Show More
@@ -1,782 +1,787 | |||
|
1 | 1 | """DEPRECATED: Input handling and transformation machinery. |
|
2 | 2 | |
|
3 | 3 | This module was deprecated in IPython 7.0, in favour of inputtransformer2. |
|
4 | 4 | |
|
5 | 5 | The first class in this module, :class:`InputSplitter`, is designed to tell when |
|
6 | 6 | input from a line-oriented frontend is complete and should be executed, and when |
|
7 | 7 | the user should be prompted for another line of code instead. The name 'input |
|
8 | 8 | splitter' is largely for historical reasons. |
|
9 | 9 | |
|
10 | 10 | A companion, :class:`IPythonInputSplitter`, provides the same functionality but |
|
11 | 11 | with full support for the extended IPython syntax (magics, system calls, etc). |
|
12 | 12 | The code to actually do these transformations is in :mod:`IPython.core.inputtransformer`. |
|
13 | 13 | :class:`IPythonInputSplitter` feeds the raw code to the transformers in order |
|
14 | 14 | and stores the results. |
|
15 | 15 | |
|
16 | 16 | For more details, see the class docstrings below. |
|
17 | 17 | """ |
|
18 | 18 | |
|
19 | 19 | from warnings import warn |
|
20 | 20 | |
|
21 | 21 | warn('IPython.core.inputsplitter is deprecated since IPython 7 in favor of `IPython.core.inputtransformer2`', |
|
22 | 22 | DeprecationWarning) |
|
23 | 23 | |
|
24 | 24 | # Copyright (c) IPython Development Team. |
|
25 | 25 | # Distributed under the terms of the Modified BSD License. |
|
26 | 26 | import ast |
|
27 | 27 | import codeop |
|
28 | 28 | import io |
|
29 | 29 | import re |
|
30 | 30 | import sys |
|
31 | 31 | import tokenize |
|
32 | 32 | import warnings |
|
33 | 33 | |
|
34 | 34 | from typing import List |
|
35 | 35 | |
|
36 | 36 | from IPython.core.inputtransformer import (leading_indent, |
|
37 | 37 | classic_prompt, |
|
38 | 38 | ipy_prompt, |
|
39 | 39 | cellmagic, |
|
40 | 40 | assemble_logical_lines, |
|
41 | 41 | help_end, |
|
42 | 42 | escaped_commands, |
|
43 | 43 | assign_from_magic, |
|
44 | 44 | assign_from_system, |
|
45 | 45 | assemble_python_lines, |
|
46 | 46 | ) |
|
47 | 47 | from IPython.utils import tokenutil |
|
48 | 48 | |
|
49 | 49 | # These are available in this module for backwards compatibility. |
|
50 | 50 | from IPython.core.inputtransformer import (ESC_SHELL, ESC_SH_CAP, ESC_HELP, |
|
51 | 51 | ESC_HELP2, ESC_MAGIC, ESC_MAGIC2, |
|
52 | 52 | ESC_QUOTE, ESC_QUOTE2, ESC_PAREN, ESC_SEQUENCES) |
|
53 | 53 | |
|
54 | 54 | #----------------------------------------------------------------------------- |
|
55 | 55 | # Utilities |
|
56 | 56 | #----------------------------------------------------------------------------- |
|
57 | 57 | |
|
58 | 58 | # FIXME: These are general-purpose utilities that later can be moved to the |
|
59 | 59 | # general ward. Kept here for now because we're being very strict about test |
|
60 | 60 | # coverage with this code, and this lets us ensure that we keep 100% coverage |
|
61 | 61 | # while developing. |
|
62 | 62 | |
|
63 | 63 | # compiled regexps for autoindent management |
|
64 | 64 | dedent_re = re.compile('|'.join([ |
|
65 | 65 | r'^\s+raise(\s.*)?$', # raise statement (+ space + other stuff, maybe) |
|
66 | 66 | r'^\s+raise\([^\)]*\).*$', # wacky raise with immediate open paren |
|
67 | 67 | r'^\s+return(\s.*)?$', # normal return (+ space + other stuff, maybe) |
|
68 | 68 | r'^\s+return\([^\)]*\).*$', # wacky return with immediate open paren |
|
69 | 69 | r'^\s+pass\s*$', # pass (optionally followed by trailing spaces) |
|
70 | 70 | r'^\s+break\s*$', # break (optionally followed by trailing spaces) |
|
71 | 71 | r'^\s+continue\s*$', # continue (optionally followed by trailing spaces) |
|
72 | 72 | ])) |
|
73 | 73 | ini_spaces_re = re.compile(r'^([ \t\r\f\v]+)') |
|
74 | 74 | |
|
75 | 75 | # regexp to match pure comment lines so we don't accidentally insert 'if 1:' |
|
76 | 76 | # before pure comments |
|
77 | 77 | comment_line_re = re.compile(r'^\s*\#') |
|
78 | 78 | |
|
79 | 79 | |
|
80 | 80 | def num_ini_spaces(s): |
|
81 | 81 | """Return the number of initial spaces in a string. |
|
82 | 82 | |
|
83 | 83 | Note that tabs are counted as a single space. For now, we do *not* support |
|
84 | 84 | mixing of tabs and spaces in the user's input. |
|
85 | 85 | |
|
86 | 86 | Parameters |
|
87 | 87 | ---------- |
|
88 | 88 | s : string |
|
89 | 89 | |
|
90 | 90 | Returns |
|
91 | 91 | ------- |
|
92 | 92 | n : int |
|
93 | 93 | """ |
|
94 | 94 | |
|
95 | 95 | ini_spaces = ini_spaces_re.match(s) |
|
96 | 96 | if ini_spaces: |
|
97 | 97 | return ini_spaces.end() |
|
98 | 98 | else: |
|
99 | 99 | return 0 |
|
100 | 100 | |
|
101 | 101 | # Fake token types for partial_tokenize: |
|
102 | 102 | INCOMPLETE_STRING = tokenize.N_TOKENS |
|
103 | 103 | IN_MULTILINE_STATEMENT = tokenize.N_TOKENS + 1 |
|
104 | 104 | |
|
105 | 105 | # The 2 classes below have the same API as TokenInfo, but don't try to look up |
|
106 | 106 | # a token type name that they won't find. |
|
107 | 107 | class IncompleteString: |
|
108 | 108 | type = exact_type = INCOMPLETE_STRING |
|
109 | 109 | def __init__(self, s, start, end, line): |
|
110 | 110 | self.s = s |
|
111 | 111 | self.start = start |
|
112 | 112 | self.end = end |
|
113 | 113 | self.line = line |
|
114 | 114 | |
|
115 | 115 | class InMultilineStatement: |
|
116 | 116 | type = exact_type = IN_MULTILINE_STATEMENT |
|
117 | 117 | def __init__(self, pos, line): |
|
118 | 118 | self.s = '' |
|
119 | 119 | self.start = self.end = pos |
|
120 | 120 | self.line = line |
|
121 | 121 | |
|
122 | 122 | def partial_tokens(s): |
|
123 | 123 | """Iterate over tokens from a possibly-incomplete string of code. |
|
124 | 124 | |
|
125 | 125 | This adds two special token types: INCOMPLETE_STRING and |
|
126 | 126 | IN_MULTILINE_STATEMENT. These can only occur as the last token yielded, and |
|
127 | 127 | represent the two main ways for code to be incomplete. |
|
128 | 128 | """ |
|
129 | 129 | readline = io.StringIO(s).readline |
|
130 | 130 | token = tokenize.TokenInfo(tokenize.NEWLINE, '', (1, 0), (1, 0), '') |
|
131 | 131 | try: |
|
132 | 132 | for token in tokenutil.generate_tokens_catch_errors(readline): |
|
133 | 133 | yield token |
|
134 | 134 | except tokenize.TokenError as e: |
|
135 | 135 | # catch EOF error |
|
136 | 136 | lines = s.splitlines(keepends=True) |
|
137 | 137 | end = len(lines), len(lines[-1]) |
|
138 | 138 | if 'multi-line string' in e.args[0]: |
|
139 | 139 | l, c = start = token.end |
|
140 | 140 | s = lines[l-1][c:] + ''.join(lines[l:]) |
|
141 | 141 | yield IncompleteString(s, start, end, lines[-1]) |
|
142 | 142 | elif 'multi-line statement' in e.args[0]: |
|
143 | 143 | yield InMultilineStatement(end, lines[-1]) |
|
144 | 144 | else: |
|
145 | 145 | raise |
|
146 | 146 | |
|
147 | 147 | def find_next_indent(code): |
|
148 | 148 | """Find the number of spaces for the next line of indentation""" |
|
149 | 149 | tokens = list(partial_tokens(code)) |
|
150 | 150 | if tokens[-1].type == tokenize.ENDMARKER: |
|
151 | 151 | tokens.pop() |
|
152 | 152 | if not tokens: |
|
153 | 153 | return 0 |
|
154 | 154 | |
|
155 | while (tokens[-1].type in {tokenize.DEDENT, tokenize.NEWLINE, tokenize.COMMENT, tokenize.ERRORTOKEN}): | |
|
155 | while tokens[-1].type in { | |
|
156 | tokenize.DEDENT, | |
|
157 | tokenize.NEWLINE, | |
|
158 | tokenize.COMMENT, | |
|
159 | tokenize.ERRORTOKEN, | |
|
160 | }: | |
|
156 | 161 | tokens.pop() |
|
157 | 162 | |
|
158 | 163 | # Starting in Python 3.12, the tokenize module adds implicit newlines at the end |
|
159 | 164 | # of input. We need to remove those if we're in a multiline statement |
|
160 | 165 | if tokens[-1].type == IN_MULTILINE_STATEMENT: |
|
161 | 166 | while tokens[-2].type in {tokenize.NL}: |
|
162 | 167 | tokens.pop(-2) |
|
163 | 168 | |
|
164 | 169 | |
|
165 | 170 | if tokens[-1].type == INCOMPLETE_STRING: |
|
166 | 171 | # Inside a multiline string |
|
167 | 172 | return 0 |
|
168 | 173 | |
|
169 | 174 | # Find the indents used before |
|
170 | 175 | prev_indents = [0] |
|
171 | 176 | def _add_indent(n): |
|
172 | 177 | if n != prev_indents[-1]: |
|
173 | 178 | prev_indents.append(n) |
|
174 | 179 | |
|
175 | 180 | tokiter = iter(tokens) |
|
176 | 181 | for tok in tokiter: |
|
177 | 182 | if tok.type in {tokenize.INDENT, tokenize.DEDENT}: |
|
178 | 183 | _add_indent(tok.end[1]) |
|
179 | 184 | elif (tok.type == tokenize.NL): |
|
180 | 185 | try: |
|
181 | 186 | _add_indent(next(tokiter).start[1]) |
|
182 | 187 | except StopIteration: |
|
183 | 188 | break |
|
184 | 189 | |
|
185 | 190 | last_indent = prev_indents.pop() |
|
186 | 191 | |
|
187 | 192 | # If we've just opened a multiline statement (e.g. 'a = ['), indent more |
|
188 | 193 | if tokens[-1].type == IN_MULTILINE_STATEMENT: |
|
189 | 194 | if tokens[-2].exact_type in {tokenize.LPAR, tokenize.LSQB, tokenize.LBRACE}: |
|
190 | 195 | return last_indent + 4 |
|
191 | 196 | return last_indent |
|
192 | 197 | |
|
193 | 198 | if tokens[-1].exact_type == tokenize.COLON: |
|
194 | 199 | # Line ends with colon - indent |
|
195 | 200 | return last_indent + 4 |
|
196 | 201 | |
|
197 | 202 | if last_indent: |
|
198 | 203 | # Examine the last line for dedent cues - statements like return or |
|
199 | 204 | # raise which normally end a block of code. |
|
200 | 205 | last_line_starts = 0 |
|
201 | 206 | for i, tok in enumerate(tokens): |
|
202 | 207 | if tok.type == tokenize.NEWLINE: |
|
203 | 208 | last_line_starts = i + 1 |
|
204 | 209 | |
|
205 | 210 | last_line_tokens = tokens[last_line_starts:] |
|
206 | 211 | names = [t.string for t in last_line_tokens if t.type == tokenize.NAME] |
|
207 | 212 | if names and names[0] in {'raise', 'return', 'pass', 'break', 'continue'}: |
|
208 | 213 | # Find the most recent indentation less than the current level |
|
209 | 214 | for indent in reversed(prev_indents): |
|
210 | 215 | if indent < last_indent: |
|
211 | 216 | return indent |
|
212 | 217 | |
|
213 | 218 | return last_indent |
|
214 | 219 | |
|
215 | 220 | |
|
216 | 221 | def last_blank(src): |
|
217 | 222 | """Determine if the input source ends in a blank. |
|
218 | 223 | |
|
219 | 224 | A blank is either a newline or a line consisting of whitespace. |
|
220 | 225 | |
|
221 | 226 | Parameters |
|
222 | 227 | ---------- |
|
223 | 228 | src : string |
|
224 | 229 | A single or multiline string. |
|
225 | 230 | """ |
|
226 | 231 | if not src: return False |
|
227 | 232 | ll = src.splitlines()[-1] |
|
228 | 233 | return (ll == '') or ll.isspace() |
|
229 | 234 | |
|
230 | 235 | |
|
231 | 236 | last_two_blanks_re = re.compile(r'\n\s*\n\s*$', re.MULTILINE) |
|
232 | 237 | last_two_blanks_re2 = re.compile(r'.+\n\s*\n\s+$', re.MULTILINE) |
|
233 | 238 | |
|
234 | 239 | def last_two_blanks(src): |
|
235 | 240 | """Determine if the input source ends in two blanks. |
|
236 | 241 | |
|
237 | 242 | A blank is either a newline or a line consisting of whitespace. |
|
238 | 243 | |
|
239 | 244 | Parameters |
|
240 | 245 | ---------- |
|
241 | 246 | src : string |
|
242 | 247 | A single or multiline string. |
|
243 | 248 | """ |
|
244 | 249 | if not src: return False |
|
245 | 250 | # The logic here is tricky: I couldn't get a regexp to work and pass all |
|
246 | 251 | # the tests, so I took a different approach: split the source by lines, |
|
247 | 252 | # grab the last two and prepend '###\n' as a stand-in for whatever was in |
|
248 | 253 | # the body before the last two lines. Then, with that structure, it's |
|
249 | 254 | # possible to analyze with two regexps. Not the most elegant solution, but |
|
250 | 255 | # it works. If anyone tries to change this logic, make sure to validate |
|
251 | 256 | # the whole test suite first! |
|
252 | 257 | new_src = '\n'.join(['###\n'] + src.splitlines()[-2:]) |
|
253 | 258 | return (bool(last_two_blanks_re.match(new_src)) or |
|
254 | 259 | bool(last_two_blanks_re2.match(new_src)) ) |
|
255 | 260 | |
|
256 | 261 | |
|
257 | 262 | def remove_comments(src): |
|
258 | 263 | """Remove all comments from input source. |
|
259 | 264 | |
|
260 | 265 | Note: comments are NOT recognized inside of strings! |
|
261 | 266 | |
|
262 | 267 | Parameters |
|
263 | 268 | ---------- |
|
264 | 269 | src : string |
|
265 | 270 | A single or multiline input string. |
|
266 | 271 | |
|
267 | 272 | Returns |
|
268 | 273 | ------- |
|
269 | 274 | String with all Python comments removed. |
|
270 | 275 | """ |
|
271 | 276 | |
|
272 | 277 | return re.sub('#.*', '', src) |
|
273 | 278 | |
|
274 | 279 | |
|
275 | 280 | def get_input_encoding(): |
|
276 | 281 | """Return the default standard input encoding. |
|
277 | 282 | |
|
278 | 283 | If sys.stdin has no encoding, 'ascii' is returned.""" |
|
279 | 284 | # There are strange environments for which sys.stdin.encoding is None. We |
|
280 | 285 | # ensure that a valid encoding is returned. |
|
281 | 286 | encoding = getattr(sys.stdin, 'encoding', None) |
|
282 | 287 | if encoding is None: |
|
283 | 288 | encoding = 'ascii' |
|
284 | 289 | return encoding |
|
285 | 290 | |
|
286 | 291 | #----------------------------------------------------------------------------- |
|
287 | 292 | # Classes and functions for normal Python syntax handling |
|
288 | 293 | #----------------------------------------------------------------------------- |
|
289 | 294 | |
|
290 | 295 | class InputSplitter(object): |
|
291 | 296 | r"""An object that can accumulate lines of Python source before execution. |
|
292 | 297 | |
|
293 | 298 | This object is designed to be fed python source line-by-line, using |
|
294 | 299 | :meth:`push`. It will return on each push whether the currently pushed |
|
295 | 300 | code could be executed already. In addition, it provides a method called |
|
296 | 301 | :meth:`push_accepts_more` that can be used to query whether more input |
|
297 | 302 | can be pushed into a single interactive block. |
|
298 | 303 | |
|
299 | 304 | This is a simple example of how an interactive terminal-based client can use |
|
300 | 305 | this tool:: |
|
301 | 306 | |
|
302 | 307 | isp = InputSplitter() |
|
303 | 308 | while isp.push_accepts_more(): |
|
304 | 309 | indent = ' '*isp.indent_spaces |
|
305 | 310 | prompt = '>>> ' + indent |
|
306 | 311 | line = indent + raw_input(prompt) |
|
307 | 312 | isp.push(line) |
|
308 | 313 | print 'Input source was:\n', isp.source_reset(), |
|
309 | 314 | """ |
|
310 | 315 | # A cache for storing the current indentation |
|
311 | 316 | # The first value stores the most recently processed source input |
|
312 | 317 | # The second value is the number of spaces for the current indentation |
|
313 | 318 | # If self.source matches the first value, the second value is a valid |
|
314 | 319 | # current indentation. Otherwise, the cache is invalid and the indentation |
|
315 | 320 | # must be recalculated. |
|
316 | 321 | _indent_spaces_cache = None, None |
|
317 | 322 | # String, indicating the default input encoding. It is computed by default |
|
318 | 323 | # at initialization time via get_input_encoding(), but it can be reset by a |
|
319 | 324 | # client with specific knowledge of the encoding. |
|
320 | 325 | encoding = '' |
|
321 | 326 | # String where the current full source input is stored, properly encoded. |
|
322 | 327 | # Reading this attribute is the normal way of querying the currently pushed |
|
323 | 328 | # source code, that has been properly encoded. |
|
324 | 329 | source = '' |
|
325 | 330 | # Code object corresponding to the current source. It is automatically |
|
326 | 331 | # synced to the source, so it can be queried at any time to obtain the code |
|
327 | 332 | # object; it will be None if the source doesn't compile to valid Python. |
|
328 | 333 | code = None |
|
329 | 334 | |
|
330 | 335 | # Private attributes |
|
331 | 336 | |
|
332 | 337 | # List with lines of input accumulated so far |
|
333 | 338 | _buffer: List[str] |
|
334 | 339 | # Command compiler |
|
335 | 340 | _compile: codeop.CommandCompiler |
|
336 | 341 | # Boolean indicating whether the current block is complete |
|
337 | 342 | _is_complete = None |
|
338 | 343 | # Boolean indicating whether the current block has an unrecoverable syntax error |
|
339 | 344 | _is_invalid = False |
|
340 | 345 | |
|
341 | 346 | def __init__(self) -> None: |
|
342 | 347 | """Create a new InputSplitter instance.""" |
|
343 | 348 | self._buffer = [] |
|
344 | 349 | self._compile = codeop.CommandCompiler() |
|
345 | 350 | self.encoding = get_input_encoding() |
|
346 | 351 | |
|
347 | 352 | def reset(self): |
|
348 | 353 | """Reset the input buffer and associated state.""" |
|
349 | 354 | self._buffer[:] = [] |
|
350 | 355 | self.source = '' |
|
351 | 356 | self.code = None |
|
352 | 357 | self._is_complete = False |
|
353 | 358 | self._is_invalid = False |
|
354 | 359 | |
|
355 | 360 | def source_reset(self): |
|
356 | 361 | """Return the input source and perform a full reset. |
|
357 | 362 | """ |
|
358 | 363 | out = self.source |
|
359 | 364 | self.reset() |
|
360 | 365 | return out |
|
361 | 366 | |
|
362 | 367 | def check_complete(self, source): |
|
363 | 368 | """Return whether a block of code is ready to execute, or should be continued |
|
364 | 369 | |
|
365 | 370 | This is a non-stateful API, and will reset the state of this InputSplitter. |
|
366 | 371 | |
|
367 | 372 | Parameters |
|
368 | 373 | ---------- |
|
369 | 374 | source : string |
|
370 | 375 | Python input code, which can be multiline. |
|
371 | 376 | |
|
372 | 377 | Returns |
|
373 | 378 | ------- |
|
374 | 379 | status : str |
|
375 | 380 | One of 'complete', 'incomplete', or 'invalid' if source is not a |
|
376 | 381 | prefix of valid code. |
|
377 | 382 | indent_spaces : int or None |
|
378 | 383 | The number of spaces by which to indent the next line of code. If |
|
379 | 384 | status is not 'incomplete', this is None. |
|
380 | 385 | """ |
|
381 | 386 | self.reset() |
|
382 | 387 | try: |
|
383 | 388 | self.push(source) |
|
384 | 389 | except SyntaxError: |
|
385 | 390 | # Transformers in IPythonInputSplitter can raise SyntaxError, |
|
386 | 391 | # which push() will not catch. |
|
387 | 392 | return 'invalid', None |
|
388 | 393 | else: |
|
389 | 394 | if self._is_invalid: |
|
390 | 395 | return 'invalid', None |
|
391 | 396 | elif self.push_accepts_more(): |
|
392 | 397 | return 'incomplete', self.get_indent_spaces() |
|
393 | 398 | else: |
|
394 | 399 | return 'complete', None |
|
395 | 400 | finally: |
|
396 | 401 | self.reset() |
|
397 | 402 | |
|
398 | 403 | def push(self, lines:str) -> bool: |
|
399 | 404 | """Push one or more lines of input. |
|
400 | 405 | |
|
401 | 406 | This stores the given lines and returns a status code indicating |
|
402 | 407 | whether the code forms a complete Python block or not. |
|
403 | 408 | |
|
404 | 409 | Any exceptions generated in compilation are swallowed, but if an |
|
405 | 410 | exception was produced, the method returns True. |
|
406 | 411 | |
|
407 | 412 | Parameters |
|
408 | 413 | ---------- |
|
409 | 414 | lines : string |
|
410 | 415 | One or more lines of Python input. |
|
411 | 416 | |
|
412 | 417 | Returns |
|
413 | 418 | ------- |
|
414 | 419 | is_complete : boolean |
|
415 | 420 | True if the current input source (the result of the current input |
|
416 | 421 | plus prior inputs) forms a complete Python execution block. Note that |
|
417 | 422 | this value is also stored as a private attribute (``_is_complete``), so it |
|
418 | 423 | can be queried at any time. |
|
419 | 424 | """ |
|
420 | 425 | assert isinstance(lines, str) |
|
421 | 426 | self._store(lines) |
|
422 | 427 | source = self.source |
|
423 | 428 | |
|
424 | 429 | # Before calling _compile(), reset the code object to None so that if an |
|
425 | 430 | # exception is raised in compilation, we don't mislead by having |
|
426 | 431 | # inconsistent code/source attributes. |
|
427 | 432 | self.code, self._is_complete = None, None |
|
428 | 433 | self._is_invalid = False |
|
429 | 434 | |
|
430 | 435 | # Honor termination lines properly |
|
431 | 436 | if source.endswith('\\\n'): |
|
432 | 437 | return False |
|
433 | 438 | |
|
434 | 439 | try: |
|
435 | 440 | with warnings.catch_warnings(): |
|
436 | 441 | warnings.simplefilter('error', SyntaxWarning) |
|
437 | 442 | self.code = self._compile(source, symbol="exec") |
|
438 | 443 | # Invalid syntax can produce any of a number of different errors from |
|
439 | 444 | # inside the compiler, so we have to catch them all. Syntax errors |
|
440 | 445 | # immediately produce a 'ready' block, so the invalid Python can be |
|
441 | 446 | # sent to the kernel for evaluation with possible ipython |
|
442 | 447 | # special-syntax conversion. |
|
443 | 448 | except (SyntaxError, OverflowError, ValueError, TypeError, |
|
444 | 449 | MemoryError, SyntaxWarning): |
|
445 | 450 | self._is_complete = True |
|
446 | 451 | self._is_invalid = True |
|
447 | 452 | else: |
|
448 | 453 | # Compilation didn't produce any exceptions (though it may not have |
|
449 | 454 | # given a complete code object) |
|
450 | 455 | self._is_complete = self.code is not None |
|
451 | 456 | |
|
452 | 457 | return self._is_complete |
|
453 | 458 | |
|
454 | 459 | def push_accepts_more(self): |
|
455 | 460 | """Return whether a block of interactive input can accept more input. |
|
456 | 461 | |
|
457 | 462 | This method is meant to be used by line-oriented frontends, who need to |
|
458 | 463 | guess whether a block is complete or not based solely on prior and |
|
459 | 464 | current input lines. The InputSplitter considers it has a complete |
|
460 | 465 | interactive block and will not accept more input when either: |
|
461 | 466 | |
|
462 | 467 | * A SyntaxError is raised |
|
463 | 468 | |
|
464 | 469 | * The code is complete and consists of a single line or a single |
|
465 | 470 | non-compound statement |
|
466 | 471 | |
|
467 | 472 | * The code is complete and has a blank line at the end |
|
468 | 473 | |
|
469 | 474 | If the current input produces a syntax error, this method immediately |
|
470 | 475 | returns False but does *not* raise the syntax error exception, as |
|
471 | 476 | typically clients will want to send invalid syntax to an execution |
|
472 | 477 | backend which might convert the invalid syntax into valid Python via |
|
473 | 478 | one of the dynamic IPython mechanisms. |
|
474 | 479 | """ |
|
475 | 480 | |
|
476 | 481 | # With incomplete input, unconditionally accept more |
|
477 | 482 | # A syntax error also sets _is_complete to True - see push() |
|
478 | 483 | if not self._is_complete: |
|
479 | 484 | #print("Not complete") # debug |
|
480 | 485 | return True |
|
481 | 486 | |
|
482 | 487 | # The user can make any (complete) input execute by leaving a blank line |
|
483 | 488 | last_line = self.source.splitlines()[-1] |
|
484 | 489 | if (not last_line) or last_line.isspace(): |
|
485 | 490 | #print("Blank line") # debug |
|
486 | 491 | return False |
|
487 | 492 | |
|
488 | 493 | # If there's just a single line or AST node, and we're flush left, as is |
|
489 | 494 | # the case after a simple statement such as 'a=1', we want to execute it |
|
490 | 495 | # straight away. |
|
491 | 496 | if self.get_indent_spaces() == 0: |
|
492 | 497 | if len(self.source.splitlines()) <= 1: |
|
493 | 498 | return False |
|
494 | 499 | |
|
495 | 500 | try: |
|
496 | 501 | code_ast = ast.parse("".join(self._buffer)) |
|
497 | 502 | except Exception: |
|
498 | 503 | #print("Can't parse AST") # debug |
|
499 | 504 | return False |
|
500 | 505 | else: |
|
501 | 506 | if len(code_ast.body) == 1 and \ |
|
502 | 507 | not hasattr(code_ast.body[0], 'body'): |
|
503 | 508 | #print("Simple statement") # debug |
|
504 | 509 | return False |
|
505 | 510 | |
|
506 | 511 | # General fallback - accept more code |
|
507 | 512 | return True |
|
508 | 513 | |
|
509 | 514 | def get_indent_spaces(self): |
|
510 | 515 | sourcefor, n = self._indent_spaces_cache |
|
511 | 516 | if sourcefor == self.source: |
|
512 | 517 | return n |
|
513 | 518 | |
|
514 | 519 | # self.source always has a trailing newline |
|
515 | 520 | n = find_next_indent(self.source[:-1]) |
|
516 | 521 | self._indent_spaces_cache = (self.source, n) |
|
517 | 522 | return n |
|
518 | 523 | |
|
519 | 524 | # Backwards compatibility. I think all code that used .indent_spaces was |
|
520 | 525 | # inside IPython, but we can leave this here until IPython 7 in case any |
|
521 | 526 | # other modules are using it. -TK, November 2017 |
|
522 | 527 | indent_spaces = property(get_indent_spaces) |
|
523 | 528 | |
|
524 | 529 | def _store(self, lines, buffer=None, store='source'): |
|
525 | 530 | """Store one or more lines of input. |
|
526 | 531 | |
|
527 | 532 | If input lines are not newline-terminated, a newline is automatically |
|
528 | 533 | appended.""" |
|
529 | 534 | |
|
530 | 535 | if buffer is None: |
|
531 | 536 | buffer = self._buffer |
|
532 | 537 | |
|
533 | 538 | if lines.endswith('\n'): |
|
534 | 539 | buffer.append(lines) |
|
535 | 540 | else: |
|
536 | 541 | buffer.append(lines+'\n') |
|
537 | 542 | setattr(self, store, self._set_source(buffer)) |
|
538 | 543 | |
|
539 | 544 | def _set_source(self, buffer): |
|
540 | 545 | return u''.join(buffer) |
|
541 | 546 | |
|
542 | 547 | |
|
543 | 548 | class IPythonInputSplitter(InputSplitter): |
|
544 | 549 | """An input splitter that recognizes all of IPython's special syntax.""" |
|
545 | 550 | |
|
546 | 551 | # String with raw, untransformed input. |
|
547 | 552 | source_raw = '' |
|
548 | 553 | |
|
549 | 554 | # Flag to track when a transformer has stored input that it hasn't given |
|
550 | 555 | # back yet. |
|
551 | 556 | transformer_accumulating = False |
|
552 | 557 | |
|
553 | 558 | # Flag to track when assemble_python_lines has stored input that it hasn't |
|
554 | 559 | # given back yet. |
|
555 | 560 | within_python_line = False |
|
556 | 561 | |
|
557 | 562 | # Private attributes |
|
558 | 563 | |
|
559 | 564 | # List with lines of raw input accumulated so far. |
|
560 | 565 | _buffer_raw = None |
|
561 | 566 | |
|
562 | 567 | def __init__(self, line_input_checker=True, physical_line_transforms=None, |
|
563 | 568 | logical_line_transforms=None, python_line_transforms=None): |
|
564 | 569 | super(IPythonInputSplitter, self).__init__() |
|
565 | 570 | self._buffer_raw = [] |
|
566 | 571 | self._validate = True |
|
567 | 572 | |
|
568 | 573 | if physical_line_transforms is not None: |
|
569 | 574 | self.physical_line_transforms = physical_line_transforms |
|
570 | 575 | else: |
|
571 | 576 | self.physical_line_transforms = [ |
|
572 | 577 | leading_indent(), |
|
573 | 578 | classic_prompt(), |
|
574 | 579 | ipy_prompt(), |
|
575 | 580 | cellmagic(end_on_blank_line=line_input_checker), |
|
576 | 581 | ] |
|
577 | 582 | |
|
578 | 583 | self.assemble_logical_lines = assemble_logical_lines() |
|
579 | 584 | if logical_line_transforms is not None: |
|
580 | 585 | self.logical_line_transforms = logical_line_transforms |
|
581 | 586 | else: |
|
582 | 587 | self.logical_line_transforms = [ |
|
583 | 588 | help_end(), |
|
584 | 589 | escaped_commands(), |
|
585 | 590 | assign_from_magic(), |
|
586 | 591 | assign_from_system(), |
|
587 | 592 | ] |
|
588 | 593 | |
|
589 | 594 | self.assemble_python_lines = assemble_python_lines() |
|
590 | 595 | if python_line_transforms is not None: |
|
591 | 596 | self.python_line_transforms = python_line_transforms |
|
592 | 597 | else: |
|
593 | 598 | # We don't use any of these at present |
|
594 | 599 | self.python_line_transforms = [] |
|
595 | 600 | |
|
596 | 601 | @property |
|
597 | 602 | def transforms(self): |
|
598 | 603 | "Quick access to all transformers." |
|
599 | 604 | return self.physical_line_transforms + \ |
|
600 | 605 | [self.assemble_logical_lines] + self.logical_line_transforms + \ |
|
601 | 606 | [self.assemble_python_lines] + self.python_line_transforms |
|
602 | 607 | |
|
603 | 608 | @property |
|
604 | 609 | def transforms_in_use(self): |
|
605 | 610 | """Transformers, excluding logical line transformers if we're in a |
|
606 | 611 | Python line.""" |
|
607 | 612 | t = self.physical_line_transforms[:] |
|
608 | 613 | if not self.within_python_line: |
|
609 | 614 | t += [self.assemble_logical_lines] + self.logical_line_transforms |
|
610 | 615 | return t + [self.assemble_python_lines] + self.python_line_transforms |
|
611 | 616 | |
|
612 | 617 | def reset(self): |
|
613 | 618 | """Reset the input buffer and associated state.""" |
|
614 | 619 | super(IPythonInputSplitter, self).reset() |
|
615 | 620 | self._buffer_raw[:] = [] |
|
616 | 621 | self.source_raw = '' |
|
617 | 622 | self.transformer_accumulating = False |
|
618 | 623 | self.within_python_line = False |
|
619 | 624 | |
|
620 | 625 | for t in self.transforms: |
|
621 | 626 | try: |
|
622 | 627 | t.reset() |
|
623 | 628 | except SyntaxError: |
|
624 | 629 | # Nothing that calls reset() expects to handle transformer |
|
625 | 630 | # errors |
|
626 | 631 | pass |
|
627 | 632 | |
|
628 | 633 | def flush_transformers(self): |
|
629 | 634 | def _flush(transform, outs): |
|
630 | 635 | """yield transformed lines |
|
631 | 636 | |
|
632 | 637 | always strings, never None |
|
633 | 638 | |
|
634 | 639 | transform: the current transform |
|
635 | 640 | outs: an iterable of previously transformed inputs. |
|
636 | 641 | Each may be multiline, which will be passed |
|
637 | 642 | one line at a time to transform. |
|
638 | 643 | """ |
|
639 | 644 | for out in outs: |
|
640 | 645 | for line in out.splitlines(): |
|
641 | 646 | # push one line at a time |
|
642 | 647 | tmp = transform.push(line) |
|
643 | 648 | if tmp is not None: |
|
644 | 649 | yield tmp |
|
645 | 650 | |
|
646 | 651 | # reset the transform |
|
647 | 652 | tmp = transform.reset() |
|
648 | 653 | if tmp is not None: |
|
649 | 654 | yield tmp |
|
650 | 655 | |
|
651 | 656 | out = [] |
|
652 | 657 | for t in self.transforms_in_use: |
|
653 | 658 | out = _flush(t, out) |
|
654 | 659 | |
|
655 | 660 | out = list(out) |
|
656 | 661 | if out: |
|
657 | 662 | self._store('\n'.join(out)) |
|
658 | 663 | |
|
659 | 664 | def raw_reset(self): |
|
660 | 665 | """Return raw input only and perform a full reset. |
|
661 | 666 | """ |
|
662 | 667 | out = self.source_raw |
|
663 | 668 | self.reset() |
|
664 | 669 | return out |
|
665 | 670 | |
|
666 | 671 | def source_reset(self): |
|
667 | 672 | try: |
|
668 | 673 | self.flush_transformers() |
|
669 | 674 | return self.source |
|
670 | 675 | finally: |
|
671 | 676 | self.reset() |
|
672 | 677 | |
|
673 | 678 | def push_accepts_more(self): |
|
674 | 679 | if self.transformer_accumulating: |
|
675 | 680 | return True |
|
676 | 681 | else: |
|
677 | 682 | return super(IPythonInputSplitter, self).push_accepts_more() |
|
678 | 683 | |
|
679 | 684 | def transform_cell(self, cell): |
|
680 | 685 | """Process and translate a cell of input. |
|
681 | 686 | """ |
|
682 | 687 | self.reset() |
|
683 | 688 | try: |
|
684 | 689 | self.push(cell) |
|
685 | 690 | self.flush_transformers() |
|
686 | 691 | return self.source |
|
687 | 692 | finally: |
|
688 | 693 | self.reset() |
|
689 | 694 | |
|
690 | 695 | def push(self, lines:str) -> bool: |
|
691 | 696 | """Push one or more lines of IPython input. |
|
692 | 697 | |
|
693 | 698 | This stores the given lines and returns a status code indicating |
|
694 | 699 | whether the code forms a complete Python block or not, after processing |
|
695 | 700 | all input lines for special IPython syntax. |
|
696 | 701 | |
|
697 | 702 | Any exceptions generated in compilation are swallowed, but if an |
|
698 | 703 | exception was produced, the method returns True. |
|
699 | 704 | |
|
700 | 705 | Parameters |
|
701 | 706 | ---------- |
|
702 | 707 | lines : string |
|
703 | 708 | One or more lines of Python input. |
|
704 | 709 | |
|
705 | 710 | Returns |
|
706 | 711 | ------- |
|
707 | 712 | is_complete : boolean |
|
708 | 713 | True if the current input source (the result of the current input |
|
709 | 714 | plus prior inputs) forms a complete Python execution block. Note that |
|
710 | 715 | this value is also stored as a private attribute (_is_complete), so it |
|
711 | 716 | can be queried at any time. |
|
712 | 717 | """ |
|
713 | 718 | assert isinstance(lines, str) |
|
714 | 719 | # We must ensure all input is pure unicode |
|
715 | 720 | # ''.splitlines() --> [], but we need to push the empty line to transformers |
|
716 | 721 | lines_list = lines.splitlines() |
|
717 | 722 | if not lines_list: |
|
718 | 723 | lines_list = [''] |
|
719 | 724 | |
|
720 | 725 | # Store raw source before applying any transformations to it. Note |
|
721 | 726 | # that this must be done *after* the reset() call that would otherwise |
|
722 | 727 | # flush the buffer. |
|
723 | 728 | self._store(lines, self._buffer_raw, 'source_raw') |
|
724 | 729 | |
|
725 | 730 | transformed_lines_list = [] |
|
726 | 731 | for line in lines_list: |
|
727 | 732 | transformed = self._transform_line(line) |
|
728 | 733 | if transformed is not None: |
|
729 | 734 | transformed_lines_list.append(transformed) |
|
730 | 735 | |
|
731 | 736 | if transformed_lines_list: |
|
732 | 737 | transformed_lines = '\n'.join(transformed_lines_list) |
|
733 | 738 | return super(IPythonInputSplitter, self).push(transformed_lines) |
|
734 | 739 | else: |
|
735 | 740 | # Got nothing back from transformers - they must be waiting for |
|
736 | 741 | # more input. |
|
737 | 742 | return False |
|
738 | 743 | |
|
739 | 744 | def _transform_line(self, line): |
|
740 | 745 | """Push a line of input code through the various transformers. |
|
741 | 746 | |
|
742 | 747 | Returns any output from the transformers, or None if a transformer |
|
743 | 748 | is accumulating lines. |
|
744 | 749 | |
|
745 | 750 | Sets self.transformer_accumulating as a side effect. |
|
746 | 751 | """ |
|
747 | 752 | def _accumulating(dbg): |
|
748 | 753 | #print(dbg) |
|
749 | 754 | self.transformer_accumulating = True |
|
750 | 755 | return None |
|
751 | 756 | |
|
752 | 757 | for transformer in self.physical_line_transforms: |
|
753 | 758 | line = transformer.push(line) |
|
754 | 759 | if line is None: |
|
755 | 760 | return _accumulating(transformer) |
|
756 | 761 | |
|
757 | 762 | if not self.within_python_line: |
|
758 | 763 | line = self.assemble_logical_lines.push(line) |
|
759 | 764 | if line is None: |
|
760 | 765 | return _accumulating('acc logical line') |
|
761 | 766 | |
|
762 | 767 | for transformer in self.logical_line_transforms: |
|
763 | 768 | line = transformer.push(line) |
|
764 | 769 | if line is None: |
|
765 | 770 | return _accumulating(transformer) |
|
766 | 771 | |
|
767 | 772 | line = self.assemble_python_lines.push(line) |
|
768 | 773 | if line is None: |
|
769 | 774 | self.within_python_line = True |
|
770 | 775 | return _accumulating('acc python line') |
|
771 | 776 | else: |
|
772 | 777 | self.within_python_line = False |
|
773 | 778 | |
|
774 | 779 | for transformer in self.python_line_transforms: |
|
775 | 780 | line = transformer.push(line) |
|
776 | 781 | if line is None: |
|
777 | 782 | return _accumulating(transformer) |
|
778 | 783 | |
|
779 | 784 | #print("transformers clear") #debug |
|
780 | 785 | self.transformer_accumulating = False |
|
781 | 786 | return line |
|
782 | 787 |
@@ -1,823 +1,827 | |||
|
1 | 1 | """Input transformer machinery to support IPython special syntax. |
|
2 | 2 | |
|
3 | 3 | This includes the machinery to recognise and transform ``%magic`` commands, |
|
4 | 4 | ``!system`` commands, ``help?`` querying, prompt stripping, and so forth. |
|
5 | 5 | |
|
6 | 6 | Added: IPython 7.0. Replaces inputsplitter and inputtransformer which were |
|
7 | 7 | deprecated in 7.0. |
|
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 ast |
|
14 | 14 | from codeop import CommandCompiler, Compile |
|
15 | 15 | import re |
|
16 | 16 | import sys |
|
17 | 17 | import tokenize |
|
18 | 18 | from typing import List, Tuple, Optional, Any |
|
19 | 19 | import warnings |
|
20 | 20 | |
|
21 | 21 | from IPython.utils import tokenutil |
|
22 | 22 | |
|
23 | 23 | _indent_re = re.compile(r'^[ \t]+') |
|
24 | 24 | |
|
25 | 25 | def leading_empty_lines(lines): |
|
26 | 26 | """Remove leading empty lines |
|
27 | 27 | |
|
28 | 28 | If the leading lines are empty or contain only whitespace, they will be |
|
29 | 29 | removed. |
|
30 | 30 | """ |
|
31 | 31 | if not lines: |
|
32 | 32 | return lines |
|
33 | 33 | for i, line in enumerate(lines): |
|
34 | 34 | if line and not line.isspace(): |
|
35 | 35 | return lines[i:] |
|
36 | 36 | return lines |
|
37 | 37 | |
|
38 | 38 | def leading_indent(lines): |
|
39 | 39 | """Remove leading indentation. |
|
40 | 40 | |
|
41 | 41 | If the first line starts with a spaces or tabs, the same whitespace will be |
|
42 | 42 | removed from each following line in the cell. |
|
43 | 43 | """ |
|
44 | 44 | if not lines: |
|
45 | 45 | return lines |
|
46 | 46 | m = _indent_re.match(lines[0]) |
|
47 | 47 | if not m: |
|
48 | 48 | return lines |
|
49 | 49 | space = m.group(0) |
|
50 | 50 | n = len(space) |
|
51 | 51 | return [l[n:] if l.startswith(space) else l |
|
52 | 52 | for l in lines] |
|
53 | 53 | |
|
54 | 54 | class PromptStripper: |
|
55 | 55 | """Remove matching input prompts from a block of input. |
|
56 | 56 | |
|
57 | 57 | Parameters |
|
58 | 58 | ---------- |
|
59 | 59 | prompt_re : regular expression |
|
60 | 60 | A regular expression matching any input prompt (including continuation, |
|
61 | 61 | e.g. ``...``) |
|
62 | 62 | initial_re : regular expression, optional |
|
63 | 63 | A regular expression matching only the initial prompt, but not continuation. |
|
64 | 64 | If no initial expression is given, prompt_re will be used everywhere. |
|
65 | 65 | Used mainly for plain Python prompts (``>>>``), where the continuation prompt |
|
66 | 66 | ``...`` is a valid Python expression in Python 3, so shouldn't be stripped. |
|
67 | 67 | |
|
68 | 68 | Notes |
|
69 | 69 | ----- |
|
70 | 70 | |
|
71 | 71 | If initial_re and prompt_re differ, |
|
72 | 72 | only initial_re will be tested against the first line. |
|
73 | 73 | If any prompt is found on the first two lines, |
|
74 | 74 | prompts will be stripped from the rest of the block. |
|
75 | 75 | """ |
|
76 | 76 | def __init__(self, prompt_re, initial_re=None): |
|
77 | 77 | self.prompt_re = prompt_re |
|
78 | 78 | self.initial_re = initial_re or prompt_re |
|
79 | 79 | |
|
80 | 80 | def _strip(self, lines): |
|
81 | 81 | return [self.prompt_re.sub('', l, count=1) for l in lines] |
|
82 | 82 | |
|
83 | 83 | def __call__(self, lines): |
|
84 | 84 | if not lines: |
|
85 | 85 | return lines |
|
86 | 86 | if self.initial_re.match(lines[0]) or \ |
|
87 | 87 | (len(lines) > 1 and self.prompt_re.match(lines[1])): |
|
88 | 88 | return self._strip(lines) |
|
89 | 89 | return lines |
|
90 | 90 | |
|
91 | 91 | classic_prompt = PromptStripper( |
|
92 | 92 | prompt_re=re.compile(r'^(>>>|\.\.\.)( |$)'), |
|
93 | 93 | initial_re=re.compile(r'^>>>( |$)') |
|
94 | 94 | ) |
|
95 | 95 | |
|
96 | 96 | ipython_prompt = PromptStripper( |
|
97 | 97 | re.compile( |
|
98 | 98 | r""" |
|
99 | 99 | ^( # Match from the beginning of a line, either: |
|
100 | 100 | |
|
101 | 101 | # 1. First-line prompt: |
|
102 | 102 | ((\[nav\]|\[ins\])?\ )? # Vi editing mode prompt, if it's there |
|
103 | 103 | In\ # The 'In' of the prompt, with a space |
|
104 | 104 | \[\d+\]: # Command index, as displayed in the prompt |
|
105 | 105 | \ # With a mandatory trailing space |
|
106 | 106 | |
|
107 | 107 | | # ... or ... |
|
108 | 108 | |
|
109 | 109 | # 2. The three dots of the multiline prompt |
|
110 | 110 | \s* # All leading whitespace characters |
|
111 | 111 | \.{3,}: # The three (or more) dots |
|
112 | 112 | \ ? # With an optional trailing space |
|
113 | 113 | |
|
114 | 114 | ) |
|
115 | 115 | """, |
|
116 | 116 | re.VERBOSE, |
|
117 | 117 | ) |
|
118 | 118 | ) |
|
119 | 119 | |
|
120 | 120 | |
|
121 | 121 | def cell_magic(lines): |
|
122 | 122 | if not lines or not lines[0].startswith('%%'): |
|
123 | 123 | return lines |
|
124 | 124 | if re.match(r'%%\w+\?', lines[0]): |
|
125 | 125 | # This case will be handled by help_end |
|
126 | 126 | return lines |
|
127 | 127 | magic_name, _, first_line = lines[0][2:].rstrip().partition(' ') |
|
128 | 128 | body = ''.join(lines[1:]) |
|
129 | 129 | return ['get_ipython().run_cell_magic(%r, %r, %r)\n' |
|
130 | 130 | % (magic_name, first_line, body)] |
|
131 | 131 | |
|
132 | 132 | |
|
133 | 133 | def _find_assign_op(token_line) -> Optional[int]: |
|
134 | 134 | """Get the index of the first assignment in the line ('=' not inside brackets) |
|
135 | 135 | |
|
136 | 136 | Note: We don't try to support multiple special assignment (a = b = %foo) |
|
137 | 137 | """ |
|
138 | 138 | paren_level = 0 |
|
139 | 139 | for i, ti in enumerate(token_line): |
|
140 | 140 | s = ti.string |
|
141 | 141 | if s == '=' and paren_level == 0: |
|
142 | 142 | return i |
|
143 | 143 | if s in {'(','[','{'}: |
|
144 | 144 | paren_level += 1 |
|
145 | 145 | elif s in {')', ']', '}'}: |
|
146 | 146 | if paren_level > 0: |
|
147 | 147 | paren_level -= 1 |
|
148 | 148 | return None |
|
149 | 149 | |
|
150 | 150 | def find_end_of_continued_line(lines, start_line: int): |
|
151 | 151 | """Find the last line of a line explicitly extended using backslashes. |
|
152 | 152 | |
|
153 | 153 | Uses 0-indexed line numbers. |
|
154 | 154 | """ |
|
155 | 155 | end_line = start_line |
|
156 | 156 | while lines[end_line].endswith('\\\n'): |
|
157 | 157 | end_line += 1 |
|
158 | 158 | if end_line >= len(lines): |
|
159 | 159 | break |
|
160 | 160 | return end_line |
|
161 | 161 | |
|
162 | 162 | def assemble_continued_line(lines, start: Tuple[int, int], end_line: int): |
|
163 | 163 | r"""Assemble a single line from multiple continued line pieces |
|
164 | 164 | |
|
165 | 165 | Continued lines are lines ending in ``\``, and the line following the last |
|
166 | 166 | ``\`` in the block. |
|
167 | 167 | |
|
168 | 168 | For example, this code continues over multiple lines:: |
|
169 | 169 | |
|
170 | 170 | if (assign_ix is not None) \ |
|
171 | 171 | and (len(line) >= assign_ix + 2) \ |
|
172 | 172 | and (line[assign_ix+1].string == '%') \ |
|
173 | 173 | and (line[assign_ix+2].type == tokenize.NAME): |
|
174 | 174 | |
|
175 | 175 | This statement contains four continued line pieces. |
|
176 | 176 | Assembling these pieces into a single line would give:: |
|
177 | 177 | |
|
178 | 178 | if (assign_ix is not None) and (len(line) >= assign_ix + 2) and (line[... |
|
179 | 179 | |
|
180 | 180 | This uses 0-indexed line numbers. *start* is (lineno, colno). |
|
181 | 181 | |
|
182 | 182 | Used to allow ``%magic`` and ``!system`` commands to be continued over |
|
183 | 183 | multiple lines. |
|
184 | 184 | """ |
|
185 | 185 | parts = [lines[start[0]][start[1]:]] + lines[start[0]+1:end_line+1] |
|
186 | 186 | return ' '.join([p.rstrip()[:-1] for p in parts[:-1]] # Strip backslash+newline |
|
187 | 187 | + [parts[-1].rstrip()]) # Strip newline from last line |
|
188 | 188 | |
|
189 | 189 | class TokenTransformBase: |
|
190 | 190 | """Base class for transformations which examine tokens. |
|
191 | 191 | |
|
192 | 192 | Special syntax should not be transformed when it occurs inside strings or |
|
193 | 193 | comments. This is hard to reliably avoid with regexes. The solution is to |
|
194 | 194 | tokenise the code as Python, and recognise the special syntax in the tokens. |
|
195 | 195 | |
|
196 | 196 | IPython's special syntax is not valid Python syntax, so tokenising may go |
|
197 | 197 | wrong after the special syntax starts. These classes therefore find and |
|
198 | 198 | transform *one* instance of special syntax at a time into regular Python |
|
199 | 199 | syntax. After each transformation, tokens are regenerated to find the next |
|
200 | 200 | piece of special syntax. |
|
201 | 201 | |
|
202 | 202 | Subclasses need to implement one class method (find) |
|
203 | 203 | and one regular method (transform). |
|
204 | 204 | |
|
205 | 205 | The priority attribute can select which transformation to apply if multiple |
|
206 | 206 | transformers match in the same place. Lower numbers have higher priority. |
|
207 | 207 | This allows "%magic?" to be turned into a help call rather than a magic call. |
|
208 | 208 | """ |
|
209 | 209 | # Lower numbers -> higher priority (for matches in the same location) |
|
210 | 210 | priority = 10 |
|
211 | 211 | |
|
212 | 212 | def sortby(self): |
|
213 | 213 | return self.start_line, self.start_col, self.priority |
|
214 | 214 | |
|
215 | 215 | def __init__(self, start): |
|
216 | 216 | self.start_line = start[0] - 1 # Shift from 1-index to 0-index |
|
217 | 217 | self.start_col = start[1] |
|
218 | 218 | |
|
219 | 219 | @classmethod |
|
220 | 220 | def find(cls, tokens_by_line): |
|
221 | 221 | """Find one instance of special syntax in the provided tokens. |
|
222 | 222 | |
|
223 | 223 | Tokens are grouped into logical lines for convenience, |
|
224 | 224 | so it is easy to e.g. look at the first token of each line. |
|
225 | 225 | *tokens_by_line* is a list of lists of tokenize.TokenInfo objects. |
|
226 | 226 | |
|
227 | 227 | This should return an instance of its class, pointing to the start |
|
228 | 228 | position it has found, or None if it found no match. |
|
229 | 229 | """ |
|
230 | 230 | raise NotImplementedError |
|
231 | 231 | |
|
232 | 232 | def transform(self, lines: List[str]): |
|
233 | 233 | """Transform one instance of special syntax found by ``find()`` |
|
234 | 234 | |
|
235 | 235 | Takes a list of strings representing physical lines, |
|
236 | 236 | returns a similar list of transformed lines. |
|
237 | 237 | """ |
|
238 | 238 | raise NotImplementedError |
|
239 | 239 | |
|
240 | 240 | class MagicAssign(TokenTransformBase): |
|
241 | 241 | """Transformer for assignments from magics (a = %foo)""" |
|
242 | 242 | @classmethod |
|
243 | 243 | def find(cls, tokens_by_line): |
|
244 | 244 | """Find the first magic assignment (a = %foo) in the cell. |
|
245 | 245 | """ |
|
246 | 246 | for line in tokens_by_line: |
|
247 | 247 | assign_ix = _find_assign_op(line) |
|
248 | 248 | if (assign_ix is not None) \ |
|
249 | 249 | and (len(line) >= assign_ix + 2) \ |
|
250 | 250 | and (line[assign_ix+1].string == '%') \ |
|
251 | 251 | and (line[assign_ix+2].type == tokenize.NAME): |
|
252 | 252 | return cls(line[assign_ix+1].start) |
|
253 | 253 | |
|
254 | 254 | def transform(self, lines: List[str]): |
|
255 | 255 | """Transform a magic assignment found by the ``find()`` classmethod. |
|
256 | 256 | """ |
|
257 | 257 | start_line, start_col = self.start_line, self.start_col |
|
258 | 258 | lhs = lines[start_line][:start_col] |
|
259 | 259 | end_line = find_end_of_continued_line(lines, start_line) |
|
260 | 260 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
261 | 261 | assert rhs.startswith('%'), rhs |
|
262 | 262 | magic_name, _, args = rhs[1:].partition(' ') |
|
263 | 263 | |
|
264 | 264 | lines_before = lines[:start_line] |
|
265 | 265 | call = "get_ipython().run_line_magic({!r}, {!r})".format(magic_name, args) |
|
266 | 266 | new_line = lhs + call + '\n' |
|
267 | 267 | lines_after = lines[end_line+1:] |
|
268 | 268 | |
|
269 | 269 | return lines_before + [new_line] + lines_after |
|
270 | 270 | |
|
271 | 271 | |
|
272 | 272 | class SystemAssign(TokenTransformBase): |
|
273 | 273 | """Transformer for assignments from system commands (a = !foo)""" |
|
274 | 274 | @classmethod |
|
275 | 275 | def find_pre_312(cls, tokens_by_line): |
|
276 | 276 | for line in tokens_by_line: |
|
277 | 277 | assign_ix = _find_assign_op(line) |
|
278 | 278 | if (assign_ix is not None) \ |
|
279 | 279 | and not line[assign_ix].line.strip().startswith('=') \ |
|
280 | 280 | and (len(line) >= assign_ix + 2) \ |
|
281 | 281 | and (line[assign_ix + 1].type == tokenize.ERRORTOKEN): |
|
282 | 282 | ix = assign_ix + 1 |
|
283 | 283 | |
|
284 | 284 | while ix < len(line) and line[ix].type == tokenize.ERRORTOKEN: |
|
285 | 285 | if line[ix].string == '!': |
|
286 | 286 | return cls(line[ix].start) |
|
287 | 287 | elif not line[ix].string.isspace(): |
|
288 | 288 | break |
|
289 | 289 | ix += 1 |
|
290 | 290 | |
|
291 | 291 | @classmethod |
|
292 | 292 | def find_post_312(cls, tokens_by_line): |
|
293 | 293 | for line in tokens_by_line: |
|
294 | 294 | assign_ix = _find_assign_op(line) |
|
295 | if (assign_ix is not None) \ | |
|
296 | and not line[assign_ix].line.strip().startswith('=') \ | |
|
297 | and (len(line) >= assign_ix + 2) \ | |
|
298 |
|
|
|
299 |
|
|
|
295 | if ( | |
|
296 | (assign_ix is not None) | |
|
297 | and not line[assign_ix].line.strip().startswith("=") | |
|
298 | and (len(line) >= assign_ix + 2) | |
|
299 | and (line[assign_ix + 1].type == tokenize.OP) | |
|
300 | and (line[assign_ix + 1].string == "!") | |
|
301 | ): | |
|
300 | 302 | return cls(line[assign_ix + 1].start) |
|
301 | 303 | |
|
302 | 304 | @classmethod |
|
303 | 305 | def find(cls, tokens_by_line): |
|
304 | """Find the first system assignment (a = !foo) in the cell. | |
|
305 | """ | |
|
306 | """Find the first system assignment (a = !foo) in the cell.""" | |
|
306 | 307 | if sys.version_info < (3, 12): |
|
307 | 308 | return cls.find_pre_312(tokens_by_line) |
|
308 | 309 | return cls.find_post_312(tokens_by_line) |
|
309 | 310 | |
|
310 | 311 | def transform(self, lines: List[str]): |
|
311 | 312 | """Transform a system assignment found by the ``find()`` classmethod. |
|
312 | 313 | """ |
|
313 | 314 | start_line, start_col = self.start_line, self.start_col |
|
314 | 315 | |
|
315 | 316 | lhs = lines[start_line][:start_col] |
|
316 | 317 | end_line = find_end_of_continued_line(lines, start_line) |
|
317 | 318 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
318 | 319 | assert rhs.startswith('!'), rhs |
|
319 | 320 | cmd = rhs[1:] |
|
320 | 321 | |
|
321 | 322 | lines_before = lines[:start_line] |
|
322 | 323 | call = "get_ipython().getoutput({!r})".format(cmd) |
|
323 | 324 | new_line = lhs + call + '\n' |
|
324 | 325 | lines_after = lines[end_line + 1:] |
|
325 | 326 | |
|
326 | 327 | return lines_before + [new_line] + lines_after |
|
327 | 328 | |
|
328 | 329 | # The escape sequences that define the syntax transformations IPython will |
|
329 | 330 | # apply to user input. These can NOT be just changed here: many regular |
|
330 | 331 | # expressions and other parts of the code may use their hardcoded values, and |
|
331 | 332 | # for all intents and purposes they constitute the 'IPython syntax', so they |
|
332 | 333 | # should be considered fixed. |
|
333 | 334 | |
|
334 | 335 | ESC_SHELL = '!' # Send line to underlying system shell |
|
335 | 336 | ESC_SH_CAP = '!!' # Send line to system shell and capture output |
|
336 | 337 | ESC_HELP = '?' # Find information about object |
|
337 | 338 | ESC_HELP2 = '??' # Find extra-detailed information about object |
|
338 | 339 | ESC_MAGIC = '%' # Call magic function |
|
339 | 340 | ESC_MAGIC2 = '%%' # Call cell-magic function |
|
340 | 341 | ESC_QUOTE = ',' # Split args on whitespace, quote each as string and call |
|
341 | 342 | ESC_QUOTE2 = ';' # Quote all args as a single string, call |
|
342 | 343 | ESC_PAREN = '/' # Call first argument with rest of line as arguments |
|
343 | 344 | |
|
344 | 345 | ESCAPE_SINGLES = {'!', '?', '%', ',', ';', '/'} |
|
345 | 346 | ESCAPE_DOUBLES = {'!!', '??'} # %% (cell magic) is handled separately |
|
346 | 347 | |
|
347 | 348 | def _make_help_call(target, esc): |
|
348 | 349 | """Prepares a pinfo(2)/psearch call from a target name and the escape |
|
349 | 350 | (i.e. ? or ??)""" |
|
350 | 351 | method = 'pinfo2' if esc == '??' \ |
|
351 | 352 | else 'psearch' if '*' in target \ |
|
352 | 353 | else 'pinfo' |
|
353 | 354 | arg = " ".join([method, target]) |
|
354 | 355 | #Prepare arguments for get_ipython().run_line_magic(magic_name, magic_args) |
|
355 | 356 | t_magic_name, _, t_magic_arg_s = arg.partition(' ') |
|
356 | 357 | t_magic_name = t_magic_name.lstrip(ESC_MAGIC) |
|
357 | 358 | return "get_ipython().run_line_magic(%r, %r)" % (t_magic_name, t_magic_arg_s) |
|
358 | 359 | |
|
359 | 360 | |
|
360 | 361 | def _tr_help(content): |
|
361 | 362 | """Translate lines escaped with: ? |
|
362 | 363 | |
|
363 | 364 | A naked help line should fire the intro help screen (shell.show_usage()) |
|
364 | 365 | """ |
|
365 | 366 | if not content: |
|
366 | 367 | return 'get_ipython().show_usage()' |
|
367 | 368 | |
|
368 | 369 | return _make_help_call(content, '?') |
|
369 | 370 | |
|
370 | 371 | def _tr_help2(content): |
|
371 | 372 | """Translate lines escaped with: ?? |
|
372 | 373 | |
|
373 | 374 | A naked help line should fire the intro help screen (shell.show_usage()) |
|
374 | 375 | """ |
|
375 | 376 | if not content: |
|
376 | 377 | return 'get_ipython().show_usage()' |
|
377 | 378 | |
|
378 | 379 | return _make_help_call(content, '??') |
|
379 | 380 | |
|
380 | 381 | def _tr_magic(content): |
|
381 | 382 | "Translate lines escaped with a percent sign: %" |
|
382 | 383 | name, _, args = content.partition(' ') |
|
383 | 384 | return 'get_ipython().run_line_magic(%r, %r)' % (name, args) |
|
384 | 385 | |
|
385 | 386 | def _tr_quote(content): |
|
386 | 387 | "Translate lines escaped with a comma: ," |
|
387 | 388 | name, _, args = content.partition(' ') |
|
388 | 389 | return '%s("%s")' % (name, '", "'.join(args.split()) ) |
|
389 | 390 | |
|
390 | 391 | def _tr_quote2(content): |
|
391 | 392 | "Translate lines escaped with a semicolon: ;" |
|
392 | 393 | name, _, args = content.partition(' ') |
|
393 | 394 | return '%s("%s")' % (name, args) |
|
394 | 395 | |
|
395 | 396 | def _tr_paren(content): |
|
396 | 397 | "Translate lines escaped with a slash: /" |
|
397 | 398 | name, _, args = content.partition(' ') |
|
398 | 399 | return '%s(%s)' % (name, ", ".join(args.split())) |
|
399 | 400 | |
|
400 | 401 | tr = { ESC_SHELL : 'get_ipython().system({!r})'.format, |
|
401 | 402 | ESC_SH_CAP : 'get_ipython().getoutput({!r})'.format, |
|
402 | 403 | ESC_HELP : _tr_help, |
|
403 | 404 | ESC_HELP2 : _tr_help2, |
|
404 | 405 | ESC_MAGIC : _tr_magic, |
|
405 | 406 | ESC_QUOTE : _tr_quote, |
|
406 | 407 | ESC_QUOTE2 : _tr_quote2, |
|
407 | 408 | ESC_PAREN : _tr_paren } |
|
408 | 409 | |
|
409 | 410 | class EscapedCommand(TokenTransformBase): |
|
410 | 411 | """Transformer for escaped commands like %foo, !foo, or /foo""" |
|
411 | 412 | @classmethod |
|
412 | 413 | def find(cls, tokens_by_line): |
|
413 | 414 | """Find the first escaped command (%foo, !foo, etc.) in the cell. |
|
414 | 415 | """ |
|
415 | 416 | for line in tokens_by_line: |
|
416 | 417 | if not line: |
|
417 | 418 | continue |
|
418 | 419 | ix = 0 |
|
419 | 420 | ll = len(line) |
|
420 | 421 | while ll > ix and line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
421 | 422 | ix += 1 |
|
422 | 423 | if ix >= ll: |
|
423 | 424 | continue |
|
424 | 425 | if line[ix].string in ESCAPE_SINGLES: |
|
425 | 426 | return cls(line[ix].start) |
|
426 | 427 | |
|
427 | 428 | def transform(self, lines): |
|
428 | 429 | """Transform an escaped line found by the ``find()`` classmethod. |
|
429 | 430 | """ |
|
430 | 431 | start_line, start_col = self.start_line, self.start_col |
|
431 | 432 | |
|
432 | 433 | indent = lines[start_line][:start_col] |
|
433 | 434 | end_line = find_end_of_continued_line(lines, start_line) |
|
434 | 435 | line = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
435 | 436 | |
|
436 | 437 | if len(line) > 1 and line[:2] in ESCAPE_DOUBLES: |
|
437 | 438 | escape, content = line[:2], line[2:] |
|
438 | 439 | else: |
|
439 | 440 | escape, content = line[:1], line[1:] |
|
440 | 441 | |
|
441 | 442 | if escape in tr: |
|
442 | 443 | call = tr[escape](content) |
|
443 | 444 | else: |
|
444 | 445 | call = '' |
|
445 | 446 | |
|
446 | 447 | lines_before = lines[:start_line] |
|
447 | 448 | new_line = indent + call + '\n' |
|
448 | 449 | lines_after = lines[end_line + 1:] |
|
449 | 450 | |
|
450 | 451 | return lines_before + [new_line] + lines_after |
|
451 | 452 | |
|
452 | 453 | |
|
453 | 454 | _help_end_re = re.compile( |
|
454 | 455 | r"""(%{0,2} |
|
455 | 456 | (?!\d)[\w*]+ # Variable name |
|
456 | 457 | (\.(?!\d)[\w*]+|\[-?[0-9]+\])* # .etc.etc or [0], we only support literal integers. |
|
457 | 458 | ) |
|
458 | 459 | (\?\??)$ # ? or ?? |
|
459 | 460 | """, |
|
460 | 461 | re.VERBOSE, |
|
461 | 462 | ) |
|
462 | 463 | |
|
463 | 464 | |
|
464 | 465 | class HelpEnd(TokenTransformBase): |
|
465 | 466 | """Transformer for help syntax: obj? and obj??""" |
|
466 | 467 | # This needs to be higher priority (lower number) than EscapedCommand so |
|
467 | 468 | # that inspecting magics (%foo?) works. |
|
468 | 469 | priority = 5 |
|
469 | 470 | |
|
470 | 471 | def __init__(self, start, q_locn): |
|
471 | 472 | super().__init__(start) |
|
472 | 473 | self.q_line = q_locn[0] - 1 # Shift from 1-indexed to 0-indexed |
|
473 | 474 | self.q_col = q_locn[1] |
|
474 | 475 | |
|
475 | 476 | @classmethod |
|
476 | 477 | def find(cls, tokens_by_line): |
|
477 | 478 | """Find the first help command (foo?) in the cell. |
|
478 | 479 | """ |
|
479 | 480 | for line in tokens_by_line: |
|
480 | 481 | # Last token is NEWLINE; look at last but one |
|
481 | 482 | if len(line) > 2 and line[-2].string == '?': |
|
482 | 483 | # Find the first token that's not INDENT/DEDENT |
|
483 | 484 | ix = 0 |
|
484 | 485 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
485 | 486 | ix += 1 |
|
486 | 487 | return cls(line[ix].start, line[-2].start) |
|
487 | 488 | |
|
488 | 489 | def transform(self, lines): |
|
489 | 490 | """Transform a help command found by the ``find()`` classmethod. |
|
490 | 491 | """ |
|
491 | 492 | |
|
492 | 493 | piece = "".join(lines[self.start_line : self.q_line + 1]) |
|
493 | 494 | indent, content = piece[: self.start_col], piece[self.start_col :] |
|
494 | 495 | lines_before = lines[: self.start_line] |
|
495 | 496 | lines_after = lines[self.q_line + 1 :] |
|
496 | 497 | |
|
497 | 498 | m = _help_end_re.search(content) |
|
498 | 499 | if not m: |
|
499 | 500 | raise SyntaxError(content) |
|
500 | 501 | assert m is not None, content |
|
501 | 502 | target = m.group(1) |
|
502 | 503 | esc = m.group(3) |
|
503 | 504 | |
|
504 | 505 | |
|
505 | 506 | call = _make_help_call(target, esc) |
|
506 | 507 | new_line = indent + call + '\n' |
|
507 | 508 | |
|
508 | 509 | return lines_before + [new_line] + lines_after |
|
509 | 510 | |
|
510 | 511 | def make_tokens_by_line(lines:List[str]): |
|
511 | 512 | """Tokenize a series of lines and group tokens by line. |
|
512 | 513 | |
|
513 | 514 | The tokens for a multiline Python string or expression are grouped as one |
|
514 | 515 | line. All lines except the last lines should keep their line ending ('\\n', |
|
515 | 516 | '\\r\\n') for this to properly work. Use `.splitlines(keeplineending=True)` |
|
516 | 517 | for example when passing block of text to this function. |
|
517 | 518 | |
|
518 | 519 | """ |
|
519 | 520 | # NL tokens are used inside multiline expressions, but also after blank |
|
520 | 521 | # lines or comments. This is intentional - see https://bugs.python.org/issue17061 |
|
521 | 522 | # We want to group the former case together but split the latter, so we |
|
522 | 523 | # track parentheses level, similar to the internals of tokenize. |
|
523 | 524 | |
|
524 | 525 | # reexported from token on 3.7+ |
|
525 | 526 | NEWLINE, NL = tokenize.NEWLINE, tokenize.NL # type: ignore |
|
526 | 527 | tokens_by_line: List[List[Any]] = [[]] |
|
527 | 528 | if len(lines) > 1 and not lines[0].endswith(("\n", "\r", "\r\n", "\x0b", "\x0c")): |
|
528 | 529 | warnings.warn( |
|
529 | 530 | "`make_tokens_by_line` received a list of lines which do not have lineending markers ('\\n', '\\r', '\\r\\n', '\\x0b', '\\x0c'), behavior will be unspecified", |
|
530 | 531 | stacklevel=2, |
|
531 | 532 | ) |
|
532 | 533 | parenlev = 0 |
|
533 | 534 | try: |
|
534 |
for token in tokenutil.generate_tokens_catch_errors( |
|
|
535 | extra_errors_to_catch=['expected EOF']): | |
|
535 | for token in tokenutil.generate_tokens_catch_errors( | |
|
536 | iter(lines).__next__, extra_errors_to_catch=["expected EOF"] | |
|
537 | ): | |
|
536 | 538 | tokens_by_line[-1].append(token) |
|
537 | 539 | if (token.type == NEWLINE) \ |
|
538 | 540 | or ((token.type == NL) and (parenlev <= 0)): |
|
539 | 541 | tokens_by_line.append([]) |
|
540 | 542 | elif token.string in {'(', '[', '{'}: |
|
541 | 543 | parenlev += 1 |
|
542 | 544 | elif token.string in {')', ']', '}'}: |
|
543 | 545 | if parenlev > 0: |
|
544 | 546 | parenlev -= 1 |
|
545 | 547 | except tokenize.TokenError: |
|
546 | 548 | # Input ended in a multiline string or expression. That's OK for us. |
|
547 | 549 | pass |
|
548 | 550 | |
|
549 | 551 | |
|
550 | 552 | if not tokens_by_line[-1]: |
|
551 | 553 | tokens_by_line.pop() |
|
552 | 554 | |
|
553 | 555 | |
|
554 | 556 | return tokens_by_line |
|
555 | 557 | |
|
556 | 558 | |
|
557 | 559 | def has_sunken_brackets(tokens: List[tokenize.TokenInfo]): |
|
558 | 560 | """Check if the depth of brackets in the list of tokens drops below 0""" |
|
559 | 561 | parenlev = 0 |
|
560 | 562 | for token in tokens: |
|
561 | 563 | if token.string in {"(", "[", "{"}: |
|
562 | 564 | parenlev += 1 |
|
563 | 565 | elif token.string in {")", "]", "}"}: |
|
564 | 566 | parenlev -= 1 |
|
565 | 567 | if parenlev < 0: |
|
566 | 568 | return True |
|
567 | 569 | return False |
|
568 | 570 | |
|
569 | 571 | |
|
570 | 572 | def show_linewise_tokens(s: str): |
|
571 | 573 | """For investigation and debugging""" |
|
572 | 574 | warnings.warn( |
|
573 | 575 | "show_linewise_tokens is deprecated since IPython 8.6", |
|
574 | 576 | DeprecationWarning, |
|
575 | 577 | stacklevel=2, |
|
576 | 578 | ) |
|
577 | 579 | if not s.endswith("\n"): |
|
578 | 580 | s += "\n" |
|
579 | 581 | lines = s.splitlines(keepends=True) |
|
580 | 582 | for line in make_tokens_by_line(lines): |
|
581 | 583 | print("Line -------") |
|
582 | 584 | for tokinfo in line: |
|
583 | 585 | print(" ", tokinfo) |
|
584 | 586 | |
|
585 | 587 | # Arbitrary limit to prevent getting stuck in infinite loops |
|
586 | 588 | TRANSFORM_LOOP_LIMIT = 500 |
|
587 | 589 | |
|
588 | 590 | class TransformerManager: |
|
589 | 591 | """Applies various transformations to a cell or code block. |
|
590 | 592 | |
|
591 | 593 | The key methods for external use are ``transform_cell()`` |
|
592 | 594 | and ``check_complete()``. |
|
593 | 595 | """ |
|
594 | 596 | def __init__(self): |
|
595 | 597 | self.cleanup_transforms = [ |
|
596 | 598 | leading_empty_lines, |
|
597 | 599 | leading_indent, |
|
598 | 600 | classic_prompt, |
|
599 | 601 | ipython_prompt, |
|
600 | 602 | ] |
|
601 | 603 | self.line_transforms = [ |
|
602 | 604 | cell_magic, |
|
603 | 605 | ] |
|
604 | 606 | self.token_transformers = [ |
|
605 | 607 | MagicAssign, |
|
606 | 608 | SystemAssign, |
|
607 | 609 | EscapedCommand, |
|
608 | 610 | HelpEnd, |
|
609 | 611 | ] |
|
610 | 612 | |
|
611 | 613 | def do_one_token_transform(self, lines): |
|
612 | 614 | """Find and run the transform earliest in the code. |
|
613 | 615 | |
|
614 | 616 | Returns (changed, lines). |
|
615 | 617 | |
|
616 | 618 | This method is called repeatedly until changed is False, indicating |
|
617 | 619 | that all available transformations are complete. |
|
618 | 620 | |
|
619 | 621 | The tokens following IPython special syntax might not be valid, so |
|
620 | 622 | the transformed code is retokenised every time to identify the next |
|
621 | 623 | piece of special syntax. Hopefully long code cells are mostly valid |
|
622 | 624 | Python, not using lots of IPython special syntax, so this shouldn't be |
|
623 | 625 | a performance issue. |
|
624 | 626 | """ |
|
625 | 627 | tokens_by_line = make_tokens_by_line(lines) |
|
626 | 628 | candidates = [] |
|
627 | 629 | for transformer_cls in self.token_transformers: |
|
628 | 630 | transformer = transformer_cls.find(tokens_by_line) |
|
629 | 631 | if transformer: |
|
630 | 632 | candidates.append(transformer) |
|
631 | 633 | |
|
632 | 634 | if not candidates: |
|
633 | 635 | # Nothing to transform |
|
634 | 636 | return False, lines |
|
635 | 637 | ordered_transformers = sorted(candidates, key=TokenTransformBase.sortby) |
|
636 | 638 | for transformer in ordered_transformers: |
|
637 | 639 | try: |
|
638 | 640 | return True, transformer.transform(lines) |
|
639 | 641 | except SyntaxError: |
|
640 | 642 | pass |
|
641 | 643 | return False, lines |
|
642 | 644 | |
|
643 | 645 | def do_token_transforms(self, lines): |
|
644 | 646 | for _ in range(TRANSFORM_LOOP_LIMIT): |
|
645 | 647 | changed, lines = self.do_one_token_transform(lines) |
|
646 | 648 | if not changed: |
|
647 | 649 | return lines |
|
648 | 650 | |
|
649 | 651 | raise RuntimeError("Input transformation still changing after " |
|
650 | 652 | "%d iterations. Aborting." % TRANSFORM_LOOP_LIMIT) |
|
651 | 653 | |
|
652 | 654 | def transform_cell(self, cell: str) -> str: |
|
653 | 655 | """Transforms a cell of input code""" |
|
654 | 656 | if not cell.endswith('\n'): |
|
655 | 657 | cell += '\n' # Ensure the cell has a trailing newline |
|
656 | 658 | lines = cell.splitlines(keepends=True) |
|
657 | 659 | for transform in self.cleanup_transforms + self.line_transforms: |
|
658 | 660 | lines = transform(lines) |
|
659 | 661 | |
|
660 | 662 | lines = self.do_token_transforms(lines) |
|
661 | 663 | return ''.join(lines) |
|
662 | 664 | |
|
663 | 665 | def check_complete(self, cell: str): |
|
664 | 666 | """Return whether a block of code is ready to execute, or should be continued |
|
665 | 667 | |
|
666 | 668 | Parameters |
|
667 | 669 | ---------- |
|
668 | 670 | cell : string |
|
669 | 671 | Python input code, which can be multiline. |
|
670 | 672 | |
|
671 | 673 | Returns |
|
672 | 674 | ------- |
|
673 | 675 | status : str |
|
674 | 676 | One of 'complete', 'incomplete', or 'invalid' if source is not a |
|
675 | 677 | prefix of valid code. |
|
676 | 678 | indent_spaces : int or None |
|
677 | 679 | The number of spaces by which to indent the next line of code. If |
|
678 | 680 | status is not 'incomplete', this is None. |
|
679 | 681 | """ |
|
680 | 682 | # Remember if the lines ends in a new line. |
|
681 | 683 | ends_with_newline = False |
|
682 | 684 | for character in reversed(cell): |
|
683 | 685 | if character == '\n': |
|
684 | 686 | ends_with_newline = True |
|
685 | 687 | break |
|
686 | 688 | elif character.strip(): |
|
687 | 689 | break |
|
688 | 690 | else: |
|
689 | 691 | continue |
|
690 | 692 | |
|
691 | 693 | if not ends_with_newline: |
|
692 | 694 | # Append an newline for consistent tokenization |
|
693 | 695 | # See https://bugs.python.org/issue33899 |
|
694 | 696 | cell += '\n' |
|
695 | 697 | |
|
696 | 698 | lines = cell.splitlines(keepends=True) |
|
697 | 699 | |
|
698 | 700 | if not lines: |
|
699 | 701 | return 'complete', None |
|
700 | 702 | |
|
701 | 703 | for line in reversed(lines): |
|
702 | 704 | if not line.strip(): |
|
703 | 705 | continue |
|
704 |
elif line.strip( |
|
|
705 |
return |
|
|
706 | elif line.strip("\n").endswith("\\"): | |
|
707 | return "incomplete", find_last_indent(lines) | |
|
706 | 708 | else: |
|
707 | 709 | break |
|
708 | 710 | |
|
709 | 711 | try: |
|
710 | 712 | for transform in self.cleanup_transforms: |
|
711 | 713 | if not getattr(transform, 'has_side_effects', False): |
|
712 | 714 | lines = transform(lines) |
|
713 | 715 | except SyntaxError: |
|
714 | 716 | return 'invalid', None |
|
715 | 717 | |
|
716 | 718 | if lines[0].startswith('%%'): |
|
717 | 719 | # Special case for cell magics - completion marked by blank line |
|
718 | 720 | if lines[-1].strip(): |
|
719 | 721 | return 'incomplete', find_last_indent(lines) |
|
720 | 722 | else: |
|
721 | 723 | return 'complete', None |
|
722 | 724 | |
|
723 | 725 | try: |
|
724 | 726 | for transform in self.line_transforms: |
|
725 | 727 | if not getattr(transform, 'has_side_effects', False): |
|
726 | 728 | lines = transform(lines) |
|
727 | 729 | lines = self.do_token_transforms(lines) |
|
728 | 730 | except SyntaxError: |
|
729 | 731 | return 'invalid', None |
|
730 | 732 | |
|
731 | 733 | tokens_by_line = make_tokens_by_line(lines) |
|
732 | 734 | |
|
733 | 735 | # Bail if we got one line and there are more closing parentheses than |
|
734 | 736 | # the opening ones |
|
735 | 737 | if ( |
|
736 | 738 | len(lines) == 1 |
|
737 | 739 | and tokens_by_line |
|
738 | 740 | and has_sunken_brackets(tokens_by_line[0]) |
|
739 | 741 | ): |
|
740 | 742 | return "invalid", None |
|
741 | 743 | |
|
742 | 744 | if not tokens_by_line: |
|
743 | 745 | return 'incomplete', find_last_indent(lines) |
|
744 | 746 | |
|
745 | if (tokens_by_line[-1][-1].type != tokenize.ENDMARKER | |
|
746 |
|
|
|
747 | if ( | |
|
748 | tokens_by_line[-1][-1].type != tokenize.ENDMARKER | |
|
749 | and tokens_by_line[-1][-1].type != tokenize.ERRORTOKEN | |
|
750 | ): | |
|
747 | 751 | # We're in a multiline string or expression |
|
748 | 752 | return 'incomplete', find_last_indent(lines) |
|
749 | 753 | |
|
750 | 754 | newline_types = {tokenize.NEWLINE, tokenize.COMMENT, tokenize.ENDMARKER} # type: ignore |
|
751 | 755 | |
|
752 | 756 | # Pop the last line which only contains DEDENTs and ENDMARKER |
|
753 | 757 | last_token_line = None |
|
754 | 758 | if {t.type for t in tokens_by_line[-1]} in [ |
|
755 | 759 | {tokenize.DEDENT, tokenize.ENDMARKER}, |
|
756 | 760 | {tokenize.ENDMARKER} |
|
757 | 761 | ] and len(tokens_by_line) > 1: |
|
758 | 762 | last_token_line = tokens_by_line.pop() |
|
759 | 763 | |
|
760 | 764 | while tokens_by_line[-1] and tokens_by_line[-1][-1].type in newline_types: |
|
761 | 765 | tokens_by_line[-1].pop() |
|
762 | 766 | |
|
763 | 767 | if not tokens_by_line[-1]: |
|
764 | 768 | return 'incomplete', find_last_indent(lines) |
|
765 | 769 | |
|
766 | 770 | if tokens_by_line[-1][-1].string == ':': |
|
767 | 771 | # The last line starts a block (e.g. 'if foo:') |
|
768 | 772 | ix = 0 |
|
769 | 773 | while tokens_by_line[-1][ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
770 | 774 | ix += 1 |
|
771 | 775 | |
|
772 | 776 | indent = tokens_by_line[-1][ix].start[1] |
|
773 | 777 | return 'incomplete', indent + 4 |
|
774 | 778 | |
|
775 | 779 | if tokens_by_line[-1][0].line.endswith('\\'): |
|
776 | 780 | return 'incomplete', None |
|
777 | 781 | |
|
778 | 782 | # At this point, our checks think the code is complete (or invalid). |
|
779 | 783 | # We'll use codeop.compile_command to check this with the real parser |
|
780 | 784 | try: |
|
781 | 785 | with warnings.catch_warnings(): |
|
782 | 786 | warnings.simplefilter('error', SyntaxWarning) |
|
783 | 787 | res = compile_command(''.join(lines), symbol='exec') |
|
784 | 788 | except (SyntaxError, OverflowError, ValueError, TypeError, |
|
785 | 789 | MemoryError, SyntaxWarning): |
|
786 | 790 | return 'invalid', None |
|
787 | 791 | else: |
|
788 | 792 | if res is None: |
|
789 | 793 | return 'incomplete', find_last_indent(lines) |
|
790 | 794 | |
|
791 | 795 | if last_token_line and last_token_line[0].type == tokenize.DEDENT: |
|
792 | 796 | if ends_with_newline: |
|
793 | 797 | return 'complete', None |
|
794 | 798 | return 'incomplete', find_last_indent(lines) |
|
795 | 799 | |
|
796 | 800 | # If there's a blank line at the end, assume we're ready to execute |
|
797 | 801 | if not lines[-1].strip(): |
|
798 | 802 | return 'complete', None |
|
799 | 803 | |
|
800 | 804 | return 'complete', None |
|
801 | 805 | |
|
802 | 806 | |
|
803 | 807 | def find_last_indent(lines): |
|
804 | 808 | m = _indent_re.match(lines[-1]) |
|
805 | 809 | if not m: |
|
806 | 810 | return 0 |
|
807 | 811 | return len(m.group(0).replace('\t', ' '*4)) |
|
808 | 812 | |
|
809 | 813 | |
|
810 | 814 | class MaybeAsyncCompile(Compile): |
|
811 | 815 | def __init__(self, extra_flags=0): |
|
812 | 816 | super().__init__() |
|
813 | 817 | self.flags |= extra_flags |
|
814 | 818 | |
|
815 | 819 | |
|
816 | 820 | class MaybeAsyncCommandCompiler(CommandCompiler): |
|
817 | 821 | def __init__(self, extra_flags=0): |
|
818 | 822 | self.compiler = MaybeAsyncCompile(extra_flags=extra_flags) |
|
819 | 823 | |
|
820 | 824 | |
|
821 | 825 | _extra_flags = ast.PyCF_ALLOW_TOP_LEVEL_AWAIT |
|
822 | 826 | |
|
823 | 827 | compile_command = MaybeAsyncCommandCompiler(extra_flags=_extra_flags) |
@@ -1,447 +1,448 | |||
|
1 | 1 | """Tests for the token-based transformers in IPython.core.inputtransformer2 |
|
2 | 2 | |
|
3 | 3 | Line-based transformers are the simpler ones; token-based transformers are |
|
4 | 4 | more complex. See test_inputtransformer2_line for tests for line-based |
|
5 | 5 | transformations. |
|
6 | 6 | """ |
|
7 | 7 | import platform |
|
8 | 8 | import string |
|
9 | 9 | import sys |
|
10 | 10 | from textwrap import dedent |
|
11 | 11 | |
|
12 | 12 | import pytest |
|
13 | 13 | |
|
14 | 14 | from IPython.core import inputtransformer2 as ipt2 |
|
15 | 15 | from IPython.core.inputtransformer2 import _find_assign_op, make_tokens_by_line |
|
16 | 16 | |
|
17 | 17 | MULTILINE_MAGIC = ( |
|
18 | 18 | """\ |
|
19 | 19 | a = f() |
|
20 | 20 | %foo \\ |
|
21 | 21 | bar |
|
22 | 22 | g() |
|
23 | 23 | """.splitlines( |
|
24 | 24 | keepends=True |
|
25 | 25 | ), |
|
26 | 26 | (2, 0), |
|
27 | 27 | """\ |
|
28 | 28 | a = f() |
|
29 | 29 | get_ipython().run_line_magic('foo', ' bar') |
|
30 | 30 | g() |
|
31 | 31 | """.splitlines( |
|
32 | 32 | keepends=True |
|
33 | 33 | ), |
|
34 | 34 | ) |
|
35 | 35 | |
|
36 | 36 | INDENTED_MAGIC = ( |
|
37 | 37 | """\ |
|
38 | 38 | for a in range(5): |
|
39 | 39 | %ls |
|
40 | 40 | """.splitlines( |
|
41 | 41 | keepends=True |
|
42 | 42 | ), |
|
43 | 43 | (2, 4), |
|
44 | 44 | """\ |
|
45 | 45 | for a in range(5): |
|
46 | 46 | get_ipython().run_line_magic('ls', '') |
|
47 | 47 | """.splitlines( |
|
48 | 48 | keepends=True |
|
49 | 49 | ), |
|
50 | 50 | ) |
|
51 | 51 | |
|
52 | 52 | CRLF_MAGIC = ( |
|
53 | 53 | ["a = f()\n", "%ls\r\n", "g()\n"], |
|
54 | 54 | (2, 0), |
|
55 | 55 | ["a = f()\n", "get_ipython().run_line_magic('ls', '')\n", "g()\n"], |
|
56 | 56 | ) |
|
57 | 57 | |
|
58 | 58 | MULTILINE_MAGIC_ASSIGN = ( |
|
59 | 59 | """\ |
|
60 | 60 | a = f() |
|
61 | 61 | b = %foo \\ |
|
62 | 62 | bar |
|
63 | 63 | g() |
|
64 | 64 | """.splitlines( |
|
65 | 65 | keepends=True |
|
66 | 66 | ), |
|
67 | 67 | (2, 4), |
|
68 | 68 | """\ |
|
69 | 69 | a = f() |
|
70 | 70 | b = get_ipython().run_line_magic('foo', ' bar') |
|
71 | 71 | g() |
|
72 | 72 | """.splitlines( |
|
73 | 73 | keepends=True |
|
74 | 74 | ), |
|
75 | 75 | ) |
|
76 | 76 | |
|
77 | 77 | MULTILINE_SYSTEM_ASSIGN = ("""\ |
|
78 | 78 | a = f() |
|
79 | 79 | b = !foo \\ |
|
80 | 80 | bar |
|
81 | 81 | g() |
|
82 | 82 | """.splitlines(keepends=True), (2, 4), """\ |
|
83 | 83 | a = f() |
|
84 | 84 | b = get_ipython().getoutput('foo bar') |
|
85 | 85 | g() |
|
86 | 86 | """.splitlines(keepends=True)) |
|
87 | 87 | |
|
88 | 88 | ##### |
|
89 | 89 | |
|
90 | 90 | MULTILINE_SYSTEM_ASSIGN_AFTER_DEDENT = ( |
|
91 | 91 | """\ |
|
92 | 92 | def test(): |
|
93 | 93 | for i in range(1): |
|
94 | 94 | print(i) |
|
95 | 95 | res =! ls |
|
96 | 96 | """.splitlines( |
|
97 | 97 | keepends=True |
|
98 | 98 | ), |
|
99 | 99 | (4, 7), |
|
100 | 100 | """\ |
|
101 | 101 | def test(): |
|
102 | 102 | for i in range(1): |
|
103 | 103 | print(i) |
|
104 | 104 | res =get_ipython().getoutput(\' ls\') |
|
105 | 105 | """.splitlines( |
|
106 | 106 | keepends=True |
|
107 | 107 | ), |
|
108 | 108 | ) |
|
109 | 109 | |
|
110 | 110 | ###### |
|
111 | 111 | |
|
112 | 112 | AUTOCALL_QUOTE = ([",f 1 2 3\n"], (1, 0), ['f("1", "2", "3")\n']) |
|
113 | 113 | |
|
114 | 114 | AUTOCALL_QUOTE2 = ([";f 1 2 3\n"], (1, 0), ['f("1 2 3")\n']) |
|
115 | 115 | |
|
116 | 116 | AUTOCALL_PAREN = (["/f 1 2 3\n"], (1, 0), ["f(1, 2, 3)\n"]) |
|
117 | 117 | |
|
118 | 118 | SIMPLE_HELP = (["foo?\n"], (1, 0), ["get_ipython().run_line_magic('pinfo', 'foo')\n"]) |
|
119 | 119 | |
|
120 | 120 | DETAILED_HELP = ( |
|
121 | 121 | ["foo??\n"], |
|
122 | 122 | (1, 0), |
|
123 | 123 | ["get_ipython().run_line_magic('pinfo2', 'foo')\n"], |
|
124 | 124 | ) |
|
125 | 125 | |
|
126 | 126 | MAGIC_HELP = (["%foo?\n"], (1, 0), ["get_ipython().run_line_magic('pinfo', '%foo')\n"]) |
|
127 | 127 | |
|
128 | 128 | HELP_IN_EXPR = ( |
|
129 | 129 | ["a = b + c?\n"], |
|
130 | 130 | (1, 0), |
|
131 | 131 | ["get_ipython().run_line_magic('pinfo', 'c')\n"], |
|
132 | 132 | ) |
|
133 | 133 | |
|
134 | 134 | HELP_CONTINUED_LINE = ( |
|
135 | 135 | """\ |
|
136 | 136 | a = \\ |
|
137 | 137 | zip? |
|
138 | 138 | """.splitlines( |
|
139 | 139 | keepends=True |
|
140 | 140 | ), |
|
141 | 141 | (1, 0), |
|
142 | 142 | [r"get_ipython().run_line_magic('pinfo', 'zip')" + "\n"], |
|
143 | 143 | ) |
|
144 | 144 | |
|
145 | 145 | HELP_MULTILINE = ( |
|
146 | 146 | """\ |
|
147 | 147 | (a, |
|
148 | 148 | b) = zip? |
|
149 | 149 | """.splitlines( |
|
150 | 150 | keepends=True |
|
151 | 151 | ), |
|
152 | 152 | (1, 0), |
|
153 | 153 | [r"get_ipython().run_line_magic('pinfo', 'zip')" + "\n"], |
|
154 | 154 | ) |
|
155 | 155 | |
|
156 | 156 | HELP_UNICODE = ( |
|
157 | 157 | ["Ο.foo?\n"], |
|
158 | 158 | (1, 0), |
|
159 | 159 | ["get_ipython().run_line_magic('pinfo', 'Ο.foo')\n"], |
|
160 | 160 | ) |
|
161 | 161 | |
|
162 | 162 | |
|
163 | 163 | def null_cleanup_transformer(lines): |
|
164 | 164 | """ |
|
165 | 165 | A cleanup transform that returns an empty list. |
|
166 | 166 | """ |
|
167 | 167 | return [] |
|
168 | 168 | |
|
169 | 169 | |
|
170 | 170 | def test_check_make_token_by_line_never_ends_empty(): |
|
171 | 171 | """ |
|
172 | 172 | Check that not sequence of single or double characters ends up leading to en empty list of tokens |
|
173 | 173 | """ |
|
174 | 174 | from string import printable |
|
175 | 175 | |
|
176 | 176 | for c in printable: |
|
177 | 177 | assert make_tokens_by_line(c)[-1] != [] |
|
178 | 178 | for k in printable: |
|
179 | 179 | assert make_tokens_by_line(c + k)[-1] != [] |
|
180 | 180 | |
|
181 | 181 | |
|
182 | 182 | def check_find(transformer, case, match=True): |
|
183 | 183 | sample, expected_start, _ = case |
|
184 | 184 | tbl = make_tokens_by_line(sample) |
|
185 | 185 | res = transformer.find(tbl) |
|
186 | 186 | if match: |
|
187 | 187 | # start_line is stored 0-indexed, expected values are 1-indexed |
|
188 | 188 | assert (res.start_line + 1, res.start_col) == expected_start |
|
189 | 189 | return res |
|
190 | 190 | else: |
|
191 | 191 | assert res is None |
|
192 | 192 | |
|
193 | 193 | |
|
194 | 194 | def check_transform(transformer_cls, case): |
|
195 | 195 | lines, start, expected = case |
|
196 | 196 | transformer = transformer_cls(start) |
|
197 | 197 | assert transformer.transform(lines) == expected |
|
198 | 198 | |
|
199 | 199 | |
|
200 | 200 | def test_continued_line(): |
|
201 | 201 | lines = MULTILINE_MAGIC_ASSIGN[0] |
|
202 | 202 | assert ipt2.find_end_of_continued_line(lines, 1) == 2 |
|
203 | 203 | |
|
204 | 204 | assert ipt2.assemble_continued_line(lines, (1, 5), 2) == "foo bar" |
|
205 | 205 | |
|
206 | 206 | |
|
207 | 207 | def test_find_assign_magic(): |
|
208 | 208 | check_find(ipt2.MagicAssign, MULTILINE_MAGIC_ASSIGN) |
|
209 | 209 | check_find(ipt2.MagicAssign, MULTILINE_SYSTEM_ASSIGN, match=False) |
|
210 | 210 | check_find(ipt2.MagicAssign, MULTILINE_SYSTEM_ASSIGN_AFTER_DEDENT, match=False) |
|
211 | 211 | |
|
212 | 212 | |
|
213 | 213 | def test_transform_assign_magic(): |
|
214 | 214 | check_transform(ipt2.MagicAssign, MULTILINE_MAGIC_ASSIGN) |
|
215 | 215 | |
|
216 | 216 | |
|
217 | 217 | def test_find_assign_system(): |
|
218 | 218 | check_find(ipt2.SystemAssign, MULTILINE_SYSTEM_ASSIGN) |
|
219 | 219 | check_find(ipt2.SystemAssign, MULTILINE_SYSTEM_ASSIGN_AFTER_DEDENT) |
|
220 | 220 | check_find(ipt2.SystemAssign, (["a = !ls\n"], (1, 5), None)) |
|
221 | 221 | check_find(ipt2.SystemAssign, (["a=!ls\n"], (1, 2), None)) |
|
222 | 222 | check_find(ipt2.SystemAssign, MULTILINE_MAGIC_ASSIGN, match=False) |
|
223 | 223 | |
|
224 | 224 | |
|
225 | 225 | def test_transform_assign_system(): |
|
226 | 226 | check_transform(ipt2.SystemAssign, MULTILINE_SYSTEM_ASSIGN) |
|
227 | 227 | check_transform(ipt2.SystemAssign, MULTILINE_SYSTEM_ASSIGN_AFTER_DEDENT) |
|
228 | 228 | |
|
229 | 229 | |
|
230 | 230 | def test_find_magic_escape(): |
|
231 | 231 | check_find(ipt2.EscapedCommand, MULTILINE_MAGIC) |
|
232 | 232 | check_find(ipt2.EscapedCommand, INDENTED_MAGIC) |
|
233 | 233 | check_find(ipt2.EscapedCommand, MULTILINE_MAGIC_ASSIGN, match=False) |
|
234 | 234 | |
|
235 | 235 | |
|
236 | 236 | def test_transform_magic_escape(): |
|
237 | 237 | check_transform(ipt2.EscapedCommand, MULTILINE_MAGIC) |
|
238 | 238 | check_transform(ipt2.EscapedCommand, INDENTED_MAGIC) |
|
239 | 239 | check_transform(ipt2.EscapedCommand, CRLF_MAGIC) |
|
240 | 240 | |
|
241 | 241 | |
|
242 | 242 | def test_find_autocalls(): |
|
243 | 243 | for case in [AUTOCALL_QUOTE, AUTOCALL_QUOTE2, AUTOCALL_PAREN]: |
|
244 | 244 | print("Testing %r" % case[0]) |
|
245 | 245 | check_find(ipt2.EscapedCommand, case) |
|
246 | 246 | |
|
247 | 247 | |
|
248 | 248 | def test_transform_autocall(): |
|
249 | 249 | for case in [AUTOCALL_QUOTE, AUTOCALL_QUOTE2, AUTOCALL_PAREN]: |
|
250 | 250 | print("Testing %r" % case[0]) |
|
251 | 251 | check_transform(ipt2.EscapedCommand, case) |
|
252 | 252 | |
|
253 | 253 | |
|
254 | 254 | def test_find_help(): |
|
255 | 255 | for case in [SIMPLE_HELP, DETAILED_HELP, MAGIC_HELP, HELP_IN_EXPR]: |
|
256 | 256 | check_find(ipt2.HelpEnd, case) |
|
257 | 257 | |
|
258 | 258 | tf = check_find(ipt2.HelpEnd, HELP_CONTINUED_LINE) |
|
259 | 259 | assert tf.q_line == 1 |
|
260 | 260 | assert tf.q_col == 3 |
|
261 | 261 | |
|
262 | 262 | tf = check_find(ipt2.HelpEnd, HELP_MULTILINE) |
|
263 | 263 | assert tf.q_line == 1 |
|
264 | 264 | assert tf.q_col == 8 |
|
265 | 265 | |
|
266 | 266 | # ? in a comment does not trigger help |
|
267 | 267 | check_find(ipt2.HelpEnd, (["foo # bar?\n"], None, None), match=False) |
|
268 | 268 | # Nor in a string |
|
269 | 269 | check_find(ipt2.HelpEnd, (["foo = '''bar?\n"], None, None), match=False) |
|
270 | 270 | |
|
271 | 271 | |
|
272 | 272 | def test_transform_help(): |
|
273 | 273 | tf = ipt2.HelpEnd((1, 0), (1, 9)) |
|
274 | 274 | assert tf.transform(HELP_IN_EXPR[0]) == HELP_IN_EXPR[2] |
|
275 | 275 | |
|
276 | 276 | tf = ipt2.HelpEnd((1, 0), (2, 3)) |
|
277 | 277 | assert tf.transform(HELP_CONTINUED_LINE[0]) == HELP_CONTINUED_LINE[2] |
|
278 | 278 | |
|
279 | 279 | tf = ipt2.HelpEnd((1, 0), (2, 8)) |
|
280 | 280 | assert tf.transform(HELP_MULTILINE[0]) == HELP_MULTILINE[2] |
|
281 | 281 | |
|
282 | 282 | tf = ipt2.HelpEnd((1, 0), (1, 0)) |
|
283 | 283 | assert tf.transform(HELP_UNICODE[0]) == HELP_UNICODE[2] |
|
284 | 284 | |
|
285 | 285 | |
|
286 | 286 | def test_find_assign_op_dedent(): |
|
287 | 287 | """ |
|
288 | 288 | be careful that empty token like dedent are not counted as parens |
|
289 | 289 | """ |
|
290 | 290 | |
|
291 | 291 | class Tk: |
|
292 | 292 | def __init__(self, s): |
|
293 | 293 | self.string = s |
|
294 | 294 | |
|
295 | 295 | assert _find_assign_op([Tk(s) for s in ("", "a", "=", "b")]) == 2 |
|
296 | 296 | assert ( |
|
297 | 297 | _find_assign_op([Tk(s) for s in ("", "(", "a", "=", "b", ")", "=", "5")]) == 6 |
|
298 | 298 | ) |
|
299 | 299 | |
|
300 | ||
|
300 | 301 | extra_closing_paren_param = ( |
|
301 | 302 | pytest.param("(\n))", "invalid", None) |
|
302 | 303 | if sys.version_info >= (3, 12) |
|
303 | 304 | else pytest.param("(\n))", "incomplete", 0) |
|
304 | 305 | ) |
|
305 | 306 | examples = [ |
|
306 | 307 | pytest.param("a = 1", "complete", None), |
|
307 | 308 | pytest.param("for a in range(5):", "incomplete", 4), |
|
308 | 309 | pytest.param("for a in range(5):\n if a > 0:", "incomplete", 8), |
|
309 | 310 | pytest.param("raise = 2", "invalid", None), |
|
310 | 311 | pytest.param("a = [1,\n2,", "incomplete", 0), |
|
311 | 312 | extra_closing_paren_param, |
|
312 | 313 | pytest.param("\\\r\n", "incomplete", 0), |
|
313 | 314 | pytest.param("a = '''\n hi", "incomplete", 3), |
|
314 | 315 | pytest.param("def a():\n x=1\n global x", "invalid", None), |
|
315 | 316 | pytest.param( |
|
316 | 317 | "a \\ ", |
|
317 | 318 | "invalid", |
|
318 | 319 | None, |
|
319 | 320 | marks=pytest.mark.xfail( |
|
320 | 321 | reason="Bug in python 3.9.8 βΒ bpo 45738", |
|
321 | 322 | condition=sys.version_info |
|
322 | 323 | in [(3, 9, 8, "final", 0), (3, 11, 0, "alpha", 2)], |
|
323 | 324 | raises=SystemError, |
|
324 | 325 | strict=True, |
|
325 | 326 | ), |
|
326 | 327 | ), # Nothing allowed after backslash, |
|
327 | 328 | pytest.param("1\\\n+2", "complete", None), |
|
328 | 329 | ] |
|
329 | 330 | |
|
330 | 331 | |
|
331 | 332 | @pytest.mark.parametrize("code, expected, number", examples) |
|
332 | 333 | def test_check_complete_param(code, expected, number): |
|
333 | 334 | cc = ipt2.TransformerManager().check_complete |
|
334 | 335 | assert cc(code) == (expected, number) |
|
335 | 336 | |
|
336 | 337 | |
|
337 | 338 | @pytest.mark.xfail(platform.python_implementation() == "PyPy", reason="fail on pypy") |
|
338 | 339 | @pytest.mark.xfail( |
|
339 | 340 | reason="Bug in python 3.9.8 βΒ bpo 45738", |
|
340 | 341 | condition=sys.version_info in [(3, 9, 8, "final", 0), (3, 11, 0, "alpha", 2)], |
|
341 | 342 | raises=SystemError, |
|
342 | 343 | strict=True, |
|
343 | 344 | ) |
|
344 | 345 | def test_check_complete(): |
|
345 | 346 | cc = ipt2.TransformerManager().check_complete |
|
346 | 347 | |
|
347 | 348 | example = dedent( |
|
348 | 349 | """ |
|
349 | 350 | if True: |
|
350 | 351 | a=1""" |
|
351 | 352 | ) |
|
352 | 353 | |
|
353 | 354 | assert cc(example) == ("incomplete", 4) |
|
354 | 355 | assert cc(example + "\n") == ("complete", None) |
|
355 | 356 | assert cc(example + "\n ") == ("complete", None) |
|
356 | 357 | |
|
357 | 358 | # no need to loop on all the letters/numbers. |
|
358 | 359 | short = "12abAB" + string.printable[62:] |
|
359 | 360 | for c in short: |
|
360 | 361 | # test does not raise: |
|
361 | 362 | cc(c) |
|
362 | 363 | for k in short: |
|
363 | 364 | cc(c + k) |
|
364 | 365 | |
|
365 | 366 | assert cc("def f():\n x=0\n \\\n ") == ("incomplete", 2) |
|
366 | 367 | |
|
367 | 368 | |
|
368 | 369 | @pytest.mark.xfail(platform.python_implementation() == "PyPy", reason="fail on pypy") |
|
369 | 370 | @pytest.mark.parametrize( |
|
370 | 371 | "value, expected", |
|
371 | 372 | [ |
|
372 | 373 | ('''def foo():\n """''', ("incomplete", 4)), |
|
373 | 374 | ("""async with example:\n pass""", ("incomplete", 4)), |
|
374 | 375 | ("""async with example:\n pass\n """, ("complete", None)), |
|
375 | 376 | ], |
|
376 | 377 | ) |
|
377 | 378 | def test_check_complete_II(value, expected): |
|
378 | 379 | """ |
|
379 | 380 | Test that multiple line strings are properly handled. |
|
380 | 381 | |
|
381 | 382 | Separate test function for convenience |
|
382 | 383 | |
|
383 | 384 | """ |
|
384 | 385 | cc = ipt2.TransformerManager().check_complete |
|
385 | 386 | assert cc(value) == expected |
|
386 | 387 | |
|
387 | 388 | |
|
388 | 389 | @pytest.mark.parametrize( |
|
389 | 390 | "value, expected", |
|
390 | 391 | [ |
|
391 | 392 | (")", ("invalid", None)), |
|
392 | 393 | ("]", ("invalid", None)), |
|
393 | 394 | ("}", ("invalid", None)), |
|
394 | 395 | (")(", ("invalid", None)), |
|
395 | 396 | ("][", ("invalid", None)), |
|
396 | 397 | ("}{", ("invalid", None)), |
|
397 | 398 | ("]()(", ("invalid", None)), |
|
398 | 399 | ("())(", ("invalid", None)), |
|
399 | 400 | (")[](", ("invalid", None)), |
|
400 | 401 | ("()](", ("invalid", None)), |
|
401 | 402 | ], |
|
402 | 403 | ) |
|
403 | 404 | def test_check_complete_invalidates_sunken_brackets(value, expected): |
|
404 | 405 | """ |
|
405 | 406 | Test that a single line with more closing brackets than the opening ones is |
|
406 | 407 | interpreted as invalid |
|
407 | 408 | """ |
|
408 | 409 | cc = ipt2.TransformerManager().check_complete |
|
409 | 410 | assert cc(value) == expected |
|
410 | 411 | |
|
411 | 412 | |
|
412 | 413 | def test_null_cleanup_transformer(): |
|
413 | 414 | manager = ipt2.TransformerManager() |
|
414 | 415 | manager.cleanup_transforms.insert(0, null_cleanup_transformer) |
|
415 | 416 | assert manager.transform_cell("") == "" |
|
416 | 417 | |
|
417 | 418 | |
|
418 | 419 | def test_side_effects_I(): |
|
419 | 420 | count = 0 |
|
420 | 421 | |
|
421 | 422 | def counter(lines): |
|
422 | 423 | nonlocal count |
|
423 | 424 | count += 1 |
|
424 | 425 | return lines |
|
425 | 426 | |
|
426 | 427 | counter.has_side_effects = True |
|
427 | 428 | |
|
428 | 429 | manager = ipt2.TransformerManager() |
|
429 | 430 | manager.cleanup_transforms.insert(0, counter) |
|
430 | 431 | assert manager.check_complete("a=1\n") == ("complete", None) |
|
431 | 432 | assert count == 0 |
|
432 | 433 | |
|
433 | 434 | |
|
434 | 435 | def test_side_effects_II(): |
|
435 | 436 | count = 0 |
|
436 | 437 | |
|
437 | 438 | def counter(lines): |
|
438 | 439 | nonlocal count |
|
439 | 440 | count += 1 |
|
440 | 441 | return lines |
|
441 | 442 | |
|
442 | 443 | counter.has_side_effects = True |
|
443 | 444 | |
|
444 | 445 | manager = ipt2.TransformerManager() |
|
445 | 446 | manager.line_transforms.insert(0, counter) |
|
446 | 447 | assert manager.check_complete("b=1\n") == ("complete", None) |
|
447 | 448 | assert count == 0 |
@@ -1,150 +1,155 | |||
|
1 | 1 | """Token-related utilities""" |
|
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 collections import namedtuple |
|
7 | 7 | from io import StringIO |
|
8 | 8 | from keyword import iskeyword |
|
9 | 9 | |
|
10 | 10 | import tokenize |
|
11 | 11 | |
|
12 | 12 | |
|
13 | 13 | Token = namedtuple('Token', ['token', 'text', 'start', 'end', 'line']) |
|
14 | 14 | |
|
15 | 15 | def generate_tokens(readline): |
|
16 | 16 | """wrap generate_tokens to catch EOF errors""" |
|
17 | 17 | try: |
|
18 | 18 | for token in tokenize.generate_tokens(readline): |
|
19 | 19 | yield token |
|
20 | 20 | except tokenize.TokenError: |
|
21 | 21 | # catch EOF error |
|
22 | 22 | return |
|
23 | 23 | |
|
24 | ||
|
24 | 25 | def generate_tokens_catch_errors(readline, extra_errors_to_catch=None): |
|
25 | default_errors_to_catch = ['unterminated string literal', 'invalid non-printable character', | |
|
26 | 'after line continuation character'] | |
|
26 | default_errors_to_catch = [ | |
|
27 | "unterminated string literal", | |
|
28 | "invalid non-printable character", | |
|
29 | "after line continuation character", | |
|
30 | ] | |
|
27 | 31 | assert extra_errors_to_catch is None or isinstance(extra_errors_to_catch, list) |
|
28 | 32 | errors_to_catch = default_errors_to_catch + (extra_errors_to_catch or []) |
|
29 | 33 | |
|
30 | 34 | tokens = [] |
|
31 | 35 | try: |
|
32 | 36 | for token in tokenize.generate_tokens(readline): |
|
33 | 37 | tokens.append(token) |
|
34 | 38 | yield token |
|
35 | 39 | except tokenize.TokenError as exc: |
|
36 | 40 | if any(error in exc.args[0] for error in errors_to_catch): |
|
37 | 41 | if tokens: |
|
38 | 42 | start = tokens[-1].start[0], tokens[-1].end[0] |
|
39 | 43 | end = start |
|
40 | 44 | line = tokens[-1].line |
|
41 | 45 | else: |
|
42 | 46 | start = end = (1, 0) |
|
43 |
line = |
|
|
44 |
yield tokenize.TokenInfo(tokenize.ERRORTOKEN, |
|
|
47 | line = "" | |
|
48 | yield tokenize.TokenInfo(tokenize.ERRORTOKEN, "", start, end, line) | |
|
45 | 49 | else: |
|
46 | 50 | # Catch EOF |
|
47 | 51 | raise |
|
48 | 52 | |
|
53 | ||
|
49 | 54 | def line_at_cursor(cell, cursor_pos=0): |
|
50 | 55 | """Return the line in a cell at a given cursor position |
|
51 | 56 | |
|
52 | 57 | Used for calling line-based APIs that don't support multi-line input, yet. |
|
53 | 58 | |
|
54 | 59 | Parameters |
|
55 | 60 | ---------- |
|
56 | 61 | cell : str |
|
57 | 62 | multiline block of text |
|
58 | 63 | cursor_pos : integer |
|
59 | 64 | the cursor position |
|
60 | 65 | |
|
61 | 66 | Returns |
|
62 | 67 | ------- |
|
63 | 68 | (line, offset): (string, integer) |
|
64 | 69 | The line with the current cursor, and the character offset of the start of the line. |
|
65 | 70 | """ |
|
66 | 71 | offset = 0 |
|
67 | 72 | lines = cell.splitlines(True) |
|
68 | 73 | for line in lines: |
|
69 | 74 | next_offset = offset + len(line) |
|
70 | 75 | if not line.endswith('\n'): |
|
71 | 76 | # If the last line doesn't have a trailing newline, treat it as if |
|
72 | 77 | # it does so that the cursor at the end of the line still counts |
|
73 | 78 | # as being on that line. |
|
74 | 79 | next_offset += 1 |
|
75 | 80 | if next_offset > cursor_pos: |
|
76 | 81 | break |
|
77 | 82 | offset = next_offset |
|
78 | 83 | else: |
|
79 | 84 | line = "" |
|
80 | 85 | return (line, offset) |
|
81 | 86 | |
|
82 | 87 | def token_at_cursor(cell, cursor_pos=0): |
|
83 | 88 | """Get the token at a given cursor |
|
84 | 89 | |
|
85 | 90 | Used for introspection. |
|
86 | 91 | |
|
87 | 92 | Function calls are prioritized, so the token for the callable will be returned |
|
88 | 93 | if the cursor is anywhere inside the call. |
|
89 | 94 | |
|
90 | 95 | Parameters |
|
91 | 96 | ---------- |
|
92 | 97 | cell : unicode |
|
93 | 98 | A block of Python code |
|
94 | 99 | cursor_pos : int |
|
95 | 100 | The location of the cursor in the block where the token should be found |
|
96 | 101 | """ |
|
97 | 102 | names = [] |
|
98 | 103 | tokens = [] |
|
99 | 104 | call_names = [] |
|
100 | 105 | |
|
101 | 106 | offsets = {1: 0} # lines start at 1 |
|
102 | 107 | for tup in generate_tokens(StringIO(cell).readline): |
|
103 | 108 | |
|
104 | 109 | tok = Token(*tup) |
|
105 | 110 | |
|
106 | 111 | # token, text, start, end, line = tup |
|
107 | 112 | start_line, start_col = tok.start |
|
108 | 113 | end_line, end_col = tok.end |
|
109 | 114 | if end_line + 1 not in offsets: |
|
110 | 115 | # keep track of offsets for each line |
|
111 | 116 | lines = tok.line.splitlines(True) |
|
112 | 117 | for lineno, line in enumerate(lines, start_line + 1): |
|
113 | 118 | if lineno not in offsets: |
|
114 | 119 | offsets[lineno] = offsets[lineno-1] + len(line) |
|
115 | 120 | |
|
116 | 121 | offset = offsets[start_line] |
|
117 | 122 | # allow '|foo' to find 'foo' at the beginning of a line |
|
118 | 123 | boundary = cursor_pos + 1 if start_col == 0 else cursor_pos |
|
119 | 124 | if offset + start_col >= boundary: |
|
120 | 125 | # current token starts after the cursor, |
|
121 | 126 | # don't consume it |
|
122 | 127 | break |
|
123 | 128 | |
|
124 | 129 | if tok.token == tokenize.NAME and not iskeyword(tok.text): |
|
125 | 130 | if names and tokens and tokens[-1].token == tokenize.OP and tokens[-1].text == '.': |
|
126 | 131 | names[-1] = "%s.%s" % (names[-1], tok.text) |
|
127 | 132 | else: |
|
128 | 133 | names.append(tok.text) |
|
129 | 134 | elif tok.token == tokenize.OP: |
|
130 | 135 | if tok.text == '=' and names: |
|
131 | 136 | # don't inspect the lhs of an assignment |
|
132 | 137 | names.pop(-1) |
|
133 | 138 | if tok.text == '(' and names: |
|
134 | 139 | # if we are inside a function call, inspect the function |
|
135 | 140 | call_names.append(names[-1]) |
|
136 | 141 | elif tok.text == ')' and call_names: |
|
137 | 142 | call_names.pop(-1) |
|
138 | 143 | |
|
139 | 144 | tokens.append(tok) |
|
140 | 145 | |
|
141 | 146 | if offsets[end_line] + end_col > cursor_pos: |
|
142 | 147 | # we found the cursor, stop reading |
|
143 | 148 | break |
|
144 | 149 | |
|
145 | 150 | if call_names: |
|
146 | 151 | return call_names[-1] |
|
147 | 152 | elif names: |
|
148 | 153 | return names[-1] |
|
149 | 154 | else: |
|
150 | 155 | return '' |
General Comments 0
You need to be logged in to leave comments.
Login now