Show More
@@ -1,1014 +1,1021 b'' | |||
|
1 | 1 | """Analysis of text input into executable blocks. |
|
2 | 2 | |
|
3 | 3 | The main class in this module, :class:`InputSplitter`, is designed to break |
|
4 | 4 | input from either interactive, line-by-line environments or block-based ones, |
|
5 | 5 | into standalone blocks that can be executed by Python as 'single' statements |
|
6 | 6 | (thus triggering sys.displayhook). |
|
7 | 7 | |
|
8 | 8 | A companion, :class:`IPythonInputSplitter`, provides the same functionality but |
|
9 | 9 | with full support for the extended IPython syntax (magics, system calls, etc). |
|
10 | 10 | |
|
11 | 11 | For more details, see the class docstring below. |
|
12 | 12 | |
|
13 | 13 | Syntax Transformations |
|
14 | 14 | ---------------------- |
|
15 | 15 | |
|
16 | 16 | One of the main jobs of the code in this file is to apply all syntax |
|
17 | 17 | transformations that make up 'the IPython language', i.e. magics, shell |
|
18 | 18 | escapes, etc. All transformations should be implemented as *fully stateless* |
|
19 | 19 | entities, that simply take one line as their input and return a line. |
|
20 | 20 | Internally for implementation purposes they may be a normal function or a |
|
21 | 21 | callable object, but the only input they receive will be a single line and they |
|
22 | 22 | should only return a line, without holding any data-dependent state between |
|
23 | 23 | calls. |
|
24 | 24 | |
|
25 | 25 | As an example, the EscapedTransformer is a class so we can more clearly group |
|
26 | 26 | together the functionality of dispatching to individual functions based on the |
|
27 | 27 | starting escape character, but the only method for public use is its call |
|
28 | 28 | method. |
|
29 | 29 | |
|
30 | 30 | |
|
31 | 31 | ToDo |
|
32 | 32 | ---- |
|
33 | 33 | |
|
34 | 34 | - Should we make push() actually raise an exception once push_accepts_more() |
|
35 | 35 | returns False? |
|
36 | 36 | |
|
37 | 37 | - Naming cleanups. The tr_* names aren't the most elegant, though now they are |
|
38 | 38 | at least just attributes of a class so not really very exposed. |
|
39 | 39 | |
|
40 | 40 | - Think about the best way to support dynamic things: automagic, autocall, |
|
41 | 41 | macros, etc. |
|
42 | 42 | |
|
43 | 43 | - Think of a better heuristic for the application of the transforms in |
|
44 | 44 | IPythonInputSplitter.push() than looking at the buffer ending in ':'. Idea: |
|
45 | 45 | track indentation change events (indent, dedent, nothing) and apply them only |
|
46 | 46 | if the indentation went up, but not otherwise. |
|
47 | 47 | |
|
48 | 48 | - Think of the cleanest way for supporting user-specified transformations (the |
|
49 | 49 | user prefilters we had before). |
|
50 | 50 | |
|
51 | 51 | Authors |
|
52 | 52 | ------- |
|
53 | 53 | |
|
54 | 54 | * Fernando Perez |
|
55 | 55 | * Brian Granger |
|
56 | 56 | """ |
|
57 | 57 | #----------------------------------------------------------------------------- |
|
58 | 58 | # Copyright (C) 2010 The IPython Development Team |
|
59 | 59 | # |
|
60 | 60 | # Distributed under the terms of the BSD License. The full license is in |
|
61 | 61 | # the file COPYING, distributed as part of this software. |
|
62 | 62 | #----------------------------------------------------------------------------- |
|
63 | 63 | from __future__ import print_function |
|
64 | 64 | |
|
65 | 65 | #----------------------------------------------------------------------------- |
|
66 | 66 | # Imports |
|
67 | 67 | #----------------------------------------------------------------------------- |
|
68 | 68 | # stdlib |
|
69 | 69 | import codeop |
|
70 | 70 | import re |
|
71 | 71 | import sys |
|
72 | 72 | |
|
73 | 73 | # IPython modules |
|
74 | 74 | from IPython.utils.text import make_quoted_expr |
|
75 | 75 | |
|
76 | 76 | #----------------------------------------------------------------------------- |
|
77 | 77 | # Globals |
|
78 | 78 | #----------------------------------------------------------------------------- |
|
79 | 79 | |
|
80 | 80 | # The escape sequences that define the syntax transformations IPython will |
|
81 | 81 | # apply to user input. These can NOT be just changed here: many regular |
|
82 | 82 | # expressions and other parts of the code may use their hardcoded values, and |
|
83 | 83 | # for all intents and purposes they constitute the 'IPython syntax', so they |
|
84 | 84 | # should be considered fixed. |
|
85 | 85 | |
|
86 | 86 | ESC_SHELL = '!' # Send line to underlying system shell |
|
87 | 87 | ESC_SH_CAP = '!!' # Send line to system shell and capture output |
|
88 | 88 | ESC_HELP = '?' # Find information about object |
|
89 | 89 | ESC_HELP2 = '??' # Find extra-detailed information about object |
|
90 | 90 | ESC_MAGIC = '%' # Call magic function |
|
91 | 91 | ESC_QUOTE = ',' # Split args on whitespace, quote each as string and call |
|
92 | 92 | ESC_QUOTE2 = ';' # Quote all args as a single string, call |
|
93 | 93 | ESC_PAREN = '/' # Call first argument with rest of line as arguments |
|
94 | 94 | |
|
95 | 95 | #----------------------------------------------------------------------------- |
|
96 | 96 | # Utilities |
|
97 | 97 | #----------------------------------------------------------------------------- |
|
98 | 98 | |
|
99 | 99 | # FIXME: These are general-purpose utilities that later can be moved to the |
|
100 | 100 | # general ward. Kept here for now because we're being very strict about test |
|
101 | 101 | # coverage with this code, and this lets us ensure that we keep 100% coverage |
|
102 | 102 | # while developing. |
|
103 | 103 | |
|
104 | 104 | # compiled regexps for autoindent management |
|
105 | 105 | dedent_re = re.compile(r'^\s+raise|^\s+return|^\s+pass') |
|
106 | 106 | ini_spaces_re = re.compile(r'^([ \t\r\f\v]+)') |
|
107 | 107 | |
|
108 | 108 | # regexp to match pure comment lines so we don't accidentally insert 'if 1:' |
|
109 | 109 | # before pure comments |
|
110 | 110 | comment_line_re = re.compile('^\s*\#') |
|
111 | 111 | |
|
112 | 112 | |
|
113 | 113 | def num_ini_spaces(s): |
|
114 | 114 | """Return the number of initial spaces in a string. |
|
115 | 115 | |
|
116 | 116 | Note that tabs are counted as a single space. For now, we do *not* support |
|
117 | 117 | mixing of tabs and spaces in the user's input. |
|
118 | 118 | |
|
119 | 119 | Parameters |
|
120 | 120 | ---------- |
|
121 | 121 | s : string |
|
122 | 122 | |
|
123 | 123 | Returns |
|
124 | 124 | ------- |
|
125 | 125 | n : int |
|
126 | 126 | """ |
|
127 | 127 | |
|
128 | 128 | ini_spaces = ini_spaces_re.match(s) |
|
129 | 129 | if ini_spaces: |
|
130 | 130 | return ini_spaces.end() |
|
131 | 131 | else: |
|
132 | 132 | return 0 |
|
133 | 133 | |
|
134 | 134 | |
|
135 | 135 | def remove_comments(src): |
|
136 | 136 | """Remove all comments from input source. |
|
137 | 137 | |
|
138 | 138 | Note: comments are NOT recognized inside of strings! |
|
139 | 139 | |
|
140 | 140 | Parameters |
|
141 | 141 | ---------- |
|
142 | 142 | src : string |
|
143 | 143 | A single or multiline input string. |
|
144 | 144 | |
|
145 | 145 | Returns |
|
146 | 146 | ------- |
|
147 | 147 | String with all Python comments removed. |
|
148 | 148 | """ |
|
149 | 149 | |
|
150 | 150 | return re.sub('#.*', '', src) |
|
151 | 151 | |
|
152 | 152 | |
|
153 | 153 | def get_input_encoding(): |
|
154 | 154 | """Return the default standard input encoding. |
|
155 | 155 | |
|
156 | 156 | If sys.stdin has no encoding, 'ascii' is returned.""" |
|
157 | 157 | # There are strange environments for which sys.stdin.encoding is None. We |
|
158 | 158 | # ensure that a valid encoding is returned. |
|
159 | 159 | encoding = getattr(sys.stdin, 'encoding', None) |
|
160 | 160 | if encoding is None: |
|
161 | 161 | encoding = 'ascii' |
|
162 | 162 | return encoding |
|
163 | 163 | |
|
164 | 164 | #----------------------------------------------------------------------------- |
|
165 | 165 | # Classes and functions for normal Python syntax handling |
|
166 | 166 | #----------------------------------------------------------------------------- |
|
167 | 167 | |
|
168 | 168 | # HACK! This implementation, written by Robert K a while ago using the |
|
169 | 169 | # compiler module, is more robust than the other one below, but it expects its |
|
170 | 170 | # input to be pure python (no ipython syntax). For now we're using it as a |
|
171 | 171 | # second-pass splitter after the first pass transforms the input to pure |
|
172 | 172 | # python. |
|
173 | 173 | |
|
174 | 174 | def split_blocks(python): |
|
175 | 175 | """ Split multiple lines of code into discrete commands that can be |
|
176 | 176 | executed singly. |
|
177 | 177 | |
|
178 | 178 | Parameters |
|
179 | 179 | ---------- |
|
180 | 180 | python : str |
|
181 | 181 | Pure, exec'able Python code. |
|
182 | 182 | |
|
183 | 183 | Returns |
|
184 | 184 | ------- |
|
185 | 185 | commands : list of str |
|
186 | 186 | Separate commands that can be exec'ed independently. |
|
187 | 187 | """ |
|
188 | 188 | |
|
189 | 189 | import compiler |
|
190 | 190 | |
|
191 | 191 | # compiler.parse treats trailing spaces after a newline as a |
|
192 | 192 | # SyntaxError. This is different than codeop.CommandCompiler, which |
|
193 | 193 | # will compile the trailng spaces just fine. We simply strip any |
|
194 | 194 | # trailing whitespace off. Passing a string with trailing whitespace |
|
195 | 195 | # to exec will fail however. There seems to be some inconsistency in |
|
196 | 196 | # how trailing whitespace is handled, but this seems to work. |
|
197 | 197 | python_ori = python # save original in case we bail on error |
|
198 | 198 | python = python.strip() |
|
199 | 199 | |
|
200 | 200 | # The compiler module does not like unicode. We need to convert |
|
201 | 201 | # it encode it: |
|
202 | 202 | if isinstance(python, unicode): |
|
203 | 203 | # Use the utf-8-sig BOM so the compiler detects this a UTF-8 |
|
204 | 204 | # encode string. |
|
205 | 205 | python = '\xef\xbb\xbf' + python.encode('utf-8') |
|
206 | 206 | |
|
207 | 207 | # The compiler module will parse the code into an abstract syntax tree. |
|
208 | 208 | # This has a bug with str("a\nb"), but not str("""a\nb""")!!! |
|
209 | 209 | try: |
|
210 | 210 | ast = compiler.parse(python) |
|
211 | 211 | except: |
|
212 | 212 | return [python_ori] |
|
213 | 213 | |
|
214 | 214 | # Uncomment to help debug the ast tree |
|
215 | 215 | # for n in ast.node: |
|
216 | 216 | # print n.lineno,'->',n |
|
217 | 217 | |
|
218 | 218 | # Each separate command is available by iterating over ast.node. The |
|
219 | 219 | # lineno attribute is the line number (1-indexed) beginning the commands |
|
220 | 220 | # suite. |
|
221 | 221 | # lines ending with ";" yield a Discard Node that doesn't have a lineno |
|
222 | 222 | # attribute. These nodes can and should be discarded. But there are |
|
223 | 223 | # other situations that cause Discard nodes that shouldn't be discarded. |
|
224 | 224 | # We might eventually discover other cases where lineno is None and have |
|
225 | 225 | # to put in a more sophisticated test. |
|
226 | 226 | linenos = [x.lineno-1 for x in ast.node if x.lineno is not None] |
|
227 | 227 | |
|
228 | # When we have a bare string as the first statement, it does not end up as | |
|
229 | # a Discard Node in the AST as we might expect. Instead, it gets interpreted | |
|
230 | # as the docstring of the module. Check for this case and prepend 0 (the | |
|
231 | # first line number) to the list of linenos to account for it. | |
|
232 | if ast.doc is not None: | |
|
233 | linenos.insert(0, 0) | |
|
234 | ||
|
228 | 235 | # When we finally get the slices, we will need to slice all the way to |
|
229 | 236 | # the end even though we don't have a line number for it. Fortunately, |
|
230 | 237 | # None does the job nicely. |
|
231 | 238 | linenos.append(None) |
|
232 | 239 | |
|
233 | 240 | # Same problem at the other end: sometimes the ast tree has its |
|
234 | 241 | # first complete statement not starting on line 0. In this case |
|
235 | 242 | # we might miss part of it. This fixes ticket 266993. Thanks Gael! |
|
236 | 243 | linenos[0] = 0 |
|
237 | 244 | |
|
238 | 245 | lines = python.splitlines() |
|
239 | 246 | |
|
240 | 247 | # Create a list of atomic commands. |
|
241 | 248 | cmds = [] |
|
242 | 249 | for i, j in zip(linenos[:-1], linenos[1:]): |
|
243 | 250 | cmd = lines[i:j] |
|
244 | 251 | if cmd: |
|
245 | 252 | cmds.append('\n'.join(cmd)+'\n') |
|
246 | 253 | |
|
247 | 254 | return cmds |
|
248 | 255 | |
|
249 | 256 | |
|
250 | 257 | class InputSplitter(object): |
|
251 | 258 | """An object that can split Python source input in executable blocks. |
|
252 | 259 | |
|
253 | 260 | This object is designed to be used in one of two basic modes: |
|
254 | 261 | |
|
255 | 262 | 1. By feeding it python source line-by-line, using :meth:`push`. In this |
|
256 | 263 | mode, it will return on each push whether the currently pushed code |
|
257 | 264 | could be executed already. In addition, it provides a method called |
|
258 | 265 | :meth:`push_accepts_more` that can be used to query whether more input |
|
259 | 266 | can be pushed into a single interactive block. |
|
260 | 267 | |
|
261 | 268 | 2. By calling :meth:`split_blocks` with a single, multiline Python string, |
|
262 | 269 | that is then split into blocks each of which can be executed |
|
263 | 270 | interactively as a single statement. |
|
264 | 271 | |
|
265 | 272 | This is a simple example of how an interactive terminal-based client can use |
|
266 | 273 | this tool:: |
|
267 | 274 | |
|
268 | 275 | isp = InputSplitter() |
|
269 | 276 | while isp.push_accepts_more(): |
|
270 | 277 | indent = ' '*isp.indent_spaces |
|
271 | 278 | prompt = '>>> ' + indent |
|
272 | 279 | line = indent + raw_input(prompt) |
|
273 | 280 | isp.push(line) |
|
274 | 281 | print 'Input source was:\n', isp.source_reset(), |
|
275 | 282 | """ |
|
276 | 283 | # Number of spaces of indentation computed from input that has been pushed |
|
277 | 284 | # so far. This is the attributes callers should query to get the current |
|
278 | 285 | # indentation level, in order to provide auto-indent facilities. |
|
279 | 286 | indent_spaces = 0 |
|
280 | 287 | # String, indicating the default input encoding. It is computed by default |
|
281 | 288 | # at initialization time via get_input_encoding(), but it can be reset by a |
|
282 | 289 | # client with specific knowledge of the encoding. |
|
283 | 290 | encoding = '' |
|
284 | 291 | # String where the current full source input is stored, properly encoded. |
|
285 | 292 | # Reading this attribute is the normal way of querying the currently pushed |
|
286 | 293 | # source code, that has been properly encoded. |
|
287 | 294 | source = '' |
|
288 | 295 | # Code object corresponding to the current source. It is automatically |
|
289 | 296 | # synced to the source, so it can be queried at any time to obtain the code |
|
290 | 297 | # object; it will be None if the source doesn't compile to valid Python. |
|
291 | 298 | code = None |
|
292 | 299 | # Input mode |
|
293 | 300 | input_mode = 'line' |
|
294 | 301 | |
|
295 | 302 | # Private attributes |
|
296 | 303 | |
|
297 | 304 | # List with lines of input accumulated so far |
|
298 | 305 | _buffer = None |
|
299 | 306 | # Command compiler |
|
300 | 307 | _compile = None |
|
301 | 308 | # Mark when input has changed indentation all the way back to flush-left |
|
302 | 309 | _full_dedent = False |
|
303 | 310 | # Boolean indicating whether the current block is complete |
|
304 | 311 | _is_complete = None |
|
305 | 312 | |
|
306 | 313 | def __init__(self, input_mode=None): |
|
307 | 314 | """Create a new InputSplitter instance. |
|
308 | 315 | |
|
309 | 316 | Parameters |
|
310 | 317 | ---------- |
|
311 | 318 | input_mode : str |
|
312 | 319 | |
|
313 | 320 | One of ['line', 'cell']; default is 'line'. |
|
314 | 321 | |
|
315 | 322 | The input_mode parameter controls how new inputs are used when fed via |
|
316 | 323 | the :meth:`push` method: |
|
317 | 324 | |
|
318 | 325 | - 'line': meant for line-oriented clients, inputs are appended one at a |
|
319 | 326 | time to the internal buffer and the whole buffer is compiled. |
|
320 | 327 | |
|
321 | 328 | - 'cell': meant for clients that can edit multi-line 'cells' of text at |
|
322 | 329 | a time. A cell can contain one or more blocks that can be compile in |
|
323 | 330 | 'single' mode by Python. In this mode, each new input new input |
|
324 | 331 | completely replaces all prior inputs. Cell mode is thus equivalent |
|
325 | 332 | to prepending a full reset() to every push() call. |
|
326 | 333 | """ |
|
327 | 334 | self._buffer = [] |
|
328 | 335 | self._compile = codeop.CommandCompiler() |
|
329 | 336 | self.encoding = get_input_encoding() |
|
330 | 337 | self.input_mode = InputSplitter.input_mode if input_mode is None \ |
|
331 | 338 | else input_mode |
|
332 | 339 | |
|
333 | 340 | def reset(self): |
|
334 | 341 | """Reset the input buffer and associated state.""" |
|
335 | 342 | self.indent_spaces = 0 |
|
336 | 343 | self._buffer[:] = [] |
|
337 | 344 | self.source = '' |
|
338 | 345 | self.code = None |
|
339 | 346 | self._is_complete = False |
|
340 | 347 | self._full_dedent = False |
|
341 | 348 | |
|
342 | 349 | def source_reset(self): |
|
343 | 350 | """Return the input source and perform a full reset. |
|
344 | 351 | """ |
|
345 | 352 | out = self.source |
|
346 | 353 | self.reset() |
|
347 | 354 | return out |
|
348 | 355 | |
|
349 | 356 | def push(self, lines): |
|
350 |
"""Push one or |
|
|
357 | """Push one or more lines of input. | |
|
351 | 358 | |
|
352 | 359 | This stores the given lines and returns a status code indicating |
|
353 | 360 | whether the code forms a complete Python block or not. |
|
354 | 361 | |
|
355 | 362 | Any exceptions generated in compilation are swallowed, but if an |
|
356 | 363 | exception was produced, the method returns True. |
|
357 | 364 | |
|
358 | 365 | Parameters |
|
359 | 366 | ---------- |
|
360 | 367 | lines : string |
|
361 | 368 | One or more lines of Python input. |
|
362 | 369 | |
|
363 | 370 | Returns |
|
364 | 371 | ------- |
|
365 | 372 | is_complete : boolean |
|
366 | 373 | True if the current input source (the result of the current input |
|
367 | 374 | plus prior inputs) forms a complete Python execution block. Note that |
|
368 | 375 | this value is also stored as a private attribute (_is_complete), so it |
|
369 | 376 | can be queried at any time. |
|
370 | 377 | """ |
|
371 | 378 | if self.input_mode == 'cell': |
|
372 | 379 | self.reset() |
|
373 | 380 | |
|
374 | 381 | self._store(lines) |
|
375 | 382 | source = self.source |
|
376 | 383 | |
|
377 | 384 | # Before calling _compile(), reset the code object to None so that if an |
|
378 | 385 | # exception is raised in compilation, we don't mislead by having |
|
379 | 386 | # inconsistent code/source attributes. |
|
380 | 387 | self.code, self._is_complete = None, None |
|
381 | 388 | |
|
382 | 389 | # Honor termination lines properly |
|
383 | 390 | if source.rstrip().endswith('\\'): |
|
384 | 391 | return False |
|
385 | 392 | |
|
386 | 393 | self._update_indent(lines) |
|
387 | 394 | try: |
|
388 | 395 | self.code = self._compile(source) |
|
389 | 396 | # Invalid syntax can produce any of a number of different errors from |
|
390 | 397 | # inside the compiler, so we have to catch them all. Syntax errors |
|
391 | 398 | # immediately produce a 'ready' block, so the invalid Python can be |
|
392 | 399 | # sent to the kernel for evaluation with possible ipython |
|
393 | 400 | # special-syntax conversion. |
|
394 | 401 | except (SyntaxError, OverflowError, ValueError, TypeError, |
|
395 | 402 | MemoryError): |
|
396 | 403 | self._is_complete = True |
|
397 | 404 | else: |
|
398 | 405 | # Compilation didn't produce any exceptions (though it may not have |
|
399 | 406 | # given a complete code object) |
|
400 | 407 | self._is_complete = self.code is not None |
|
401 | 408 | |
|
402 | 409 | return self._is_complete |
|
403 | 410 | |
|
404 | 411 | def push_accepts_more(self): |
|
405 | 412 | """Return whether a block of interactive input can accept more input. |
|
406 | 413 | |
|
407 | 414 | This method is meant to be used by line-oriented frontends, who need to |
|
408 | 415 | guess whether a block is complete or not based solely on prior and |
|
409 | 416 | current input lines. The InputSplitter considers it has a complete |
|
410 | 417 | interactive block and will not accept more input only when either a |
|
411 | 418 | SyntaxError is raised, or *all* of the following are true: |
|
412 | 419 | |
|
413 | 420 | 1. The input compiles to a complete statement. |
|
414 | 421 | |
|
415 | 422 | 2. The indentation level is flush-left (because if we are indented, |
|
416 | 423 | like inside a function definition or for loop, we need to keep |
|
417 | 424 | reading new input). |
|
418 | 425 | |
|
419 | 426 | 3. There is one extra line consisting only of whitespace. |
|
420 | 427 | |
|
421 | 428 | Because of condition #3, this method should be used only by |
|
422 | 429 | *line-oriented* frontends, since it means that intermediate blank lines |
|
423 | 430 | are not allowed in function definitions (or any other indented block). |
|
424 | 431 | |
|
425 | 432 | Block-oriented frontends that have a separate keyboard event to |
|
426 | 433 | indicate execution should use the :meth:`split_blocks` method instead. |
|
427 | 434 | |
|
428 | 435 | If the current input produces a syntax error, this method immediately |
|
429 | 436 | returns False but does *not* raise the syntax error exception, as |
|
430 | 437 | typically clients will want to send invalid syntax to an execution |
|
431 | 438 | backend which might convert the invalid syntax into valid Python via |
|
432 | 439 | one of the dynamic IPython mechanisms. |
|
433 | 440 | """ |
|
434 | 441 | |
|
435 | 442 | # With incomplete input, unconditionally accept more |
|
436 | 443 | if not self._is_complete: |
|
437 | 444 | return True |
|
438 | 445 | |
|
439 | 446 | # If we already have complete input and we're flush left, the answer |
|
440 | 447 | # depends. In line mode, we're done. But in cell mode, we need to |
|
441 | 448 | # check how many blocks the input so far compiles into, because if |
|
442 | 449 | # there's already more than one full independent block of input, then |
|
443 | 450 | # the client has entered full 'cell' mode and is feeding lines that |
|
444 | 451 | # each is complete. In this case we should then keep accepting. |
|
445 | 452 | # The Qt terminal-like console does precisely this, to provide the |
|
446 | 453 | # convenience of terminal-like input of single expressions, but |
|
447 | 454 | # allowing the user (with a separate keystroke) to switch to 'cell' |
|
448 | 455 | # mode and type multiple expressions in one shot. |
|
449 | 456 | if self.indent_spaces==0: |
|
450 | 457 | if self.input_mode=='line': |
|
451 | 458 | return False |
|
452 | 459 | else: |
|
453 | 460 | nblocks = len(split_blocks(''.join(self._buffer))) |
|
454 | 461 | if nblocks==1: |
|
455 | 462 | return False |
|
456 | 463 | |
|
457 | 464 | # When input is complete, then termination is marked by an extra blank |
|
458 | 465 | # line at the end. |
|
459 | 466 | last_line = self.source.splitlines()[-1] |
|
460 | 467 | return bool(last_line and not last_line.isspace()) |
|
461 | 468 | |
|
462 | 469 | def split_blocks(self, lines): |
|
463 | 470 | """Split a multiline string into multiple input blocks. |
|
464 | 471 | |
|
465 | 472 | Note: this method starts by performing a full reset(). |
|
466 | 473 | |
|
467 | 474 | Parameters |
|
468 | 475 | ---------- |
|
469 | 476 | lines : str |
|
470 | 477 | A possibly multiline string. |
|
471 | 478 | |
|
472 | 479 | Returns |
|
473 | 480 | ------- |
|
474 | 481 | blocks : list |
|
475 | 482 | A list of strings, each possibly multiline. Each string corresponds |
|
476 | 483 | to a single block that can be compiled in 'single' mode (unless it |
|
477 | 484 | has a syntax error).""" |
|
478 | 485 | |
|
479 | 486 | # This code is fairly delicate. If you make any changes here, make |
|
480 | 487 | # absolutely sure that you do run the full test suite and ALL tests |
|
481 | 488 | # pass. |
|
482 | 489 | |
|
483 | 490 | self.reset() |
|
484 | 491 | blocks = [] |
|
485 | 492 | |
|
486 | 493 | # Reversed copy so we can use pop() efficiently and consume the input |
|
487 | 494 | # as a stack |
|
488 | 495 | lines = lines.splitlines()[::-1] |
|
489 | 496 | # Outer loop over all input |
|
490 | 497 | while lines: |
|
491 | 498 | #print 'Current lines:', lines # dbg |
|
492 | 499 | # Inner loop to build each block |
|
493 | 500 | while True: |
|
494 | 501 | # Safety exit from inner loop |
|
495 | 502 | if not lines: |
|
496 | 503 | break |
|
497 | 504 | # Grab next line but don't push it yet |
|
498 | 505 | next_line = lines.pop() |
|
499 | 506 | # Blank/empty lines are pushed as-is |
|
500 | 507 | if not next_line or next_line.isspace(): |
|
501 | 508 | self.push(next_line) |
|
502 | 509 | continue |
|
503 | 510 | |
|
504 | 511 | # Check indentation changes caused by the *next* line |
|
505 | 512 | indent_spaces, _full_dedent = self._find_indent(next_line) |
|
506 | 513 | |
|
507 | 514 | # If the next line causes a dedent, it can be for two differnt |
|
508 | 515 | # reasons: either an explicit de-dent by the user or a |
|
509 | 516 | # return/raise/pass statement. These MUST be handled |
|
510 | 517 | # separately: |
|
511 | 518 | # |
|
512 | 519 | # 1. the first case is only detected when the actual explicit |
|
513 | 520 | # dedent happens, and that would be the *first* line of a *new* |
|
514 | 521 | # block. Thus, we must put the line back into the input buffer |
|
515 | 522 | # so that it starts a new block on the next pass. |
|
516 | 523 | # |
|
517 | 524 | # 2. the second case is detected in the line before the actual |
|
518 | 525 | # dedent happens, so , we consume the line and we can break out |
|
519 | 526 | # to start a new block. |
|
520 | 527 | |
|
521 | 528 | # Case 1, explicit dedent causes a break. |
|
522 | 529 | # Note: check that we weren't on the very last line, else we'll |
|
523 | 530 | # enter an infinite loop adding/removing the last line. |
|
524 | 531 | if _full_dedent and lines and not next_line.startswith(' '): |
|
525 | 532 | lines.append(next_line) |
|
526 | 533 | break |
|
527 | 534 | |
|
528 | 535 | # Otherwise any line is pushed |
|
529 | 536 | self.push(next_line) |
|
530 | 537 | |
|
531 | 538 | # Case 2, full dedent with full block ready: |
|
532 | 539 | if _full_dedent or \ |
|
533 | 540 | self.indent_spaces==0 and not self.push_accepts_more(): |
|
534 | 541 | break |
|
535 | 542 | # Form the new block with the current source input |
|
536 | 543 | blocks.append(self.source_reset()) |
|
537 | 544 | |
|
538 | 545 | #return blocks |
|
539 | 546 | # HACK!!! Now that our input is in blocks but guaranteed to be pure |
|
540 | 547 | # python syntax, feed it back a second time through the AST-based |
|
541 | 548 | # splitter, which is more accurate than ours. |
|
542 | 549 | return split_blocks(''.join(blocks)) |
|
543 | 550 | |
|
544 | 551 | #------------------------------------------------------------------------ |
|
545 | 552 | # Private interface |
|
546 | 553 | #------------------------------------------------------------------------ |
|
547 | 554 | |
|
548 | 555 | def _find_indent(self, line): |
|
549 | 556 | """Compute the new indentation level for a single line. |
|
550 | 557 | |
|
551 | 558 | Parameters |
|
552 | 559 | ---------- |
|
553 | 560 | line : str |
|
554 | 561 | A single new line of non-whitespace, non-comment Python input. |
|
555 | 562 | |
|
556 | 563 | Returns |
|
557 | 564 | ------- |
|
558 | 565 | indent_spaces : int |
|
559 | 566 | New value for the indent level (it may be equal to self.indent_spaces |
|
560 | 567 | if indentation doesn't change. |
|
561 | 568 | |
|
562 | 569 | full_dedent : boolean |
|
563 | 570 | Whether the new line causes a full flush-left dedent. |
|
564 | 571 | """ |
|
565 | 572 | indent_spaces = self.indent_spaces |
|
566 | 573 | full_dedent = self._full_dedent |
|
567 | 574 | |
|
568 | 575 | inisp = num_ini_spaces(line) |
|
569 | 576 | if inisp < indent_spaces: |
|
570 | 577 | indent_spaces = inisp |
|
571 | 578 | if indent_spaces <= 0: |
|
572 | 579 | #print 'Full dedent in text',self.source # dbg |
|
573 | 580 | full_dedent = True |
|
574 | 581 | |
|
575 | 582 | if line[-1] == ':': |
|
576 | 583 | indent_spaces += 4 |
|
577 | 584 | elif dedent_re.match(line): |
|
578 | 585 | indent_spaces -= 4 |
|
579 | 586 | if indent_spaces <= 0: |
|
580 | 587 | full_dedent = True |
|
581 | 588 | |
|
582 | 589 | # Safety |
|
583 | 590 | if indent_spaces < 0: |
|
584 | 591 | indent_spaces = 0 |
|
585 | 592 | #print 'safety' # dbg |
|
586 | 593 | |
|
587 | 594 | return indent_spaces, full_dedent |
|
588 | 595 | |
|
589 | 596 | def _update_indent(self, lines): |
|
590 | 597 | for line in remove_comments(lines).splitlines(): |
|
591 | 598 | if line and not line.isspace(): |
|
592 | 599 | self.indent_spaces, self._full_dedent = self._find_indent(line) |
|
593 | 600 | |
|
594 | 601 | def _store(self, lines, buffer=None, store='source'): |
|
595 | 602 | """Store one or more lines of input. |
|
596 | 603 | |
|
597 | 604 | If input lines are not newline-terminated, a newline is automatically |
|
598 | 605 | appended.""" |
|
599 | 606 | |
|
600 | 607 | if buffer is None: |
|
601 | 608 | buffer = self._buffer |
|
602 | 609 | |
|
603 | 610 | if lines.endswith('\n'): |
|
604 | 611 | buffer.append(lines) |
|
605 | 612 | else: |
|
606 | 613 | buffer.append(lines+'\n') |
|
607 | 614 | setattr(self, store, self._set_source(buffer)) |
|
608 | 615 | |
|
609 | 616 | def _set_source(self, buffer): |
|
610 | 617 | return ''.join(buffer).encode(self.encoding) |
|
611 | 618 | |
|
612 | 619 | |
|
613 | 620 | #----------------------------------------------------------------------------- |
|
614 | 621 | # Functions and classes for IPython-specific syntactic support |
|
615 | 622 | #----------------------------------------------------------------------------- |
|
616 | 623 | |
|
617 | 624 | # RegExp for splitting line contents into pre-char//first word-method//rest. |
|
618 | 625 | # For clarity, each group in on one line. |
|
619 | 626 | |
|
620 | 627 | line_split = re.compile(""" |
|
621 | 628 | ^(\s*) # any leading space |
|
622 | 629 | ([,;/%]|!!?|\?\??) # escape character or characters |
|
623 | 630 | \s*(%?[\w\.\*]*) # function/method, possibly with leading % |
|
624 | 631 | # to correctly treat things like '?%magic' |
|
625 | 632 | (\s+.*$|$) # rest of line |
|
626 | 633 | """, re.VERBOSE) |
|
627 | 634 | |
|
628 | 635 | |
|
629 | 636 | def split_user_input(line): |
|
630 | 637 | """Split user input into early whitespace, esc-char, function part and rest. |
|
631 | 638 | |
|
632 | 639 | This is currently handles lines with '=' in them in a very inconsistent |
|
633 | 640 | manner. |
|
634 | 641 | |
|
635 | 642 | Examples |
|
636 | 643 | ======== |
|
637 | 644 | >>> split_user_input('x=1') |
|
638 | 645 | ('', '', 'x=1', '') |
|
639 | 646 | >>> split_user_input('?') |
|
640 | 647 | ('', '?', '', '') |
|
641 | 648 | >>> split_user_input('??') |
|
642 | 649 | ('', '??', '', '') |
|
643 | 650 | >>> split_user_input(' ?') |
|
644 | 651 | (' ', '?', '', '') |
|
645 | 652 | >>> split_user_input(' ??') |
|
646 | 653 | (' ', '??', '', '') |
|
647 | 654 | >>> split_user_input('??x') |
|
648 | 655 | ('', '??', 'x', '') |
|
649 | 656 | >>> split_user_input('?x=1') |
|
650 | 657 | ('', '', '?x=1', '') |
|
651 | 658 | >>> split_user_input('!ls') |
|
652 | 659 | ('', '!', 'ls', '') |
|
653 | 660 | >>> split_user_input(' !ls') |
|
654 | 661 | (' ', '!', 'ls', '') |
|
655 | 662 | >>> split_user_input('!!ls') |
|
656 | 663 | ('', '!!', 'ls', '') |
|
657 | 664 | >>> split_user_input(' !!ls') |
|
658 | 665 | (' ', '!!', 'ls', '') |
|
659 | 666 | >>> split_user_input(',ls') |
|
660 | 667 | ('', ',', 'ls', '') |
|
661 | 668 | >>> split_user_input(';ls') |
|
662 | 669 | ('', ';', 'ls', '') |
|
663 | 670 | >>> split_user_input(' ;ls') |
|
664 | 671 | (' ', ';', 'ls', '') |
|
665 | 672 | >>> split_user_input('f.g(x)') |
|
666 | 673 | ('', '', 'f.g(x)', '') |
|
667 | 674 | >>> split_user_input('f.g (x)') |
|
668 | 675 | ('', '', 'f.g', '(x)') |
|
669 | 676 | >>> split_user_input('?%hist') |
|
670 | 677 | ('', '?', '%hist', '') |
|
671 | 678 | >>> split_user_input('?x*') |
|
672 | 679 | ('', '?', 'x*', '') |
|
673 | 680 | """ |
|
674 | 681 | match = line_split.match(line) |
|
675 | 682 | if match: |
|
676 | 683 | lspace, esc, fpart, rest = match.groups() |
|
677 | 684 | else: |
|
678 | 685 | # print "match failed for line '%s'" % line |
|
679 | 686 | try: |
|
680 | 687 | fpart, rest = line.split(None, 1) |
|
681 | 688 | except ValueError: |
|
682 | 689 | # print "split failed for line '%s'" % line |
|
683 | 690 | fpart, rest = line,'' |
|
684 | 691 | lspace = re.match('^(\s*)(.*)', line).groups()[0] |
|
685 | 692 | esc = '' |
|
686 | 693 | |
|
687 | 694 | # fpart has to be a valid python identifier, so it better be only pure |
|
688 | 695 | # ascii, no unicode: |
|
689 | 696 | try: |
|
690 | 697 | fpart = fpart.encode('ascii') |
|
691 | 698 | except UnicodeEncodeError: |
|
692 | 699 | lspace = unicode(lspace) |
|
693 | 700 | rest = fpart + u' ' + rest |
|
694 | 701 | fpart = u'' |
|
695 | 702 | |
|
696 | 703 | #print 'line:<%s>' % line # dbg |
|
697 | 704 | #print 'esc <%s> fpart <%s> rest <%s>' % (esc,fpart.strip(),rest) # dbg |
|
698 | 705 | return lspace, esc, fpart.strip(), rest.lstrip() |
|
699 | 706 | |
|
700 | 707 | |
|
701 | 708 | # The escaped translators ALL receive a line where their own escape has been |
|
702 | 709 | # stripped. Only '?' is valid at the end of the line, all others can only be |
|
703 | 710 | # placed at the start. |
|
704 | 711 | |
|
705 | 712 | class LineInfo(object): |
|
706 | 713 | """A single line of input and associated info. |
|
707 | 714 | |
|
708 | 715 | This is a utility class that mostly wraps the output of |
|
709 | 716 | :func:`split_user_input` into a convenient object to be passed around |
|
710 | 717 | during input transformations. |
|
711 | 718 | |
|
712 | 719 | Includes the following as properties: |
|
713 | 720 | |
|
714 | 721 | line |
|
715 | 722 | The original, raw line |
|
716 | 723 | |
|
717 | 724 | lspace |
|
718 | 725 | Any early whitespace before actual text starts. |
|
719 | 726 | |
|
720 | 727 | esc |
|
721 | 728 | The initial esc character (or characters, for double-char escapes like |
|
722 | 729 | '??' or '!!'). |
|
723 | 730 | |
|
724 | 731 | fpart |
|
725 | 732 | The 'function part', which is basically the maximal initial sequence |
|
726 | 733 | of valid python identifiers and the '.' character. This is what is |
|
727 | 734 | checked for alias and magic transformations, used for auto-calling, |
|
728 | 735 | etc. |
|
729 | 736 | |
|
730 | 737 | rest |
|
731 | 738 | Everything else on the line. |
|
732 | 739 | """ |
|
733 | 740 | def __init__(self, line): |
|
734 | 741 | self.line = line |
|
735 | 742 | self.lspace, self.esc, self.fpart, self.rest = \ |
|
736 | 743 | split_user_input(line) |
|
737 | 744 | |
|
738 | 745 | def __str__(self): |
|
739 | 746 | return "LineInfo [%s|%s|%s|%s]" % (self.lspace, self.esc, |
|
740 | 747 | self.fpart, self.rest) |
|
741 | 748 | |
|
742 | 749 | |
|
743 | 750 | # Transformations of the special syntaxes that don't rely on an explicit escape |
|
744 | 751 | # character but instead on patterns on the input line |
|
745 | 752 | |
|
746 | 753 | # The core transformations are implemented as standalone functions that can be |
|
747 | 754 | # tested and validated in isolation. Each of these uses a regexp, we |
|
748 | 755 | # pre-compile these and keep them close to each function definition for clarity |
|
749 | 756 | |
|
750 | 757 | _assign_system_re = re.compile(r'(?P<lhs>(\s*)([\w\.]+)((\s*,\s*[\w\.]+)*))' |
|
751 | 758 | r'\s*=\s*!\s*(?P<cmd>.*)') |
|
752 | 759 | |
|
753 | 760 | def transform_assign_system(line): |
|
754 | 761 | """Handle the `files = !ls` syntax.""" |
|
755 | 762 | m = _assign_system_re.match(line) |
|
756 | 763 | if m is not None: |
|
757 | 764 | cmd = m.group('cmd') |
|
758 | 765 | lhs = m.group('lhs') |
|
759 | 766 | expr = make_quoted_expr(cmd) |
|
760 | 767 | new_line = '%s = get_ipython().getoutput(%s)' % (lhs, expr) |
|
761 | 768 | return new_line |
|
762 | 769 | return line |
|
763 | 770 | |
|
764 | 771 | |
|
765 | 772 | _assign_magic_re = re.compile(r'(?P<lhs>(\s*)([\w\.]+)((\s*,\s*[\w\.]+)*))' |
|
766 | 773 | r'\s*=\s*%\s*(?P<cmd>.*)') |
|
767 | 774 | |
|
768 | 775 | def transform_assign_magic(line): |
|
769 | 776 | """Handle the `a = %who` syntax.""" |
|
770 | 777 | m = _assign_magic_re.match(line) |
|
771 | 778 | if m is not None: |
|
772 | 779 | cmd = m.group('cmd') |
|
773 | 780 | lhs = m.group('lhs') |
|
774 | 781 | expr = make_quoted_expr(cmd) |
|
775 | 782 | new_line = '%s = get_ipython().magic(%s)' % (lhs, expr) |
|
776 | 783 | return new_line |
|
777 | 784 | return line |
|
778 | 785 | |
|
779 | 786 | |
|
780 | 787 | _classic_prompt_re = re.compile(r'^([ \t]*>>> |^[ \t]*\.\.\. )') |
|
781 | 788 | |
|
782 | 789 | def transform_classic_prompt(line): |
|
783 | 790 | """Handle inputs that start with '>>> ' syntax.""" |
|
784 | 791 | |
|
785 | 792 | if not line or line.isspace(): |
|
786 | 793 | return line |
|
787 | 794 | m = _classic_prompt_re.match(line) |
|
788 | 795 | if m: |
|
789 | 796 | return line[len(m.group(0)):] |
|
790 | 797 | else: |
|
791 | 798 | return line |
|
792 | 799 | |
|
793 | 800 | |
|
794 | 801 | _ipy_prompt_re = re.compile(r'^([ \t]*In \[\d+\]: |^[ \t]*\ \ \ \.\.\.+: )') |
|
795 | 802 | |
|
796 | 803 | def transform_ipy_prompt(line): |
|
797 | 804 | """Handle inputs that start classic IPython prompt syntax.""" |
|
798 | 805 | |
|
799 | 806 | if not line or line.isspace(): |
|
800 | 807 | return line |
|
801 | 808 | #print 'LINE: %r' % line # dbg |
|
802 | 809 | m = _ipy_prompt_re.match(line) |
|
803 | 810 | if m: |
|
804 | 811 | #print 'MATCH! %r -> %r' % (line, line[len(m.group(0)):]) # dbg |
|
805 | 812 | return line[len(m.group(0)):] |
|
806 | 813 | else: |
|
807 | 814 | return line |
|
808 | 815 | |
|
809 | 816 | |
|
810 | 817 | class EscapedTransformer(object): |
|
811 | 818 | """Class to transform lines that are explicitly escaped out.""" |
|
812 | 819 | |
|
813 | 820 | def __init__(self): |
|
814 | 821 | tr = { ESC_SHELL : self._tr_system, |
|
815 | 822 | ESC_SH_CAP : self._tr_system2, |
|
816 | 823 | ESC_HELP : self._tr_help, |
|
817 | 824 | ESC_HELP2 : self._tr_help, |
|
818 | 825 | ESC_MAGIC : self._tr_magic, |
|
819 | 826 | ESC_QUOTE : self._tr_quote, |
|
820 | 827 | ESC_QUOTE2 : self._tr_quote2, |
|
821 | 828 | ESC_PAREN : self._tr_paren } |
|
822 | 829 | self.tr = tr |
|
823 | 830 | |
|
824 | 831 | # Support for syntax transformations that use explicit escapes typed by the |
|
825 | 832 | # user at the beginning of a line |
|
826 | 833 | @staticmethod |
|
827 | 834 | def _tr_system(line_info): |
|
828 | 835 | "Translate lines escaped with: !" |
|
829 | 836 | cmd = line_info.line.lstrip().lstrip(ESC_SHELL) |
|
830 | 837 | return '%sget_ipython().system(%s)' % (line_info.lspace, |
|
831 | 838 | make_quoted_expr(cmd)) |
|
832 | 839 | |
|
833 | 840 | @staticmethod |
|
834 | 841 | def _tr_system2(line_info): |
|
835 | 842 | "Translate lines escaped with: !!" |
|
836 | 843 | cmd = line_info.line.lstrip()[2:] |
|
837 | 844 | return '%sget_ipython().getoutput(%s)' % (line_info.lspace, |
|
838 | 845 | make_quoted_expr(cmd)) |
|
839 | 846 | |
|
840 | 847 | @staticmethod |
|
841 | 848 | def _tr_help(line_info): |
|
842 | 849 | "Translate lines escaped with: ?/??" |
|
843 | 850 | # A naked help line should just fire the intro help screen |
|
844 | 851 | if not line_info.line[1:]: |
|
845 | 852 | return 'get_ipython().show_usage()' |
|
846 | 853 | |
|
847 | 854 | # There may be one or two '?' at the end, move them to the front so that |
|
848 | 855 | # the rest of the logic can assume escapes are at the start |
|
849 | 856 | l_ori = line_info |
|
850 | 857 | line = line_info.line |
|
851 | 858 | if line.endswith('?'): |
|
852 | 859 | line = line[-1] + line[:-1] |
|
853 | 860 | if line.endswith('?'): |
|
854 | 861 | line = line[-1] + line[:-1] |
|
855 | 862 | line_info = LineInfo(line) |
|
856 | 863 | |
|
857 | 864 | # From here on, simply choose which level of detail to get, and |
|
858 | 865 | # special-case the psearch syntax |
|
859 | 866 | pinfo = 'pinfo' # default |
|
860 | 867 | if '*' in line_info.line: |
|
861 | 868 | pinfo = 'psearch' |
|
862 | 869 | elif line_info.esc == '??': |
|
863 | 870 | pinfo = 'pinfo2' |
|
864 | 871 | |
|
865 | 872 | tpl = '%sget_ipython().magic("%s %s")' |
|
866 | 873 | return tpl % (line_info.lspace, pinfo, |
|
867 | 874 | ' '.join([line_info.fpart, line_info.rest]).strip()) |
|
868 | 875 | |
|
869 | 876 | @staticmethod |
|
870 | 877 | def _tr_magic(line_info): |
|
871 | 878 | "Translate lines escaped with: %" |
|
872 | 879 | tpl = '%sget_ipython().magic(%s)' |
|
873 | 880 | cmd = make_quoted_expr(' '.join([line_info.fpart, |
|
874 | 881 | line_info.rest]).strip()) |
|
875 | 882 | return tpl % (line_info.lspace, cmd) |
|
876 | 883 | |
|
877 | 884 | @staticmethod |
|
878 | 885 | def _tr_quote(line_info): |
|
879 | 886 | "Translate lines escaped with: ," |
|
880 | 887 | return '%s%s("%s")' % (line_info.lspace, line_info.fpart, |
|
881 | 888 | '", "'.join(line_info.rest.split()) ) |
|
882 | 889 | |
|
883 | 890 | @staticmethod |
|
884 | 891 | def _tr_quote2(line_info): |
|
885 | 892 | "Translate lines escaped with: ;" |
|
886 | 893 | return '%s%s("%s")' % (line_info.lspace, line_info.fpart, |
|
887 | 894 | line_info.rest) |
|
888 | 895 | |
|
889 | 896 | @staticmethod |
|
890 | 897 | def _tr_paren(line_info): |
|
891 | 898 | "Translate lines escaped with: /" |
|
892 | 899 | return '%s%s(%s)' % (line_info.lspace, line_info.fpart, |
|
893 | 900 | ", ".join(line_info.rest.split())) |
|
894 | 901 | |
|
895 | 902 | def __call__(self, line): |
|
896 | 903 | """Class to transform lines that are explicitly escaped out. |
|
897 | 904 | |
|
898 | 905 | This calls the above _tr_* static methods for the actual line |
|
899 | 906 | translations.""" |
|
900 | 907 | |
|
901 | 908 | # Empty lines just get returned unmodified |
|
902 | 909 | if not line or line.isspace(): |
|
903 | 910 | return line |
|
904 | 911 | |
|
905 | 912 | # Get line endpoints, where the escapes can be |
|
906 | 913 | line_info = LineInfo(line) |
|
907 | 914 | |
|
908 | 915 | # If the escape is not at the start, only '?' needs to be special-cased. |
|
909 | 916 | # All other escapes are only valid at the start |
|
910 | 917 | if not line_info.esc in self.tr: |
|
911 | 918 | if line.endswith(ESC_HELP): |
|
912 | 919 | return self._tr_help(line_info) |
|
913 | 920 | else: |
|
914 | 921 | # If we don't recognize the escape, don't modify the line |
|
915 | 922 | return line |
|
916 | 923 | |
|
917 | 924 | return self.tr[line_info.esc](line_info) |
|
918 | 925 | |
|
919 | 926 | |
|
920 | 927 | # A function-looking object to be used by the rest of the code. The purpose of |
|
921 | 928 | # the class in this case is to organize related functionality, more than to |
|
922 | 929 | # manage state. |
|
923 | 930 | transform_escaped = EscapedTransformer() |
|
924 | 931 | |
|
925 | 932 | |
|
926 | 933 | class IPythonInputSplitter(InputSplitter): |
|
927 | 934 | """An input splitter that recognizes all of IPython's special syntax.""" |
|
928 | 935 | |
|
929 | 936 | # String with raw, untransformed input. |
|
930 | 937 | source_raw = '' |
|
931 | 938 | |
|
932 | 939 | # Private attributes |
|
933 | 940 | |
|
934 | 941 | # List with lines of raw input accumulated so far. |
|
935 | 942 | _buffer_raw = None |
|
936 | 943 | |
|
937 | 944 | def __init__(self, input_mode=None): |
|
938 | 945 | InputSplitter.__init__(self, input_mode) |
|
939 | 946 | self._buffer_raw = [] |
|
940 | 947 | |
|
941 | 948 | def reset(self): |
|
942 | 949 | """Reset the input buffer and associated state.""" |
|
943 | 950 | InputSplitter.reset(self) |
|
944 | 951 | self._buffer_raw[:] = [] |
|
945 | 952 | self.source_raw = '' |
|
946 | 953 | |
|
947 | 954 | def source_raw_reset(self): |
|
948 | 955 | """Return input and raw source and perform a full reset. |
|
949 | 956 | """ |
|
950 | 957 | out = self.source |
|
951 | 958 | out_r = self.source_raw |
|
952 | 959 | self.reset() |
|
953 | 960 | return out, out_r |
|
954 | 961 | |
|
955 | 962 | def push(self, lines): |
|
956 | 963 | """Push one or more lines of IPython input. |
|
957 | 964 | """ |
|
958 | 965 | if not lines: |
|
959 | 966 | return super(IPythonInputSplitter, self).push(lines) |
|
960 | 967 | |
|
961 | 968 | # We must ensure all input is pure unicode |
|
962 | 969 | if type(lines)==str: |
|
963 | 970 | lines = lines.decode(self.encoding) |
|
964 | 971 | |
|
965 | 972 | lines_list = lines.splitlines() |
|
966 | 973 | |
|
967 | 974 | transforms = [transform_escaped, transform_assign_system, |
|
968 | 975 | transform_assign_magic, transform_ipy_prompt, |
|
969 | 976 | transform_classic_prompt] |
|
970 | 977 | |
|
971 | 978 | # Transform logic |
|
972 | 979 | # |
|
973 | 980 | # We only apply the line transformers to the input if we have either no |
|
974 | 981 | # input yet, or complete input, or if the last line of the buffer ends |
|
975 | 982 | # with ':' (opening an indented block). This prevents the accidental |
|
976 | 983 | # transformation of escapes inside multiline expressions like |
|
977 | 984 | # triple-quoted strings or parenthesized expressions. |
|
978 | 985 | # |
|
979 | 986 | # The last heuristic, while ugly, ensures that the first line of an |
|
980 | 987 | # indented block is correctly transformed. |
|
981 | 988 | # |
|
982 | 989 | # FIXME: try to find a cleaner approach for this last bit. |
|
983 | 990 | |
|
984 | 991 | # If we were in 'block' mode, since we're going to pump the parent |
|
985 | 992 | # class by hand line by line, we need to temporarily switch out to |
|
986 | 993 | # 'line' mode, do a single manual reset and then feed the lines one |
|
987 | 994 | # by one. Note that this only matters if the input has more than one |
|
988 | 995 | # line. |
|
989 | 996 | changed_input_mode = False |
|
990 | 997 | |
|
991 | 998 | if self.input_mode == 'cell': |
|
992 | 999 | self.reset() |
|
993 | 1000 | changed_input_mode = True |
|
994 | 1001 | saved_input_mode = 'cell' |
|
995 | 1002 | self.input_mode = 'line' |
|
996 | 1003 | |
|
997 | 1004 | # Store raw source before applying any transformations to it. Note |
|
998 | 1005 | # that this must be done *after* the reset() call that would otherwise |
|
999 | 1006 | # flush the buffer. |
|
1000 | 1007 | self._store(lines, self._buffer_raw, 'source_raw') |
|
1001 | 1008 | |
|
1002 | 1009 | try: |
|
1003 | 1010 | push = super(IPythonInputSplitter, self).push |
|
1004 | 1011 | for line in lines_list: |
|
1005 | 1012 | if self._is_complete or not self._buffer or \ |
|
1006 | 1013 | (self._buffer and self._buffer[-1].rstrip().endswith(':')): |
|
1007 | 1014 | for f in transforms: |
|
1008 | 1015 | line = f(line) |
|
1009 | 1016 | |
|
1010 | 1017 | out = push(line) |
|
1011 | 1018 | finally: |
|
1012 | 1019 | if changed_input_mode: |
|
1013 | 1020 | self.input_mode = saved_input_mode |
|
1014 | 1021 | return out |
@@ -1,679 +1,688 b'' | |||
|
1 | 1 | # -*- coding: utf-8 -*- |
|
2 | 2 | """Tests for the inputsplitter module. |
|
3 | 3 | """ |
|
4 | 4 | #----------------------------------------------------------------------------- |
|
5 | 5 | # Copyright (C) 2010 The IPython Development Team |
|
6 | 6 | # |
|
7 | 7 | # Distributed under the terms of the BSD License. The full license is in |
|
8 | 8 | # the file COPYING, distributed as part of this software. |
|
9 | 9 | #----------------------------------------------------------------------------- |
|
10 | 10 | |
|
11 | 11 | #----------------------------------------------------------------------------- |
|
12 | 12 | # Imports |
|
13 | 13 | #----------------------------------------------------------------------------- |
|
14 | 14 | # stdlib |
|
15 | 15 | import unittest |
|
16 | 16 | import sys |
|
17 | 17 | |
|
18 | 18 | # Third party |
|
19 | 19 | import nose.tools as nt |
|
20 | 20 | |
|
21 | 21 | # Our own |
|
22 | 22 | from IPython.core import inputsplitter as isp |
|
23 | 23 | |
|
24 | 24 | #----------------------------------------------------------------------------- |
|
25 | 25 | # Semi-complete examples (also used as tests) |
|
26 | 26 | #----------------------------------------------------------------------------- |
|
27 | 27 | |
|
28 | 28 | # Note: at the bottom, there's a slightly more complete version of this that |
|
29 | 29 | # can be useful during development of code here. |
|
30 | 30 | |
|
31 | 31 | def mini_interactive_loop(input_func): |
|
32 | 32 | """Minimal example of the logic of an interactive interpreter loop. |
|
33 | 33 | |
|
34 | 34 | This serves as an example, and it is used by the test system with a fake |
|
35 | 35 | raw_input that simulates interactive input.""" |
|
36 | 36 | |
|
37 | 37 | from IPython.core.inputsplitter import InputSplitter |
|
38 | 38 | |
|
39 | 39 | isp = InputSplitter() |
|
40 | 40 | # In practice, this input loop would be wrapped in an outside loop to read |
|
41 | 41 | # input indefinitely, until some exit/quit command was issued. Here we |
|
42 | 42 | # only illustrate the basic inner loop. |
|
43 | 43 | while isp.push_accepts_more(): |
|
44 | 44 | indent = ' '*isp.indent_spaces |
|
45 | 45 | prompt = '>>> ' + indent |
|
46 | 46 | line = indent + input_func(prompt) |
|
47 | 47 | isp.push(line) |
|
48 | 48 | |
|
49 | 49 | # Here we just return input so we can use it in a test suite, but a real |
|
50 | 50 | # interpreter would instead send it for execution somewhere. |
|
51 | 51 | src = isp.source_reset() |
|
52 | 52 | #print 'Input source was:\n', src # dbg |
|
53 | 53 | return src |
|
54 | 54 | |
|
55 | 55 | #----------------------------------------------------------------------------- |
|
56 | 56 | # Test utilities, just for local use |
|
57 | 57 | #----------------------------------------------------------------------------- |
|
58 | 58 | |
|
59 | 59 | def assemble(block): |
|
60 | 60 | """Assemble a block into multi-line sub-blocks.""" |
|
61 | 61 | return ['\n'.join(sub_block)+'\n' for sub_block in block] |
|
62 | 62 | |
|
63 | 63 | |
|
64 | 64 | def pseudo_input(lines): |
|
65 | 65 | """Return a function that acts like raw_input but feeds the input list.""" |
|
66 | 66 | ilines = iter(lines) |
|
67 | 67 | def raw_in(prompt): |
|
68 | 68 | try: |
|
69 | 69 | return next(ilines) |
|
70 | 70 | except StopIteration: |
|
71 | 71 | return '' |
|
72 | 72 | return raw_in |
|
73 | 73 | |
|
74 | 74 | #----------------------------------------------------------------------------- |
|
75 | 75 | # Tests |
|
76 | 76 | #----------------------------------------------------------------------------- |
|
77 | 77 | def test_spaces(): |
|
78 | 78 | tests = [('', 0), |
|
79 | 79 | (' ', 1), |
|
80 | 80 | ('\n', 0), |
|
81 | 81 | (' \n', 1), |
|
82 | 82 | ('x', 0), |
|
83 | 83 | (' x', 1), |
|
84 | 84 | (' x',2), |
|
85 | 85 | (' x',4), |
|
86 | 86 | # Note: tabs are counted as a single whitespace! |
|
87 | 87 | ('\tx', 1), |
|
88 | 88 | ('\t x', 2), |
|
89 | 89 | ] |
|
90 | 90 | |
|
91 | 91 | for s, nsp in tests: |
|
92 | 92 | nt.assert_equal(isp.num_ini_spaces(s), nsp) |
|
93 | 93 | |
|
94 | 94 | |
|
95 | 95 | def test_remove_comments(): |
|
96 | 96 | tests = [('text', 'text'), |
|
97 | 97 | ('text # comment', 'text '), |
|
98 | 98 | ('text # comment\n', 'text \n'), |
|
99 | 99 | ('text # comment \n', 'text \n'), |
|
100 | 100 | ('line # c \nline\n','line \nline\n'), |
|
101 | 101 | ('line # c \nline#c2 \nline\nline #c\n\n', |
|
102 | 102 | 'line \nline\nline\nline \n\n'), |
|
103 | 103 | ] |
|
104 | 104 | |
|
105 | 105 | for inp, out in tests: |
|
106 | 106 | nt.assert_equal(isp.remove_comments(inp), out) |
|
107 | 107 | |
|
108 | 108 | |
|
109 | 109 | def test_get_input_encoding(): |
|
110 | 110 | encoding = isp.get_input_encoding() |
|
111 | 111 | nt.assert_true(isinstance(encoding, basestring)) |
|
112 | 112 | # simple-minded check that at least encoding a simple string works with the |
|
113 | 113 | # encoding we got. |
|
114 | 114 | nt.assert_equal('test'.encode(encoding), 'test') |
|
115 | 115 | |
|
116 | 116 | |
|
117 | 117 | class NoInputEncodingTestCase(unittest.TestCase): |
|
118 | 118 | def setUp(self): |
|
119 | 119 | self.old_stdin = sys.stdin |
|
120 | 120 | class X: pass |
|
121 | 121 | fake_stdin = X() |
|
122 | 122 | sys.stdin = fake_stdin |
|
123 | 123 | |
|
124 | 124 | def test(self): |
|
125 | 125 | # Verify that if sys.stdin has no 'encoding' attribute we do the right |
|
126 | 126 | # thing |
|
127 | 127 | enc = isp.get_input_encoding() |
|
128 | 128 | self.assertEqual(enc, 'ascii') |
|
129 | 129 | |
|
130 | 130 | def tearDown(self): |
|
131 | 131 | sys.stdin = self.old_stdin |
|
132 | 132 | |
|
133 | 133 | |
|
134 | 134 | class InputSplitterTestCase(unittest.TestCase): |
|
135 | 135 | def setUp(self): |
|
136 | 136 | self.isp = isp.InputSplitter() |
|
137 | 137 | |
|
138 | 138 | def test_reset(self): |
|
139 | 139 | isp = self.isp |
|
140 | 140 | isp.push('x=1') |
|
141 | 141 | isp.reset() |
|
142 | 142 | self.assertEqual(isp._buffer, []) |
|
143 | 143 | self.assertEqual(isp.indent_spaces, 0) |
|
144 | 144 | self.assertEqual(isp.source, '') |
|
145 | 145 | self.assertEqual(isp.code, None) |
|
146 | 146 | self.assertEqual(isp._is_complete, False) |
|
147 | 147 | |
|
148 | 148 | def test_source(self): |
|
149 | 149 | self.isp._store('1') |
|
150 | 150 | self.isp._store('2') |
|
151 | 151 | self.assertEqual(self.isp.source, '1\n2\n') |
|
152 | 152 | self.assertTrue(len(self.isp._buffer)>0) |
|
153 | 153 | self.assertEqual(self.isp.source_reset(), '1\n2\n') |
|
154 | 154 | self.assertEqual(self.isp._buffer, []) |
|
155 | 155 | self.assertEqual(self.isp.source, '') |
|
156 | 156 | |
|
157 | 157 | def test_indent(self): |
|
158 | 158 | isp = self.isp # shorthand |
|
159 | 159 | isp.push('x=1') |
|
160 | 160 | self.assertEqual(isp.indent_spaces, 0) |
|
161 | 161 | isp.push('if 1:\n x=1') |
|
162 | 162 | self.assertEqual(isp.indent_spaces, 4) |
|
163 | 163 | isp.push('y=2\n') |
|
164 | 164 | self.assertEqual(isp.indent_spaces, 0) |
|
165 | 165 | |
|
166 | 166 | def test_indent2(self): |
|
167 | 167 | # In cell mode, inputs must be fed in whole blocks, so skip this test |
|
168 | 168 | if self.isp.input_mode == 'cell': return |
|
169 | 169 | |
|
170 | 170 | isp = self.isp |
|
171 | 171 | isp.push('if 1:') |
|
172 | 172 | self.assertEqual(isp.indent_spaces, 4) |
|
173 | 173 | isp.push(' x=1') |
|
174 | 174 | self.assertEqual(isp.indent_spaces, 4) |
|
175 | 175 | # Blank lines shouldn't change the indent level |
|
176 | 176 | isp.push(' '*2) |
|
177 | 177 | self.assertEqual(isp.indent_spaces, 4) |
|
178 | 178 | |
|
179 | 179 | def test_indent3(self): |
|
180 | 180 | # In cell mode, inputs must be fed in whole blocks, so skip this test |
|
181 | 181 | if self.isp.input_mode == 'cell': return |
|
182 | 182 | |
|
183 | 183 | isp = self.isp |
|
184 | 184 | # When a multiline statement contains parens or multiline strings, we |
|
185 | 185 | # shouldn't get confused. |
|
186 | 186 | isp.push("if 1:") |
|
187 | 187 | isp.push(" x = (1+\n 2)") |
|
188 | 188 | self.assertEqual(isp.indent_spaces, 4) |
|
189 | 189 | |
|
190 | 190 | def test_dedent(self): |
|
191 | 191 | isp = self.isp # shorthand |
|
192 | 192 | isp.push('if 1:') |
|
193 | 193 | self.assertEqual(isp.indent_spaces, 4) |
|
194 | 194 | isp.push(' pass') |
|
195 | 195 | self.assertEqual(isp.indent_spaces, 0) |
|
196 | 196 | |
|
197 | 197 | def test_push(self): |
|
198 | 198 | isp = self.isp |
|
199 | 199 | self.assertTrue(isp.push('x=1')) |
|
200 | 200 | |
|
201 | 201 | def test_push2(self): |
|
202 | 202 | isp = self.isp |
|
203 | 203 | self.assertFalse(isp.push('if 1:')) |
|
204 | 204 | for line in [' x=1', '# a comment', ' y=2']: |
|
205 | 205 | self.assertTrue(isp.push(line)) |
|
206 | 206 | |
|
207 | 207 | def test_replace_mode(self): |
|
208 | 208 | isp = self.isp |
|
209 | 209 | isp.input_mode = 'cell' |
|
210 | 210 | isp.push('x=1') |
|
211 | 211 | self.assertEqual(isp.source, 'x=1\n') |
|
212 | 212 | isp.push('x=2') |
|
213 | 213 | self.assertEqual(isp.source, 'x=2\n') |
|
214 | 214 | |
|
215 | 215 | def test_push_accepts_more(self): |
|
216 | 216 | isp = self.isp |
|
217 | 217 | isp.push('x=1') |
|
218 | 218 | self.assertFalse(isp.push_accepts_more()) |
|
219 | 219 | |
|
220 | 220 | def test_push_accepts_more2(self): |
|
221 | 221 | # In cell mode, inputs must be fed in whole blocks, so skip this test |
|
222 | 222 | if self.isp.input_mode == 'cell': return |
|
223 | 223 | |
|
224 | 224 | isp = self.isp |
|
225 | 225 | isp.push('if 1:') |
|
226 | 226 | self.assertTrue(isp.push_accepts_more()) |
|
227 | 227 | isp.push(' x=1') |
|
228 | 228 | self.assertTrue(isp.push_accepts_more()) |
|
229 | 229 | isp.push('') |
|
230 | 230 | self.assertFalse(isp.push_accepts_more()) |
|
231 | 231 | |
|
232 | 232 | def test_push_accepts_more3(self): |
|
233 | 233 | isp = self.isp |
|
234 | 234 | isp.push("x = (2+\n3)") |
|
235 | 235 | self.assertFalse(isp.push_accepts_more()) |
|
236 | 236 | |
|
237 | 237 | def test_push_accepts_more4(self): |
|
238 | 238 | # In cell mode, inputs must be fed in whole blocks, so skip this test |
|
239 | 239 | if self.isp.input_mode == 'cell': return |
|
240 | 240 | |
|
241 | 241 | isp = self.isp |
|
242 | 242 | # When a multiline statement contains parens or multiline strings, we |
|
243 | 243 | # shouldn't get confused. |
|
244 | 244 | # FIXME: we should be able to better handle de-dents in statements like |
|
245 | 245 | # multiline strings and multiline expressions (continued with \ or |
|
246 | 246 | # parens). Right now we aren't handling the indentation tracking quite |
|
247 | 247 | # correctly with this, though in practice it may not be too much of a |
|
248 | 248 | # problem. We'll need to see. |
|
249 | 249 | isp.push("if 1:") |
|
250 | 250 | isp.push(" x = (2+") |
|
251 | 251 | isp.push(" 3)") |
|
252 | 252 | self.assertTrue(isp.push_accepts_more()) |
|
253 | 253 | isp.push(" y = 3") |
|
254 | 254 | self.assertTrue(isp.push_accepts_more()) |
|
255 | 255 | isp.push('') |
|
256 | 256 | self.assertFalse(isp.push_accepts_more()) |
|
257 | 257 | |
|
258 | 258 | def test_continuation(self): |
|
259 | 259 | isp = self.isp |
|
260 | 260 | isp.push("import os, \\") |
|
261 | 261 | self.assertTrue(isp.push_accepts_more()) |
|
262 | 262 | isp.push("sys") |
|
263 | 263 | self.assertFalse(isp.push_accepts_more()) |
|
264 | 264 | |
|
265 | 265 | def test_syntax_error(self): |
|
266 | 266 | isp = self.isp |
|
267 | 267 | # Syntax errors immediately produce a 'ready' block, so the invalid |
|
268 | 268 | # Python can be sent to the kernel for evaluation with possible ipython |
|
269 | 269 | # special-syntax conversion. |
|
270 | 270 | isp.push('run foo') |
|
271 | 271 | self.assertFalse(isp.push_accepts_more()) |
|
272 | 272 | |
|
273 | 273 | def check_split(self, block_lines, compile=True): |
|
274 | 274 | blocks = assemble(block_lines) |
|
275 | 275 | lines = ''.join(blocks) |
|
276 | 276 | oblock = self.isp.split_blocks(lines) |
|
277 | 277 | self.assertEqual(oblock, blocks) |
|
278 | 278 | if compile: |
|
279 | 279 | for block in blocks: |
|
280 | 280 | self.isp._compile(block) |
|
281 | 281 | |
|
282 | 282 | def test_split(self): |
|
283 | 283 | # All blocks of input we want to test in a list. The format for each |
|
284 | 284 | # block is a list of lists, with each inner lists consisting of all the |
|
285 | 285 | # lines (as single-lines) that should make up a sub-block. |
|
286 | 286 | |
|
287 | 287 | # Note: do NOT put here sub-blocks that don't compile, as the |
|
288 | 288 | # check_split() routine makes a final verification pass to check that |
|
289 | 289 | # each sub_block, as returned by split_blocks(), does compile |
|
290 | 290 | # correctly. |
|
291 | 291 | all_blocks = [ [['x=1']], |
|
292 | 292 | |
|
293 | 293 | [['x=1'], |
|
294 | 294 | ['y=2']], |
|
295 | 295 | |
|
296 | 296 | [['x=1', |
|
297 | 297 | '# a comment'], |
|
298 | 298 | ['y=11']], |
|
299 | 299 | |
|
300 | 300 | [['if 1:', |
|
301 | 301 | ' x=1'], |
|
302 | 302 | ['y=3']], |
|
303 | 303 | |
|
304 | 304 | [['def f(x):', |
|
305 | 305 | ' return x'], |
|
306 | 306 | ['x=1']], |
|
307 | 307 | |
|
308 | 308 | [['def f(x):', |
|
309 | 309 | ' x+=1', |
|
310 | 310 | ' ', |
|
311 | 311 | ' return x'], |
|
312 | 312 | ['x=1']], |
|
313 | 313 | |
|
314 | 314 | [['def f(x):', |
|
315 | 315 | ' if x>0:', |
|
316 | 316 | ' y=1', |
|
317 | 317 | ' # a comment', |
|
318 | 318 | ' else:', |
|
319 | 319 | ' y=4', |
|
320 | 320 | ' ', |
|
321 | 321 | ' return y'], |
|
322 | 322 | ['x=1'], |
|
323 | 323 | ['if 1:', |
|
324 | 324 | ' y=11'] ], |
|
325 | 325 | |
|
326 | 326 | [['for i in range(10):' |
|
327 | 327 | ' x=i**2']], |
|
328 | 328 | |
|
329 | 329 | [['for i in range(10):' |
|
330 | 330 | ' x=i**2'], |
|
331 | 331 | ['z = 1']], |
|
332 | ||
|
333 | [['"asdf"']], | |
|
334 | ||
|
335 | [['"asdf"'], | |
|
336 | ['10'], | |
|
337 | ], | |
|
338 | ||
|
339 | [['"""foo', | |
|
340 | 'bar"""']], | |
|
332 | 341 | ] |
|
333 | 342 | for block_lines in all_blocks: |
|
334 | 343 | self.check_split(block_lines) |
|
335 | 344 | |
|
336 | 345 | def test_split_syntax_errors(self): |
|
337 | 346 | # Block splitting with invalid syntax |
|
338 | 347 | all_blocks = [ [['a syntax error']], |
|
339 | 348 | |
|
340 | 349 | [['x=1', |
|
341 | 350 | 'another syntax error']], |
|
342 | 351 | |
|
343 | 352 | [['for i in range(10):' |
|
344 | 353 | ' yet another error']], |
|
345 | 354 | |
|
346 | 355 | ] |
|
347 | 356 | for block_lines in all_blocks: |
|
348 | 357 | self.check_split(block_lines, compile=False) |
|
349 | 358 | |
|
350 | 359 | def test_unicode(self): |
|
351 | 360 | self.isp.push(u"PΓ©rez") |
|
352 | 361 | self.isp.push(u'\xc3\xa9') |
|
353 | 362 | self.isp.push("u'\xc3\xa9'") |
|
354 | 363 | |
|
355 | 364 | class InteractiveLoopTestCase(unittest.TestCase): |
|
356 | 365 | """Tests for an interactive loop like a python shell. |
|
357 | 366 | """ |
|
358 | 367 | def check_ns(self, lines, ns): |
|
359 | 368 | """Validate that the given input lines produce the resulting namespace. |
|
360 | 369 | |
|
361 | 370 | Note: the input lines are given exactly as they would be typed in an |
|
362 | 371 | auto-indenting environment, as mini_interactive_loop above already does |
|
363 | 372 | auto-indenting and prepends spaces to the input. |
|
364 | 373 | """ |
|
365 | 374 | src = mini_interactive_loop(pseudo_input(lines)) |
|
366 | 375 | test_ns = {} |
|
367 | 376 | exec src in test_ns |
|
368 | 377 | # We can't check that the provided ns is identical to the test_ns, |
|
369 | 378 | # because Python fills test_ns with extra keys (copyright, etc). But |
|
370 | 379 | # we can check that the given dict is *contained* in test_ns |
|
371 | 380 | for k,v in ns.iteritems(): |
|
372 | 381 | self.assertEqual(test_ns[k], v) |
|
373 | 382 | |
|
374 | 383 | def test_simple(self): |
|
375 | 384 | self.check_ns(['x=1'], dict(x=1)) |
|
376 | 385 | |
|
377 | 386 | def test_simple2(self): |
|
378 | 387 | self.check_ns(['if 1:', 'x=2'], dict(x=2)) |
|
379 | 388 | |
|
380 | 389 | def test_xy(self): |
|
381 | 390 | self.check_ns(['x=1; y=2'], dict(x=1, y=2)) |
|
382 | 391 | |
|
383 | 392 | def test_abc(self): |
|
384 | 393 | self.check_ns(['if 1:','a=1','b=2','c=3'], dict(a=1, b=2, c=3)) |
|
385 | 394 | |
|
386 | 395 | def test_multi(self): |
|
387 | 396 | self.check_ns(['x =(1+','1+','2)'], dict(x=4)) |
|
388 | 397 | |
|
389 | 398 | |
|
390 | 399 | def test_LineInfo(): |
|
391 | 400 | """Simple test for LineInfo construction and str()""" |
|
392 | 401 | linfo = isp.LineInfo(' %cd /home') |
|
393 | 402 | nt.assert_equals(str(linfo), 'LineInfo [ |%|cd|/home]') |
|
394 | 403 | |
|
395 | 404 | |
|
396 | 405 | def test_split_user_input(): |
|
397 | 406 | """Unicode test - split_user_input already has good doctests""" |
|
398 | 407 | line = u"PΓ©rez Fernando" |
|
399 | 408 | parts = isp.split_user_input(line) |
|
400 | 409 | parts_expected = (u'', u'', u'', line) |
|
401 | 410 | nt.assert_equal(parts, parts_expected) |
|
402 | 411 | |
|
403 | 412 | |
|
404 | 413 | # Transformer tests |
|
405 | 414 | def transform_checker(tests, func): |
|
406 | 415 | """Utility to loop over test inputs""" |
|
407 | 416 | for inp, tr in tests: |
|
408 | 417 | nt.assert_equals(func(inp), tr) |
|
409 | 418 | |
|
410 | 419 | # Data for all the syntax tests in the form of lists of pairs of |
|
411 | 420 | # raw/transformed input. We store it here as a global dict so that we can use |
|
412 | 421 | # it both within single-function tests and also to validate the behavior of the |
|
413 | 422 | # larger objects |
|
414 | 423 | |
|
415 | 424 | syntax = \ |
|
416 | 425 | dict(assign_system = |
|
417 | 426 | [('a =! ls', 'a = get_ipython().getoutput("ls")'), |
|
418 | 427 | ('b = !ls', 'b = get_ipython().getoutput("ls")'), |
|
419 | 428 | ('x=1', 'x=1'), # normal input is unmodified |
|
420 | 429 | (' ',' '), # blank lines are kept intact |
|
421 | 430 | ], |
|
422 | 431 | |
|
423 | 432 | assign_magic = |
|
424 | 433 | [('a =% who', 'a = get_ipython().magic("who")'), |
|
425 | 434 | ('b = %who', 'b = get_ipython().magic("who")'), |
|
426 | 435 | ('x=1', 'x=1'), # normal input is unmodified |
|
427 | 436 | (' ',' '), # blank lines are kept intact |
|
428 | 437 | ], |
|
429 | 438 | |
|
430 | 439 | classic_prompt = |
|
431 | 440 | [('>>> x=1', 'x=1'), |
|
432 | 441 | ('x=1', 'x=1'), # normal input is unmodified |
|
433 | 442 | (' ', ' '), # blank lines are kept intact |
|
434 | 443 | ('... ', ''), # continuation prompts |
|
435 | 444 | ], |
|
436 | 445 | |
|
437 | 446 | ipy_prompt = |
|
438 | 447 | [('In [1]: x=1', 'x=1'), |
|
439 | 448 | ('x=1', 'x=1'), # normal input is unmodified |
|
440 | 449 | (' ',' '), # blank lines are kept intact |
|
441 | 450 | (' ....: ', ''), # continuation prompts |
|
442 | 451 | ], |
|
443 | 452 | |
|
444 | 453 | # Tests for the escape transformer to leave normal code alone |
|
445 | 454 | escaped_noesc = |
|
446 | 455 | [ (' ', ' '), |
|
447 | 456 | ('x=1', 'x=1'), |
|
448 | 457 | ], |
|
449 | 458 | |
|
450 | 459 | # System calls |
|
451 | 460 | escaped_shell = |
|
452 | 461 | [ ('!ls', 'get_ipython().system("ls")'), |
|
453 | 462 | # Double-escape shell, this means to capture the output of the |
|
454 | 463 | # subprocess and return it |
|
455 | 464 | ('!!ls', 'get_ipython().getoutput("ls")'), |
|
456 | 465 | ], |
|
457 | 466 | |
|
458 | 467 | # Help/object info |
|
459 | 468 | escaped_help = |
|
460 | 469 | [ ('?', 'get_ipython().show_usage()'), |
|
461 | 470 | ('?x1', 'get_ipython().magic("pinfo x1")'), |
|
462 | 471 | ('??x2', 'get_ipython().magic("pinfo2 x2")'), |
|
463 | 472 | ('x3?', 'get_ipython().magic("pinfo x3")'), |
|
464 | 473 | ('x4??', 'get_ipython().magic("pinfo2 x4")'), |
|
465 | 474 | ('%hist?', 'get_ipython().magic("pinfo %hist")'), |
|
466 | 475 | ('f*?', 'get_ipython().magic("psearch f*")'), |
|
467 | 476 | ('ax.*aspe*?', 'get_ipython().magic("psearch ax.*aspe*")'), |
|
468 | 477 | ], |
|
469 | 478 | |
|
470 | 479 | # Explicit magic calls |
|
471 | 480 | escaped_magic = |
|
472 | 481 | [ ('%cd', 'get_ipython().magic("cd")'), |
|
473 | 482 | ('%cd /home', 'get_ipython().magic("cd /home")'), |
|
474 | 483 | (' %magic', ' get_ipython().magic("magic")'), |
|
475 | 484 | ], |
|
476 | 485 | |
|
477 | 486 | # Quoting with separate arguments |
|
478 | 487 | escaped_quote = |
|
479 | 488 | [ (',f', 'f("")'), |
|
480 | 489 | (',f x', 'f("x")'), |
|
481 | 490 | (' ,f y', ' f("y")'), |
|
482 | 491 | (',f a b', 'f("a", "b")'), |
|
483 | 492 | ], |
|
484 | 493 | |
|
485 | 494 | # Quoting with single argument |
|
486 | 495 | escaped_quote2 = |
|
487 | 496 | [ (';f', 'f("")'), |
|
488 | 497 | (';f x', 'f("x")'), |
|
489 | 498 | (' ;f y', ' f("y")'), |
|
490 | 499 | (';f a b', 'f("a b")'), |
|
491 | 500 | ], |
|
492 | 501 | |
|
493 | 502 | # Simply apply parens |
|
494 | 503 | escaped_paren = |
|
495 | 504 | [ ('/f', 'f()'), |
|
496 | 505 | ('/f x', 'f(x)'), |
|
497 | 506 | (' /f y', ' f(y)'), |
|
498 | 507 | ('/f a b', 'f(a, b)'), |
|
499 | 508 | ], |
|
500 | 509 | |
|
501 | 510 | ) |
|
502 | 511 | |
|
503 | 512 | # multiline syntax examples. Each of these should be a list of lists, with |
|
504 | 513 | # each entry itself having pairs of raw/transformed input. The union (with |
|
505 | 514 | # '\n'.join() of the transformed inputs is what the splitter should produce |
|
506 | 515 | # when fed the raw lines one at a time via push. |
|
507 | 516 | syntax_ml = \ |
|
508 | 517 | dict(classic_prompt = |
|
509 | 518 | [ [('>>> for i in range(10):','for i in range(10):'), |
|
510 | 519 | ('... print i',' print i'), |
|
511 | 520 | ('... ', ''), |
|
512 | 521 | ], |
|
513 | 522 | ], |
|
514 | 523 | |
|
515 | 524 | ipy_prompt = |
|
516 | 525 | [ [('In [24]: for i in range(10):','for i in range(10):'), |
|
517 | 526 | (' ....: print i',' print i'), |
|
518 | 527 | (' ....: ', ''), |
|
519 | 528 | ], |
|
520 | 529 | ], |
|
521 | 530 | ) |
|
522 | 531 | |
|
523 | 532 | |
|
524 | 533 | def test_assign_system(): |
|
525 | 534 | transform_checker(syntax['assign_system'], isp.transform_assign_system) |
|
526 | 535 | |
|
527 | 536 | |
|
528 | 537 | def test_assign_magic(): |
|
529 | 538 | transform_checker(syntax['assign_magic'], isp.transform_assign_magic) |
|
530 | 539 | |
|
531 | 540 | |
|
532 | 541 | def test_classic_prompt(): |
|
533 | 542 | transform_checker(syntax['classic_prompt'], isp.transform_classic_prompt) |
|
534 | 543 | for example in syntax_ml['classic_prompt']: |
|
535 | 544 | transform_checker(example, isp.transform_classic_prompt) |
|
536 | 545 | |
|
537 | 546 | |
|
538 | 547 | def test_ipy_prompt(): |
|
539 | 548 | transform_checker(syntax['ipy_prompt'], isp.transform_ipy_prompt) |
|
540 | 549 | for example in syntax_ml['ipy_prompt']: |
|
541 | 550 | transform_checker(example, isp.transform_ipy_prompt) |
|
542 | 551 | |
|
543 | 552 | |
|
544 | 553 | def test_escaped_noesc(): |
|
545 | 554 | transform_checker(syntax['escaped_noesc'], isp.transform_escaped) |
|
546 | 555 | |
|
547 | 556 | |
|
548 | 557 | def test_escaped_shell(): |
|
549 | 558 | transform_checker(syntax['escaped_shell'], isp.transform_escaped) |
|
550 | 559 | |
|
551 | 560 | |
|
552 | 561 | def test_escaped_help(): |
|
553 | 562 | transform_checker(syntax['escaped_help'], isp.transform_escaped) |
|
554 | 563 | |
|
555 | 564 | |
|
556 | 565 | def test_escaped_magic(): |
|
557 | 566 | transform_checker(syntax['escaped_magic'], isp.transform_escaped) |
|
558 | 567 | |
|
559 | 568 | |
|
560 | 569 | def test_escaped_quote(): |
|
561 | 570 | transform_checker(syntax['escaped_quote'], isp.transform_escaped) |
|
562 | 571 | |
|
563 | 572 | |
|
564 | 573 | def test_escaped_quote2(): |
|
565 | 574 | transform_checker(syntax['escaped_quote2'], isp.transform_escaped) |
|
566 | 575 | |
|
567 | 576 | |
|
568 | 577 | def test_escaped_paren(): |
|
569 | 578 | transform_checker(syntax['escaped_paren'], isp.transform_escaped) |
|
570 | 579 | |
|
571 | 580 | |
|
572 | 581 | class IPythonInputTestCase(InputSplitterTestCase): |
|
573 | 582 | """By just creating a new class whose .isp is a different instance, we |
|
574 | 583 | re-run the same test battery on the new input splitter. |
|
575 | 584 | |
|
576 | 585 | In addition, this runs the tests over the syntax and syntax_ml dicts that |
|
577 | 586 | were tested by individual functions, as part of the OO interface. |
|
578 | 587 | |
|
579 | 588 | It also makes some checks on the raw buffer storage. |
|
580 | 589 | """ |
|
581 | 590 | |
|
582 | 591 | def setUp(self): |
|
583 | 592 | self.isp = isp.IPythonInputSplitter(input_mode='line') |
|
584 | 593 | |
|
585 | 594 | def test_syntax(self): |
|
586 | 595 | """Call all single-line syntax tests from the main object""" |
|
587 | 596 | isp = self.isp |
|
588 | 597 | for example in syntax.itervalues(): |
|
589 | 598 | for raw, out_t in example: |
|
590 | 599 | if raw.startswith(' '): |
|
591 | 600 | continue |
|
592 | 601 | |
|
593 | 602 | isp.push(raw) |
|
594 | 603 | out, out_raw = isp.source_raw_reset() |
|
595 | 604 | self.assertEqual(out.rstrip(), out_t) |
|
596 | 605 | self.assertEqual(out_raw.rstrip(), raw.rstrip()) |
|
597 | 606 | |
|
598 | 607 | def test_syntax_multiline(self): |
|
599 | 608 | isp = self.isp |
|
600 | 609 | for example in syntax_ml.itervalues(): |
|
601 | 610 | out_t_parts = [] |
|
602 | 611 | raw_parts = [] |
|
603 | 612 | for line_pairs in example: |
|
604 | 613 | for lraw, out_t_part in line_pairs: |
|
605 | 614 | isp.push(lraw) |
|
606 | 615 | out_t_parts.append(out_t_part) |
|
607 | 616 | raw_parts.append(lraw) |
|
608 | 617 | |
|
609 | 618 | out, out_raw = isp.source_raw_reset() |
|
610 | 619 | out_t = '\n'.join(out_t_parts).rstrip() |
|
611 | 620 | raw = '\n'.join(raw_parts).rstrip() |
|
612 | 621 | self.assertEqual(out.rstrip(), out_t) |
|
613 | 622 | self.assertEqual(out_raw.rstrip(), raw) |
|
614 | 623 | |
|
615 | 624 | |
|
616 | 625 | class BlockIPythonInputTestCase(IPythonInputTestCase): |
|
617 | 626 | |
|
618 | 627 | # Deactivate tests that don't make sense for the block mode |
|
619 | 628 | test_push3 = test_split = lambda s: None |
|
620 | 629 | |
|
621 | 630 | def setUp(self): |
|
622 | 631 | self.isp = isp.IPythonInputSplitter(input_mode='cell') |
|
623 | 632 | |
|
624 | 633 | def test_syntax_multiline(self): |
|
625 | 634 | isp = self.isp |
|
626 | 635 | for example in syntax_ml.itervalues(): |
|
627 | 636 | raw_parts = [] |
|
628 | 637 | out_t_parts = [] |
|
629 | 638 | for line_pairs in example: |
|
630 | 639 | for raw, out_t_part in line_pairs: |
|
631 | 640 | raw_parts.append(raw) |
|
632 | 641 | out_t_parts.append(out_t_part) |
|
633 | 642 | |
|
634 | 643 | raw = '\n'.join(raw_parts) |
|
635 | 644 | out_t = '\n'.join(out_t_parts) |
|
636 | 645 | |
|
637 | 646 | isp.push(raw) |
|
638 | 647 | out, out_raw = isp.source_raw_reset() |
|
639 | 648 | # Match ignoring trailing whitespace |
|
640 | 649 | self.assertEqual(out.rstrip(), out_t.rstrip()) |
|
641 | 650 | self.assertEqual(out_raw.rstrip(), raw.rstrip()) |
|
642 | 651 | |
|
643 | 652 | |
|
644 | 653 | #----------------------------------------------------------------------------- |
|
645 | 654 | # Main - use as a script, mostly for developer experiments |
|
646 | 655 | #----------------------------------------------------------------------------- |
|
647 | 656 | |
|
648 | 657 | if __name__ == '__main__': |
|
649 | 658 | # A simple demo for interactive experimentation. This code will not get |
|
650 | 659 | # picked up by any test suite. |
|
651 | 660 | from IPython.core.inputsplitter import InputSplitter, IPythonInputSplitter |
|
652 | 661 | |
|
653 | 662 | # configure here the syntax to use, prompt and whether to autoindent |
|
654 | 663 | #isp, start_prompt = InputSplitter(), '>>> ' |
|
655 | 664 | isp, start_prompt = IPythonInputSplitter(), 'In> ' |
|
656 | 665 | |
|
657 | 666 | autoindent = True |
|
658 | 667 | #autoindent = False |
|
659 | 668 | |
|
660 | 669 | try: |
|
661 | 670 | while True: |
|
662 | 671 | prompt = start_prompt |
|
663 | 672 | while isp.push_accepts_more(): |
|
664 | 673 | indent = ' '*isp.indent_spaces |
|
665 | 674 | if autoindent: |
|
666 | 675 | line = indent + raw_input(prompt+indent) |
|
667 | 676 | else: |
|
668 | 677 | line = raw_input(prompt) |
|
669 | 678 | isp.push(line) |
|
670 | 679 | prompt = '... ' |
|
671 | 680 | |
|
672 | 681 | # Here we just return input so we can use it in a test suite, but a |
|
673 | 682 | # real interpreter would instead send it for execution somewhere. |
|
674 | 683 | #src = isp.source; raise EOFError # dbg |
|
675 | 684 | src, raw = isp.source_raw_reset() |
|
676 | 685 | print 'Input source was:\n', src |
|
677 | 686 | print 'Raw source was:\n', raw |
|
678 | 687 | except EOFError: |
|
679 | 688 | print 'Bye' |
General Comments 0
You need to be logged in to leave comments.
Login now