, *args, **kwargs) |
|
82 | 82 | finally: |
|
83 | 83 | view._in_sync_results = False |
|
84 | 84 | view._sync_results() |
|
85 | 85 | return ret |
|
86 | 86 | |
|
87 | 87 | #-------------------------------------------------------------------------- |
|
88 | 88 | # Classes |
|
89 | 89 | #-------------------------------------------------------------------------- |
|
90 | 90 | |
|
91 | 91 | class RemoteFunction(object): |
|
92 | 92 | """Turn an existing function into a remote function. |
|
93 | 93 | |
|
94 | 94 | Parameters |
|
95 | 95 | ---------- |
|
96 | 96 | |
|
97 | 97 | view : View instance |
|
98 | 98 | The view to be used for execution |
|
99 | 99 | f : callable |
|
100 | 100 | The function to be wrapped into a remote function |
|
101 | 101 | block : bool [default: None] |
|
102 | 102 | Whether to wait for results or not. The default behavior is |
|
103 | 103 | to use the current `block` attribute of `view` |
|
104 | 104 | |
|
105 | 105 | **flags : remaining kwargs are passed to View.temp_flags |
|
106 | 106 | """ |
|
107 | 107 | |
|
108 | 108 | view = None # the remote connection |
|
109 | 109 | func = None # the wrapped function |
|
110 | 110 | block = None # whether to block |
|
111 | 111 | flags = None # dict of extra kwargs for temp_flags |
|
112 | 112 | |
|
113 | 113 | def __init__(self, view, f, block=None, **flags): |
|
114 | 114 | self.view = view |
|
115 | 115 | self.func = f |
|
116 | 116 | self.block=block |
|
117 | 117 | self.flags=flags |
|
118 | 118 | |
|
119 | 119 | def __call__(self, *args, **kwargs): |
|
120 | 120 | block = self.view.block if self.block is None else self.block |
|
121 | 121 | with self.view.temp_flags(block=block, **self.flags): |
|
122 | 122 | return self.view.apply(self.func, *args, **kwargs) |
|
123 | 123 | |
|
124 | 124 | |
|
125 | 125 | class ParallelFunction(RemoteFunction): |
|
126 | 126 | """Class for mapping a function to sequences. |
|
127 | 127 | |
|
128 | 128 | This will distribute the sequences according the a mapper, and call |
|
129 | 129 | the function on each sub-sequence. If called via map, then the function |
|
130 | 130 | will be called once on each element, rather that each sub-sequence. |
|
131 | 131 | |
|
132 | 132 | Parameters |
|
133 | 133 | ---------- |
|
134 | 134 | |
|
135 | 135 | view : View instance |
|
136 | 136 | The view to be used for execution |
|
137 | 137 | f : callable |
|
138 | 138 | The function to be wrapped into a remote function |
|
139 | 139 | dist : str [default: 'b'] |
|
140 | 140 | The key for which mapObject to use to distribute sequences |
|
141 | 141 | options are: |
|
142 | 142 | |
|
143 | 143 | * 'b' : use contiguous chunks in order |
|
144 | 144 | * 'r' : use round-robin striping |
|
145 | 145 | |
|
146 | 146 | block : bool [default: None] |
|
147 | 147 | Whether to wait for results or not. The default behavior is |
|
148 | 148 | to use the current `block` attribute of `view` |
|
149 | 149 | chunksize : int or None |
|
150 | 150 | The size of chunk to use when breaking up sequences in a load-balanced manner |
|
151 | 151 | ordered : bool [default: True] |
|
152 | 152 | Whether the result should be kept in order. If False, |
|
153 | 153 | results become available as they arrive, regardless of submission order. |
|
154 | 154 | **flags |
|
155 | 155 | remaining kwargs are passed to View.temp_flags |
|
156 | 156 | """ |
|
157 | 157 | |
|
158 | 158 | chunksize = None |
|
159 | 159 | ordered = None |
|
160 | 160 | mapObject = None |
|
161 | 161 | _mapping = False |
|
162 | 162 | |
|
163 | 163 | def __init__(self, view, f, dist='b', block=None, chunksize=None, ordered=True, **flags): |
|
164 | 164 | super(ParallelFunction, self).__init__(view, f, block=block, **flags) |
|
165 | 165 | self.chunksize = chunksize |
|
166 | 166 | self.ordered = ordered |
|
167 | 167 | |
|
168 | 168 | mapClass = Map.dists[dist] |
|
169 | 169 | self.mapObject = mapClass() |
|
170 | 170 | |
|
171 | 171 | @sync_view_results |
|
172 | 172 | def __call__(self, *sequences): |
|
173 | 173 | client = self.view.client |
|
174 | 174 | |
|
175 | 175 | lens = [] |
|
176 | 176 | maxlen = minlen = -1 |
|
177 | 177 | for i, seq in enumerate(sequences): |
|
178 | 178 | try: |
|
179 | 179 | n = len(seq) |
|
180 | 180 | except Exception: |
|
181 | 181 | seq = list(seq) |
|
182 | 182 | if isinstance(sequences, tuple): |
|
183 | 183 | # can't alter a tuple |
|
184 | 184 | sequences = list(sequences) |
|
185 | 185 | sequences[i] = seq |
|
186 | 186 | n = len(seq) |
|
187 | 187 | if n > maxlen: |
|
188 | 188 | maxlen = n |
|
189 | 189 | if minlen == -1 or n < minlen: |
|
190 | 190 | minlen = n |
|
191 | 191 | lens.append(n) |
|
192 | 192 | |
|
193 | 193 | if maxlen == 0: |
|
194 | 194 | # nothing to iterate over |
|
195 | 195 | return [] |
|
196 | 196 | |
|
197 | 197 | # check that the length of sequences match |
|
198 | 198 | if not self._mapping and minlen != maxlen: |
|
199 | 199 | msg = 'all sequences must have equal length, but have %s' % lens |
|
200 | 200 | raise ValueError(msg) |
|
201 | 201 | |
|
202 | 202 | balanced = 'Balanced' in self.view.__class__.__name__ |
|
203 | 203 | if balanced: |
|
204 | 204 | if self.chunksize: |
|
205 | 205 | nparts = maxlen // self.chunksize + int(maxlen % self.chunksize > 0) |
|
206 | 206 | else: |
|
207 | 207 | nparts = maxlen |
|
208 | 208 | targets = [None]*nparts |
|
209 | 209 | else: |
|
210 | 210 | if self.chunksize: |
|
211 | 211 | warnings.warn("`chunksize` is ignored unless load balancing", UserWarning) |
|
212 | 212 | # multiplexed: |
|
213 | 213 | targets = self.view.targets |
|
214 | 214 | # 'all' is lazily evaluated at execution time, which is now: |
|
215 | 215 | if targets == 'all': |
|
216 | 216 | targets = client._build_targets(targets)[1] |
|
217 | 217 | elif isinstance(targets, int): |
|
218 | 218 | # single-engine view, targets must be iterable |
|
219 | 219 | targets = [targets] |
|
220 | 220 | nparts = len(targets) |
|
221 | 221 | |
|
222 | 222 | msg_ids = [] |
|
223 | 223 | for index, t in enumerate(targets): |
|
224 | 224 | args = [] |
|
225 | 225 | for seq in sequences: |
|
226 | 226 | part = self.mapObject.getPartition(seq, index, nparts, maxlen) |
|
227 | 227 | args.append(part) |
|
228 | 228 | |
|
229 | 229 | if sum([len(arg) for arg in args]) == 0: |
|
230 | 230 | continue |
|
231 | 231 | |
|
232 | 232 | if self._mapping: |
|
233 | 233 | if sys.version_info[0] >= 3: |
|
234 | 234 | f = lambda f, *sequences: list(map(f, *sequences)) |
|
235 | 235 | else: |
|
236 | 236 | f = map |
|
237 | 237 | args = [self.func] + args |
|
238 | 238 | else: |
|
239 | 239 | f=self.func |
|
240 | 240 | |
|
241 | 241 | view = self.view if balanced else client[t] |
|
242 | 242 | with view.temp_flags(block=False, **self.flags): |
|
243 | 243 | ar = view.apply(f, *args) |
|
244 | 244 | |
|
245 | 245 | msg_ids.extend(ar.msg_ids) |
|
246 | 246 | |
|
247 | 247 | r = AsyncMapResult(self.view.client, msg_ids, self.mapObject, |
|
248 | 248 | fname=getname(self.func), |
|
249 | 249 | ordered=self.ordered |
|
250 | 250 | ) |
|
251 | 251 | |
|
252 | 252 | if self.block: |
|
253 | 253 | try: |
|
254 | 254 | return r.get() |
|
255 | 255 | except KeyboardInterrupt: |
|
256 | 256 | return r |
|
257 | 257 | else: |
|
258 | 258 | return r |
|
259 | 259 | |
|
260 | 260 | def map(self, *sequences): |
|
261 | 261 | """call a function on each element of one or more sequence(s) remotely. |
|
262 | 262 | This should behave very much like the builtin map, but return an AsyncMapResult |
|
263 | 263 | if self.block is False. |
|
264 | 264 | |
|
265 | 265 | That means it can take generators (will be cast to lists locally), |
|
266 | 266 | and mismatched sequence lengths will be padded with None. |
|
267 | 267 | """ |
|
268 | 268 | # set _mapping as a flag for use inside self.__call__ |
|
269 | 269 | self._mapping = True |
|
270 | 270 | try: |
|
271 | 271 | ret = self(*sequences) |
|
272 | 272 | finally: |
|
273 | 273 | self._mapping = False |
|
274 | 274 | return ret |
|
275 | 275 | |
|
276 | 276 | __all__ = ['remote', 'parallel', 'RemoteFunction', 'ParallelFunction'] |
@@ -1,1125 +1,1125 b'' | |||
|
1 | 1 | """Views of remote engines.""" |
|
2 | 2 | |
|
3 | 3 | # Copyright (c) IPython Development Team. |
|
4 | 4 | # Distributed under the terms of the Modified BSD License. |
|
5 | 5 | |
|
6 | 6 | from __future__ import print_function |
|
7 | 7 | |
|
8 | 8 | import imp |
|
9 | 9 | import sys |
|
10 | 10 | import warnings |
|
11 | 11 | from contextlib import contextmanager |
|
12 | 12 | from types import ModuleType |
|
13 | 13 | |
|
14 | 14 | import zmq |
|
15 | 15 | |
|
16 | 16 | from IPython.testing.skipdoctest import skip_doctest |
|
17 | 17 | from IPython.utils import pickleutil |
|
18 | 18 | from IPython.utils.traitlets import ( |
|
19 | 19 | HasTraits, Any, Bool, List, Dict, Set, Instance, CFloat, Integer |
|
20 | 20 | ) |
|
21 |
from |
|
|
21 | from decorator import decorator | |
|
22 | 22 | |
|
23 | 23 | from IPython.parallel import util |
|
24 | 24 | from IPython.parallel.controller.dependency import Dependency, dependent |
|
25 | 25 | from IPython.utils.py3compat import string_types, iteritems, PY3 |
|
26 | 26 | |
|
27 | 27 | from . import map as Map |
|
28 | 28 | from .asyncresult import AsyncResult, AsyncMapResult |
|
29 | 29 | from .remotefunction import ParallelFunction, parallel, remote, getname |
|
30 | 30 | |
|
31 | 31 | #----------------------------------------------------------------------------- |
|
32 | 32 | # Decorators |
|
33 | 33 | #----------------------------------------------------------------------------- |
|
34 | 34 | |
|
35 | 35 | @decorator |
|
36 | 36 | def save_ids(f, self, *args, **kwargs): |
|
37 | 37 | """Keep our history and outstanding attributes up to date after a method call.""" |
|
38 | 38 | n_previous = len(self.client.history) |
|
39 | 39 | try: |
|
40 | 40 | ret = f(self, *args, **kwargs) |
|
41 | 41 | finally: |
|
42 | 42 | nmsgs = len(self.client.history) - n_previous |
|
43 | 43 | msg_ids = self.client.history[-nmsgs:] |
|
44 | 44 | self.history.extend(msg_ids) |
|
45 | 45 | self.outstanding.update(msg_ids) |
|
46 | 46 | return ret |
|
47 | 47 | |
|
48 | 48 | @decorator |
|
49 | 49 | def sync_results(f, self, *args, **kwargs): |
|
50 | 50 | """sync relevant results from self.client to our results attribute.""" |
|
51 | 51 | if self._in_sync_results: |
|
52 | 52 | return f(self, *args, **kwargs) |
|
53 | 53 | self._in_sync_results = True |
|
54 | 54 | try: |
|
55 | 55 | ret = f(self, *args, **kwargs) |
|
56 | 56 | finally: |
|
57 | 57 | self._in_sync_results = False |
|
58 | 58 | self._sync_results() |
|
59 | 59 | return ret |
|
60 | 60 | |
|
61 | 61 | @decorator |
|
62 | 62 | def spin_after(f, self, *args, **kwargs): |
|
63 | 63 | """call spin after the method.""" |
|
64 | 64 | ret = f(self, *args, **kwargs) |
|
65 | 65 | self.spin() |
|
66 | 66 | return ret |
|
67 | 67 | |
|
68 | 68 | #----------------------------------------------------------------------------- |
|
69 | 69 | # Classes |
|
70 | 70 | #----------------------------------------------------------------------------- |
|
71 | 71 | |
|
72 | 72 | @skip_doctest |
|
73 | 73 | class View(HasTraits): |
|
74 | 74 | """Base View class for more convenint apply(f,*args,**kwargs) syntax via attributes. |
|
75 | 75 | |
|
76 | 76 | Don't use this class, use subclasses. |
|
77 | 77 | |
|
78 | 78 | Methods |
|
79 | 79 | ------- |
|
80 | 80 | |
|
81 | 81 | spin |
|
82 | 82 | flushes incoming results and registration state changes |
|
83 | 83 | control methods spin, and requesting `ids` also ensures up to date |
|
84 | 84 | |
|
85 | 85 | wait |
|
86 | 86 | wait on one or more msg_ids |
|
87 | 87 | |
|
88 | 88 | execution methods |
|
89 | 89 | apply |
|
90 | 90 | legacy: execute, run |
|
91 | 91 | |
|
92 | 92 | data movement |
|
93 | 93 | push, pull, scatter, gather |
|
94 | 94 | |
|
95 | 95 | query methods |
|
96 | 96 | get_result, queue_status, purge_results, result_status |
|
97 | 97 | |
|
98 | 98 | control methods |
|
99 | 99 | abort, shutdown |
|
100 | 100 | |
|
101 | 101 | """ |
|
102 | 102 | # flags |
|
103 | 103 | block=Bool(False) |
|
104 | 104 | track=Bool(True) |
|
105 | 105 | targets = Any() |
|
106 | 106 | |
|
107 | 107 | history=List() |
|
108 | 108 | outstanding = Set() |
|
109 | 109 | results = Dict() |
|
110 | 110 | client = Instance('IPython.parallel.Client') |
|
111 | 111 | |
|
112 | 112 | _socket = Instance('zmq.Socket') |
|
113 | 113 | _flag_names = List(['targets', 'block', 'track']) |
|
114 | 114 | _in_sync_results = Bool(False) |
|
115 | 115 | _targets = Any() |
|
116 | 116 | _idents = Any() |
|
117 | 117 | |
|
118 | 118 | def __init__(self, client=None, socket=None, **flags): |
|
119 | 119 | super(View, self).__init__(client=client, _socket=socket) |
|
120 | 120 | self.results = client.results |
|
121 | 121 | self.block = client.block |
|
122 | 122 | |
|
123 | 123 | self.set_flags(**flags) |
|
124 | 124 | |
|
125 | 125 | assert not self.__class__ is View, "Don't use base View objects, use subclasses" |
|
126 | 126 | |
|
127 | 127 | def __repr__(self): |
|
128 | 128 | strtargets = str(self.targets) |
|
129 | 129 | if len(strtargets) > 16: |
|
130 | 130 | strtargets = strtargets[:12]+'...]' |
|
131 | 131 | return "<%s %s>"%(self.__class__.__name__, strtargets) |
|
132 | 132 | |
|
133 | 133 | def __len__(self): |
|
134 | 134 | if isinstance(self.targets, list): |
|
135 | 135 | return len(self.targets) |
|
136 | 136 | elif isinstance(self.targets, int): |
|
137 | 137 | return 1 |
|
138 | 138 | else: |
|
139 | 139 | return len(self.client) |
|
140 | 140 | |
|
141 | 141 | def set_flags(self, **kwargs): |
|
142 | 142 | """set my attribute flags by keyword. |
|
143 | 143 | |
|
144 | 144 | Views determine behavior with a few attributes (`block`, `track`, etc.). |
|
145 | 145 | These attributes can be set all at once by name with this method. |
|
146 | 146 | |
|
147 | 147 | Parameters |
|
148 | 148 | ---------- |
|
149 | 149 | |
|
150 | 150 | block : bool |
|
151 | 151 | whether to wait for results |
|
152 | 152 | track : bool |
|
153 | 153 | whether to create a MessageTracker to allow the user to |
|
154 | 154 | safely edit after arrays and buffers during non-copying |
|
155 | 155 | sends. |
|
156 | 156 | """ |
|
157 | 157 | for name, value in iteritems(kwargs): |
|
158 | 158 | if name not in self._flag_names: |
|
159 | 159 | raise KeyError("Invalid name: %r"%name) |
|
160 | 160 | else: |
|
161 | 161 | setattr(self, name, value) |
|
162 | 162 | |
|
163 | 163 | @contextmanager |
|
164 | 164 | def temp_flags(self, **kwargs): |
|
165 | 165 | """temporarily set flags, for use in `with` statements. |
|
166 | 166 | |
|
167 | 167 | See set_flags for permanent setting of flags |
|
168 | 168 | |
|
169 | 169 | Examples |
|
170 | 170 | -------- |
|
171 | 171 | |
|
172 | 172 | >>> view.track=False |
|
173 | 173 | ... |
|
174 | 174 | >>> with view.temp_flags(track=True): |
|
175 | 175 | ... ar = view.apply(dostuff, my_big_array) |
|
176 | 176 | ... ar.tracker.wait() # wait for send to finish |
|
177 | 177 | >>> view.track |
|
178 | 178 | False |
|
179 | 179 | |
|
180 | 180 | """ |
|
181 | 181 | # preflight: save flags, and set temporaries |
|
182 | 182 | saved_flags = {} |
|
183 | 183 | for f in self._flag_names: |
|
184 | 184 | saved_flags[f] = getattr(self, f) |
|
185 | 185 | self.set_flags(**kwargs) |
|
186 | 186 | # yield to the with-statement block |
|
187 | 187 | try: |
|
188 | 188 | yield |
|
189 | 189 | finally: |
|
190 | 190 | # postflight: restore saved flags |
|
191 | 191 | self.set_flags(**saved_flags) |
|
192 | 192 | |
|
193 | 193 | |
|
194 | 194 | #---------------------------------------------------------------- |
|
195 | 195 | # apply |
|
196 | 196 | #---------------------------------------------------------------- |
|
197 | 197 | |
|
198 | 198 | def _sync_results(self): |
|
199 | 199 | """to be called by @sync_results decorator |
|
200 | 200 | |
|
201 | 201 | after submitting any tasks. |
|
202 | 202 | """ |
|
203 | 203 | delta = self.outstanding.difference(self.client.outstanding) |
|
204 | 204 | completed = self.outstanding.intersection(delta) |
|
205 | 205 | self.outstanding = self.outstanding.difference(completed) |
|
206 | 206 | |
|
207 | 207 | @sync_results |
|
208 | 208 | @save_ids |
|
209 | 209 | def _really_apply(self, f, args, kwargs, block=None, **options): |
|
210 | 210 | """wrapper for client.send_apply_request""" |
|
211 | 211 | raise NotImplementedError("Implement in subclasses") |
|
212 | 212 | |
|
213 | 213 | def apply(self, f, *args, **kwargs): |
|
214 | 214 | """calls ``f(*args, **kwargs)`` on remote engines, returning the result. |
|
215 | 215 | |
|
216 | 216 | This method sets all apply flags via this View's attributes. |
|
217 | 217 | |
|
218 | 218 | Returns :class:`~IPython.parallel.client.asyncresult.AsyncResult` |
|
219 | 219 | instance if ``self.block`` is False, otherwise the return value of |
|
220 | 220 | ``f(*args, **kwargs)``. |
|
221 | 221 | """ |
|
222 | 222 | return self._really_apply(f, args, kwargs) |
|
223 | 223 | |
|
224 | 224 | def apply_async(self, f, *args, **kwargs): |
|
225 | 225 | """calls ``f(*args, **kwargs)`` on remote engines in a nonblocking manner. |
|
226 | 226 | |
|
227 | 227 | Returns :class:`~IPython.parallel.client.asyncresult.AsyncResult` instance. |
|
228 | 228 | """ |
|
229 | 229 | return self._really_apply(f, args, kwargs, block=False) |
|
230 | 230 | |
|
231 | 231 | @spin_after |
|
232 | 232 | def apply_sync(self, f, *args, **kwargs): |
|
233 | 233 | """calls ``f(*args, **kwargs)`` on remote engines in a blocking manner, |
|
234 | 234 | returning the result. |
|
235 | 235 | """ |
|
236 | 236 | return self._really_apply(f, args, kwargs, block=True) |
|
237 | 237 | |
|
238 | 238 | #---------------------------------------------------------------- |
|
239 | 239 | # wrappers for client and control methods |
|
240 | 240 | #---------------------------------------------------------------- |
|
241 | 241 | @sync_results |
|
242 | 242 | def spin(self): |
|
243 | 243 | """spin the client, and sync""" |
|
244 | 244 | self.client.spin() |
|
245 | 245 | |
|
246 | 246 | @sync_results |
|
247 | 247 | def wait(self, jobs=None, timeout=-1): |
|
248 | 248 | """waits on one or more `jobs`, for up to `timeout` seconds. |
|
249 | 249 | |
|
250 | 250 | Parameters |
|
251 | 251 | ---------- |
|
252 | 252 | |
|
253 | 253 | jobs : int, str, or list of ints and/or strs, or one or more AsyncResult objects |
|
254 | 254 | ints are indices to self.history |
|
255 | 255 | strs are msg_ids |
|
256 | 256 | default: wait on all outstanding messages |
|
257 | 257 | timeout : float |
|
258 | 258 | a time in seconds, after which to give up. |
|
259 | 259 | default is -1, which means no timeout |
|
260 | 260 | |
|
261 | 261 | Returns |
|
262 | 262 | ------- |
|
263 | 263 | |
|
264 | 264 | True : when all msg_ids are done |
|
265 | 265 | False : timeout reached, some msg_ids still outstanding |
|
266 | 266 | """ |
|
267 | 267 | if jobs is None: |
|
268 | 268 | jobs = self.history |
|
269 | 269 | return self.client.wait(jobs, timeout) |
|
270 | 270 | |
|
271 | 271 | def abort(self, jobs=None, targets=None, block=None): |
|
272 | 272 | """Abort jobs on my engines. |
|
273 | 273 | |
|
274 | 274 | Parameters |
|
275 | 275 | ---------- |
|
276 | 276 | |
|
277 | 277 | jobs : None, str, list of strs, optional |
|
278 | 278 | if None: abort all jobs. |
|
279 | 279 | else: abort specific msg_id(s). |
|
280 | 280 | """ |
|
281 | 281 | block = block if block is not None else self.block |
|
282 | 282 | targets = targets if targets is not None else self.targets |
|
283 | 283 | jobs = jobs if jobs is not None else list(self.outstanding) |
|
284 | 284 | |
|
285 | 285 | return self.client.abort(jobs=jobs, targets=targets, block=block) |
|
286 | 286 | |
|
287 | 287 | def queue_status(self, targets=None, verbose=False): |
|
288 | 288 | """Fetch the Queue status of my engines""" |
|
289 | 289 | targets = targets if targets is not None else self.targets |
|
290 | 290 | return self.client.queue_status(targets=targets, verbose=verbose) |
|
291 | 291 | |
|
292 | 292 | def purge_results(self, jobs=[], targets=[]): |
|
293 | 293 | """Instruct the controller to forget specific results.""" |
|
294 | 294 | if targets is None or targets == 'all': |
|
295 | 295 | targets = self.targets |
|
296 | 296 | return self.client.purge_results(jobs=jobs, targets=targets) |
|
297 | 297 | |
|
298 | 298 | def shutdown(self, targets=None, restart=False, hub=False, block=None): |
|
299 | 299 | """Terminates one or more engine processes, optionally including the hub. |
|
300 | 300 | """ |
|
301 | 301 | block = self.block if block is None else block |
|
302 | 302 | if targets is None or targets == 'all': |
|
303 | 303 | targets = self.targets |
|
304 | 304 | return self.client.shutdown(targets=targets, restart=restart, hub=hub, block=block) |
|
305 | 305 | |
|
306 | 306 | @spin_after |
|
307 | 307 | def get_result(self, indices_or_msg_ids=None, block=None, owner=True): |
|
308 | 308 | """return one or more results, specified by history index or msg_id. |
|
309 | 309 | |
|
310 | 310 | See :meth:`IPython.parallel.client.client.Client.get_result` for details. |
|
311 | 311 | """ |
|
312 | 312 | |
|
313 | 313 | if indices_or_msg_ids is None: |
|
314 | 314 | indices_or_msg_ids = -1 |
|
315 | 315 | if isinstance(indices_or_msg_ids, int): |
|
316 | 316 | indices_or_msg_ids = self.history[indices_or_msg_ids] |
|
317 | 317 | elif isinstance(indices_or_msg_ids, (list,tuple,set)): |
|
318 | 318 | indices_or_msg_ids = list(indices_or_msg_ids) |
|
319 | 319 | for i,index in enumerate(indices_or_msg_ids): |
|
320 | 320 | if isinstance(index, int): |
|
321 | 321 | indices_or_msg_ids[i] = self.history[index] |
|
322 | 322 | return self.client.get_result(indices_or_msg_ids, block=block, owner=owner) |
|
323 | 323 | |
|
324 | 324 | #------------------------------------------------------------------- |
|
325 | 325 | # Map |
|
326 | 326 | #------------------------------------------------------------------- |
|
327 | 327 | |
|
328 | 328 | @sync_results |
|
329 | 329 | def map(self, f, *sequences, **kwargs): |
|
330 | 330 | """override in subclasses""" |
|
331 | 331 | raise NotImplementedError |
|
332 | 332 | |
|
333 | 333 | def map_async(self, f, *sequences, **kwargs): |
|
334 | 334 | """Parallel version of builtin :func:`python:map`, using this view's engines. |
|
335 | 335 | |
|
336 | 336 | This is equivalent to ``map(...block=False)``. |
|
337 | 337 | |
|
338 | 338 | See `self.map` for details. |
|
339 | 339 | """ |
|
340 | 340 | if 'block' in kwargs: |
|
341 | 341 | raise TypeError("map_async doesn't take a `block` keyword argument.") |
|
342 | 342 | kwargs['block'] = False |
|
343 | 343 | return self.map(f,*sequences,**kwargs) |
|
344 | 344 | |
|
345 | 345 | def map_sync(self, f, *sequences, **kwargs): |
|
346 | 346 | """Parallel version of builtin :func:`python:map`, using this view's engines. |
|
347 | 347 | |
|
348 | 348 | This is equivalent to ``map(...block=True)``. |
|
349 | 349 | |
|
350 | 350 | See `self.map` for details. |
|
351 | 351 | """ |
|
352 | 352 | if 'block' in kwargs: |
|
353 | 353 | raise TypeError("map_sync doesn't take a `block` keyword argument.") |
|
354 | 354 | kwargs['block'] = True |
|
355 | 355 | return self.map(f,*sequences,**kwargs) |
|
356 | 356 | |
|
357 | 357 | def imap(self, f, *sequences, **kwargs): |
|
358 | 358 | """Parallel version of :func:`itertools.imap`. |
|
359 | 359 | |
|
360 | 360 | See `self.map` for details. |
|
361 | 361 | |
|
362 | 362 | """ |
|
363 | 363 | |
|
364 | 364 | return iter(self.map_async(f,*sequences, **kwargs)) |
|
365 | 365 | |
|
366 | 366 | #------------------------------------------------------------------- |
|
367 | 367 | # Decorators |
|
368 | 368 | #------------------------------------------------------------------- |
|
369 | 369 | |
|
370 | 370 | def remote(self, block=None, **flags): |
|
371 | 371 | """Decorator for making a RemoteFunction""" |
|
372 | 372 | block = self.block if block is None else block |
|
373 | 373 | return remote(self, block=block, **flags) |
|
374 | 374 | |
|
375 | 375 | def parallel(self, dist='b', block=None, **flags): |
|
376 | 376 | """Decorator for making a ParallelFunction""" |
|
377 | 377 | block = self.block if block is None else block |
|
378 | 378 | return parallel(self, dist=dist, block=block, **flags) |
|
379 | 379 | |
|
380 | 380 | @skip_doctest |
|
381 | 381 | class DirectView(View): |
|
382 | 382 | """Direct Multiplexer View of one or more engines. |
|
383 | 383 | |
|
384 | 384 | These are created via indexed access to a client: |
|
385 | 385 | |
|
386 | 386 | >>> dv_1 = client[1] |
|
387 | 387 | >>> dv_all = client[:] |
|
388 | 388 | >>> dv_even = client[::2] |
|
389 | 389 | >>> dv_some = client[1:3] |
|
390 | 390 | |
|
391 | 391 | This object provides dictionary access to engine namespaces: |
|
392 | 392 | |
|
393 | 393 | # push a=5: |
|
394 | 394 | >>> dv['a'] = 5 |
|
395 | 395 | # pull 'foo': |
|
396 | 396 | >>> dv['foo'] |
|
397 | 397 | |
|
398 | 398 | """ |
|
399 | 399 | |
|
400 | 400 | def __init__(self, client=None, socket=None, targets=None): |
|
401 | 401 | super(DirectView, self).__init__(client=client, socket=socket, targets=targets) |
|
402 | 402 | |
|
403 | 403 | @property |
|
404 | 404 | def importer(self): |
|
405 | 405 | """sync_imports(local=True) as a property. |
|
406 | 406 | |
|
407 | 407 | See sync_imports for details. |
|
408 | 408 | |
|
409 | 409 | """ |
|
410 | 410 | return self.sync_imports(True) |
|
411 | 411 | |
|
412 | 412 | @contextmanager |
|
413 | 413 | def sync_imports(self, local=True, quiet=False): |
|
414 | 414 | """Context Manager for performing simultaneous local and remote imports. |
|
415 | 415 | |
|
416 | 416 | 'import x as y' will *not* work. The 'as y' part will simply be ignored. |
|
417 | 417 | |
|
418 | 418 | If `local=True`, then the package will also be imported locally. |
|
419 | 419 | |
|
420 | 420 | If `quiet=True`, no output will be produced when attempting remote |
|
421 | 421 | imports. |
|
422 | 422 | |
|
423 | 423 | Note that remote-only (`local=False`) imports have not been implemented. |
|
424 | 424 | |
|
425 | 425 | >>> with view.sync_imports(): |
|
426 | 426 | ... from numpy import recarray |
|
427 | 427 | importing recarray from numpy on engine(s) |
|
428 | 428 | |
|
429 | 429 | """ |
|
430 | 430 | from IPython.utils.py3compat import builtin_mod |
|
431 | 431 | local_import = builtin_mod.__import__ |
|
432 | 432 | modules = set() |
|
433 | 433 | results = [] |
|
434 | 434 | @util.interactive |
|
435 | 435 | def remote_import(name, fromlist, level): |
|
436 | 436 | """the function to be passed to apply, that actually performs the import |
|
437 | 437 | on the engine, and loads up the user namespace. |
|
438 | 438 | """ |
|
439 | 439 | import sys |
|
440 | 440 | user_ns = globals() |
|
441 | 441 | mod = __import__(name, fromlist=fromlist, level=level) |
|
442 | 442 | if fromlist: |
|
443 | 443 | for key in fromlist: |
|
444 | 444 | user_ns[key] = getattr(mod, key) |
|
445 | 445 | else: |
|
446 | 446 | user_ns[name] = sys.modules[name] |
|
447 | 447 | |
|
448 | 448 | def view_import(name, globals={}, locals={}, fromlist=[], level=0): |
|
449 | 449 | """the drop-in replacement for __import__, that optionally imports |
|
450 | 450 | locally as well. |
|
451 | 451 | """ |
|
452 | 452 | # don't override nested imports |
|
453 | 453 | save_import = builtin_mod.__import__ |
|
454 | 454 | builtin_mod.__import__ = local_import |
|
455 | 455 | |
|
456 | 456 | if imp.lock_held(): |
|
457 | 457 | # this is a side-effect import, don't do it remotely, or even |
|
458 | 458 | # ignore the local effects |
|
459 | 459 | return local_import(name, globals, locals, fromlist, level) |
|
460 | 460 | |
|
461 | 461 | imp.acquire_lock() |
|
462 | 462 | if local: |
|
463 | 463 | mod = local_import(name, globals, locals, fromlist, level) |
|
464 | 464 | else: |
|
465 | 465 | raise NotImplementedError("remote-only imports not yet implemented") |
|
466 | 466 | imp.release_lock() |
|
467 | 467 | |
|
468 | 468 | key = name+':'+','.join(fromlist or []) |
|
469 | 469 | if level <= 0 and key not in modules: |
|
470 | 470 | modules.add(key) |
|
471 | 471 | if not quiet: |
|
472 | 472 | if fromlist: |
|
473 | 473 | print("importing %s from %s on engine(s)"%(','.join(fromlist), name)) |
|
474 | 474 | else: |
|
475 | 475 | print("importing %s on engine(s)"%name) |
|
476 | 476 | results.append(self.apply_async(remote_import, name, fromlist, level)) |
|
477 | 477 | # restore override |
|
478 | 478 | builtin_mod.__import__ = save_import |
|
479 | 479 | |
|
480 | 480 | return mod |
|
481 | 481 | |
|
482 | 482 | # override __import__ |
|
483 | 483 | builtin_mod.__import__ = view_import |
|
484 | 484 | try: |
|
485 | 485 | # enter the block |
|
486 | 486 | yield |
|
487 | 487 | except ImportError: |
|
488 | 488 | if local: |
|
489 | 489 | raise |
|
490 | 490 | else: |
|
491 | 491 | # ignore import errors if not doing local imports |
|
492 | 492 | pass |
|
493 | 493 | finally: |
|
494 | 494 | # always restore __import__ |
|
495 | 495 | builtin_mod.__import__ = local_import |
|
496 | 496 | |
|
497 | 497 | for r in results: |
|
498 | 498 | # raise possible remote ImportErrors here |
|
499 | 499 | r.get() |
|
500 | 500 | |
|
501 | 501 | def use_dill(self): |
|
502 | 502 | """Expand serialization support with dill |
|
503 | 503 | |
|
504 | 504 | adds support for closures, etc. |
|
505 | 505 | |
|
506 | 506 | This calls IPython.utils.pickleutil.use_dill() here and on each engine. |
|
507 | 507 | """ |
|
508 | 508 | pickleutil.use_dill() |
|
509 | 509 | return self.apply(pickleutil.use_dill) |
|
510 | 510 | |
|
511 | 511 | def use_cloudpickle(self): |
|
512 | 512 | """Expand serialization support with cloudpickle. |
|
513 | 513 | """ |
|
514 | 514 | pickleutil.use_cloudpickle() |
|
515 | 515 | return self.apply(pickleutil.use_cloudpickle) |
|
516 | 516 | |
|
517 | 517 | |
|
518 | 518 | @sync_results |
|
519 | 519 | @save_ids |
|
520 | 520 | def _really_apply(self, f, args=None, kwargs=None, targets=None, block=None, track=None): |
|
521 | 521 | """calls f(*args, **kwargs) on remote engines, returning the result. |
|
522 | 522 | |
|
523 | 523 | This method sets all of `apply`'s flags via this View's attributes. |
|
524 | 524 | |
|
525 | 525 | Parameters |
|
526 | 526 | ---------- |
|
527 | 527 | |
|
528 | 528 | f : callable |
|
529 | 529 | |
|
530 | 530 | args : list [default: empty] |
|
531 | 531 | |
|
532 | 532 | kwargs : dict [default: empty] |
|
533 | 533 | |
|
534 | 534 | targets : target list [default: self.targets] |
|
535 | 535 | where to run |
|
536 | 536 | block : bool [default: self.block] |
|
537 | 537 | whether to block |
|
538 | 538 | track : bool [default: self.track] |
|
539 | 539 | whether to ask zmq to track the message, for safe non-copying sends |
|
540 | 540 | |
|
541 | 541 | Returns |
|
542 | 542 | ------- |
|
543 | 543 | |
|
544 | 544 | if self.block is False: |
|
545 | 545 | returns AsyncResult |
|
546 | 546 | else: |
|
547 | 547 | returns actual result of f(*args, **kwargs) on the engine(s) |
|
548 | 548 | This will be a list of self.targets is also a list (even length 1), or |
|
549 | 549 | the single result if self.targets is an integer engine id |
|
550 | 550 | """ |
|
551 | 551 | args = [] if args is None else args |
|
552 | 552 | kwargs = {} if kwargs is None else kwargs |
|
553 | 553 | block = self.block if block is None else block |
|
554 | 554 | track = self.track if track is None else track |
|
555 | 555 | targets = self.targets if targets is None else targets |
|
556 | 556 | |
|
557 | 557 | _idents, _targets = self.client._build_targets(targets) |
|
558 | 558 | msg_ids = [] |
|
559 | 559 | trackers = [] |
|
560 | 560 | for ident in _idents: |
|
561 | 561 | msg = self.client.send_apply_request(self._socket, f, args, kwargs, track=track, |
|
562 | 562 | ident=ident) |
|
563 | 563 | if track: |
|
564 | 564 | trackers.append(msg['tracker']) |
|
565 | 565 | msg_ids.append(msg['header']['msg_id']) |
|
566 | 566 | if isinstance(targets, int): |
|
567 | 567 | msg_ids = msg_ids[0] |
|
568 | 568 | tracker = None if track is False else zmq.MessageTracker(*trackers) |
|
569 | 569 | ar = AsyncResult(self.client, msg_ids, fname=getname(f), targets=_targets, |
|
570 | 570 | tracker=tracker, owner=True, |
|
571 | 571 | ) |
|
572 | 572 | if block: |
|
573 | 573 | try: |
|
574 | 574 | return ar.get() |
|
575 | 575 | except KeyboardInterrupt: |
|
576 | 576 | pass |
|
577 | 577 | return ar |
|
578 | 578 | |
|
579 | 579 | |
|
580 | 580 | @sync_results |
|
581 | 581 | def map(self, f, *sequences, **kwargs): |
|
582 | 582 | """``view.map(f, *sequences, block=self.block)`` => list|AsyncMapResult |
|
583 | 583 | |
|
584 | 584 | Parallel version of builtin `map`, using this View's `targets`. |
|
585 | 585 | |
|
586 | 586 | There will be one task per target, so work will be chunked |
|
587 | 587 | if the sequences are longer than `targets`. |
|
588 | 588 | |
|
589 | 589 | Results can be iterated as they are ready, but will become available in chunks. |
|
590 | 590 | |
|
591 | 591 | Parameters |
|
592 | 592 | ---------- |
|
593 | 593 | |
|
594 | 594 | f : callable |
|
595 | 595 | function to be mapped |
|
596 | 596 | *sequences: one or more sequences of matching length |
|
597 | 597 | the sequences to be distributed and passed to `f` |
|
598 | 598 | block : bool |
|
599 | 599 | whether to wait for the result or not [default self.block] |
|
600 | 600 | |
|
601 | 601 | Returns |
|
602 | 602 | ------- |
|
603 | 603 | |
|
604 | 604 | |
|
605 | 605 | If block=False |
|
606 | 606 | An :class:`~IPython.parallel.client.asyncresult.AsyncMapResult` instance. |
|
607 | 607 | An object like AsyncResult, but which reassembles the sequence of results |
|
608 | 608 | into a single list. AsyncMapResults can be iterated through before all |
|
609 | 609 | results are complete. |
|
610 | 610 | else |
|
611 | 611 | A list, the result of ``map(f,*sequences)`` |
|
612 | 612 | """ |
|
613 | 613 | |
|
614 | 614 | block = kwargs.pop('block', self.block) |
|
615 | 615 | for k in kwargs.keys(): |
|
616 | 616 | if k not in ['block', 'track']: |
|
617 | 617 | raise TypeError("invalid keyword arg, %r"%k) |
|
618 | 618 | |
|
619 | 619 | assert len(sequences) > 0, "must have some sequences to map onto!" |
|
620 | 620 | pf = ParallelFunction(self, f, block=block, **kwargs) |
|
621 | 621 | return pf.map(*sequences) |
|
622 | 622 | |
|
623 | 623 | @sync_results |
|
624 | 624 | @save_ids |
|
625 | 625 | def execute(self, code, silent=True, targets=None, block=None): |
|
626 | 626 | """Executes `code` on `targets` in blocking or nonblocking manner. |
|
627 | 627 | |
|
628 | 628 | ``execute`` is always `bound` (affects engine namespace) |
|
629 | 629 | |
|
630 | 630 | Parameters |
|
631 | 631 | ---------- |
|
632 | 632 | |
|
633 | 633 | code : str |
|
634 | 634 | the code string to be executed |
|
635 | 635 | block : bool |
|
636 | 636 | whether or not to wait until done to return |
|
637 | 637 | default: self.block |
|
638 | 638 | """ |
|
639 | 639 | block = self.block if block is None else block |
|
640 | 640 | targets = self.targets if targets is None else targets |
|
641 | 641 | |
|
642 | 642 | _idents, _targets = self.client._build_targets(targets) |
|
643 | 643 | msg_ids = [] |
|
644 | 644 | trackers = [] |
|
645 | 645 | for ident in _idents: |
|
646 | 646 | msg = self.client.send_execute_request(self._socket, code, silent=silent, ident=ident) |
|
647 | 647 | msg_ids.append(msg['header']['msg_id']) |
|
648 | 648 | if isinstance(targets, int): |
|
649 | 649 | msg_ids = msg_ids[0] |
|
650 | 650 | ar = AsyncResult(self.client, msg_ids, fname='execute', targets=_targets, owner=True) |
|
651 | 651 | if block: |
|
652 | 652 | try: |
|
653 | 653 | ar.get() |
|
654 | 654 | except KeyboardInterrupt: |
|
655 | 655 | pass |
|
656 | 656 | return ar |
|
657 | 657 | |
|
658 | 658 | def run(self, filename, targets=None, block=None): |
|
659 | 659 | """Execute contents of `filename` on my engine(s). |
|
660 | 660 | |
|
661 | 661 | This simply reads the contents of the file and calls `execute`. |
|
662 | 662 | |
|
663 | 663 | Parameters |
|
664 | 664 | ---------- |
|
665 | 665 | |
|
666 | 666 | filename : str |
|
667 | 667 | The path to the file |
|
668 | 668 | targets : int/str/list of ints/strs |
|
669 | 669 | the engines on which to execute |
|
670 | 670 | default : all |
|
671 | 671 | block : bool |
|
672 | 672 | whether or not to wait until done |
|
673 | 673 | default: self.block |
|
674 | 674 | |
|
675 | 675 | """ |
|
676 | 676 | with open(filename, 'r') as f: |
|
677 | 677 | # add newline in case of trailing indented whitespace |
|
678 | 678 | # which will cause SyntaxError |
|
679 | 679 | code = f.read()+'\n' |
|
680 | 680 | return self.execute(code, block=block, targets=targets) |
|
681 | 681 | |
|
682 | 682 | def update(self, ns): |
|
683 | 683 | """update remote namespace with dict `ns` |
|
684 | 684 | |
|
685 | 685 | See `push` for details. |
|
686 | 686 | """ |
|
687 | 687 | return self.push(ns, block=self.block, track=self.track) |
|
688 | 688 | |
|
689 | 689 | def push(self, ns, targets=None, block=None, track=None): |
|
690 | 690 | """update remote namespace with dict `ns` |
|
691 | 691 | |
|
692 | 692 | Parameters |
|
693 | 693 | ---------- |
|
694 | 694 | |
|
695 | 695 | ns : dict |
|
696 | 696 | dict of keys with which to update engine namespace(s) |
|
697 | 697 | block : bool [default : self.block] |
|
698 | 698 | whether to wait to be notified of engine receipt |
|
699 | 699 | |
|
700 | 700 | """ |
|
701 | 701 | |
|
702 | 702 | block = block if block is not None else self.block |
|
703 | 703 | track = track if track is not None else self.track |
|
704 | 704 | targets = targets if targets is not None else self.targets |
|
705 | 705 | # applier = self.apply_sync if block else self.apply_async |
|
706 | 706 | if not isinstance(ns, dict): |
|
707 | 707 | raise TypeError("Must be a dict, not %s"%type(ns)) |
|
708 | 708 | return self._really_apply(util._push, kwargs=ns, block=block, track=track, targets=targets) |
|
709 | 709 | |
|
710 | 710 | def get(self, key_s): |
|
711 | 711 | """get object(s) by `key_s` from remote namespace |
|
712 | 712 | |
|
713 | 713 | see `pull` for details. |
|
714 | 714 | """ |
|
715 | 715 | # block = block if block is not None else self.block |
|
716 | 716 | return self.pull(key_s, block=True) |
|
717 | 717 | |
|
718 | 718 | def pull(self, names, targets=None, block=None): |
|
719 | 719 | """get object(s) by `name` from remote namespace |
|
720 | 720 | |
|
721 | 721 | will return one object if it is a key. |
|
722 | 722 | can also take a list of keys, in which case it will return a list of objects. |
|
723 | 723 | """ |
|
724 | 724 | block = block if block is not None else self.block |
|
725 | 725 | targets = targets if targets is not None else self.targets |
|
726 | 726 | applier = self.apply_sync if block else self.apply_async |
|
727 | 727 | if isinstance(names, string_types): |
|
728 | 728 | pass |
|
729 | 729 | elif isinstance(names, (list,tuple,set)): |
|
730 | 730 | for key in names: |
|
731 | 731 | if not isinstance(key, string_types): |
|
732 | 732 | raise TypeError("keys must be str, not type %r"%type(key)) |
|
733 | 733 | else: |
|
734 | 734 | raise TypeError("names must be strs, not %r"%names) |
|
735 | 735 | return self._really_apply(util._pull, (names,), block=block, targets=targets) |
|
736 | 736 | |
|
737 | 737 | def scatter(self, key, seq, dist='b', flatten=False, targets=None, block=None, track=None): |
|
738 | 738 | """ |
|
739 | 739 | Partition a Python sequence and send the partitions to a set of engines. |
|
740 | 740 | """ |
|
741 | 741 | block = block if block is not None else self.block |
|
742 | 742 | track = track if track is not None else self.track |
|
743 | 743 | targets = targets if targets is not None else self.targets |
|
744 | 744 | |
|
745 | 745 | # construct integer ID list: |
|
746 | 746 | targets = self.client._build_targets(targets)[1] |
|
747 | 747 | |
|
748 | 748 | mapObject = Map.dists[dist]() |
|
749 | 749 | nparts = len(targets) |
|
750 | 750 | msg_ids = [] |
|
751 | 751 | trackers = [] |
|
752 | 752 | for index, engineid in enumerate(targets): |
|
753 | 753 | partition = mapObject.getPartition(seq, index, nparts) |
|
754 | 754 | if flatten and len(partition) == 1: |
|
755 | 755 | ns = {key: partition[0]} |
|
756 | 756 | else: |
|
757 | 757 | ns = {key: partition} |
|
758 | 758 | r = self.push(ns, block=False, track=track, targets=engineid) |
|
759 | 759 | msg_ids.extend(r.msg_ids) |
|
760 | 760 | if track: |
|
761 | 761 | trackers.append(r._tracker) |
|
762 | 762 | |
|
763 | 763 | if track: |
|
764 | 764 | tracker = zmq.MessageTracker(*trackers) |
|
765 | 765 | else: |
|
766 | 766 | tracker = None |
|
767 | 767 | |
|
768 | 768 | r = AsyncResult(self.client, msg_ids, fname='scatter', targets=targets, |
|
769 | 769 | tracker=tracker, owner=True, |
|
770 | 770 | ) |
|
771 | 771 | if block: |
|
772 | 772 | r.wait() |
|
773 | 773 | else: |
|
774 | 774 | return r |
|
775 | 775 | |
|
776 | 776 | @sync_results |
|
777 | 777 | @save_ids |
|
778 | 778 | def gather(self, key, dist='b', targets=None, block=None): |
|
779 | 779 | """ |
|
780 | 780 | Gather a partitioned sequence on a set of engines as a single local seq. |
|
781 | 781 | """ |
|
782 | 782 | block = block if block is not None else self.block |
|
783 | 783 | targets = targets if targets is not None else self.targets |
|
784 | 784 | mapObject = Map.dists[dist]() |
|
785 | 785 | msg_ids = [] |
|
786 | 786 | |
|
787 | 787 | # construct integer ID list: |
|
788 | 788 | targets = self.client._build_targets(targets)[1] |
|
789 | 789 | |
|
790 | 790 | for index, engineid in enumerate(targets): |
|
791 | 791 | msg_ids.extend(self.pull(key, block=False, targets=engineid).msg_ids) |
|
792 | 792 | |
|
793 | 793 | r = AsyncMapResult(self.client, msg_ids, mapObject, fname='gather') |
|
794 | 794 | |
|
795 | 795 | if block: |
|
796 | 796 | try: |
|
797 | 797 | return r.get() |
|
798 | 798 | except KeyboardInterrupt: |
|
799 | 799 | pass |
|
800 | 800 | return r |
|
801 | 801 | |
|
802 | 802 | def __getitem__(self, key): |
|
803 | 803 | return self.get(key) |
|
804 | 804 | |
|
805 | 805 | def __setitem__(self,key, value): |
|
806 | 806 | self.update({key:value}) |
|
807 | 807 | |
|
808 | 808 | def clear(self, targets=None, block=None): |
|
809 | 809 | """Clear the remote namespaces on my engines.""" |
|
810 | 810 | block = block if block is not None else self.block |
|
811 | 811 | targets = targets if targets is not None else self.targets |
|
812 | 812 | return self.client.clear(targets=targets, block=block) |
|
813 | 813 | |
|
814 | 814 | #---------------------------------------- |
|
815 | 815 | # activate for %px, %autopx, etc. magics |
|
816 | 816 | #---------------------------------------- |
|
817 | 817 | |
|
818 | 818 | def activate(self, suffix=''): |
|
819 | 819 | """Activate IPython magics associated with this View |
|
820 | 820 | |
|
821 | 821 | Defines the magics `%px, %autopx, %pxresult, %%px, %pxconfig` |
|
822 | 822 | |
|
823 | 823 | Parameters |
|
824 | 824 | ---------- |
|
825 | 825 | |
|
826 | 826 | suffix: str [default: ''] |
|
827 | 827 | The suffix, if any, for the magics. This allows you to have |
|
828 | 828 | multiple views associated with parallel magics at the same time. |
|
829 | 829 | |
|
830 | 830 | e.g. ``rc[::2].activate(suffix='_even')`` will give you |
|
831 | 831 | the magics ``%px_even``, ``%pxresult_even``, etc. for running magics |
|
832 | 832 | on the even engines. |
|
833 | 833 | """ |
|
834 | 834 | |
|
835 | 835 | from IPython.parallel.client.magics import ParallelMagics |
|
836 | 836 | |
|
837 | 837 | try: |
|
838 | 838 | # This is injected into __builtins__. |
|
839 | 839 | ip = get_ipython() |
|
840 | 840 | except NameError: |
|
841 | 841 | print("The IPython parallel magics (%px, etc.) only work within IPython.") |
|
842 | 842 | return |
|
843 | 843 | |
|
844 | 844 | M = ParallelMagics(ip, self, suffix) |
|
845 | 845 | ip.magics_manager.register(M) |
|
846 | 846 | |
|
847 | 847 | |
|
848 | 848 | @skip_doctest |
|
849 | 849 | class LoadBalancedView(View): |
|
850 | 850 | """An load-balancing View that only executes via the Task scheduler. |
|
851 | 851 | |
|
852 | 852 | Load-balanced views can be created with the client's `view` method: |
|
853 | 853 | |
|
854 | 854 | >>> v = client.load_balanced_view() |
|
855 | 855 | |
|
856 | 856 | or targets can be specified, to restrict the potential destinations: |
|
857 | 857 | |
|
858 | 858 | >>> v = client.load_balanced_view([1,3]) |
|
859 | 859 | |
|
860 | 860 | which would restrict loadbalancing to between engines 1 and 3. |
|
861 | 861 | |
|
862 | 862 | """ |
|
863 | 863 | |
|
864 | 864 | follow=Any() |
|
865 | 865 | after=Any() |
|
866 | 866 | timeout=CFloat() |
|
867 | 867 | retries = Integer(0) |
|
868 | 868 | |
|
869 | 869 | _task_scheme = Any() |
|
870 | 870 | _flag_names = List(['targets', 'block', 'track', 'follow', 'after', 'timeout', 'retries']) |
|
871 | 871 | |
|
872 | 872 | def __init__(self, client=None, socket=None, **flags): |
|
873 | 873 | super(LoadBalancedView, self).__init__(client=client, socket=socket, **flags) |
|
874 | 874 | self._task_scheme=client._task_scheme |
|
875 | 875 | |
|
876 | 876 | def _validate_dependency(self, dep): |
|
877 | 877 | """validate a dependency. |
|
878 | 878 | |
|
879 | 879 | For use in `set_flags`. |
|
880 | 880 | """ |
|
881 | 881 | if dep is None or isinstance(dep, string_types + (AsyncResult, Dependency)): |
|
882 | 882 | return True |
|
883 | 883 | elif isinstance(dep, (list,set, tuple)): |
|
884 | 884 | for d in dep: |
|
885 | 885 | if not isinstance(d, string_types + (AsyncResult,)): |
|
886 | 886 | return False |
|
887 | 887 | elif isinstance(dep, dict): |
|
888 | 888 | if set(dep.keys()) != set(Dependency().as_dict().keys()): |
|
889 | 889 | return False |
|
890 | 890 | if not isinstance(dep['msg_ids'], list): |
|
891 | 891 | return False |
|
892 | 892 | for d in dep['msg_ids']: |
|
893 | 893 | if not isinstance(d, string_types): |
|
894 | 894 | return False |
|
895 | 895 | else: |
|
896 | 896 | return False |
|
897 | 897 | |
|
898 | 898 | return True |
|
899 | 899 | |
|
900 | 900 | def _render_dependency(self, dep): |
|
901 | 901 | """helper for building jsonable dependencies from various input forms.""" |
|
902 | 902 | if isinstance(dep, Dependency): |
|
903 | 903 | return dep.as_dict() |
|
904 | 904 | elif isinstance(dep, AsyncResult): |
|
905 | 905 | return dep.msg_ids |
|
906 | 906 | elif dep is None: |
|
907 | 907 | return [] |
|
908 | 908 | else: |
|
909 | 909 | # pass to Dependency constructor |
|
910 | 910 | return list(Dependency(dep)) |
|
911 | 911 | |
|
912 | 912 | def set_flags(self, **kwargs): |
|
913 | 913 | """set my attribute flags by keyword. |
|
914 | 914 | |
|
915 | 915 | A View is a wrapper for the Client's apply method, but with attributes |
|
916 | 916 | that specify keyword arguments, those attributes can be set by keyword |
|
917 | 917 | argument with this method. |
|
918 | 918 | |
|
919 | 919 | Parameters |
|
920 | 920 | ---------- |
|
921 | 921 | |
|
922 | 922 | block : bool |
|
923 | 923 | whether to wait for results |
|
924 | 924 | track : bool |
|
925 | 925 | whether to create a MessageTracker to allow the user to |
|
926 | 926 | safely edit after arrays and buffers during non-copying |
|
927 | 927 | sends. |
|
928 | 928 | |
|
929 | 929 | after : Dependency or collection of msg_ids |
|
930 | 930 | Only for load-balanced execution (targets=None) |
|
931 | 931 | Specify a list of msg_ids as a time-based dependency. |
|
932 | 932 | This job will only be run *after* the dependencies |
|
933 | 933 | have been met. |
|
934 | 934 | |
|
935 | 935 | follow : Dependency or collection of msg_ids |
|
936 | 936 | Only for load-balanced execution (targets=None) |
|
937 | 937 | Specify a list of msg_ids as a location-based dependency. |
|
938 | 938 | This job will only be run on an engine where this dependency |
|
939 | 939 | is met. |
|
940 | 940 | |
|
941 | 941 | timeout : float/int or None |
|
942 | 942 | Only for load-balanced execution (targets=None) |
|
943 | 943 | Specify an amount of time (in seconds) for the scheduler to |
|
944 | 944 | wait for dependencies to be met before failing with a |
|
945 | 945 | DependencyTimeout. |
|
946 | 946 | |
|
947 | 947 | retries : int |
|
948 | 948 | Number of times a task will be retried on failure. |
|
949 | 949 | """ |
|
950 | 950 | |
|
951 | 951 | super(LoadBalancedView, self).set_flags(**kwargs) |
|
952 | 952 | for name in ('follow', 'after'): |
|
953 | 953 | if name in kwargs: |
|
954 | 954 | value = kwargs[name] |
|
955 | 955 | if self._validate_dependency(value): |
|
956 | 956 | setattr(self, name, value) |
|
957 | 957 | else: |
|
958 | 958 | raise ValueError("Invalid dependency: %r"%value) |
|
959 | 959 | if 'timeout' in kwargs: |
|
960 | 960 | t = kwargs['timeout'] |
|
961 | 961 | if not isinstance(t, (int, float, type(None))): |
|
962 | 962 | if (not PY3) and (not isinstance(t, long)): |
|
963 | 963 | raise TypeError("Invalid type for timeout: %r"%type(t)) |
|
964 | 964 | if t is not None: |
|
965 | 965 | if t < 0: |
|
966 | 966 | raise ValueError("Invalid timeout: %s"%t) |
|
967 | 967 | self.timeout = t |
|
968 | 968 | |
|
969 | 969 | @sync_results |
|
970 | 970 | @save_ids |
|
971 | 971 | def _really_apply(self, f, args=None, kwargs=None, block=None, track=None, |
|
972 | 972 | after=None, follow=None, timeout=None, |
|
973 | 973 | targets=None, retries=None): |
|
974 | 974 | """calls f(*args, **kwargs) on a remote engine, returning the result. |
|
975 | 975 | |
|
976 | 976 | This method temporarily sets all of `apply`'s flags for a single call. |
|
977 | 977 | |
|
978 | 978 | Parameters |
|
979 | 979 | ---------- |
|
980 | 980 | |
|
981 | 981 | f : callable |
|
982 | 982 | |
|
983 | 983 | args : list [default: empty] |
|
984 | 984 | |
|
985 | 985 | kwargs : dict [default: empty] |
|
986 | 986 | |
|
987 | 987 | block : bool [default: self.block] |
|
988 | 988 | whether to block |
|
989 | 989 | track : bool [default: self.track] |
|
990 | 990 | whether to ask zmq to track the message, for safe non-copying sends |
|
991 | 991 | |
|
992 | 992 | !!!!!! TODO: THE REST HERE !!!! |
|
993 | 993 | |
|
994 | 994 | Returns |
|
995 | 995 | ------- |
|
996 | 996 | |
|
997 | 997 | if self.block is False: |
|
998 | 998 | returns AsyncResult |
|
999 | 999 | else: |
|
1000 | 1000 | returns actual result of f(*args, **kwargs) on the engine(s) |
|
1001 | 1001 | This will be a list of self.targets is also a list (even length 1), or |
|
1002 | 1002 | the single result if self.targets is an integer engine id |
|
1003 | 1003 | """ |
|
1004 | 1004 | |
|
1005 | 1005 | # validate whether we can run |
|
1006 | 1006 | if self._socket.closed: |
|
1007 | 1007 | msg = "Task farming is disabled" |
|
1008 | 1008 | if self._task_scheme == 'pure': |
|
1009 | 1009 | msg += " because the pure ZMQ scheduler cannot handle" |
|
1010 | 1010 | msg += " disappearing engines." |
|
1011 | 1011 | raise RuntimeError(msg) |
|
1012 | 1012 | |
|
1013 | 1013 | if self._task_scheme == 'pure': |
|
1014 | 1014 | # pure zmq scheme doesn't support extra features |
|
1015 | 1015 | msg = "Pure ZMQ scheduler doesn't support the following flags:" |
|
1016 | 1016 | "follow, after, retries, targets, timeout" |
|
1017 | 1017 | if (follow or after or retries or targets or timeout): |
|
1018 | 1018 | # hard fail on Scheduler flags |
|
1019 | 1019 | raise RuntimeError(msg) |
|
1020 | 1020 | if isinstance(f, dependent): |
|
1021 | 1021 | # soft warn on functional dependencies |
|
1022 | 1022 | warnings.warn(msg, RuntimeWarning) |
|
1023 | 1023 | |
|
1024 | 1024 | # build args |
|
1025 | 1025 | args = [] if args is None else args |
|
1026 | 1026 | kwargs = {} if kwargs is None else kwargs |
|
1027 | 1027 | block = self.block if block is None else block |
|
1028 | 1028 | track = self.track if track is None else track |
|
1029 | 1029 | after = self.after if after is None else after |
|
1030 | 1030 | retries = self.retries if retries is None else retries |
|
1031 | 1031 | follow = self.follow if follow is None else follow |
|
1032 | 1032 | timeout = self.timeout if timeout is None else timeout |
|
1033 | 1033 | targets = self.targets if targets is None else targets |
|
1034 | 1034 | |
|
1035 | 1035 | if not isinstance(retries, int): |
|
1036 | 1036 | raise TypeError('retries must be int, not %r'%type(retries)) |
|
1037 | 1037 | |
|
1038 | 1038 | if targets is None: |
|
1039 | 1039 | idents = [] |
|
1040 | 1040 | else: |
|
1041 | 1041 | idents = self.client._build_targets(targets)[0] |
|
1042 | 1042 | # ensure *not* bytes |
|
1043 | 1043 | idents = [ ident.decode() for ident in idents ] |
|
1044 | 1044 | |
|
1045 | 1045 | after = self._render_dependency(after) |
|
1046 | 1046 | follow = self._render_dependency(follow) |
|
1047 | 1047 | metadata = dict(after=after, follow=follow, timeout=timeout, targets=idents, retries=retries) |
|
1048 | 1048 | |
|
1049 | 1049 | msg = self.client.send_apply_request(self._socket, f, args, kwargs, track=track, |
|
1050 | 1050 | metadata=metadata) |
|
1051 | 1051 | tracker = None if track is False else msg['tracker'] |
|
1052 | 1052 | |
|
1053 | 1053 | ar = AsyncResult(self.client, msg['header']['msg_id'], fname=getname(f), |
|
1054 | 1054 | targets=None, tracker=tracker, owner=True, |
|
1055 | 1055 | ) |
|
1056 | 1056 | if block: |
|
1057 | 1057 | try: |
|
1058 | 1058 | return ar.get() |
|
1059 | 1059 | except KeyboardInterrupt: |
|
1060 | 1060 | pass |
|
1061 | 1061 | return ar |
|
1062 | 1062 | |
|
1063 | 1063 | @sync_results |
|
1064 | 1064 | @save_ids |
|
1065 | 1065 | def map(self, f, *sequences, **kwargs): |
|
1066 | 1066 | """``view.map(f, *sequences, block=self.block, chunksize=1, ordered=True)`` => list|AsyncMapResult |
|
1067 | 1067 | |
|
1068 | 1068 | Parallel version of builtin `map`, load-balanced by this View. |
|
1069 | 1069 | |
|
1070 | 1070 | `block`, and `chunksize` can be specified by keyword only. |
|
1071 | 1071 | |
|
1072 | 1072 | Each `chunksize` elements will be a separate task, and will be |
|
1073 | 1073 | load-balanced. This lets individual elements be available for iteration |
|
1074 | 1074 | as soon as they arrive. |
|
1075 | 1075 | |
|
1076 | 1076 | Parameters |
|
1077 | 1077 | ---------- |
|
1078 | 1078 | |
|
1079 | 1079 | f : callable |
|
1080 | 1080 | function to be mapped |
|
1081 | 1081 | *sequences: one or more sequences of matching length |
|
1082 | 1082 | the sequences to be distributed and passed to `f` |
|
1083 | 1083 | block : bool [default self.block] |
|
1084 | 1084 | whether to wait for the result or not |
|
1085 | 1085 | track : bool |
|
1086 | 1086 | whether to create a MessageTracker to allow the user to |
|
1087 | 1087 | safely edit after arrays and buffers during non-copying |
|
1088 | 1088 | sends. |
|
1089 | 1089 | chunksize : int [default 1] |
|
1090 | 1090 | how many elements should be in each task. |
|
1091 | 1091 | ordered : bool [default True] |
|
1092 | 1092 | Whether the results should be gathered as they arrive, or enforce |
|
1093 | 1093 | the order of submission. |
|
1094 | 1094 | |
|
1095 | 1095 | Only applies when iterating through AsyncMapResult as results arrive. |
|
1096 | 1096 | Has no effect when block=True. |
|
1097 | 1097 | |
|
1098 | 1098 | Returns |
|
1099 | 1099 | ------- |
|
1100 | 1100 | |
|
1101 | 1101 | if block=False |
|
1102 | 1102 | An :class:`~IPython.parallel.client.asyncresult.AsyncMapResult` instance. |
|
1103 | 1103 | An object like AsyncResult, but which reassembles the sequence of results |
|
1104 | 1104 | into a single list. AsyncMapResults can be iterated through before all |
|
1105 | 1105 | results are complete. |
|
1106 | 1106 | else |
|
1107 | 1107 | A list, the result of ``map(f,*sequences)`` |
|
1108 | 1108 | """ |
|
1109 | 1109 | |
|
1110 | 1110 | # default |
|
1111 | 1111 | block = kwargs.get('block', self.block) |
|
1112 | 1112 | chunksize = kwargs.get('chunksize', 1) |
|
1113 | 1113 | ordered = kwargs.get('ordered', True) |
|
1114 | 1114 | |
|
1115 | 1115 | keyset = set(kwargs.keys()) |
|
1116 | 1116 | extra_keys = keyset.difference_update(set(['block', 'chunksize'])) |
|
1117 | 1117 | if extra_keys: |
|
1118 | 1118 | raise TypeError("Invalid kwargs: %s"%list(extra_keys)) |
|
1119 | 1119 | |
|
1120 | 1120 | assert len(sequences) > 0, "must have some sequences to map onto!" |
|
1121 | 1121 | |
|
1122 | 1122 | pf = ParallelFunction(self, f, block=block, chunksize=chunksize, ordered=ordered) |
|
1123 | 1123 | return pf.map(*sequences) |
|
1124 | 1124 | |
|
1125 | 1125 | __all__ = ['LoadBalancedView', 'DirectView'] |
@@ -1,849 +1,849 b'' | |||
|
1 | 1 | """The Python scheduler for rich scheduling. |
|
2 | 2 | |
|
3 | 3 | The Pure ZMQ scheduler does not allow routing schemes other than LRU, |
|
4 | 4 | nor does it check msg_id DAG dependencies. For those, a slightly slower |
|
5 | 5 | Python Scheduler exists. |
|
6 | 6 | """ |
|
7 | 7 | |
|
8 | 8 | # Copyright (c) IPython Development Team. |
|
9 | 9 | # Distributed under the terms of the Modified BSD License. |
|
10 | 10 | |
|
11 | 11 | import logging |
|
12 | 12 | import sys |
|
13 | 13 | import time |
|
14 | 14 | |
|
15 | 15 | from collections import deque |
|
16 | 16 | from datetime import datetime |
|
17 | 17 | from random import randint, random |
|
18 | 18 | from types import FunctionType |
|
19 | 19 | |
|
20 | 20 | try: |
|
21 | 21 | import numpy |
|
22 | 22 | except ImportError: |
|
23 | 23 | numpy = None |
|
24 | 24 | |
|
25 | 25 | import zmq |
|
26 | 26 | from zmq.eventloop import ioloop, zmqstream |
|
27 | 27 | |
|
28 | 28 | # local imports |
|
29 |
from |
|
|
29 | from decorator import decorator | |
|
30 | 30 | from IPython.config.application import Application |
|
31 | 31 | from IPython.config.loader import Config |
|
32 | 32 | from IPython.utils.traitlets import Instance, Dict, List, Set, Integer, Enum, CBytes |
|
33 | 33 | from IPython.utils.py3compat import cast_bytes |
|
34 | 34 | |
|
35 | 35 | from IPython.parallel import error, util |
|
36 | 36 | from IPython.parallel.factory import SessionFactory |
|
37 | 37 | from IPython.parallel.util import connect_logger, local_logger |
|
38 | 38 | |
|
39 | 39 | from .dependency import Dependency |
|
40 | 40 | |
|
41 | 41 | @decorator |
|
42 | 42 | def logged(f,self,*args,**kwargs): |
|
43 | 43 | # print ("#--------------------") |
|
44 | 44 | self.log.debug("scheduler::%s(*%s,**%s)", f.__name__, args, kwargs) |
|
45 | 45 | # print ("#--") |
|
46 | 46 | return f(self,*args, **kwargs) |
|
47 | 47 | |
|
48 | 48 | #---------------------------------------------------------------------- |
|
49 | 49 | # Chooser functions |
|
50 | 50 | #---------------------------------------------------------------------- |
|
51 | 51 | |
|
52 | 52 | def plainrandom(loads): |
|
53 | 53 | """Plain random pick.""" |
|
54 | 54 | n = len(loads) |
|
55 | 55 | return randint(0,n-1) |
|
56 | 56 | |
|
57 | 57 | def lru(loads): |
|
58 | 58 | """Always pick the front of the line. |
|
59 | 59 | |
|
60 | 60 | The content of `loads` is ignored. |
|
61 | 61 | |
|
62 | 62 | Assumes LRU ordering of loads, with oldest first. |
|
63 | 63 | """ |
|
64 | 64 | return 0 |
|
65 | 65 | |
|
66 | 66 | def twobin(loads): |
|
67 | 67 | """Pick two at random, use the LRU of the two. |
|
68 | 68 | |
|
69 | 69 | The content of loads is ignored. |
|
70 | 70 | |
|
71 | 71 | Assumes LRU ordering of loads, with oldest first. |
|
72 | 72 | """ |
|
73 | 73 | n = len(loads) |
|
74 | 74 | a = randint(0,n-1) |
|
75 | 75 | b = randint(0,n-1) |
|
76 | 76 | return min(a,b) |
|
77 | 77 | |
|
78 | 78 | def weighted(loads): |
|
79 | 79 | """Pick two at random using inverse load as weight. |
|
80 | 80 | |
|
81 | 81 | Return the less loaded of the two. |
|
82 | 82 | """ |
|
83 | 83 | # weight 0 a million times more than 1: |
|
84 | 84 | weights = 1./(1e-6+numpy.array(loads)) |
|
85 | 85 | sums = weights.cumsum() |
|
86 | 86 | t = sums[-1] |
|
87 | 87 | x = random()*t |
|
88 | 88 | y = random()*t |
|
89 | 89 | idx = 0 |
|
90 | 90 | idy = 0 |
|
91 | 91 | while sums[idx] < x: |
|
92 | 92 | idx += 1 |
|
93 | 93 | while sums[idy] < y: |
|
94 | 94 | idy += 1 |
|
95 | 95 | if weights[idy] > weights[idx]: |
|
96 | 96 | return idy |
|
97 | 97 | else: |
|
98 | 98 | return idx |
|
99 | 99 | |
|
100 | 100 | def leastload(loads): |
|
101 | 101 | """Always choose the lowest load. |
|
102 | 102 | |
|
103 | 103 | If the lowest load occurs more than once, the first |
|
104 | 104 | occurance will be used. If loads has LRU ordering, this means |
|
105 | 105 | the LRU of those with the lowest load is chosen. |
|
106 | 106 | """ |
|
107 | 107 | return loads.index(min(loads)) |
|
108 | 108 | |
|
109 | 109 | #--------------------------------------------------------------------- |
|
110 | 110 | # Classes |
|
111 | 111 | #--------------------------------------------------------------------- |
|
112 | 112 | |
|
113 | 113 | |
|
114 | 114 | # store empty default dependency: |
|
115 | 115 | MET = Dependency([]) |
|
116 | 116 | |
|
117 | 117 | |
|
118 | 118 | class Job(object): |
|
119 | 119 | """Simple container for a job""" |
|
120 | 120 | def __init__(self, msg_id, raw_msg, idents, msg, header, metadata, |
|
121 | 121 | targets, after, follow, timeout): |
|
122 | 122 | self.msg_id = msg_id |
|
123 | 123 | self.raw_msg = raw_msg |
|
124 | 124 | self.idents = idents |
|
125 | 125 | self.msg = msg |
|
126 | 126 | self.header = header |
|
127 | 127 | self.metadata = metadata |
|
128 | 128 | self.targets = targets |
|
129 | 129 | self.after = after |
|
130 | 130 | self.follow = follow |
|
131 | 131 | self.timeout = timeout |
|
132 | 132 | |
|
133 | 133 | self.removed = False # used for lazy-delete from sorted queue |
|
134 | 134 | self.timestamp = time.time() |
|
135 | 135 | self.timeout_id = 0 |
|
136 | 136 | self.blacklist = set() |
|
137 | 137 | |
|
138 | 138 | def __lt__(self, other): |
|
139 | 139 | return self.timestamp < other.timestamp |
|
140 | 140 | |
|
141 | 141 | def __cmp__(self, other): |
|
142 | 142 | return cmp(self.timestamp, other.timestamp) |
|
143 | 143 | |
|
144 | 144 | @property |
|
145 | 145 | def dependents(self): |
|
146 | 146 | return self.follow.union(self.after) |
|
147 | 147 | |
|
148 | 148 | |
|
149 | 149 | class TaskScheduler(SessionFactory): |
|
150 | 150 | """Python TaskScheduler object. |
|
151 | 151 | |
|
152 | 152 | This is the simplest object that supports msg_id based |
|
153 | 153 | DAG dependencies. *Only* task msg_ids are checked, not |
|
154 | 154 | msg_ids of jobs submitted via the MUX queue. |
|
155 | 155 | |
|
156 | 156 | """ |
|
157 | 157 | |
|
158 | 158 | hwm = Integer(1, config=True, |
|
159 | 159 | help="""specify the High Water Mark (HWM) for the downstream |
|
160 | 160 | socket in the Task scheduler. This is the maximum number |
|
161 | 161 | of allowed outstanding tasks on each engine. |
|
162 | 162 | |
|
163 | 163 | The default (1) means that only one task can be outstanding on each |
|
164 | 164 | engine. Setting TaskScheduler.hwm=0 means there is no limit, and the |
|
165 | 165 | engines continue to be assigned tasks while they are working, |
|
166 | 166 | effectively hiding network latency behind computation, but can result |
|
167 | 167 | in an imbalance of work when submitting many heterogenous tasks all at |
|
168 | 168 | once. Any positive value greater than one is a compromise between the |
|
169 | 169 | two. |
|
170 | 170 | |
|
171 | 171 | """ |
|
172 | 172 | ) |
|
173 | 173 | scheme_name = Enum(('leastload', 'pure', 'lru', 'plainrandom', 'weighted', 'twobin'), |
|
174 | 174 | 'leastload', config=True, |
|
175 | 175 | help="""select the task scheduler scheme [default: Python LRU] |
|
176 | 176 | Options are: 'pure', 'lru', 'plainrandom', 'weighted', 'twobin','leastload'""" |
|
177 | 177 | ) |
|
178 | 178 | def _scheme_name_changed(self, old, new): |
|
179 | 179 | self.log.debug("Using scheme %r"%new) |
|
180 | 180 | self.scheme = globals()[new] |
|
181 | 181 | |
|
182 | 182 | # input arguments: |
|
183 | 183 | scheme = Instance(FunctionType) # function for determining the destination |
|
184 | 184 | def _scheme_default(self): |
|
185 | 185 | return leastload |
|
186 | 186 | client_stream = Instance(zmqstream.ZMQStream) # client-facing stream |
|
187 | 187 | engine_stream = Instance(zmqstream.ZMQStream) # engine-facing stream |
|
188 | 188 | notifier_stream = Instance(zmqstream.ZMQStream) # hub-facing sub stream |
|
189 | 189 | mon_stream = Instance(zmqstream.ZMQStream) # hub-facing pub stream |
|
190 | 190 | query_stream = Instance(zmqstream.ZMQStream) # hub-facing DEALER stream |
|
191 | 191 | |
|
192 | 192 | # internals: |
|
193 | 193 | queue = Instance(deque) # sorted list of Jobs |
|
194 | 194 | def _queue_default(self): |
|
195 | 195 | return deque() |
|
196 | 196 | queue_map = Dict() # dict by msg_id of Jobs (for O(1) access to the Queue) |
|
197 | 197 | graph = Dict() # dict by msg_id of [ msg_ids that depend on key ] |
|
198 | 198 | retries = Dict() # dict by msg_id of retries remaining (non-neg ints) |
|
199 | 199 | # waiting = List() # list of msg_ids ready to run, but haven't due to HWM |
|
200 | 200 | pending = Dict() # dict by engine_uuid of submitted tasks |
|
201 | 201 | completed = Dict() # dict by engine_uuid of completed tasks |
|
202 | 202 | failed = Dict() # dict by engine_uuid of failed tasks |
|
203 | 203 | destinations = Dict() # dict by msg_id of engine_uuids where jobs ran (reverse of completed+failed) |
|
204 | 204 | clients = Dict() # dict by msg_id for who submitted the task |
|
205 | 205 | targets = List() # list of target IDENTs |
|
206 | 206 | loads = List() # list of engine loads |
|
207 | 207 | # full = Set() # set of IDENTs that have HWM outstanding tasks |
|
208 | 208 | all_completed = Set() # set of all completed tasks |
|
209 | 209 | all_failed = Set() # set of all failed tasks |
|
210 | 210 | all_done = Set() # set of all finished tasks=union(completed,failed) |
|
211 | 211 | all_ids = Set() # set of all submitted task IDs |
|
212 | 212 | |
|
213 | 213 | ident = CBytes() # ZMQ identity. This should just be self.session.session |
|
214 | 214 | # but ensure Bytes |
|
215 | 215 | def _ident_default(self): |
|
216 | 216 | return self.session.bsession |
|
217 | 217 | |
|
218 | 218 | def start(self): |
|
219 | 219 | self.query_stream.on_recv(self.dispatch_query_reply) |
|
220 | 220 | self.session.send(self.query_stream, "connection_request", {}) |
|
221 | 221 | |
|
222 | 222 | self.engine_stream.on_recv(self.dispatch_result, copy=False) |
|
223 | 223 | self.client_stream.on_recv(self.dispatch_submission, copy=False) |
|
224 | 224 | |
|
225 | 225 | self._notification_handlers = dict( |
|
226 | 226 | registration_notification = self._register_engine, |
|
227 | 227 | unregistration_notification = self._unregister_engine |
|
228 | 228 | ) |
|
229 | 229 | self.notifier_stream.on_recv(self.dispatch_notification) |
|
230 | 230 | self.log.info("Scheduler started [%s]" % self.scheme_name) |
|
231 | 231 | |
|
232 | 232 | def resume_receiving(self): |
|
233 | 233 | """Resume accepting jobs.""" |
|
234 | 234 | self.client_stream.on_recv(self.dispatch_submission, copy=False) |
|
235 | 235 | |
|
236 | 236 | def stop_receiving(self): |
|
237 | 237 | """Stop accepting jobs while there are no engines. |
|
238 | 238 | Leave them in the ZMQ queue.""" |
|
239 | 239 | self.client_stream.on_recv(None) |
|
240 | 240 | |
|
241 | 241 | #----------------------------------------------------------------------- |
|
242 | 242 | # [Un]Registration Handling |
|
243 | 243 | #----------------------------------------------------------------------- |
|
244 | 244 | |
|
245 | 245 | |
|
246 | 246 | def dispatch_query_reply(self, msg): |
|
247 | 247 | """handle reply to our initial connection request""" |
|
248 | 248 | try: |
|
249 | 249 | idents,msg = self.session.feed_identities(msg) |
|
250 | 250 | except ValueError: |
|
251 | 251 | self.log.warn("task::Invalid Message: %r",msg) |
|
252 | 252 | return |
|
253 | 253 | try: |
|
254 | 254 | msg = self.session.deserialize(msg) |
|
255 | 255 | except ValueError: |
|
256 | 256 | self.log.warn("task::Unauthorized message from: %r"%idents) |
|
257 | 257 | return |
|
258 | 258 | |
|
259 | 259 | content = msg['content'] |
|
260 | 260 | for uuid in content.get('engines', {}).values(): |
|
261 | 261 | self._register_engine(cast_bytes(uuid)) |
|
262 | 262 | |
|
263 | 263 | |
|
264 | 264 | @util.log_errors |
|
265 | 265 | def dispatch_notification(self, msg): |
|
266 | 266 | """dispatch register/unregister events.""" |
|
267 | 267 | try: |
|
268 | 268 | idents,msg = self.session.feed_identities(msg) |
|
269 | 269 | except ValueError: |
|
270 | 270 | self.log.warn("task::Invalid Message: %r",msg) |
|
271 | 271 | return |
|
272 | 272 | try: |
|
273 | 273 | msg = self.session.deserialize(msg) |
|
274 | 274 | except ValueError: |
|
275 | 275 | self.log.warn("task::Unauthorized message from: %r"%idents) |
|
276 | 276 | return |
|
277 | 277 | |
|
278 | 278 | msg_type = msg['header']['msg_type'] |
|
279 | 279 | |
|
280 | 280 | handler = self._notification_handlers.get(msg_type, None) |
|
281 | 281 | if handler is None: |
|
282 | 282 | self.log.error("Unhandled message type: %r"%msg_type) |
|
283 | 283 | else: |
|
284 | 284 | try: |
|
285 | 285 | handler(cast_bytes(msg['content']['uuid'])) |
|
286 | 286 | except Exception: |
|
287 | 287 | self.log.error("task::Invalid notification msg: %r", msg, exc_info=True) |
|
288 | 288 | |
|
289 | 289 | def _register_engine(self, uid): |
|
290 | 290 | """New engine with ident `uid` became available.""" |
|
291 | 291 | # head of the line: |
|
292 | 292 | self.targets.insert(0,uid) |
|
293 | 293 | self.loads.insert(0,0) |
|
294 | 294 | |
|
295 | 295 | # initialize sets |
|
296 | 296 | self.completed[uid] = set() |
|
297 | 297 | self.failed[uid] = set() |
|
298 | 298 | self.pending[uid] = {} |
|
299 | 299 | |
|
300 | 300 | # rescan the graph: |
|
301 | 301 | self.update_graph(None) |
|
302 | 302 | |
|
303 | 303 | def _unregister_engine(self, uid): |
|
304 | 304 | """Existing engine with ident `uid` became unavailable.""" |
|
305 | 305 | if len(self.targets) == 1: |
|
306 | 306 | # this was our only engine |
|
307 | 307 | pass |
|
308 | 308 | |
|
309 | 309 | # handle any potentially finished tasks: |
|
310 | 310 | self.engine_stream.flush() |
|
311 | 311 | |
|
312 | 312 | # don't pop destinations, because they might be used later |
|
313 | 313 | # map(self.destinations.pop, self.completed.pop(uid)) |
|
314 | 314 | # map(self.destinations.pop, self.failed.pop(uid)) |
|
315 | 315 | |
|
316 | 316 | # prevent this engine from receiving work |
|
317 | 317 | idx = self.targets.index(uid) |
|
318 | 318 | self.targets.pop(idx) |
|
319 | 319 | self.loads.pop(idx) |
|
320 | 320 | |
|
321 | 321 | # wait 5 seconds before cleaning up pending jobs, since the results might |
|
322 | 322 | # still be incoming |
|
323 | 323 | if self.pending[uid]: |
|
324 | 324 | self.loop.add_timeout(self.loop.time() + 5, |
|
325 | 325 | lambda : self.handle_stranded_tasks(uid), |
|
326 | 326 | ) |
|
327 | 327 | else: |
|
328 | 328 | self.completed.pop(uid) |
|
329 | 329 | self.failed.pop(uid) |
|
330 | 330 | |
|
331 | 331 | |
|
332 | 332 | def handle_stranded_tasks(self, engine): |
|
333 | 333 | """Deal with jobs resident in an engine that died.""" |
|
334 | 334 | lost = self.pending[engine] |
|
335 | 335 | for msg_id in lost.keys(): |
|
336 | 336 | if msg_id not in self.pending[engine]: |
|
337 | 337 | # prevent double-handling of messages |
|
338 | 338 | continue |
|
339 | 339 | |
|
340 | 340 | raw_msg = lost[msg_id].raw_msg |
|
341 | 341 | idents,msg = self.session.feed_identities(raw_msg, copy=False) |
|
342 | 342 | parent = self.session.unpack(msg[1].bytes) |
|
343 | 343 | idents = [engine, idents[0]] |
|
344 | 344 | |
|
345 | 345 | # build fake error reply |
|
346 | 346 | try: |
|
347 | 347 | raise error.EngineError("Engine %r died while running task %r"%(engine, msg_id)) |
|
348 | 348 | except: |
|
349 | 349 | content = error.wrap_exception() |
|
350 | 350 | # build fake metadata |
|
351 | 351 | md = dict( |
|
352 | 352 | status=u'error', |
|
353 | 353 | engine=engine.decode('ascii'), |
|
354 | 354 | date=datetime.now(), |
|
355 | 355 | ) |
|
356 | 356 | msg = self.session.msg('apply_reply', content, parent=parent, metadata=md) |
|
357 | 357 | raw_reply = list(map(zmq.Message, self.session.serialize(msg, ident=idents))) |
|
358 | 358 | # and dispatch it |
|
359 | 359 | self.dispatch_result(raw_reply) |
|
360 | 360 | |
|
361 | 361 | # finally scrub completed/failed lists |
|
362 | 362 | self.completed.pop(engine) |
|
363 | 363 | self.failed.pop(engine) |
|
364 | 364 | |
|
365 | 365 | |
|
366 | 366 | #----------------------------------------------------------------------- |
|
367 | 367 | # Job Submission |
|
368 | 368 | #----------------------------------------------------------------------- |
|
369 | 369 | |
|
370 | 370 | |
|
371 | 371 | @util.log_errors |
|
372 | 372 | def dispatch_submission(self, raw_msg): |
|
373 | 373 | """Dispatch job submission to appropriate handlers.""" |
|
374 | 374 | # ensure targets up to date: |
|
375 | 375 | self.notifier_stream.flush() |
|
376 | 376 | try: |
|
377 | 377 | idents, msg = self.session.feed_identities(raw_msg, copy=False) |
|
378 | 378 | msg = self.session.deserialize(msg, content=False, copy=False) |
|
379 | 379 | except Exception: |
|
380 | 380 | self.log.error("task::Invaid task msg: %r"%raw_msg, exc_info=True) |
|
381 | 381 | return |
|
382 | 382 | |
|
383 | 383 | |
|
384 | 384 | # send to monitor |
|
385 | 385 | self.mon_stream.send_multipart([b'intask']+raw_msg, copy=False) |
|
386 | 386 | |
|
387 | 387 | header = msg['header'] |
|
388 | 388 | md = msg['metadata'] |
|
389 | 389 | msg_id = header['msg_id'] |
|
390 | 390 | self.all_ids.add(msg_id) |
|
391 | 391 | |
|
392 | 392 | # get targets as a set of bytes objects |
|
393 | 393 | # from a list of unicode objects |
|
394 | 394 | targets = md.get('targets', []) |
|
395 | 395 | targets = set(map(cast_bytes, targets)) |
|
396 | 396 | |
|
397 | 397 | retries = md.get('retries', 0) |
|
398 | 398 | self.retries[msg_id] = retries |
|
399 | 399 | |
|
400 | 400 | # time dependencies |
|
401 | 401 | after = md.get('after', None) |
|
402 | 402 | if after: |
|
403 | 403 | after = Dependency(after) |
|
404 | 404 | if after.all: |
|
405 | 405 | if after.success: |
|
406 | 406 | after = Dependency(after.difference(self.all_completed), |
|
407 | 407 | success=after.success, |
|
408 | 408 | failure=after.failure, |
|
409 | 409 | all=after.all, |
|
410 | 410 | ) |
|
411 | 411 | if after.failure: |
|
412 | 412 | after = Dependency(after.difference(self.all_failed), |
|
413 | 413 | success=after.success, |
|
414 | 414 | failure=after.failure, |
|
415 | 415 | all=after.all, |
|
416 | 416 | ) |
|
417 | 417 | if after.check(self.all_completed, self.all_failed): |
|
418 | 418 | # recast as empty set, if `after` already met, |
|
419 | 419 | # to prevent unnecessary set comparisons |
|
420 | 420 | after = MET |
|
421 | 421 | else: |
|
422 | 422 | after = MET |
|
423 | 423 | |
|
424 | 424 | # location dependencies |
|
425 | 425 | follow = Dependency(md.get('follow', [])) |
|
426 | 426 | |
|
427 | 427 | timeout = md.get('timeout', None) |
|
428 | 428 | if timeout: |
|
429 | 429 | timeout = float(timeout) |
|
430 | 430 | |
|
431 | 431 | job = Job(msg_id=msg_id, raw_msg=raw_msg, idents=idents, msg=msg, |
|
432 | 432 | header=header, targets=targets, after=after, follow=follow, |
|
433 | 433 | timeout=timeout, metadata=md, |
|
434 | 434 | ) |
|
435 | 435 | # validate and reduce dependencies: |
|
436 | 436 | for dep in after,follow: |
|
437 | 437 | if not dep: # empty dependency |
|
438 | 438 | continue |
|
439 | 439 | # check valid: |
|
440 | 440 | if msg_id in dep or dep.difference(self.all_ids): |
|
441 | 441 | self.queue_map[msg_id] = job |
|
442 | 442 | return self.fail_unreachable(msg_id, error.InvalidDependency) |
|
443 | 443 | # check if unreachable: |
|
444 | 444 | if dep.unreachable(self.all_completed, self.all_failed): |
|
445 | 445 | self.queue_map[msg_id] = job |
|
446 | 446 | return self.fail_unreachable(msg_id) |
|
447 | 447 | |
|
448 | 448 | if after.check(self.all_completed, self.all_failed): |
|
449 | 449 | # time deps already met, try to run |
|
450 | 450 | if not self.maybe_run(job): |
|
451 | 451 | # can't run yet |
|
452 | 452 | if msg_id not in self.all_failed: |
|
453 | 453 | # could have failed as unreachable |
|
454 | 454 | self.save_unmet(job) |
|
455 | 455 | else: |
|
456 | 456 | self.save_unmet(job) |
|
457 | 457 | |
|
458 | 458 | def job_timeout(self, job, timeout_id): |
|
459 | 459 | """callback for a job's timeout. |
|
460 | 460 | |
|
461 | 461 | The job may or may not have been run at this point. |
|
462 | 462 | """ |
|
463 | 463 | if job.timeout_id != timeout_id: |
|
464 | 464 | # not the most recent call |
|
465 | 465 | return |
|
466 | 466 | now = time.time() |
|
467 | 467 | if job.timeout >= (now + 1): |
|
468 | 468 | self.log.warn("task %s timeout fired prematurely: %s > %s", |
|
469 | 469 | job.msg_id, job.timeout, now |
|
470 | 470 | ) |
|
471 | 471 | if job.msg_id in self.queue_map: |
|
472 | 472 | # still waiting, but ran out of time |
|
473 | 473 | self.log.info("task %r timed out", job.msg_id) |
|
474 | 474 | self.fail_unreachable(job.msg_id, error.TaskTimeout) |
|
475 | 475 | |
|
476 | 476 | def fail_unreachable(self, msg_id, why=error.ImpossibleDependency): |
|
477 | 477 | """a task has become unreachable, send a reply with an ImpossibleDependency |
|
478 | 478 | error.""" |
|
479 | 479 | if msg_id not in self.queue_map: |
|
480 | 480 | self.log.error("task %r already failed!", msg_id) |
|
481 | 481 | return |
|
482 | 482 | job = self.queue_map.pop(msg_id) |
|
483 | 483 | # lazy-delete from the queue |
|
484 | 484 | job.removed = True |
|
485 | 485 | for mid in job.dependents: |
|
486 | 486 | if mid in self.graph: |
|
487 | 487 | self.graph[mid].remove(msg_id) |
|
488 | 488 | |
|
489 | 489 | try: |
|
490 | 490 | raise why() |
|
491 | 491 | except: |
|
492 | 492 | content = error.wrap_exception() |
|
493 | 493 | self.log.debug("task %r failing as unreachable with: %s", msg_id, content['ename']) |
|
494 | 494 | |
|
495 | 495 | self.all_done.add(msg_id) |
|
496 | 496 | self.all_failed.add(msg_id) |
|
497 | 497 | |
|
498 | 498 | msg = self.session.send(self.client_stream, 'apply_reply', content, |
|
499 | 499 | parent=job.header, ident=job.idents) |
|
500 | 500 | self.session.send(self.mon_stream, msg, ident=[b'outtask']+job.idents) |
|
501 | 501 | |
|
502 | 502 | self.update_graph(msg_id, success=False) |
|
503 | 503 | |
|
504 | 504 | def available_engines(self): |
|
505 | 505 | """return a list of available engine indices based on HWM""" |
|
506 | 506 | if not self.hwm: |
|
507 | 507 | return list(range(len(self.targets))) |
|
508 | 508 | available = [] |
|
509 | 509 | for idx in range(len(self.targets)): |
|
510 | 510 | if self.loads[idx] < self.hwm: |
|
511 | 511 | available.append(idx) |
|
512 | 512 | return available |
|
513 | 513 | |
|
514 | 514 | def maybe_run(self, job): |
|
515 | 515 | """check location dependencies, and run if they are met.""" |
|
516 | 516 | msg_id = job.msg_id |
|
517 | 517 | self.log.debug("Attempting to assign task %s", msg_id) |
|
518 | 518 | available = self.available_engines() |
|
519 | 519 | if not available: |
|
520 | 520 | # no engines, definitely can't run |
|
521 | 521 | return False |
|
522 | 522 | |
|
523 | 523 | if job.follow or job.targets or job.blacklist or self.hwm: |
|
524 | 524 | # we need a can_run filter |
|
525 | 525 | def can_run(idx): |
|
526 | 526 | # check hwm |
|
527 | 527 | if self.hwm and self.loads[idx] == self.hwm: |
|
528 | 528 | return False |
|
529 | 529 | target = self.targets[idx] |
|
530 | 530 | # check blacklist |
|
531 | 531 | if target in job.blacklist: |
|
532 | 532 | return False |
|
533 | 533 | # check targets |
|
534 | 534 | if job.targets and target not in job.targets: |
|
535 | 535 | return False |
|
536 | 536 | # check follow |
|
537 | 537 | return job.follow.check(self.completed[target], self.failed[target]) |
|
538 | 538 | |
|
539 | 539 | indices = list(filter(can_run, available)) |
|
540 | 540 | |
|
541 | 541 | if not indices: |
|
542 | 542 | # couldn't run |
|
543 | 543 | if job.follow.all: |
|
544 | 544 | # check follow for impossibility |
|
545 | 545 | dests = set() |
|
546 | 546 | relevant = set() |
|
547 | 547 | if job.follow.success: |
|
548 | 548 | relevant = self.all_completed |
|
549 | 549 | if job.follow.failure: |
|
550 | 550 | relevant = relevant.union(self.all_failed) |
|
551 | 551 | for m in job.follow.intersection(relevant): |
|
552 | 552 | dests.add(self.destinations[m]) |
|
553 | 553 | if len(dests) > 1: |
|
554 | 554 | self.queue_map[msg_id] = job |
|
555 | 555 | self.fail_unreachable(msg_id) |
|
556 | 556 | return False |
|
557 | 557 | if job.targets: |
|
558 | 558 | # check blacklist+targets for impossibility |
|
559 | 559 | job.targets.difference_update(job.blacklist) |
|
560 | 560 | if not job.targets or not job.targets.intersection(self.targets): |
|
561 | 561 | self.queue_map[msg_id] = job |
|
562 | 562 | self.fail_unreachable(msg_id) |
|
563 | 563 | return False |
|
564 | 564 | return False |
|
565 | 565 | else: |
|
566 | 566 | indices = None |
|
567 | 567 | |
|
568 | 568 | self.submit_task(job, indices) |
|
569 | 569 | return True |
|
570 | 570 | |
|
571 | 571 | def save_unmet(self, job): |
|
572 | 572 | """Save a message for later submission when its dependencies are met.""" |
|
573 | 573 | msg_id = job.msg_id |
|
574 | 574 | self.log.debug("Adding task %s to the queue", msg_id) |
|
575 | 575 | self.queue_map[msg_id] = job |
|
576 | 576 | self.queue.append(job) |
|
577 | 577 | # track the ids in follow or after, but not those already finished |
|
578 | 578 | for dep_id in job.after.union(job.follow).difference(self.all_done): |
|
579 | 579 | if dep_id not in self.graph: |
|
580 | 580 | self.graph[dep_id] = set() |
|
581 | 581 | self.graph[dep_id].add(msg_id) |
|
582 | 582 | |
|
583 | 583 | # schedule timeout callback |
|
584 | 584 | if job.timeout: |
|
585 | 585 | timeout_id = job.timeout_id = job.timeout_id + 1 |
|
586 | 586 | self.loop.add_timeout(time.time() + job.timeout, |
|
587 | 587 | lambda : self.job_timeout(job, timeout_id) |
|
588 | 588 | ) |
|
589 | 589 | |
|
590 | 590 | |
|
591 | 591 | def submit_task(self, job, indices=None): |
|
592 | 592 | """Submit a task to any of a subset of our targets.""" |
|
593 | 593 | if indices: |
|
594 | 594 | loads = [self.loads[i] for i in indices] |
|
595 | 595 | else: |
|
596 | 596 | loads = self.loads |
|
597 | 597 | idx = self.scheme(loads) |
|
598 | 598 | if indices: |
|
599 | 599 | idx = indices[idx] |
|
600 | 600 | target = self.targets[idx] |
|
601 | 601 | # print (target, map(str, msg[:3])) |
|
602 | 602 | # send job to the engine |
|
603 | 603 | self.engine_stream.send(target, flags=zmq.SNDMORE, copy=False) |
|
604 | 604 | self.engine_stream.send_multipart(job.raw_msg, copy=False) |
|
605 | 605 | # update load |
|
606 | 606 | self.add_job(idx) |
|
607 | 607 | self.pending[target][job.msg_id] = job |
|
608 | 608 | # notify Hub |
|
609 | 609 | content = dict(msg_id=job.msg_id, engine_id=target.decode('ascii')) |
|
610 | 610 | self.session.send(self.mon_stream, 'task_destination', content=content, |
|
611 | 611 | ident=[b'tracktask',self.ident]) |
|
612 | 612 | |
|
613 | 613 | |
|
614 | 614 | #----------------------------------------------------------------------- |
|
615 | 615 | # Result Handling |
|
616 | 616 | #----------------------------------------------------------------------- |
|
617 | 617 | |
|
618 | 618 | |
|
619 | 619 | @util.log_errors |
|
620 | 620 | def dispatch_result(self, raw_msg): |
|
621 | 621 | """dispatch method for result replies""" |
|
622 | 622 | try: |
|
623 | 623 | idents,msg = self.session.feed_identities(raw_msg, copy=False) |
|
624 | 624 | msg = self.session.deserialize(msg, content=False, copy=False) |
|
625 | 625 | engine = idents[0] |
|
626 | 626 | try: |
|
627 | 627 | idx = self.targets.index(engine) |
|
628 | 628 | except ValueError: |
|
629 | 629 | pass # skip load-update for dead engines |
|
630 | 630 | else: |
|
631 | 631 | self.finish_job(idx) |
|
632 | 632 | except Exception: |
|
633 | 633 | self.log.error("task::Invalid result: %r", raw_msg, exc_info=True) |
|
634 | 634 | return |
|
635 | 635 | |
|
636 | 636 | md = msg['metadata'] |
|
637 | 637 | parent = msg['parent_header'] |
|
638 | 638 | if md.get('dependencies_met', True): |
|
639 | 639 | success = (md['status'] == 'ok') |
|
640 | 640 | msg_id = parent['msg_id'] |
|
641 | 641 | retries = self.retries[msg_id] |
|
642 | 642 | if not success and retries > 0: |
|
643 | 643 | # failed |
|
644 | 644 | self.retries[msg_id] = retries - 1 |
|
645 | 645 | self.handle_unmet_dependency(idents, parent) |
|
646 | 646 | else: |
|
647 | 647 | del self.retries[msg_id] |
|
648 | 648 | # relay to client and update graph |
|
649 | 649 | self.handle_result(idents, parent, raw_msg, success) |
|
650 | 650 | # send to Hub monitor |
|
651 | 651 | self.mon_stream.send_multipart([b'outtask']+raw_msg, copy=False) |
|
652 | 652 | else: |
|
653 | 653 | self.handle_unmet_dependency(idents, parent) |
|
654 | 654 | |
|
655 | 655 | def handle_result(self, idents, parent, raw_msg, success=True): |
|
656 | 656 | """handle a real task result, either success or failure""" |
|
657 | 657 | # first, relay result to client |
|
658 | 658 | engine = idents[0] |
|
659 | 659 | client = idents[1] |
|
660 | 660 | # swap_ids for ROUTER-ROUTER mirror |
|
661 | 661 | raw_msg[:2] = [client,engine] |
|
662 | 662 | # print (map(str, raw_msg[:4])) |
|
663 | 663 | self.client_stream.send_multipart(raw_msg, copy=False) |
|
664 | 664 | # now, update our data structures |
|
665 | 665 | msg_id = parent['msg_id'] |
|
666 | 666 | self.pending[engine].pop(msg_id) |
|
667 | 667 | if success: |
|
668 | 668 | self.completed[engine].add(msg_id) |
|
669 | 669 | self.all_completed.add(msg_id) |
|
670 | 670 | else: |
|
671 | 671 | self.failed[engine].add(msg_id) |
|
672 | 672 | self.all_failed.add(msg_id) |
|
673 | 673 | self.all_done.add(msg_id) |
|
674 | 674 | self.destinations[msg_id] = engine |
|
675 | 675 | |
|
676 | 676 | self.update_graph(msg_id, success) |
|
677 | 677 | |
|
678 | 678 | def handle_unmet_dependency(self, idents, parent): |
|
679 | 679 | """handle an unmet dependency""" |
|
680 | 680 | engine = idents[0] |
|
681 | 681 | msg_id = parent['msg_id'] |
|
682 | 682 | |
|
683 | 683 | job = self.pending[engine].pop(msg_id) |
|
684 | 684 | job.blacklist.add(engine) |
|
685 | 685 | |
|
686 | 686 | if job.blacklist == job.targets: |
|
687 | 687 | self.queue_map[msg_id] = job |
|
688 | 688 | self.fail_unreachable(msg_id) |
|
689 | 689 | elif not self.maybe_run(job): |
|
690 | 690 | # resubmit failed |
|
691 | 691 | if msg_id not in self.all_failed: |
|
692 | 692 | # put it back in our dependency tree |
|
693 | 693 | self.save_unmet(job) |
|
694 | 694 | |
|
695 | 695 | if self.hwm: |
|
696 | 696 | try: |
|
697 | 697 | idx = self.targets.index(engine) |
|
698 | 698 | except ValueError: |
|
699 | 699 | pass # skip load-update for dead engines |
|
700 | 700 | else: |
|
701 | 701 | if self.loads[idx] == self.hwm-1: |
|
702 | 702 | self.update_graph(None) |
|
703 | 703 | |
|
704 | 704 | def update_graph(self, dep_id=None, success=True): |
|
705 | 705 | """dep_id just finished. Update our dependency |
|
706 | 706 | graph and submit any jobs that just became runnable. |
|
707 | 707 | |
|
708 | 708 | Called with dep_id=None to update entire graph for hwm, but without finishing a task. |
|
709 | 709 | """ |
|
710 | 710 | # print ("\n\n***********") |
|
711 | 711 | # pprint (dep_id) |
|
712 | 712 | # pprint (self.graph) |
|
713 | 713 | # pprint (self.queue_map) |
|
714 | 714 | # pprint (self.all_completed) |
|
715 | 715 | # pprint (self.all_failed) |
|
716 | 716 | # print ("\n\n***********\n\n") |
|
717 | 717 | # update any jobs that depended on the dependency |
|
718 | 718 | msg_ids = self.graph.pop(dep_id, []) |
|
719 | 719 | |
|
720 | 720 | # recheck *all* jobs if |
|
721 | 721 | # a) we have HWM and an engine just become no longer full |
|
722 | 722 | # or b) dep_id was given as None |
|
723 | 723 | |
|
724 | 724 | if dep_id is None or self.hwm and any( [ load==self.hwm-1 for load in self.loads ]): |
|
725 | 725 | jobs = self.queue |
|
726 | 726 | using_queue = True |
|
727 | 727 | else: |
|
728 | 728 | using_queue = False |
|
729 | 729 | jobs = deque(sorted( self.queue_map[msg_id] for msg_id in msg_ids )) |
|
730 | 730 | |
|
731 | 731 | to_restore = [] |
|
732 | 732 | while jobs: |
|
733 | 733 | job = jobs.popleft() |
|
734 | 734 | if job.removed: |
|
735 | 735 | continue |
|
736 | 736 | msg_id = job.msg_id |
|
737 | 737 | |
|
738 | 738 | put_it_back = True |
|
739 | 739 | |
|
740 | 740 | if job.after.unreachable(self.all_completed, self.all_failed)\ |
|
741 | 741 | or job.follow.unreachable(self.all_completed, self.all_failed): |
|
742 | 742 | self.fail_unreachable(msg_id) |
|
743 | 743 | put_it_back = False |
|
744 | 744 | |
|
745 | 745 | elif job.after.check(self.all_completed, self.all_failed): # time deps met, maybe run |
|
746 | 746 | if self.maybe_run(job): |
|
747 | 747 | put_it_back = False |
|
748 | 748 | self.queue_map.pop(msg_id) |
|
749 | 749 | for mid in job.dependents: |
|
750 | 750 | if mid in self.graph: |
|
751 | 751 | self.graph[mid].remove(msg_id) |
|
752 | 752 | |
|
753 | 753 | # abort the loop if we just filled up all of our engines. |
|
754 | 754 | # avoids an O(N) operation in situation of full queue, |
|
755 | 755 | # where graph update is triggered as soon as an engine becomes |
|
756 | 756 | # non-full, and all tasks after the first are checked, |
|
757 | 757 | # even though they can't run. |
|
758 | 758 | if not self.available_engines(): |
|
759 | 759 | break |
|
760 | 760 | |
|
761 | 761 | if using_queue and put_it_back: |
|
762 | 762 | # popped a job from the queue but it neither ran nor failed, |
|
763 | 763 | # so we need to put it back when we are done |
|
764 | 764 | # make sure to_restore preserves the same ordering |
|
765 | 765 | to_restore.append(job) |
|
766 | 766 | |
|
767 | 767 | # put back any tasks we popped but didn't run |
|
768 | 768 | if using_queue: |
|
769 | 769 | self.queue.extendleft(to_restore) |
|
770 | 770 | |
|
771 | 771 | #---------------------------------------------------------------------- |
|
772 | 772 | # methods to be overridden by subclasses |
|
773 | 773 | #---------------------------------------------------------------------- |
|
774 | 774 | |
|
775 | 775 | def add_job(self, idx): |
|
776 | 776 | """Called after self.targets[idx] just got the job with header. |
|
777 | 777 | Override with subclasses. The default ordering is simple LRU. |
|
778 | 778 | The default loads are the number of outstanding jobs.""" |
|
779 | 779 | self.loads[idx] += 1 |
|
780 | 780 | for lis in (self.targets, self.loads): |
|
781 | 781 | lis.append(lis.pop(idx)) |
|
782 | 782 | |
|
783 | 783 | |
|
784 | 784 | def finish_job(self, idx): |
|
785 | 785 | """Called after self.targets[idx] just finished a job. |
|
786 | 786 | Override with subclasses.""" |
|
787 | 787 | self.loads[idx] -= 1 |
|
788 | 788 | |
|
789 | 789 | |
|
790 | 790 | |
|
791 | 791 | def launch_scheduler(in_addr, out_addr, mon_addr, not_addr, reg_addr, config=None, |
|
792 | 792 | logname='root', log_url=None, loglevel=logging.DEBUG, |
|
793 | 793 | identity=b'task', in_thread=False): |
|
794 | 794 | |
|
795 | 795 | ZMQStream = zmqstream.ZMQStream |
|
796 | 796 | |
|
797 | 797 | if config: |
|
798 | 798 | # unwrap dict back into Config |
|
799 | 799 | config = Config(config) |
|
800 | 800 | |
|
801 | 801 | if in_thread: |
|
802 | 802 | # use instance() to get the same Context/Loop as our parent |
|
803 | 803 | ctx = zmq.Context.instance() |
|
804 | 804 | loop = ioloop.IOLoop.instance() |
|
805 | 805 | else: |
|
806 | 806 | # in a process, don't use instance() |
|
807 | 807 | # for safety with multiprocessing |
|
808 | 808 | ctx = zmq.Context() |
|
809 | 809 | loop = ioloop.IOLoop() |
|
810 | 810 | ins = ZMQStream(ctx.socket(zmq.ROUTER),loop) |
|
811 | 811 | util.set_hwm(ins, 0) |
|
812 | 812 | ins.setsockopt(zmq.IDENTITY, identity + b'_in') |
|
813 | 813 | ins.bind(in_addr) |
|
814 | 814 | |
|
815 | 815 | outs = ZMQStream(ctx.socket(zmq.ROUTER),loop) |
|
816 | 816 | util.set_hwm(outs, 0) |
|
817 | 817 | outs.setsockopt(zmq.IDENTITY, identity + b'_out') |
|
818 | 818 | outs.bind(out_addr) |
|
819 | 819 | mons = zmqstream.ZMQStream(ctx.socket(zmq.PUB),loop) |
|
820 | 820 | util.set_hwm(mons, 0) |
|
821 | 821 | mons.connect(mon_addr) |
|
822 | 822 | nots = zmqstream.ZMQStream(ctx.socket(zmq.SUB),loop) |
|
823 | 823 | nots.setsockopt(zmq.SUBSCRIBE, b'') |
|
824 | 824 | nots.connect(not_addr) |
|
825 | 825 | |
|
826 | 826 | querys = ZMQStream(ctx.socket(zmq.DEALER),loop) |
|
827 | 827 | querys.connect(reg_addr) |
|
828 | 828 | |
|
829 | 829 | # setup logging. |
|
830 | 830 | if in_thread: |
|
831 | 831 | log = Application.instance().log |
|
832 | 832 | else: |
|
833 | 833 | if log_url: |
|
834 | 834 | log = connect_logger(logname, ctx, log_url, root="scheduler", loglevel=loglevel) |
|
835 | 835 | else: |
|
836 | 836 | log = local_logger(logname, loglevel) |
|
837 | 837 | |
|
838 | 838 | scheduler = TaskScheduler(client_stream=ins, engine_stream=outs, |
|
839 | 839 | mon_stream=mons, notifier_stream=nots, |
|
840 | 840 | query_stream=querys, |
|
841 | 841 | loop=loop, log=log, |
|
842 | 842 | config=config) |
|
843 | 843 | scheduler.start() |
|
844 | 844 | if not in_thread: |
|
845 | 845 | try: |
|
846 | 846 | loop.start() |
|
847 | 847 | except KeyboardInterrupt: |
|
848 | 848 | scheduler.log.critical("Interrupted, exiting...") |
|
849 | 849 |
@@ -1,192 +1,192 b'' | |||
|
1 | 1 | """base class for parallel client tests |
|
2 | 2 | |
|
3 | 3 | Authors: |
|
4 | 4 | |
|
5 | 5 | * Min RK |
|
6 | 6 | """ |
|
7 | 7 | |
|
8 | 8 | #------------------------------------------------------------------------------- |
|
9 | 9 | # Copyright (C) 2011 The IPython Development Team |
|
10 | 10 | # |
|
11 | 11 | # Distributed under the terms of the BSD License. The full license is in |
|
12 | 12 | # the file COPYING, distributed as part of this software. |
|
13 | 13 | #------------------------------------------------------------------------------- |
|
14 | 14 | from __future__ import print_function |
|
15 | 15 | |
|
16 | 16 | import sys |
|
17 | 17 | import tempfile |
|
18 | 18 | import time |
|
19 | 19 | |
|
20 | 20 | from nose import SkipTest |
|
21 | 21 | |
|
22 | 22 | import zmq |
|
23 | 23 | from zmq.tests import BaseZMQTestCase |
|
24 | 24 | |
|
25 |
from |
|
|
25 | from decorator import decorator | |
|
26 | 26 | |
|
27 | 27 | from IPython.parallel import error |
|
28 | 28 | from IPython.parallel import Client |
|
29 | 29 | |
|
30 | 30 | from IPython.parallel.tests import launchers, add_engines |
|
31 | 31 | |
|
32 | 32 | # simple tasks for use in apply tests |
|
33 | 33 | |
|
34 | 34 | def segfault(): |
|
35 | 35 | """this will segfault""" |
|
36 | 36 | import ctypes |
|
37 | 37 | ctypes.memset(-1,0,1) |
|
38 | 38 | |
|
39 | 39 | def crash(): |
|
40 | 40 | """from stdlib crashers in the test suite""" |
|
41 | 41 | import types |
|
42 | 42 | if sys.platform.startswith('win'): |
|
43 | 43 | import ctypes |
|
44 | 44 | ctypes.windll.kernel32.SetErrorMode(0x0002); |
|
45 | 45 | args = [ 0, 0, 0, 0, b'\x04\x71\x00\x00', (), (), (), '', '', 1, b''] |
|
46 | 46 | if sys.version_info[0] >= 3: |
|
47 | 47 | # Python3 adds 'kwonlyargcount' as the second argument to Code |
|
48 | 48 | args.insert(1, 0) |
|
49 | 49 | |
|
50 | 50 | co = types.CodeType(*args) |
|
51 | 51 | exec(co) |
|
52 | 52 | |
|
53 | 53 | def wait(n): |
|
54 | 54 | """sleep for a time""" |
|
55 | 55 | import time |
|
56 | 56 | time.sleep(n) |
|
57 | 57 | return n |
|
58 | 58 | |
|
59 | 59 | def raiser(eclass): |
|
60 | 60 | """raise an exception""" |
|
61 | 61 | raise eclass() |
|
62 | 62 | |
|
63 | 63 | def generate_output(): |
|
64 | 64 | """function for testing output |
|
65 | 65 | |
|
66 | 66 | publishes two outputs of each type, and returns |
|
67 | 67 | a rich displayable object. |
|
68 | 68 | """ |
|
69 | 69 | |
|
70 | 70 | import sys |
|
71 | 71 | from IPython.core.display import display, HTML, Math |
|
72 | 72 | |
|
73 | 73 | print("stdout") |
|
74 | 74 | print("stderr", file=sys.stderr) |
|
75 | 75 | |
|
76 | 76 | display(HTML("<b>HTML</b>")) |
|
77 | 77 | |
|
78 | 78 | print("stdout2") |
|
79 | 79 | print("stderr2", file=sys.stderr) |
|
80 | 80 | |
|
81 | 81 | display(Math(r"\alpha=\beta")) |
|
82 | 82 | |
|
83 | 83 | return Math("42") |
|
84 | 84 | |
|
85 | 85 | # test decorator for skipping tests when libraries are unavailable |
|
86 | 86 | def skip_without(*names): |
|
87 | 87 | """skip a test if some names are not importable""" |
|
88 | 88 | @decorator |
|
89 | 89 | def skip_without_names(f, *args, **kwargs): |
|
90 | 90 | """decorator to skip tests in the absence of numpy.""" |
|
91 | 91 | for name in names: |
|
92 | 92 | try: |
|
93 | 93 | __import__(name) |
|
94 | 94 | except ImportError: |
|
95 | 95 | raise SkipTest |
|
96 | 96 | return f(*args, **kwargs) |
|
97 | 97 | return skip_without_names |
|
98 | 98 | |
|
99 | 99 | #------------------------------------------------------------------------------- |
|
100 | 100 | # Classes |
|
101 | 101 | #------------------------------------------------------------------------------- |
|
102 | 102 | |
|
103 | 103 | |
|
104 | 104 | class ClusterTestCase(BaseZMQTestCase): |
|
105 | 105 | timeout = 10 |
|
106 | 106 | |
|
107 | 107 | def add_engines(self, n=1, block=True): |
|
108 | 108 | """add multiple engines to our cluster""" |
|
109 | 109 | self.engines.extend(add_engines(n)) |
|
110 | 110 | if block: |
|
111 | 111 | self.wait_on_engines() |
|
112 | 112 | |
|
113 | 113 | def minimum_engines(self, n=1, block=True): |
|
114 | 114 | """add engines until there are at least n connected""" |
|
115 | 115 | self.engines.extend(add_engines(n, total=True)) |
|
116 | 116 | if block: |
|
117 | 117 | self.wait_on_engines() |
|
118 | 118 | |
|
119 | 119 | |
|
120 | 120 | def wait_on_engines(self, timeout=5): |
|
121 | 121 | """wait for our engines to connect.""" |
|
122 | 122 | n = len(self.engines)+self.base_engine_count |
|
123 | 123 | tic = time.time() |
|
124 | 124 | while time.time()-tic < timeout and len(self.client.ids) < n: |
|
125 | 125 | time.sleep(0.1) |
|
126 | 126 | |
|
127 | 127 | assert not len(self.client.ids) < n, "waiting for engines timed out" |
|
128 | 128 | |
|
129 | 129 | def client_wait(self, client, jobs=None, timeout=-1): |
|
130 | 130 | """my wait wrapper, sets a default finite timeout to avoid hangs""" |
|
131 | 131 | if timeout < 0: |
|
132 | 132 | timeout = self.timeout |
|
133 | 133 | return Client.wait(client, jobs, timeout) |
|
134 | 134 | |
|
135 | 135 | def connect_client(self): |
|
136 | 136 | """connect a client with my Context, and track its sockets for cleanup""" |
|
137 | 137 | c = Client(profile='iptest', context=self.context) |
|
138 | 138 | c.wait = lambda *a, **kw: self.client_wait(c, *a, **kw) |
|
139 | 139 | |
|
140 | 140 | for name in filter(lambda n:n.endswith('socket'), dir(c)): |
|
141 | 141 | s = getattr(c, name) |
|
142 | 142 | s.setsockopt(zmq.LINGER, 0) |
|
143 | 143 | self.sockets.append(s) |
|
144 | 144 | return c |
|
145 | 145 | |
|
146 | 146 | def assertRaisesRemote(self, etype, f, *args, **kwargs): |
|
147 | 147 | try: |
|
148 | 148 | try: |
|
149 | 149 | f(*args, **kwargs) |
|
150 | 150 | except error.CompositeError as e: |
|
151 | 151 | e.raise_exception() |
|
152 | 152 | except error.RemoteError as e: |
|
153 | 153 | self.assertEqual(etype.__name__, e.ename, "Should have raised %r, but raised %r"%(etype.__name__, e.ename)) |
|
154 | 154 | else: |
|
155 | 155 | self.fail("should have raised a RemoteError") |
|
156 | 156 | |
|
157 | 157 | def _wait_for(self, f, timeout=10): |
|
158 | 158 | """wait for a condition""" |
|
159 | 159 | tic = time.time() |
|
160 | 160 | while time.time() <= tic + timeout: |
|
161 | 161 | if f(): |
|
162 | 162 | return |
|
163 | 163 | time.sleep(0.1) |
|
164 | 164 | self.client.spin() |
|
165 | 165 | if not f(): |
|
166 | 166 | print("Warning: Awaited condition never arrived") |
|
167 | 167 | |
|
168 | 168 | def setUp(self): |
|
169 | 169 | BaseZMQTestCase.setUp(self) |
|
170 | 170 | self.client = self.connect_client() |
|
171 | 171 | # start every test with clean engine namespaces: |
|
172 | 172 | self.client.clear(block=True) |
|
173 | 173 | self.base_engine_count=len(self.client.ids) |
|
174 | 174 | self.engines=[] |
|
175 | 175 | |
|
176 | 176 | def tearDown(self): |
|
177 | 177 | # self.client.clear(block=True) |
|
178 | 178 | # close fds: |
|
179 | 179 | for e in filter(lambda e: e.poll() is not None, launchers): |
|
180 | 180 | launchers.remove(e) |
|
181 | 181 | |
|
182 | 182 | # allow flushing of incoming messages to prevent crash on socket close |
|
183 | 183 | self.client.wait(timeout=2) |
|
184 | 184 | # time.sleep(2) |
|
185 | 185 | self.client.spin() |
|
186 | 186 | self.client.close() |
|
187 | 187 | BaseZMQTestCase.tearDown(self) |
|
188 | 188 | # this will be redundant when pyzmq merges PR #88 |
|
189 | 189 | # self.context.term() |
|
190 | 190 | # print tempfile.TemporaryFile().fileno(), |
|
191 | 191 | # sys.stdout.flush() |
|
192 | 192 |
@@ -1,389 +1,389 b'' | |||
|
1 | 1 | """Some generic utilities for dealing with classes, urls, and serialization.""" |
|
2 | 2 | |
|
3 | 3 | # Copyright (c) IPython Development Team. |
|
4 | 4 | # Distributed under the terms of the Modified BSD License. |
|
5 | 5 | |
|
6 | 6 | import logging |
|
7 | 7 | import os |
|
8 | 8 | import re |
|
9 | 9 | import stat |
|
10 | 10 | import socket |
|
11 | 11 | import sys |
|
12 | 12 | import warnings |
|
13 | 13 | from signal import signal, SIGINT, SIGABRT, SIGTERM |
|
14 | 14 | try: |
|
15 | 15 | from signal import SIGKILL |
|
16 | 16 | except ImportError: |
|
17 | 17 | SIGKILL=None |
|
18 | 18 | from types import FunctionType |
|
19 | 19 | |
|
20 | 20 | try: |
|
21 | 21 | import cPickle |
|
22 | 22 | pickle = cPickle |
|
23 | 23 | except: |
|
24 | 24 | cPickle = None |
|
25 | 25 | import pickle |
|
26 | 26 | |
|
27 | 27 | import zmq |
|
28 | 28 | from zmq.log import handlers |
|
29 | 29 | |
|
30 | 30 | from IPython.utils.log import get_logger |
|
31 |
from |
|
|
31 | from decorator import decorator | |
|
32 | 32 | |
|
33 | 33 | from IPython.config.application import Application |
|
34 | 34 | from IPython.utils.localinterfaces import localhost, is_public_ip, public_ips |
|
35 | 35 | from IPython.utils.py3compat import string_types, iteritems, itervalues |
|
36 | 36 | from IPython.kernel.zmq.log import EnginePUBHandler |
|
37 | 37 | |
|
38 | 38 | |
|
39 | 39 | #----------------------------------------------------------------------------- |
|
40 | 40 | # Classes |
|
41 | 41 | #----------------------------------------------------------------------------- |
|
42 | 42 | |
|
43 | 43 | class Namespace(dict): |
|
44 | 44 | """Subclass of dict for attribute access to keys.""" |
|
45 | 45 | |
|
46 | 46 | def __getattr__(self, key): |
|
47 | 47 | """getattr aliased to getitem""" |
|
48 | 48 | if key in self: |
|
49 | 49 | return self[key] |
|
50 | 50 | else: |
|
51 | 51 | raise NameError(key) |
|
52 | 52 | |
|
53 | 53 | def __setattr__(self, key, value): |
|
54 | 54 | """setattr aliased to setitem, with strict""" |
|
55 | 55 | if hasattr(dict, key): |
|
56 | 56 | raise KeyError("Cannot override dict keys %r"%key) |
|
57 | 57 | self[key] = value |
|
58 | 58 | |
|
59 | 59 | |
|
60 | 60 | class ReverseDict(dict): |
|
61 | 61 | """simple double-keyed subset of dict methods.""" |
|
62 | 62 | |
|
63 | 63 | def __init__(self, *args, **kwargs): |
|
64 | 64 | dict.__init__(self, *args, **kwargs) |
|
65 | 65 | self._reverse = dict() |
|
66 | 66 | for key, value in iteritems(self): |
|
67 | 67 | self._reverse[value] = key |
|
68 | 68 | |
|
69 | 69 | def __getitem__(self, key): |
|
70 | 70 | try: |
|
71 | 71 | return dict.__getitem__(self, key) |
|
72 | 72 | except KeyError: |
|
73 | 73 | return self._reverse[key] |
|
74 | 74 | |
|
75 | 75 | def __setitem__(self, key, value): |
|
76 | 76 | if key in self._reverse: |
|
77 | 77 | raise KeyError("Can't have key %r on both sides!"%key) |
|
78 | 78 | dict.__setitem__(self, key, value) |
|
79 | 79 | self._reverse[value] = key |
|
80 | 80 | |
|
81 | 81 | def pop(self, key): |
|
82 | 82 | value = dict.pop(self, key) |
|
83 | 83 | self._reverse.pop(value) |
|
84 | 84 | return value |
|
85 | 85 | |
|
86 | 86 | def get(self, key, default=None): |
|
87 | 87 | try: |
|
88 | 88 | return self[key] |
|
89 | 89 | except KeyError: |
|
90 | 90 | return default |
|
91 | 91 | |
|
92 | 92 | #----------------------------------------------------------------------------- |
|
93 | 93 | # Functions |
|
94 | 94 | #----------------------------------------------------------------------------- |
|
95 | 95 | |
|
96 | 96 | @decorator |
|
97 | 97 | def log_errors(f, self, *args, **kwargs): |
|
98 | 98 | """decorator to log unhandled exceptions raised in a method. |
|
99 | 99 | |
|
100 | 100 | For use wrapping on_recv callbacks, so that exceptions |
|
101 | 101 | do not cause the stream to be closed. |
|
102 | 102 | """ |
|
103 | 103 | try: |
|
104 | 104 | return f(self, *args, **kwargs) |
|
105 | 105 | except Exception: |
|
106 | 106 | self.log.error("Uncaught exception in %r" % f, exc_info=True) |
|
107 | 107 | |
|
108 | 108 | |
|
109 | 109 | def is_url(url): |
|
110 | 110 | """boolean check for whether a string is a zmq url""" |
|
111 | 111 | if '://' not in url: |
|
112 | 112 | return False |
|
113 | 113 | proto, addr = url.split('://', 1) |
|
114 | 114 | if proto.lower() not in ['tcp','pgm','epgm','ipc','inproc']: |
|
115 | 115 | return False |
|
116 | 116 | return True |
|
117 | 117 | |
|
118 | 118 | def validate_url(url): |
|
119 | 119 | """validate a url for zeromq""" |
|
120 | 120 | if not isinstance(url, string_types): |
|
121 | 121 | raise TypeError("url must be a string, not %r"%type(url)) |
|
122 | 122 | url = url.lower() |
|
123 | 123 | |
|
124 | 124 | proto_addr = url.split('://') |
|
125 | 125 | assert len(proto_addr) == 2, 'Invalid url: %r'%url |
|
126 | 126 | proto, addr = proto_addr |
|
127 | 127 | assert proto in ['tcp','pgm','epgm','ipc','inproc'], "Invalid protocol: %r"%proto |
|
128 | 128 | |
|
129 | 129 | # domain pattern adapted from http://www.regexlib.com/REDetails.aspx?regexp_id=391 |
|
130 | 130 | # author: Remi Sabourin |
|
131 | 131 | pat = re.compile(r'^([\w\d]([\w\d\-]{0,61}[\w\d])?\.)*[\w\d]([\w\d\-]{0,61}[\w\d])?$') |
|
132 | 132 | |
|
133 | 133 | if proto == 'tcp': |
|
134 | 134 | lis = addr.split(':') |
|
135 | 135 | assert len(lis) == 2, 'Invalid url: %r'%url |
|
136 | 136 | addr,s_port = lis |
|
137 | 137 | try: |
|
138 | 138 | port = int(s_port) |
|
139 | 139 | except ValueError: |
|
140 | 140 | raise AssertionError("Invalid port %r in url: %r"%(port, url)) |
|
141 | 141 | |
|
142 | 142 | assert addr == '*' or pat.match(addr) is not None, 'Invalid url: %r'%url |
|
143 | 143 | |
|
144 | 144 | else: |
|
145 | 145 | # only validate tcp urls currently |
|
146 | 146 | pass |
|
147 | 147 | |
|
148 | 148 | return True |
|
149 | 149 | |
|
150 | 150 | |
|
151 | 151 | def validate_url_container(container): |
|
152 | 152 | """validate a potentially nested collection of urls.""" |
|
153 | 153 | if isinstance(container, string_types): |
|
154 | 154 | url = container |
|
155 | 155 | return validate_url(url) |
|
156 | 156 | elif isinstance(container, dict): |
|
157 | 157 | container = itervalues(container) |
|
158 | 158 | |
|
159 | 159 | for element in container: |
|
160 | 160 | validate_url_container(element) |
|
161 | 161 | |
|
162 | 162 | |
|
163 | 163 | def split_url(url): |
|
164 | 164 | """split a zmq url (tcp://ip:port) into ('tcp','ip','port').""" |
|
165 | 165 | proto_addr = url.split('://') |
|
166 | 166 | assert len(proto_addr) == 2, 'Invalid url: %r'%url |
|
167 | 167 | proto, addr = proto_addr |
|
168 | 168 | lis = addr.split(':') |
|
169 | 169 | assert len(lis) == 2, 'Invalid url: %r'%url |
|
170 | 170 | addr,s_port = lis |
|
171 | 171 | return proto,addr,s_port |
|
172 | 172 | |
|
173 | 173 | |
|
174 | 174 | def disambiguate_ip_address(ip, location=None): |
|
175 | 175 | """turn multi-ip interfaces '0.0.0.0' and '*' into a connectable address |
|
176 | 176 | |
|
177 | 177 | Explicit IP addresses are returned unmodified. |
|
178 | 178 | |
|
179 | 179 | Parameters |
|
180 | 180 | ---------- |
|
181 | 181 | |
|
182 | 182 | ip : IP address |
|
183 | 183 | An IP address, or the special values 0.0.0.0, or * |
|
184 | 184 | location: IP address, optional |
|
185 | 185 | A public IP of the target machine. |
|
186 | 186 | If location is an IP of the current machine, |
|
187 | 187 | localhost will be returned, |
|
188 | 188 | otherwise location will be returned. |
|
189 | 189 | """ |
|
190 | 190 | if ip in {'0.0.0.0', '*'}: |
|
191 | 191 | if not location: |
|
192 | 192 | # unspecified location, localhost is the only choice |
|
193 | 193 | ip = localhost() |
|
194 | 194 | elif is_public_ip(location): |
|
195 | 195 | # location is a public IP on this machine, use localhost |
|
196 | 196 | ip = localhost() |
|
197 | 197 | elif not public_ips(): |
|
198 | 198 | # this machine's public IPs cannot be determined, |
|
199 | 199 | # assume `location` is not this machine |
|
200 | 200 | warnings.warn("IPython could not determine public IPs", RuntimeWarning) |
|
201 | 201 | ip = location |
|
202 | 202 | else: |
|
203 | 203 | # location is not this machine, do not use loopback |
|
204 | 204 | ip = location |
|
205 | 205 | return ip |
|
206 | 206 | |
|
207 | 207 | |
|
208 | 208 | def disambiguate_url(url, location=None): |
|
209 | 209 | """turn multi-ip interfaces '0.0.0.0' and '*' into connectable |
|
210 | 210 | ones, based on the location (default interpretation is localhost). |
|
211 | 211 | |
|
212 | 212 | This is for zeromq urls, such as ``tcp://*:10101``. |
|
213 | 213 | """ |
|
214 | 214 | try: |
|
215 | 215 | proto,ip,port = split_url(url) |
|
216 | 216 | except AssertionError: |
|
217 | 217 | # probably not tcp url; could be ipc, etc. |
|
218 | 218 | return url |
|
219 | 219 | |
|
220 | 220 | ip = disambiguate_ip_address(ip,location) |
|
221 | 221 | |
|
222 | 222 | return "%s://%s:%s"%(proto,ip,port) |
|
223 | 223 | |
|
224 | 224 | |
|
225 | 225 | #-------------------------------------------------------------------------- |
|
226 | 226 | # helpers for implementing old MEC API via view.apply |
|
227 | 227 | #-------------------------------------------------------------------------- |
|
228 | 228 | |
|
229 | 229 | def interactive(f): |
|
230 | 230 | """decorator for making functions appear as interactively defined. |
|
231 | 231 | This results in the function being linked to the user_ns as globals() |
|
232 | 232 | instead of the module globals(). |
|
233 | 233 | """ |
|
234 | 234 | |
|
235 | 235 | # build new FunctionType, so it can have the right globals |
|
236 | 236 | # interactive functions never have closures, that's kind of the point |
|
237 | 237 | if isinstance(f, FunctionType): |
|
238 | 238 | mainmod = __import__('__main__') |
|
239 | 239 | f = FunctionType(f.__code__, mainmod.__dict__, |
|
240 | 240 | f.__name__, f.__defaults__, |
|
241 | 241 | ) |
|
242 | 242 | # associate with __main__ for uncanning |
|
243 | 243 | f.__module__ = '__main__' |
|
244 | 244 | return f |
|
245 | 245 | |
|
246 | 246 | @interactive |
|
247 | 247 | def _push(**ns): |
|
248 | 248 | """helper method for implementing `client.push` via `client.apply`""" |
|
249 | 249 | user_ns = globals() |
|
250 | 250 | tmp = '_IP_PUSH_TMP_' |
|
251 | 251 | while tmp in user_ns: |
|
252 | 252 | tmp = tmp + '_' |
|
253 | 253 | try: |
|
254 | 254 | for name, value in ns.items(): |
|
255 | 255 | user_ns[tmp] = value |
|
256 | 256 | exec("%s = %s" % (name, tmp), user_ns) |
|
257 | 257 | finally: |
|
258 | 258 | user_ns.pop(tmp, None) |
|
259 | 259 | |
|
260 | 260 | @interactive |
|
261 | 261 | def _pull(keys): |
|
262 | 262 | """helper method for implementing `client.pull` via `client.apply`""" |
|
263 | 263 | if isinstance(keys, (list,tuple, set)): |
|
264 | 264 | return [eval(key, globals()) for key in keys] |
|
265 | 265 | else: |
|
266 | 266 | return eval(keys, globals()) |
|
267 | 267 | |
|
268 | 268 | @interactive |
|
269 | 269 | def _execute(code): |
|
270 | 270 | """helper method for implementing `client.execute` via `client.apply`""" |
|
271 | 271 | exec(code, globals()) |
|
272 | 272 | |
|
273 | 273 | #-------------------------------------------------------------------------- |
|
274 | 274 | # extra process management utilities |
|
275 | 275 | #-------------------------------------------------------------------------- |
|
276 | 276 | |
|
277 | 277 | _random_ports = set() |
|
278 | 278 | |
|
279 | 279 | def select_random_ports(n): |
|
280 | 280 | """Selects and return n random ports that are available.""" |
|
281 | 281 | ports = [] |
|
282 | 282 | for i in range(n): |
|
283 | 283 | sock = socket.socket() |
|
284 | 284 | sock.bind(('', 0)) |
|
285 | 285 | while sock.getsockname()[1] in _random_ports: |
|
286 | 286 | sock.close() |
|
287 | 287 | sock = socket.socket() |
|
288 | 288 | sock.bind(('', 0)) |
|
289 | 289 | ports.append(sock) |
|
290 | 290 | for i, sock in enumerate(ports): |
|
291 | 291 | port = sock.getsockname()[1] |
|
292 | 292 | sock.close() |
|
293 | 293 | ports[i] = port |
|
294 | 294 | _random_ports.add(port) |
|
295 | 295 | return ports |
|
296 | 296 | |
|
297 | 297 | def signal_children(children): |
|
298 | 298 | """Relay interupt/term signals to children, for more solid process cleanup.""" |
|
299 | 299 | def terminate_children(sig, frame): |
|
300 | 300 | log = get_logger() |
|
301 | 301 | log.critical("Got signal %i, terminating children..."%sig) |
|
302 | 302 | for child in children: |
|
303 | 303 | child.terminate() |
|
304 | 304 | |
|
305 | 305 | sys.exit(sig != SIGINT) |
|
306 | 306 | # sys.exit(sig) |
|
307 | 307 | for sig in (SIGINT, SIGABRT, SIGTERM): |
|
308 | 308 | signal(sig, terminate_children) |
|
309 | 309 | |
|
310 | 310 | def generate_exec_key(keyfile): |
|
311 | 311 | import uuid |
|
312 | 312 | newkey = str(uuid.uuid4()) |
|
313 | 313 | with open(keyfile, 'w') as f: |
|
314 | 314 | # f.write('ipython-key ') |
|
315 | 315 | f.write(newkey+'\n') |
|
316 | 316 | # set user-only RW permissions (0600) |
|
317 | 317 | # this will have no effect on Windows |
|
318 | 318 | os.chmod(keyfile, stat.S_IRUSR|stat.S_IWUSR) |
|
319 | 319 | |
|
320 | 320 | |
|
321 | 321 | def integer_loglevel(loglevel): |
|
322 | 322 | try: |
|
323 | 323 | loglevel = int(loglevel) |
|
324 | 324 | except ValueError: |
|
325 | 325 | if isinstance(loglevel, str): |
|
326 | 326 | loglevel = getattr(logging, loglevel) |
|
327 | 327 | return loglevel |
|
328 | 328 | |
|
329 | 329 | def connect_logger(logname, context, iface, root="ip", loglevel=logging.DEBUG): |
|
330 | 330 | logger = logging.getLogger(logname) |
|
331 | 331 | if any([isinstance(h, handlers.PUBHandler) for h in logger.handlers]): |
|
332 | 332 | # don't add a second PUBHandler |
|
333 | 333 | return |
|
334 | 334 | loglevel = integer_loglevel(loglevel) |
|
335 | 335 | lsock = context.socket(zmq.PUB) |
|
336 | 336 | lsock.connect(iface) |
|
337 | 337 | handler = handlers.PUBHandler(lsock) |
|
338 | 338 | handler.setLevel(loglevel) |
|
339 | 339 | handler.root_topic = root |
|
340 | 340 | logger.addHandler(handler) |
|
341 | 341 | logger.setLevel(loglevel) |
|
342 | 342 | |
|
343 | 343 | def connect_engine_logger(context, iface, engine, loglevel=logging.DEBUG): |
|
344 | 344 | logger = logging.getLogger() |
|
345 | 345 | if any([isinstance(h, handlers.PUBHandler) for h in logger.handlers]): |
|
346 | 346 | # don't add a second PUBHandler |
|
347 | 347 | return |
|
348 | 348 | loglevel = integer_loglevel(loglevel) |
|
349 | 349 | lsock = context.socket(zmq.PUB) |
|
350 | 350 | lsock.connect(iface) |
|
351 | 351 | handler = EnginePUBHandler(engine, lsock) |
|
352 | 352 | handler.setLevel(loglevel) |
|
353 | 353 | logger.addHandler(handler) |
|
354 | 354 | logger.setLevel(loglevel) |
|
355 | 355 | return logger |
|
356 | 356 | |
|
357 | 357 | def local_logger(logname, loglevel=logging.DEBUG): |
|
358 | 358 | loglevel = integer_loglevel(loglevel) |
|
359 | 359 | logger = logging.getLogger(logname) |
|
360 | 360 | if any([isinstance(h, logging.StreamHandler) for h in logger.handlers]): |
|
361 | 361 | # don't add a second StreamHandler |
|
362 | 362 | return |
|
363 | 363 | handler = logging.StreamHandler() |
|
364 | 364 | handler.setLevel(loglevel) |
|
365 | 365 | formatter = logging.Formatter("%(asctime)s.%(msecs).03d [%(name)s] %(message)s", |
|
366 | 366 | datefmt="%Y-%m-%d %H:%M:%S") |
|
367 | 367 | handler.setFormatter(formatter) |
|
368 | 368 | |
|
369 | 369 | logger.addHandler(handler) |
|
370 | 370 | logger.setLevel(loglevel) |
|
371 | 371 | return logger |
|
372 | 372 | |
|
373 | 373 | def set_hwm(sock, hwm=0): |
|
374 | 374 | """set zmq High Water Mark on a socket |
|
375 | 375 | |
|
376 | 376 | in a way that always works for various pyzmq / libzmq versions. |
|
377 | 377 | """ |
|
378 | 378 | import zmq |
|
379 | 379 | |
|
380 | 380 | for key in ('HWM', 'SNDHWM', 'RCVHWM'): |
|
381 | 381 | opt = getattr(zmq, key, None) |
|
382 | 382 | if opt is None: |
|
383 | 383 | continue |
|
384 | 384 | try: |
|
385 | 385 | sock.setsockopt(opt, hwm) |
|
386 | 386 | except zmq.ZMQError: |
|
387 | 387 | pass |
|
388 | 388 | |
|
389 | 389 |
@@ -1,400 +1,400 b'' | |||
|
1 | 1 | # -*- coding: utf-8 -*- |
|
2 | 2 | """Decorators for labeling test objects. |
|
3 | 3 | |
|
4 | 4 | Decorators that merely return a modified version of the original function |
|
5 | 5 | object are straightforward. Decorators that return a new function object need |
|
6 | 6 | to use nose.tools.make_decorator(original_function)(decorator) in returning the |
|
7 | 7 | decorator, in order to preserve metadata such as function name, setup and |
|
8 | 8 | teardown functions and so on - see nose.tools for more information. |
|
9 | 9 | |
|
10 | 10 | This module provides a set of useful decorators meant to be ready to use in |
|
11 | 11 | your own tests. See the bottom of the file for the ready-made ones, and if you |
|
12 | 12 | find yourself writing a new one that may be of generic use, add it here. |
|
13 | 13 | |
|
14 | 14 | Included decorators: |
|
15 | 15 | |
|
16 | 16 | |
|
17 | 17 | Lightweight testing that remains unittest-compatible. |
|
18 | 18 | |
|
19 | 19 | - An @as_unittest decorator can be used to tag any normal parameter-less |
|
20 | 20 | function as a unittest TestCase. Then, both nose and normal unittest will |
|
21 | 21 | recognize it as such. This will make it easier to migrate away from Nose if |
|
22 | 22 | we ever need/want to while maintaining very lightweight tests. |
|
23 | 23 | |
|
24 | 24 | NOTE: This file contains IPython-specific decorators. Using the machinery in |
|
25 | 25 | IPython.external.decorators, we import either numpy.testing.decorators if numpy is |
|
26 | 26 | available, OR use equivalent code in IPython.external._decorators, which |
|
27 | 27 | we've copied verbatim from numpy. |
|
28 | 28 | |
|
29 | 29 | Authors |
|
30 | 30 | ------- |
|
31 | 31 | |
|
32 | 32 | - Fernando Perez <Fernando.Perez@berkeley.edu> |
|
33 | 33 | """ |
|
34 | 34 | |
|
35 | 35 | #----------------------------------------------------------------------------- |
|
36 | 36 | # Copyright (C) 2009-2011 The IPython Development Team |
|
37 | 37 | # |
|
38 | 38 | # Distributed under the terms of the BSD License. The full license is in |
|
39 | 39 | # the file COPYING, distributed as part of this software. |
|
40 | 40 | #----------------------------------------------------------------------------- |
|
41 | 41 | |
|
42 | 42 | #----------------------------------------------------------------------------- |
|
43 | 43 | # Imports |
|
44 | 44 | #----------------------------------------------------------------------------- |
|
45 | 45 | |
|
46 | 46 | # Stdlib imports |
|
47 | 47 | import sys |
|
48 | 48 | import os |
|
49 | 49 | import tempfile |
|
50 | 50 | import unittest |
|
51 | 51 | |
|
52 | 52 | # Third-party imports |
|
53 | 53 | |
|
54 | 54 | # This is Michele Simionato's decorator module, kept verbatim. |
|
55 |
from |
|
|
55 | from decorator import decorator | |
|
56 | 56 | |
|
57 | 57 | # Expose the unittest-driven decorators |
|
58 | 58 | from .ipunittest import ipdoctest, ipdocstring |
|
59 | 59 | |
|
60 | 60 | # Grab the numpy-specific decorators which we keep in a file that we |
|
61 | 61 | # occasionally update from upstream: decorators.py is a copy of |
|
62 | 62 | # numpy.testing.decorators, we expose all of it here. |
|
63 | 63 | from IPython.external.decorators import * |
|
64 | 64 | |
|
65 | 65 | # For onlyif_cmd_exists decorator |
|
66 | 66 | from IPython.utils.process import is_cmd_found |
|
67 | 67 | from IPython.utils.py3compat import string_types |
|
68 | 68 | |
|
69 | 69 | #----------------------------------------------------------------------------- |
|
70 | 70 | # Classes and functions |
|
71 | 71 | #----------------------------------------------------------------------------- |
|
72 | 72 | |
|
73 | 73 | # Simple example of the basic idea |
|
74 | 74 | def as_unittest(func): |
|
75 | 75 | """Decorator to make a simple function into a normal test via unittest.""" |
|
76 | 76 | class Tester(unittest.TestCase): |
|
77 | 77 | def test(self): |
|
78 | 78 | func() |
|
79 | 79 | |
|
80 | 80 | Tester.__name__ = func.__name__ |
|
81 | 81 | |
|
82 | 82 | return Tester |
|
83 | 83 | |
|
84 | 84 | # Utility functions |
|
85 | 85 | |
|
86 | 86 | def apply_wrapper(wrapper,func): |
|
87 | 87 | """Apply a wrapper to a function for decoration. |
|
88 | 88 | |
|
89 | 89 | This mixes Michele Simionato's decorator tool with nose's make_decorator, |
|
90 | 90 | to apply a wrapper in a decorator so that all nose attributes, as well as |
|
91 | 91 | function signature and other properties, survive the decoration cleanly. |
|
92 | 92 | This will ensure that wrapped functions can still be well introspected via |
|
93 | 93 | IPython, for example. |
|
94 | 94 | """ |
|
95 | 95 | import nose.tools |
|
96 | 96 | |
|
97 | 97 | return decorator(wrapper,nose.tools.make_decorator(func)(wrapper)) |
|
98 | 98 | |
|
99 | 99 | |
|
100 | 100 | def make_label_dec(label,ds=None): |
|
101 | 101 | """Factory function to create a decorator that applies one or more labels. |
|
102 | 102 | |
|
103 | 103 | Parameters |
|
104 | 104 | ---------- |
|
105 | 105 | label : string or sequence |
|
106 | 106 | One or more labels that will be applied by the decorator to the functions |
|
107 | 107 | it decorates. Labels are attributes of the decorated function with their |
|
108 | 108 | value set to True. |
|
109 | 109 | |
|
110 | 110 | ds : string |
|
111 | 111 | An optional docstring for the resulting decorator. If not given, a |
|
112 | 112 | default docstring is auto-generated. |
|
113 | 113 | |
|
114 | 114 | Returns |
|
115 | 115 | ------- |
|
116 | 116 | A decorator. |
|
117 | 117 | |
|
118 | 118 | Examples |
|
119 | 119 | -------- |
|
120 | 120 | |
|
121 | 121 | A simple labeling decorator: |
|
122 | 122 | |
|
123 | 123 | >>> slow = make_label_dec('slow') |
|
124 | 124 | >>> slow.__doc__ |
|
125 | 125 | "Labels a test as 'slow'." |
|
126 | 126 | |
|
127 | 127 | And one that uses multiple labels and a custom docstring: |
|
128 | 128 | |
|
129 | 129 | >>> rare = make_label_dec(['slow','hard'], |
|
130 | 130 | ... "Mix labels 'slow' and 'hard' for rare tests.") |
|
131 | 131 | >>> rare.__doc__ |
|
132 | 132 | "Mix labels 'slow' and 'hard' for rare tests." |
|
133 | 133 | |
|
134 | 134 | Now, let's test using this one: |
|
135 | 135 | >>> @rare |
|
136 | 136 | ... def f(): pass |
|
137 | 137 | ... |
|
138 | 138 | >>> |
|
139 | 139 | >>> f.slow |
|
140 | 140 | True |
|
141 | 141 | >>> f.hard |
|
142 | 142 | True |
|
143 | 143 | """ |
|
144 | 144 | |
|
145 | 145 | if isinstance(label, string_types): |
|
146 | 146 | labels = [label] |
|
147 | 147 | else: |
|
148 | 148 | labels = label |
|
149 | 149 | |
|
150 | 150 | # Validate that the given label(s) are OK for use in setattr() by doing a |
|
151 | 151 | # dry run on a dummy function. |
|
152 | 152 | tmp = lambda : None |
|
153 | 153 | for label in labels: |
|
154 | 154 | setattr(tmp,label,True) |
|
155 | 155 | |
|
156 | 156 | # This is the actual decorator we'll return |
|
157 | 157 | def decor(f): |
|
158 | 158 | for label in labels: |
|
159 | 159 | setattr(f,label,True) |
|
160 | 160 | return f |
|
161 | 161 | |
|
162 | 162 | # Apply the user's docstring, or autogenerate a basic one |
|
163 | 163 | if ds is None: |
|
164 | 164 | ds = "Labels a test as %r." % label |
|
165 | 165 | decor.__doc__ = ds |
|
166 | 166 | |
|
167 | 167 | return decor |
|
168 | 168 | |
|
169 | 169 | |
|
170 | 170 | # Inspired by numpy's skipif, but uses the full apply_wrapper utility to |
|
171 | 171 | # preserve function metadata better and allows the skip condition to be a |
|
172 | 172 | # callable. |
|
173 | 173 | def skipif(skip_condition, msg=None): |
|
174 | 174 | ''' Make function raise SkipTest exception if skip_condition is true |
|
175 | 175 | |
|
176 | 176 | Parameters |
|
177 | 177 | ---------- |
|
178 | 178 | |
|
179 | 179 | skip_condition : bool or callable |
|
180 | 180 | Flag to determine whether to skip test. If the condition is a |
|
181 | 181 | callable, it is used at runtime to dynamically make the decision. This |
|
182 | 182 | is useful for tests that may require costly imports, to delay the cost |
|
183 | 183 | until the test suite is actually executed. |
|
184 | 184 | msg : string |
|
185 | 185 | Message to give on raising a SkipTest exception. |
|
186 | 186 | |
|
187 | 187 | Returns |
|
188 | 188 | ------- |
|
189 | 189 | decorator : function |
|
190 | 190 | Decorator, which, when applied to a function, causes SkipTest |
|
191 | 191 | to be raised when the skip_condition was True, and the function |
|
192 | 192 | to be called normally otherwise. |
|
193 | 193 | |
|
194 | 194 | Notes |
|
195 | 195 | ----- |
|
196 | 196 | You will see from the code that we had to further decorate the |
|
197 | 197 | decorator with the nose.tools.make_decorator function in order to |
|
198 | 198 | transmit function name, and various other metadata. |
|
199 | 199 | ''' |
|
200 | 200 | |
|
201 | 201 | def skip_decorator(f): |
|
202 | 202 | # Local import to avoid a hard nose dependency and only incur the |
|
203 | 203 | # import time overhead at actual test-time. |
|
204 | 204 | import nose |
|
205 | 205 | |
|
206 | 206 | # Allow for both boolean or callable skip conditions. |
|
207 | 207 | if callable(skip_condition): |
|
208 | 208 | skip_val = skip_condition |
|
209 | 209 | else: |
|
210 | 210 | skip_val = lambda : skip_condition |
|
211 | 211 | |
|
212 | 212 | def get_msg(func,msg=None): |
|
213 | 213 | """Skip message with information about function being skipped.""" |
|
214 | 214 | if msg is None: out = 'Test skipped due to test condition.' |
|
215 | 215 | else: out = msg |
|
216 | 216 | return "Skipping test: %s. %s" % (func.__name__,out) |
|
217 | 217 | |
|
218 | 218 | # We need to define *two* skippers because Python doesn't allow both |
|
219 | 219 | # return with value and yield inside the same function. |
|
220 | 220 | def skipper_func(*args, **kwargs): |
|
221 | 221 | """Skipper for normal test functions.""" |
|
222 | 222 | if skip_val(): |
|
223 | 223 | raise nose.SkipTest(get_msg(f,msg)) |
|
224 | 224 | else: |
|
225 | 225 | return f(*args, **kwargs) |
|
226 | 226 | |
|
227 | 227 | def skipper_gen(*args, **kwargs): |
|
228 | 228 | """Skipper for test generators.""" |
|
229 | 229 | if skip_val(): |
|
230 | 230 | raise nose.SkipTest(get_msg(f,msg)) |
|
231 | 231 | else: |
|
232 | 232 | for x in f(*args, **kwargs): |
|
233 | 233 | yield x |
|
234 | 234 | |
|
235 | 235 | # Choose the right skipper to use when building the actual generator. |
|
236 | 236 | if nose.util.isgenerator(f): |
|
237 | 237 | skipper = skipper_gen |
|
238 | 238 | else: |
|
239 | 239 | skipper = skipper_func |
|
240 | 240 | |
|
241 | 241 | return nose.tools.make_decorator(f)(skipper) |
|
242 | 242 | |
|
243 | 243 | return skip_decorator |
|
244 | 244 | |
|
245 | 245 | # A version with the condition set to true, common case just to attach a message |
|
246 | 246 | # to a skip decorator |
|
247 | 247 | def skip(msg=None): |
|
248 | 248 | """Decorator factory - mark a test function for skipping from test suite. |
|
249 | 249 | |
|
250 | 250 | Parameters |
|
251 | 251 | ---------- |
|
252 | 252 | msg : string |
|
253 | 253 | Optional message to be added. |
|
254 | 254 | |
|
255 | 255 | Returns |
|
256 | 256 | ------- |
|
257 | 257 | decorator : function |
|
258 | 258 | Decorator, which, when applied to a function, causes SkipTest |
|
259 | 259 | to be raised, with the optional message added. |
|
260 | 260 | """ |
|
261 | 261 | |
|
262 | 262 | return skipif(True,msg) |
|
263 | 263 | |
|
264 | 264 | |
|
265 | 265 | def onlyif(condition, msg): |
|
266 | 266 | """The reverse from skipif, see skipif for details.""" |
|
267 | 267 | |
|
268 | 268 | if callable(condition): |
|
269 | 269 | skip_condition = lambda : not condition() |
|
270 | 270 | else: |
|
271 | 271 | skip_condition = lambda : not condition |
|
272 | 272 | |
|
273 | 273 | return skipif(skip_condition, msg) |
|
274 | 274 | |
|
275 | 275 | #----------------------------------------------------------------------------- |
|
276 | 276 | # Utility functions for decorators |
|
277 | 277 | def module_not_available(module): |
|
278 | 278 | """Can module be imported? Returns true if module does NOT import. |
|
279 | 279 | |
|
280 | 280 | This is used to make a decorator to skip tests that require module to be |
|
281 | 281 | available, but delay the 'import numpy' to test execution time. |
|
282 | 282 | """ |
|
283 | 283 | try: |
|
284 | 284 | mod = __import__(module) |
|
285 | 285 | mod_not_avail = False |
|
286 | 286 | except ImportError: |
|
287 | 287 | mod_not_avail = True |
|
288 | 288 | |
|
289 | 289 | return mod_not_avail |
|
290 | 290 | |
|
291 | 291 | |
|
292 | 292 | def decorated_dummy(dec, name): |
|
293 | 293 | """Return a dummy function decorated with dec, with the given name. |
|
294 | 294 | |
|
295 | 295 | Examples |
|
296 | 296 | -------- |
|
297 | 297 | import IPython.testing.decorators as dec |
|
298 | 298 | setup = dec.decorated_dummy(dec.skip_if_no_x11, __name__) |
|
299 | 299 | """ |
|
300 | 300 | dummy = lambda: None |
|
301 | 301 | dummy.__name__ = name |
|
302 | 302 | return dec(dummy) |
|
303 | 303 | |
|
304 | 304 | #----------------------------------------------------------------------------- |
|
305 | 305 | # Decorators for public use |
|
306 | 306 | |
|
307 | 307 | # Decorators to skip certain tests on specific platforms. |
|
308 | 308 | skip_win32 = skipif(sys.platform == 'win32', |
|
309 | 309 | "This test does not run under Windows") |
|
310 | 310 | skip_linux = skipif(sys.platform.startswith('linux'), |
|
311 | 311 | "This test does not run under Linux") |
|
312 | 312 | skip_osx = skipif(sys.platform == 'darwin',"This test does not run under OS X") |
|
313 | 313 | |
|
314 | 314 | |
|
315 | 315 | # Decorators to skip tests if not on specific platforms. |
|
316 | 316 | skip_if_not_win32 = skipif(sys.platform != 'win32', |
|
317 | 317 | "This test only runs under Windows") |
|
318 | 318 | skip_if_not_linux = skipif(not sys.platform.startswith('linux'), |
|
319 | 319 | "This test only runs under Linux") |
|
320 | 320 | skip_if_not_osx = skipif(sys.platform != 'darwin', |
|
321 | 321 | "This test only runs under OSX") |
|
322 | 322 | |
|
323 | 323 | |
|
324 | 324 | _x11_skip_cond = (sys.platform not in ('darwin', 'win32') and |
|
325 | 325 | os.environ.get('DISPLAY', '') == '') |
|
326 | 326 | _x11_skip_msg = "Skipped under *nix when X11/XOrg not available" |
|
327 | 327 | |
|
328 | 328 | skip_if_no_x11 = skipif(_x11_skip_cond, _x11_skip_msg) |
|
329 | 329 | |
|
330 | 330 | # not a decorator itself, returns a dummy function to be used as setup |
|
331 | 331 | def skip_file_no_x11(name): |
|
332 | 332 | return decorated_dummy(skip_if_no_x11, name) if _x11_skip_cond else None |
|
333 | 333 | |
|
334 | 334 | # Other skip decorators |
|
335 | 335 | |
|
336 | 336 | # generic skip without module |
|
337 | 337 | skip_without = lambda mod: skipif(module_not_available(mod), "This test requires %s" % mod) |
|
338 | 338 | |
|
339 | 339 | skipif_not_numpy = skip_without('numpy') |
|
340 | 340 | |
|
341 | 341 | skipif_not_matplotlib = skip_without('matplotlib') |
|
342 | 342 | |
|
343 | 343 | skipif_not_sympy = skip_without('sympy') |
|
344 | 344 | |
|
345 | 345 | skip_known_failure = knownfailureif(True,'This test is known to fail') |
|
346 | 346 | |
|
347 | 347 | known_failure_py3 = knownfailureif(sys.version_info[0] >= 3, |
|
348 | 348 | 'This test is known to fail on Python 3.') |
|
349 | 349 | |
|
350 | 350 | # A null 'decorator', useful to make more readable code that needs to pick |
|
351 | 351 | # between different decorators based on OS or other conditions |
|
352 | 352 | null_deco = lambda f: f |
|
353 | 353 | |
|
354 | 354 | # Some tests only run where we can use unicode paths. Note that we can't just |
|
355 | 355 | # check os.path.supports_unicode_filenames, which is always False on Linux. |
|
356 | 356 | try: |
|
357 | 357 | f = tempfile.NamedTemporaryFile(prefix=u"tmp€") |
|
358 | 358 | except UnicodeEncodeError: |
|
359 | 359 | unicode_paths = False |
|
360 | 360 | else: |
|
361 | 361 | unicode_paths = True |
|
362 | 362 | f.close() |
|
363 | 363 | |
|
364 | 364 | onlyif_unicode_paths = onlyif(unicode_paths, ("This test is only applicable " |
|
365 | 365 | "where we can use unicode in filenames.")) |
|
366 | 366 | |
|
367 | 367 | |
|
368 | 368 | def onlyif_cmds_exist(*commands): |
|
369 | 369 | """ |
|
370 | 370 | Decorator to skip test when at least one of `commands` is not found. |
|
371 | 371 | """ |
|
372 | 372 | for cmd in commands: |
|
373 | 373 | try: |
|
374 | 374 | if not is_cmd_found(cmd): |
|
375 | 375 | return skip("This test runs only if command '{0}' " |
|
376 | 376 | "is installed".format(cmd)) |
|
377 | 377 | except ImportError as e: |
|
378 | 378 | # is_cmd_found uses pywin32 on windows, which might not be available |
|
379 | 379 | if sys.platform == 'win32' and 'pywin32' in str(e): |
|
380 | 380 | return skip("This test runs only if pywin32 and command '{0}' " |
|
381 | 381 | "is installed".format(cmd)) |
|
382 | 382 | raise e |
|
383 | 383 | return null_deco |
|
384 | 384 | |
|
385 | 385 | def onlyif_any_cmd_exists(*commands): |
|
386 | 386 | """ |
|
387 | 387 | Decorator to skip test unless at least one of `commands` is found. |
|
388 | 388 | """ |
|
389 | 389 | for cmd in commands: |
|
390 | 390 | try: |
|
391 | 391 | if is_cmd_found(cmd): |
|
392 | 392 | return null_deco |
|
393 | 393 | except ImportError as e: |
|
394 | 394 | # is_cmd_found uses pywin32 on windows, which might not be available |
|
395 | 395 | if sys.platform == 'win32' and 'pywin32' in str(e): |
|
396 | 396 | return skip("This test runs only if pywin32 and commands '{0}' " |
|
397 | 397 | "are installed".format(commands)) |
|
398 | 398 | raise e |
|
399 | 399 | return skip("This test runs only if one of the commands {0} " |
|
400 | 400 | "is installed".format(commands)) |
@@ -1,344 +1,345 b'' | |||
|
1 | 1 | #!/usr/bin/env python |
|
2 | 2 | # -*- coding: utf-8 -*- |
|
3 | 3 | """Setup script for IPython. |
|
4 | 4 | |
|
5 | 5 | Under Posix environments it works like a typical setup.py script. |
|
6 | 6 | Under Windows, the command sdist is not supported, since IPython |
|
7 | 7 | requires utilities which are not available under Windows.""" |
|
8 | 8 | |
|
9 | 9 | #----------------------------------------------------------------------------- |
|
10 | 10 | # Copyright (c) 2008-2011, IPython Development Team. |
|
11 | 11 | # Copyright (c) 2001-2007, Fernando Perez <fernando.perez@colorado.edu> |
|
12 | 12 | # Copyright (c) 2001, Janko Hauser <jhauser@zscout.de> |
|
13 | 13 | # Copyright (c) 2001, Nathaniel Gray <n8gray@caltech.edu> |
|
14 | 14 | # |
|
15 | 15 | # Distributed under the terms of the Modified BSD License. |
|
16 | 16 | # |
|
17 | 17 | # The full license is in the file COPYING.rst, distributed with this software. |
|
18 | 18 | #----------------------------------------------------------------------------- |
|
19 | 19 | |
|
20 | 20 | #----------------------------------------------------------------------------- |
|
21 | 21 | # Minimal Python version sanity check |
|
22 | 22 | #----------------------------------------------------------------------------- |
|
23 | 23 | from __future__ import print_function |
|
24 | 24 | |
|
25 | 25 | import sys |
|
26 | 26 | |
|
27 | 27 | # This check is also made in IPython/__init__, don't forget to update both when |
|
28 | 28 | # changing Python version requirements. |
|
29 | 29 | v = sys.version_info |
|
30 | 30 | if v[:2] < (2,7) or (v[0] >= 3 and v[:2] < (3,3)): |
|
31 | 31 | error = "ERROR: IPython requires Python version 2.7 or 3.3 or above." |
|
32 | 32 | print(error, file=sys.stderr) |
|
33 | 33 | sys.exit(1) |
|
34 | 34 | |
|
35 | 35 | PY3 = (sys.version_info[0] >= 3) |
|
36 | 36 | |
|
37 | 37 | # At least we're on the python version we need, move on. |
|
38 | 38 | |
|
39 | 39 | #------------------------------------------------------------------------------- |
|
40 | 40 | # Imports |
|
41 | 41 | #------------------------------------------------------------------------------- |
|
42 | 42 | |
|
43 | 43 | # Stdlib imports |
|
44 | 44 | import os |
|
45 | 45 | import shutil |
|
46 | 46 | |
|
47 | 47 | from glob import glob |
|
48 | 48 | |
|
49 | 49 | # BEFORE importing distutils, remove MANIFEST. distutils doesn't properly |
|
50 | 50 | # update it when the contents of directories change. |
|
51 | 51 | if os.path.exists('MANIFEST'): os.remove('MANIFEST') |
|
52 | 52 | |
|
53 | 53 | from distutils.core import setup |
|
54 | 54 | |
|
55 | 55 | # Our own imports |
|
56 | 56 | from setupbase import target_update |
|
57 | 57 | |
|
58 | 58 | from setupbase import ( |
|
59 | 59 | setup_args, |
|
60 | 60 | find_packages, |
|
61 | 61 | find_package_data, |
|
62 | 62 | check_package_data_first, |
|
63 | 63 | find_entry_points, |
|
64 | 64 | build_scripts_entrypt, |
|
65 | 65 | find_data_files, |
|
66 | 66 | check_for_dependencies, |
|
67 | 67 | git_prebuild, |
|
68 | 68 | check_submodule_status, |
|
69 | 69 | update_submodules, |
|
70 | 70 | require_submodules, |
|
71 | 71 | UpdateSubmodules, |
|
72 | 72 | get_bdist_wheel, |
|
73 | 73 | CompileCSS, |
|
74 | 74 | JavascriptVersion, |
|
75 | 75 | css_js_prerelease, |
|
76 | 76 | install_symlinked, |
|
77 | 77 | install_lib_symlink, |
|
78 | 78 | install_scripts_for_symlink, |
|
79 | 79 | unsymlink, |
|
80 | 80 | ) |
|
81 | 81 | from setupext import setupext |
|
82 | 82 | |
|
83 | 83 | isfile = os.path.isfile |
|
84 | 84 | pjoin = os.path.join |
|
85 | 85 | |
|
86 | 86 | #------------------------------------------------------------------------------- |
|
87 | 87 | # Handle OS specific things |
|
88 | 88 | #------------------------------------------------------------------------------- |
|
89 | 89 | |
|
90 | 90 | if os.name in ('nt','dos'): |
|
91 | 91 | os_name = 'windows' |
|
92 | 92 | else: |
|
93 | 93 | os_name = os.name |
|
94 | 94 | |
|
95 | 95 | # Under Windows, 'sdist' has not been supported. Now that the docs build with |
|
96 | 96 | # Sphinx it might work, but let's not turn it on until someone confirms that it |
|
97 | 97 | # actually works. |
|
98 | 98 | if os_name == 'windows' and 'sdist' in sys.argv: |
|
99 | 99 | print('The sdist command is not available under Windows. Exiting.') |
|
100 | 100 | sys.exit(1) |
|
101 | 101 | |
|
102 | 102 | #------------------------------------------------------------------------------- |
|
103 | 103 | # Make sure we aren't trying to run without submodules |
|
104 | 104 | #------------------------------------------------------------------------------- |
|
105 | 105 | here = os.path.abspath(os.path.dirname(__file__)) |
|
106 | 106 | |
|
107 | 107 | def require_clean_submodules(): |
|
108 | 108 | """Check on git submodules before distutils can do anything |
|
109 | 109 | |
|
110 | 110 | Since distutils cannot be trusted to update the tree |
|
111 | 111 | after everything has been set in motion, |
|
112 | 112 | this is not a distutils command. |
|
113 | 113 | """ |
|
114 | 114 | # PACKAGERS: Add a return here to skip checks for git submodules |
|
115 | 115 | |
|
116 | 116 | # don't do anything if nothing is actually supposed to happen |
|
117 | 117 | for do_nothing in ('-h', '--help', '--help-commands', 'clean', 'submodule'): |
|
118 | 118 | if do_nothing in sys.argv: |
|
119 | 119 | return |
|
120 | 120 | |
|
121 | 121 | status = check_submodule_status(here) |
|
122 | 122 | |
|
123 | 123 | if status == "missing": |
|
124 | 124 | print("checking out submodules for the first time") |
|
125 | 125 | update_submodules(here) |
|
126 | 126 | elif status == "unclean": |
|
127 | 127 | print('\n'.join([ |
|
128 | 128 | "Cannot build / install IPython with unclean submodules", |
|
129 | 129 | "Please update submodules with", |
|
130 | 130 | " python setup.py submodule", |
|
131 | 131 | "or", |
|
132 | 132 | " git submodule update", |
|
133 | 133 | "or commit any submodule changes you have made." |
|
134 | 134 | ])) |
|
135 | 135 | sys.exit(1) |
|
136 | 136 | |
|
137 | 137 | require_clean_submodules() |
|
138 | 138 | |
|
139 | 139 | #------------------------------------------------------------------------------- |
|
140 | 140 | # Things related to the IPython documentation |
|
141 | 141 | #------------------------------------------------------------------------------- |
|
142 | 142 | |
|
143 | 143 | # update the manuals when building a source dist |
|
144 | 144 | if len(sys.argv) >= 2 and sys.argv[1] in ('sdist','bdist_rpm'): |
|
145 | 145 | |
|
146 | 146 | # List of things to be updated. Each entry is a triplet of args for |
|
147 | 147 | # target_update() |
|
148 | 148 | to_update = [ |
|
149 | 149 | # FIXME - Disabled for now: we need to redo an automatic way |
|
150 | 150 | # of generating the magic info inside the rst. |
|
151 | 151 | #('docs/magic.tex', |
|
152 | 152 | #['IPython/Magic.py'], |
|
153 | 153 | #"cd doc && ./update_magic.sh" ), |
|
154 | 154 | |
|
155 | 155 | ('docs/man/ipcluster.1.gz', |
|
156 | 156 | ['docs/man/ipcluster.1'], |
|
157 | 157 | 'cd docs/man && gzip -9c ipcluster.1 > ipcluster.1.gz'), |
|
158 | 158 | |
|
159 | 159 | ('docs/man/ipcontroller.1.gz', |
|
160 | 160 | ['docs/man/ipcontroller.1'], |
|
161 | 161 | 'cd docs/man && gzip -9c ipcontroller.1 > ipcontroller.1.gz'), |
|
162 | 162 | |
|
163 | 163 | ('docs/man/ipengine.1.gz', |
|
164 | 164 | ['docs/man/ipengine.1'], |
|
165 | 165 | 'cd docs/man && gzip -9c ipengine.1 > ipengine.1.gz'), |
|
166 | 166 | |
|
167 | 167 | ('docs/man/ipython.1.gz', |
|
168 | 168 | ['docs/man/ipython.1'], |
|
169 | 169 | 'cd docs/man && gzip -9c ipython.1 > ipython.1.gz'), |
|
170 | 170 | |
|
171 | 171 | ] |
|
172 | 172 | |
|
173 | 173 | |
|
174 | 174 | [ target_update(*t) for t in to_update ] |
|
175 | 175 | |
|
176 | 176 | #--------------------------------------------------------------------------- |
|
177 | 177 | # Find all the packages, package data, and data_files |
|
178 | 178 | #--------------------------------------------------------------------------- |
|
179 | 179 | |
|
180 | 180 | packages = find_packages() |
|
181 | 181 | package_data = find_package_data() |
|
182 | 182 | |
|
183 | 183 | data_files = find_data_files() |
|
184 | 184 | |
|
185 | 185 | setup_args['packages'] = packages |
|
186 | 186 | setup_args['package_data'] = package_data |
|
187 | 187 | setup_args['data_files'] = data_files |
|
188 | 188 | |
|
189 | 189 | #--------------------------------------------------------------------------- |
|
190 | 190 | # custom distutils commands |
|
191 | 191 | #--------------------------------------------------------------------------- |
|
192 | 192 | # imports here, so they are after setuptools import if there was one |
|
193 | 193 | from distutils.command.sdist import sdist |
|
194 | 194 | from distutils.command.upload import upload |
|
195 | 195 | |
|
196 | 196 | class UploadWindowsInstallers(upload): |
|
197 | 197 | |
|
198 | 198 | description = "Upload Windows installers to PyPI (only used from tools/release_windows.py)" |
|
199 | 199 | user_options = upload.user_options + [ |
|
200 | 200 | ('files=', 'f', 'exe file (or glob) to upload') |
|
201 | 201 | ] |
|
202 | 202 | def initialize_options(self): |
|
203 | 203 | upload.initialize_options(self) |
|
204 | 204 | meta = self.distribution.metadata |
|
205 | 205 | base = '{name}-{version}'.format( |
|
206 | 206 | name=meta.get_name(), |
|
207 | 207 | version=meta.get_version() |
|
208 | 208 | ) |
|
209 | 209 | self.files = os.path.join('dist', '%s.*.exe' % base) |
|
210 | 210 | |
|
211 | 211 | def run(self): |
|
212 | 212 | for dist_file in glob(self.files): |
|
213 | 213 | self.upload_file('bdist_wininst', 'any', dist_file) |
|
214 | 214 | |
|
215 | 215 | setup_args['cmdclass'] = { |
|
216 | 216 | 'build_py': css_js_prerelease( |
|
217 | 217 | check_package_data_first(git_prebuild('IPython'))), |
|
218 | 218 | 'sdist' : css_js_prerelease(git_prebuild('IPython', sdist)), |
|
219 | 219 | 'upload_wininst' : UploadWindowsInstallers, |
|
220 | 220 | 'submodule' : UpdateSubmodules, |
|
221 | 221 | 'css' : CompileCSS, |
|
222 | 222 | 'symlink': install_symlinked, |
|
223 | 223 | 'install_lib_symlink': install_lib_symlink, |
|
224 | 224 | 'install_scripts_sym': install_scripts_for_symlink, |
|
225 | 225 | 'unsymlink': unsymlink, |
|
226 | 226 | 'jsversion' : JavascriptVersion, |
|
227 | 227 | } |
|
228 | 228 | |
|
229 | 229 | #--------------------------------------------------------------------------- |
|
230 | 230 | # Handle scripts, dependencies, and setuptools specific things |
|
231 | 231 | #--------------------------------------------------------------------------- |
|
232 | 232 | |
|
233 | 233 | # For some commands, use setuptools. Note that we do NOT list install here! |
|
234 | 234 | # If you want a setuptools-enhanced install, just run 'setupegg.py install' |
|
235 | 235 | needs_setuptools = set(('develop', 'release', 'bdist_egg', 'bdist_rpm', |
|
236 | 236 | 'bdist', 'bdist_dumb', 'bdist_wininst', 'bdist_wheel', |
|
237 | 237 | 'egg_info', 'easy_install', 'upload', 'install_egg_info', |
|
238 | 238 | )) |
|
239 | 239 | |
|
240 | 240 | if len(needs_setuptools.intersection(sys.argv)) > 0: |
|
241 | 241 | import setuptools |
|
242 | 242 | |
|
243 | 243 | # This dict is used for passing extra arguments that are setuptools |
|
244 | 244 | # specific to setup |
|
245 | 245 | setuptools_extra_args = {} |
|
246 | 246 | |
|
247 | 247 | # setuptools requirements |
|
248 | 248 | |
|
249 | 249 | pyzmq = 'pyzmq>=13' |
|
250 | 250 | |
|
251 | 251 | extras_require = dict( |
|
252 | 252 | parallel = [pyzmq], |
|
253 | 253 | qtconsole = [pyzmq, 'pygments'], |
|
254 | 254 | doc = ['Sphinx>=1.1', 'numpydoc'], |
|
255 | 255 | test = ['nose>=0.10.1', 'requests'], |
|
256 | 256 | terminal = [], |
|
257 | 257 | nbformat = ['jsonschema>=2.0'], |
|
258 | 258 | notebook = ['tornado>=4.0', pyzmq, 'jinja2', 'pygments', 'mistune>=0.5'], |
|
259 | 259 | nbconvert = ['pygments', 'jinja2', 'mistune>=0.3.1'] |
|
260 | 260 | ) |
|
261 | 261 | |
|
262 | 262 | if not sys.platform.startswith('win'): |
|
263 | 263 | extras_require['notebook'].append('terminado>=0.3.3') |
|
264 | 264 | |
|
265 | 265 | if sys.version_info < (3, 3): |
|
266 | 266 | extras_require['test'].append('mock') |
|
267 | 267 | |
|
268 | 268 | extras_require['notebook'].extend(extras_require['nbformat']) |
|
269 | 269 | extras_require['nbconvert'].extend(extras_require['nbformat']) |
|
270 | 270 | |
|
271 | 271 | install_requires = [ |
|
272 | 'decorator', | |
|
272 | 273 | 'path.py', # required by pickleshare, remove when pickleshare is added here |
|
273 | 274 | ] |
|
274 | 275 | |
|
275 | 276 | # add readline |
|
276 | 277 | if sys.platform == 'darwin': |
|
277 | 278 | if 'bdist_wheel' in sys.argv[1:] or not setupext.check_for_readline(): |
|
278 | 279 | install_requires.append('gnureadline') |
|
279 | 280 | elif sys.platform.startswith('win'): |
|
280 | 281 | extras_require['terminal'].append('pyreadline>=2.0') |
|
281 | 282 | |
|
282 | 283 | everything = set() |
|
283 | 284 | for deps in extras_require.values(): |
|
284 | 285 | everything.update(deps) |
|
285 | 286 | extras_require['all'] = everything |
|
286 | 287 | |
|
287 | 288 | if 'setuptools' in sys.modules: |
|
288 | 289 | # setup.py develop should check for submodules |
|
289 | 290 | from setuptools.command.develop import develop |
|
290 | 291 | setup_args['cmdclass']['develop'] = require_submodules(develop) |
|
291 | 292 | setup_args['cmdclass']['bdist_wheel'] = css_js_prerelease(get_bdist_wheel()) |
|
292 | 293 | |
|
293 | 294 | setuptools_extra_args['zip_safe'] = False |
|
294 | 295 | setuptools_extra_args['entry_points'] = { |
|
295 | 296 | 'console_scripts': find_entry_points(), |
|
296 | 297 | 'pygments.lexers': [ |
|
297 | 298 | 'ipythonconsole = IPython.lib.lexers:IPythonConsoleLexer', |
|
298 | 299 | 'ipython = IPython.lib.lexers:IPythonLexer', |
|
299 | 300 | 'ipython3 = IPython.lib.lexers:IPython3Lexer', |
|
300 | 301 | ], |
|
301 | 302 | } |
|
302 | 303 | setup_args['extras_require'] = extras_require |
|
303 | 304 | requires = setup_args['install_requires'] = install_requires |
|
304 | 305 | |
|
305 | 306 | # Script to be run by the windows binary installer after the default setup |
|
306 | 307 | # routine, to add shortcuts and similar windows-only things. Windows |
|
307 | 308 | # post-install scripts MUST reside in the scripts/ dir, otherwise distutils |
|
308 | 309 | # doesn't find them. |
|
309 | 310 | if 'bdist_wininst' in sys.argv: |
|
310 | 311 | if len(sys.argv) > 2 and \ |
|
311 | 312 | ('sdist' in sys.argv or 'bdist_rpm' in sys.argv): |
|
312 | 313 | print("ERROR: bdist_wininst must be run alone. Exiting.", file=sys.stderr) |
|
313 | 314 | sys.exit(1) |
|
314 | 315 | setup_args['data_files'].append( |
|
315 | 316 | ['Scripts', ('scripts/ipython.ico', 'scripts/ipython_nb.ico')]) |
|
316 | 317 | setup_args['scripts'] = [pjoin('scripts','ipython_win_post_install.py')] |
|
317 | 318 | setup_args['options'] = {"bdist_wininst": |
|
318 | 319 | {"install_script": |
|
319 | 320 | "ipython_win_post_install.py"}} |
|
320 | 321 | |
|
321 | 322 | else: |
|
322 | 323 | # If we are installing without setuptools, call this function which will |
|
323 | 324 | # check for dependencies an inform the user what is needed. This is |
|
324 | 325 | # just to make life easy for users. |
|
325 | 326 | for install_cmd in ('install', 'symlink'): |
|
326 | 327 | if install_cmd in sys.argv: |
|
327 | 328 | check_for_dependencies() |
|
328 | 329 | break |
|
329 | 330 | # scripts has to be a non-empty list, or install_scripts isn't called |
|
330 | 331 | setup_args['scripts'] = [e.split('=')[0].strip() for e in find_entry_points()] |
|
331 | 332 | |
|
332 | 333 | setup_args['cmdclass']['build_scripts'] = build_scripts_entrypt |
|
333 | 334 | |
|
334 | 335 | #--------------------------------------------------------------------------- |
|
335 | 336 | # Do the actual setup now |
|
336 | 337 | #--------------------------------------------------------------------------- |
|
337 | 338 | |
|
338 | 339 | setup_args.update(setuptools_extra_args) |
|
339 | 340 | |
|
340 | 341 | def main(): |
|
341 | 342 | setup(**setup_args) |
|
342 | 343 | |
|
343 | 344 | if __name__ == '__main__': |
|
344 | 345 | main() |
|
1 | NO CONTENT: file was removed |
|
1 | NO CONTENT: file was removed |
General Comments 0
You need to be logged in to leave comments.
Login now