Show More
@@ -1,432 +1,525 | |||
|
1 | from codeop import compile_command | |
|
1 | 2 | import re |
|
2 | 3 | from typing import List, Tuple |
|
3 | 4 | from IPython.utils import tokenize2 |
|
4 | 5 | from IPython.utils.tokenutil import generate_tokens |
|
5 | 6 | |
|
7 | _indent_re = re.compile(r'^[ \t]+') | |
|
8 | ||
|
6 | 9 | def leading_indent(lines): |
|
7 | 10 | """Remove leading indentation. |
|
8 | 11 | |
|
9 | 12 | If the first line starts with a spaces or tabs, the same whitespace will be |
|
10 | 13 | removed from each following line. |
|
11 | 14 | """ |
|
12 |
m = re.match( |
|
|
15 | m = _indent_re.match(lines[0]) | |
|
13 | 16 | if not m: |
|
14 | 17 | return lines |
|
15 | 18 | space = m.group(0) |
|
16 | 19 | n = len(space) |
|
17 | 20 | return [l[n:] if l.startswith(space) else l |
|
18 | 21 | for l in lines] |
|
19 | 22 | |
|
20 | 23 | class PromptStripper: |
|
21 | 24 | """Remove matching input prompts from a block of input. |
|
22 | 25 | |
|
23 | 26 | Parameters |
|
24 | 27 | ---------- |
|
25 | 28 | prompt_re : regular expression |
|
26 | 29 | A regular expression matching any input prompt (including continuation) |
|
27 | 30 | initial_re : regular expression, optional |
|
28 | 31 | A regular expression matching only the initial prompt, but not continuation. |
|
29 | 32 | If no initial expression is given, prompt_re will be used everywhere. |
|
30 | 33 | Used mainly for plain Python prompts, where the continuation prompt |
|
31 | 34 | ``...`` is a valid Python expression in Python 3, so shouldn't be stripped. |
|
32 | 35 | |
|
33 | 36 | If initial_re and prompt_re differ, |
|
34 | 37 | only initial_re will be tested against the first line. |
|
35 | 38 | If any prompt is found on the first two lines, |
|
36 | 39 | prompts will be stripped from the rest of the block. |
|
37 | 40 | """ |
|
38 | 41 | def __init__(self, prompt_re, initial_re=None): |
|
39 | 42 | self.prompt_re = prompt_re |
|
40 | 43 | self.initial_re = initial_re or prompt_re |
|
41 | 44 | |
|
42 | 45 | def _strip(self, lines): |
|
43 | 46 | return [self.prompt_re.sub('', l, count=1) for l in lines] |
|
44 | 47 | |
|
45 | 48 | def __call__(self, lines): |
|
46 | 49 | if self.initial_re.match(lines[0]) or \ |
|
47 | 50 | (len(lines) > 1 and self.prompt_re.match(lines[1])): |
|
48 | 51 | return self._strip(lines) |
|
49 | 52 | return lines |
|
50 | 53 | |
|
51 | 54 | classic_prompt = PromptStripper( |
|
52 | 55 | prompt_re=re.compile(r'^(>>>|\.\.\.)( |$)'), |
|
53 | 56 | initial_re=re.compile(r'^>>>( |$)') |
|
54 | 57 | ) |
|
55 | 58 | |
|
56 | 59 | ipython_prompt = PromptStripper(re.compile(r'^(In \[\d+\]: |\s*\.{3,}: ?)')) |
|
57 | 60 | |
|
58 | 61 | def cell_magic(lines): |
|
59 | 62 | if not lines[0].startswith('%%'): |
|
60 | 63 | return lines |
|
61 | 64 | if re.match('%%\w+\?', lines[0]): |
|
62 | 65 | # This case will be handled by help_end |
|
63 | 66 | return lines |
|
64 | 67 | magic_name, _, first_line = lines[0][2:-1].partition(' ') |
|
65 | 68 | body = ''.join(lines[1:]) |
|
66 | 69 | return ['get_ipython().run_cell_magic(%r, %r, %r)\n' |
|
67 | 70 | % (magic_name, first_line, body)] |
|
68 | 71 | |
|
69 | 72 | # ----- |
|
70 | 73 | |
|
71 | 74 | def _find_assign_op(token_line): |
|
72 | 75 | # Find the first assignment in the line ('=' not inside brackets) |
|
73 | 76 | # We don't try to support multiple special assignment (a = b = %foo) |
|
74 | 77 | paren_level = 0 |
|
75 | 78 | for i, ti in enumerate(token_line): |
|
76 | 79 | s = ti.string |
|
77 | 80 | if s == '=' and paren_level == 0: |
|
78 | 81 | return i |
|
79 | 82 | if s in '([{': |
|
80 | 83 | paren_level += 1 |
|
81 | 84 | elif s in ')]}': |
|
82 | 85 | paren_level -= 1 |
|
83 | 86 | |
|
84 | 87 | def find_end_of_continued_line(lines, start_line: int): |
|
85 | 88 | """Find the last line of a line explicitly extended using backslashes. |
|
86 | 89 | |
|
87 | 90 | Uses 0-indexed line numbers. |
|
88 | 91 | """ |
|
89 | 92 | end_line = start_line |
|
90 | 93 | while lines[end_line].endswith('\\\n'): |
|
91 | 94 | end_line += 1 |
|
92 | 95 | if end_line >= len(lines): |
|
93 | 96 | break |
|
94 | 97 | return end_line |
|
95 | 98 | |
|
96 | 99 | def assemble_continued_line(lines, start: Tuple[int, int], end_line: int): |
|
97 | 100 | """Assemble pieces of a continued line into a single line. |
|
98 | 101 | |
|
99 | 102 | Uses 0-indexed line numbers. *start* is (lineno, colno). |
|
100 | 103 | """ |
|
101 | 104 | parts = [lines[start[0]][start[1]:]] + lines[start[0]+1:end_line+1] |
|
102 | 105 | return ' '.join([p[:-2] for p in parts[:-1]] # Strip backslash+newline |
|
103 | 106 | + [parts[-1][:-1]]) # Strip newline from last line |
|
104 | 107 | |
|
105 | 108 | class TokenTransformBase: |
|
106 | 109 | # Lower numbers -> higher priority (for matches in the same location) |
|
107 | 110 | priority = 10 |
|
108 | 111 | |
|
109 | 112 | def sortby(self): |
|
110 | 113 | return self.start_line, self.start_col, self.priority |
|
111 | 114 | |
|
112 | 115 | def __init__(self, start): |
|
113 | 116 | self.start_line = start[0] - 1 # Shift from 1-index to 0-index |
|
114 | 117 | self.start_col = start[1] |
|
115 | 118 | |
|
116 | 119 | def transform(self, lines: List[str]): |
|
117 | 120 | raise NotImplementedError |
|
118 | 121 | |
|
119 | 122 | class MagicAssign(TokenTransformBase): |
|
120 | 123 | @classmethod |
|
121 | 124 | def find(cls, tokens_by_line): |
|
122 | 125 | """Find the first magic assignment (a = %foo) in the cell. |
|
123 | 126 | |
|
124 | 127 | Returns (line, column) of the % if found, or None. *line* is 1-indexed. |
|
125 | 128 | """ |
|
126 | 129 | for line in tokens_by_line: |
|
127 | 130 | assign_ix = _find_assign_op(line) |
|
128 | 131 | if (assign_ix is not None) \ |
|
129 | 132 | and (len(line) >= assign_ix + 2) \ |
|
130 | 133 | and (line[assign_ix+1].string == '%') \ |
|
131 | 134 | and (line[assign_ix+2].type == tokenize2.NAME): |
|
132 | 135 | return cls(line[assign_ix+1].start) |
|
133 | 136 | |
|
134 | 137 | def transform(self, lines: List[str]): |
|
135 | 138 | """Transform a magic assignment found by find |
|
136 | 139 | """ |
|
137 | 140 | start_line, start_col = self.start_line, self.start_col |
|
138 | 141 | lhs = lines[start_line][:start_col] |
|
139 | 142 | end_line = find_end_of_continued_line(lines, start_line) |
|
140 | 143 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
141 | 144 | assert rhs.startswith('%'), rhs |
|
142 | 145 | magic_name, _, args = rhs[1:].partition(' ') |
|
143 | 146 | |
|
144 | 147 | lines_before = lines[:start_line] |
|
145 | 148 | call = "get_ipython().run_line_magic({!r}, {!r})".format(magic_name, args) |
|
146 | 149 | new_line = lhs + call + '\n' |
|
147 | 150 | lines_after = lines[end_line+1:] |
|
148 | 151 | |
|
149 | 152 | return lines_before + [new_line] + lines_after |
|
150 | 153 | |
|
151 | 154 | |
|
152 | 155 | class SystemAssign(TokenTransformBase): |
|
153 | 156 | @classmethod |
|
154 | 157 | def find(cls, tokens_by_line): |
|
155 | 158 | """Find the first system assignment (a = !foo) in the cell. |
|
156 | 159 | |
|
157 | 160 | Returns (line, column) of the ! if found, or None. *line* is 1-indexed. |
|
158 | 161 | """ |
|
159 | 162 | for line in tokens_by_line: |
|
160 | 163 | assign_ix = _find_assign_op(line) |
|
161 | 164 | if (assign_ix is not None) \ |
|
162 | 165 | and (len(line) >= assign_ix + 2) \ |
|
163 | 166 | and (line[assign_ix + 1].type == tokenize2.ERRORTOKEN): |
|
164 | 167 | ix = assign_ix + 1 |
|
165 | 168 | |
|
166 | 169 | while ix < len(line) and line[ix].type == tokenize2.ERRORTOKEN: |
|
167 | 170 | if line[ix].string == '!': |
|
168 | 171 | return cls(line[ix].start) |
|
169 | 172 | elif not line[ix].string.isspace(): |
|
170 | 173 | break |
|
171 | 174 | ix += 1 |
|
172 | 175 | |
|
173 | 176 | def transform(self, lines: List[str]): |
|
174 | 177 | """Transform a system assignment found by find |
|
175 | 178 | """ |
|
176 | 179 | start_line, start_col = self.start_line, self.start_col |
|
177 | 180 | |
|
178 | 181 | lhs = lines[start_line][:start_col] |
|
179 | 182 | end_line = find_end_of_continued_line(lines, start_line) |
|
180 | 183 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
181 | 184 | assert rhs.startswith('!'), rhs |
|
182 | 185 | cmd = rhs[1:] |
|
183 | 186 | |
|
184 | 187 | lines_before = lines[:start_line] |
|
185 | 188 | call = "get_ipython().getoutput({!r})".format(cmd) |
|
186 | 189 | new_line = lhs + call + '\n' |
|
187 | 190 | lines_after = lines[end_line + 1:] |
|
188 | 191 | |
|
189 | 192 | return lines_before + [new_line] + lines_after |
|
190 | 193 | |
|
191 | 194 | # The escape sequences that define the syntax transformations IPython will |
|
192 | 195 | # apply to user input. These can NOT be just changed here: many regular |
|
193 | 196 | # expressions and other parts of the code may use their hardcoded values, and |
|
194 | 197 | # for all intents and purposes they constitute the 'IPython syntax', so they |
|
195 | 198 | # should be considered fixed. |
|
196 | 199 | |
|
197 | 200 | ESC_SHELL = '!' # Send line to underlying system shell |
|
198 | 201 | ESC_SH_CAP = '!!' # Send line to system shell and capture output |
|
199 | 202 | ESC_HELP = '?' # Find information about object |
|
200 | 203 | ESC_HELP2 = '??' # Find extra-detailed information about object |
|
201 | 204 | ESC_MAGIC = '%' # Call magic function |
|
202 | 205 | ESC_MAGIC2 = '%%' # Call cell-magic function |
|
203 | 206 | ESC_QUOTE = ',' # Split args on whitespace, quote each as string and call |
|
204 | 207 | ESC_QUOTE2 = ';' # Quote all args as a single string, call |
|
205 | 208 | ESC_PAREN = '/' # Call first argument with rest of line as arguments |
|
206 | 209 | |
|
207 | 210 | ESCAPE_SINGLES = {'!', '?', '%', ',', ';', '/'} |
|
208 | 211 | ESCAPE_DOUBLES = {'!!', '??'} # %% (cell magic) is handled separately |
|
209 | 212 | |
|
210 | 213 | def _make_help_call(target, esc, next_input=None): |
|
211 | 214 | """Prepares a pinfo(2)/psearch call from a target name and the escape |
|
212 | 215 | (i.e. ? or ??)""" |
|
213 | 216 | method = 'pinfo2' if esc == '??' \ |
|
214 | 217 | else 'psearch' if '*' in target \ |
|
215 | 218 | else 'pinfo' |
|
216 | 219 | arg = " ".join([method, target]) |
|
217 | 220 | #Prepare arguments for get_ipython().run_line_magic(magic_name, magic_args) |
|
218 | 221 | t_magic_name, _, t_magic_arg_s = arg.partition(' ') |
|
219 | 222 | t_magic_name = t_magic_name.lstrip(ESC_MAGIC) |
|
220 | 223 | if next_input is None: |
|
221 | 224 | return 'get_ipython().run_line_magic(%r, %r)' % (t_magic_name, t_magic_arg_s) |
|
222 | 225 | else: |
|
223 | 226 | return 'get_ipython().set_next_input(%r);get_ipython().run_line_magic(%r, %r)' % \ |
|
224 | 227 | (next_input, t_magic_name, t_magic_arg_s) |
|
225 | 228 | |
|
226 | 229 | def _tr_help(content): |
|
227 | 230 | "Translate lines escaped with: ?" |
|
228 | 231 | # A naked help line should just fire the intro help screen |
|
229 | 232 | if not content: |
|
230 | 233 | return 'get_ipython().show_usage()' |
|
231 | 234 | |
|
232 | 235 | return _make_help_call(content, '?') |
|
233 | 236 | |
|
234 | 237 | def _tr_help2(content): |
|
235 | 238 | "Translate lines escaped with: ??" |
|
236 | 239 | # A naked help line should just fire the intro help screen |
|
237 | 240 | if not content: |
|
238 | 241 | return 'get_ipython().show_usage()' |
|
239 | 242 | |
|
240 | 243 | return _make_help_call(content, '??') |
|
241 | 244 | |
|
242 | 245 | def _tr_magic(content): |
|
243 | 246 | "Translate lines escaped with: %" |
|
244 | 247 | name, _, args = content.partition(' ') |
|
245 | 248 | return 'get_ipython().run_line_magic(%r, %r)' % (name, args) |
|
246 | 249 | |
|
247 | 250 | def _tr_quote(content): |
|
248 | 251 | "Translate lines escaped with: ," |
|
249 | 252 | name, _, args = content.partition(' ') |
|
250 | 253 | return '%s("%s")' % (name, '", "'.join(args.split()) ) |
|
251 | 254 | |
|
252 | 255 | def _tr_quote2(content): |
|
253 | 256 | "Translate lines escaped with: ;" |
|
254 | 257 | name, _, args = content.partition(' ') |
|
255 | 258 | return '%s("%s")' % (name, args) |
|
256 | 259 | |
|
257 | 260 | def _tr_paren(content): |
|
258 | 261 | "Translate lines escaped with: /" |
|
259 | 262 | name, _, args = content.partition(' ') |
|
260 | 263 | return '%s(%s)' % (name, ", ".join(args.split())) |
|
261 | 264 | |
|
262 | 265 | tr = { ESC_SHELL : 'get_ipython().system({!r})'.format, |
|
263 | 266 | ESC_SH_CAP : 'get_ipython().getoutput({!r})'.format, |
|
264 | 267 | ESC_HELP : _tr_help, |
|
265 | 268 | ESC_HELP2 : _tr_help2, |
|
266 | 269 | ESC_MAGIC : _tr_magic, |
|
267 | 270 | ESC_QUOTE : _tr_quote, |
|
268 | 271 | ESC_QUOTE2 : _tr_quote2, |
|
269 | 272 | ESC_PAREN : _tr_paren } |
|
270 | 273 | |
|
271 | 274 | class EscapedCommand(TokenTransformBase): |
|
272 | 275 | @classmethod |
|
273 | 276 | def find(cls, tokens_by_line): |
|
274 | 277 | """Find the first escaped command (%foo, !foo, etc.) in the cell. |
|
275 | 278 | |
|
276 | 279 | Returns (line, column) of the escape if found, or None. *line* is 1-indexed. |
|
277 | 280 | """ |
|
278 | 281 | for line in tokens_by_line: |
|
279 | 282 | ix = 0 |
|
280 | 283 | while line[ix].type in {tokenize2.INDENT, tokenize2.DEDENT}: |
|
281 | 284 | ix += 1 |
|
282 | 285 | if line[ix].string in ESCAPE_SINGLES: |
|
283 | 286 | return cls(line[ix].start) |
|
284 | 287 | |
|
285 | 288 | def transform(self, lines): |
|
286 | 289 | start_line, start_col = self.start_line, self.start_col |
|
287 | 290 | |
|
288 | 291 | indent = lines[start_line][:start_col] |
|
289 | 292 | end_line = find_end_of_continued_line(lines, start_line) |
|
290 | 293 | line = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
291 | 294 | |
|
292 | 295 | if line[:2] in ESCAPE_DOUBLES: |
|
293 | 296 | escape, content = line[:2], line[2:] |
|
294 | 297 | else: |
|
295 | 298 | escape, content = line[:1], line[1:] |
|
296 | 299 | call = tr[escape](content) |
|
297 | 300 | |
|
298 | 301 | lines_before = lines[:start_line] |
|
299 | 302 | new_line = indent + call + '\n' |
|
300 | 303 | lines_after = lines[end_line + 1:] |
|
301 | 304 | |
|
302 | 305 | return lines_before + [new_line] + lines_after |
|
303 | 306 | |
|
304 | 307 | _help_end_re = re.compile(r"""(%{0,2} |
|
305 | 308 | [a-zA-Z_*][\w*]* # Variable name |
|
306 | 309 | (\.[a-zA-Z_*][\w*]*)* # .etc.etc |
|
307 | 310 | ) |
|
308 | 311 | (\?\??)$ # ? or ?? |
|
309 | 312 | """, |
|
310 | 313 | re.VERBOSE) |
|
311 | 314 | |
|
312 | 315 | class HelpEnd(TokenTransformBase): |
|
313 | 316 | # This needs to be higher priority (lower number) than EscapedCommand so |
|
314 | 317 | # that inspecting magics (%foo?) works. |
|
315 | 318 | priority = 5 |
|
316 | 319 | |
|
317 | 320 | def __init__(self, start, q_locn): |
|
318 | 321 | super().__init__(start) |
|
319 | 322 | self.q_line = q_locn[0] - 1 # Shift from 1-indexed to 0-indexed |
|
320 | 323 | self.q_col = q_locn[1] |
|
321 | 324 | |
|
322 | 325 | @classmethod |
|
323 | 326 | def find(cls, tokens_by_line): |
|
324 | 327 | for line in tokens_by_line: |
|
325 | 328 | # Last token is NEWLINE; look at last but one |
|
326 | 329 | if len(line) > 2 and line[-2].string == '?': |
|
327 | 330 | # Find the first token that's not INDENT/DEDENT |
|
328 | 331 | ix = 0 |
|
329 | 332 | while line[ix].type in {tokenize2.INDENT, tokenize2.DEDENT}: |
|
330 | 333 | ix += 1 |
|
331 | 334 | return cls(line[ix].start, line[-2].start) |
|
332 | 335 | |
|
333 | 336 | def transform(self, lines): |
|
334 | 337 | piece = ''.join(lines[self.start_line:self.q_line+1]) |
|
335 | 338 | indent, content = piece[:self.start_col], piece[self.start_col:] |
|
336 | 339 | lines_before = lines[:self.start_line] |
|
337 | 340 | lines_after = lines[self.q_line + 1:] |
|
338 | 341 | |
|
339 | 342 | m = _help_end_re.search(content) |
|
340 | 343 | assert m is not None, content |
|
341 | 344 | target = m.group(1) |
|
342 | 345 | esc = m.group(3) |
|
343 | 346 | |
|
344 | 347 | # If we're mid-command, put it back on the next prompt for the user. |
|
345 | 348 | next_input = None |
|
346 | 349 | if (not lines_before) and (not lines_after) \ |
|
347 | 350 | and content.strip() != m.group(0): |
|
348 | 351 | next_input = content.rstrip('?\n') |
|
349 | 352 | |
|
350 | 353 | call = _make_help_call(target, esc, next_input=next_input) |
|
351 | 354 | new_line = indent + call + '\n' |
|
352 | 355 | |
|
353 | 356 | return lines_before + [new_line] + lines_after |
|
354 | 357 | |
|
355 | 358 | def make_tokens_by_line(lines): |
|
356 | 359 | tokens_by_line = [[]] |
|
357 | 360 | for token in generate_tokens(iter(lines).__next__): |
|
358 | 361 | tokens_by_line[-1].append(token) |
|
359 | 362 | if token.type == tokenize2.NEWLINE: |
|
360 | 363 | tokens_by_line.append([]) |
|
361 | 364 | |
|
362 | 365 | return tokens_by_line |
|
363 | 366 | |
|
364 | 367 | def show_linewise_tokens(s: str): |
|
365 | 368 | """For investigation""" |
|
366 | 369 | if not s.endswith('\n'): |
|
367 | 370 | s += '\n' |
|
368 | 371 | lines = s.splitlines(keepends=True) |
|
369 | 372 | for line in make_tokens_by_line(lines): |
|
370 | 373 | print("Line -------") |
|
371 | 374 | for tokinfo in line: |
|
372 | 375 | print(" ", tokinfo) |
|
373 | 376 | |
|
374 | 377 | class TransformerManager: |
|
375 | 378 | def __init__(self): |
|
376 |
self. |
|
|
379 | self.cleanup_transforms = [ | |
|
377 | 380 | leading_indent, |
|
378 | 381 | classic_prompt, |
|
379 | 382 | ipython_prompt, |
|
383 | ] | |
|
384 | self.line_transforms = [ | |
|
380 | 385 | cell_magic, |
|
381 | 386 | ] |
|
382 | 387 | self.token_transformers = [ |
|
383 | 388 | MagicAssign, |
|
384 | 389 | SystemAssign, |
|
385 | 390 | EscapedCommand, |
|
386 | 391 | HelpEnd, |
|
387 | 392 | ] |
|
388 | 393 | |
|
389 | 394 | def do_one_token_transform(self, lines): |
|
390 | 395 | """Find and run the transform earliest in the code. |
|
391 | 396 | |
|
392 | 397 | Returns (changed, lines). |
|
393 | 398 | |
|
394 | 399 | This method is called repeatedly until changed is False, indicating |
|
395 | 400 | that all available transformations are complete. |
|
396 | 401 | |
|
397 | 402 | The tokens following IPython special syntax might not be valid, so |
|
398 | 403 | the transformed code is retokenised every time to identify the next |
|
399 | 404 | piece of special syntax. Hopefully long code cells are mostly valid |
|
400 | 405 | Python, not using lots of IPython special syntax, so this shouldn't be |
|
401 | 406 | a performance issue. |
|
402 | 407 | """ |
|
403 | 408 | tokens_by_line = make_tokens_by_line(lines) |
|
404 | 409 | candidates = [] |
|
405 | 410 | for transformer_cls in self.token_transformers: |
|
406 | 411 | transformer = transformer_cls.find(tokens_by_line) |
|
407 | 412 | if transformer: |
|
408 | 413 | candidates.append(transformer) |
|
409 | 414 | |
|
410 | 415 | if not candidates: |
|
411 | 416 | # Nothing to transform |
|
412 | 417 | return False, lines |
|
413 | 418 | |
|
414 | 419 | transformer = min(candidates, key=TokenTransformBase.sortby) |
|
415 | 420 | return True, transformer.transform(lines) |
|
416 | 421 | |
|
417 | 422 | def do_token_transforms(self, lines): |
|
418 | 423 | while True: |
|
419 | 424 | changed, lines = self.do_one_token_transform(lines) |
|
420 | 425 | if not changed: |
|
421 | 426 | return lines |
|
422 | 427 | |
|
423 | 428 | def transform_cell(self, cell: str): |
|
424 | 429 | if not cell.endswith('\n'): |
|
425 | 430 | cell += '\n' # Ensure every line has a newline |
|
426 | 431 | lines = cell.splitlines(keepends=True) |
|
427 | for transform in self.line_transforms: | |
|
432 | for transform in self.cleanup_transforms + self.line_transforms: | |
|
428 | 433 | #print(transform, lines) |
|
429 | 434 | lines = transform(lines) |
|
430 | 435 | |
|
431 | 436 | lines = self.do_token_transforms(lines) |
|
432 | 437 | return ''.join(lines) |
|
438 | ||
|
439 | def check_complete(self, cell: str): | |
|
440 | """Return whether a block of code is ready to execute, or should be continued | |
|
441 | ||
|
442 | Parameters | |
|
443 | ---------- | |
|
444 | source : string | |
|
445 | Python input code, which can be multiline. | |
|
446 | ||
|
447 | Returns | |
|
448 | ------- | |
|
449 | status : str | |
|
450 | One of 'complete', 'incomplete', or 'invalid' if source is not a | |
|
451 | prefix of valid code. | |
|
452 | indent_spaces : int or None | |
|
453 | The number of spaces by which to indent the next line of code. If | |
|
454 | status is not 'incomplete', this is None. | |
|
455 | """ | |
|
456 | if not cell.endswith('\n'): | |
|
457 | cell += '\n' # Ensure every line has a newline | |
|
458 | lines = cell.splitlines(keepends=True) | |
|
459 | if cell.rstrip().endswith('\\'): | |
|
460 | # Explicit backslash continuation | |
|
461 | return 'incomplete', find_last_indent(lines) | |
|
462 | ||
|
463 | try: | |
|
464 | for transform in self.cleanup_transforms: | |
|
465 | lines = transform(lines) | |
|
466 | except SyntaxError: | |
|
467 | return 'invalid', None | |
|
468 | ||
|
469 | if lines[0].startswith('%%'): | |
|
470 | # Special case for cell magics - completion marked by blank line | |
|
471 | if lines[-1].strip(): | |
|
472 | return 'incomplete', find_last_indent(lines) | |
|
473 | else: | |
|
474 | return 'complete', None | |
|
475 | ||
|
476 | try: | |
|
477 | for transform in self.line_transforms: | |
|
478 | lines = transform(lines) | |
|
479 | lines = self.do_token_transforms(lines) | |
|
480 | except SyntaxError: | |
|
481 | return 'invalid', None | |
|
482 | ||
|
483 | tokens_by_line = make_tokens_by_line(lines) | |
|
484 | if tokens_by_line[-1][-1].type != tokenize2.ENDMARKER: | |
|
485 | # We're in a multiline string or expression | |
|
486 | return 'incomplete', find_last_indent(lines) | |
|
487 | ||
|
488 | # Find the last token on the previous line that's not NEWLINE or COMMENT | |
|
489 | toks_last_line = tokens_by_line[-2] | |
|
490 | ix = len(toks_last_line) - 1 | |
|
491 | while ix >= 0 and toks_last_line[ix].type in {tokenize2.NEWLINE, | |
|
492 | tokenize2.COMMENT}: | |
|
493 | ix -= 1 | |
|
494 | ||
|
495 | if toks_last_line[ix].string == ':': | |
|
496 | # The last line starts a block (e.g. 'if foo:') | |
|
497 | ix = 0 | |
|
498 | while toks_last_line[ix].type in {tokenize2.INDENT, tokenize2.DEDENT}: | |
|
499 | ix += 1 | |
|
500 | indent = toks_last_line[ix].start[1] | |
|
501 | return 'incomplete', indent + 4 | |
|
502 | ||
|
503 | # If there's a blank line at the end, assume we're ready to execute. | |
|
504 | if not lines[-1].strip(): | |
|
505 | return 'complete', None | |
|
506 | ||
|
507 | # At this point, our checks think the code is complete (or invalid). | |
|
508 | # We'll use codeop.compile_command to check this with the real parser. | |
|
509 | ||
|
510 | try: | |
|
511 | res = compile_command(''.join(lines), symbol='exec') | |
|
512 | except (SyntaxError, OverflowError, ValueError, TypeError, | |
|
513 | MemoryError, SyntaxWarning): | |
|
514 | return 'invalid', None | |
|
515 | else: | |
|
516 | if res is None: | |
|
517 | return 'incomplete', find_last_indent(lines) | |
|
518 | return 'complete', None | |
|
519 | ||
|
520 | ||
|
521 | def find_last_indent(lines): | |
|
522 | m = _indent_re.match(lines[-1]) | |
|
523 | if not m: | |
|
524 | return 0 | |
|
525 | return len(m.group(0).replace('\t', ' '*4)) |
@@ -1,179 +1,188 | |||
|
1 | 1 | import nose.tools as nt |
|
2 | 2 | |
|
3 | 3 | from IPython.core import inputtransformer2 as ipt2 |
|
4 | 4 | from IPython.core.inputtransformer2 import make_tokens_by_line |
|
5 | 5 | |
|
6 | 6 | MULTILINE_MAGIC = ("""\ |
|
7 | 7 | a = f() |
|
8 | 8 | %foo \\ |
|
9 | 9 | bar |
|
10 | 10 | g() |
|
11 | 11 | """.splitlines(keepends=True), (2, 0), """\ |
|
12 | 12 | a = f() |
|
13 | 13 | get_ipython().run_line_magic('foo', ' bar') |
|
14 | 14 | g() |
|
15 | 15 | """.splitlines(keepends=True)) |
|
16 | 16 | |
|
17 | 17 | INDENTED_MAGIC = ("""\ |
|
18 | 18 | for a in range(5): |
|
19 | 19 | %ls |
|
20 | 20 | """.splitlines(keepends=True), (2, 4), """\ |
|
21 | 21 | for a in range(5): |
|
22 | 22 | get_ipython().run_line_magic('ls', '') |
|
23 | 23 | """.splitlines(keepends=True)) |
|
24 | 24 | |
|
25 | 25 | MULTILINE_MAGIC_ASSIGN = ("""\ |
|
26 | 26 | a = f() |
|
27 | 27 | b = %foo \\ |
|
28 | 28 | bar |
|
29 | 29 | g() |
|
30 | 30 | """.splitlines(keepends=True), (2, 4), """\ |
|
31 | 31 | a = f() |
|
32 | 32 | b = get_ipython().run_line_magic('foo', ' bar') |
|
33 | 33 | g() |
|
34 | 34 | """.splitlines(keepends=True)) |
|
35 | 35 | |
|
36 | 36 | MULTILINE_SYSTEM_ASSIGN = ("""\ |
|
37 | 37 | a = f() |
|
38 | 38 | b = !foo \\ |
|
39 | 39 | bar |
|
40 | 40 | g() |
|
41 | 41 | """.splitlines(keepends=True), (2, 4), """\ |
|
42 | 42 | a = f() |
|
43 | 43 | b = get_ipython().getoutput('foo bar') |
|
44 | 44 | g() |
|
45 | 45 | """.splitlines(keepends=True)) |
|
46 | 46 | |
|
47 | 47 | AUTOCALL_QUOTE = ( |
|
48 | 48 | [",f 1 2 3\n"], (1, 0), |
|
49 | 49 | ['f("1", "2", "3")\n'] |
|
50 | 50 | ) |
|
51 | 51 | |
|
52 | 52 | AUTOCALL_QUOTE2 = ( |
|
53 | 53 | [";f 1 2 3\n"], (1, 0), |
|
54 | 54 | ['f("1 2 3")\n'] |
|
55 | 55 | ) |
|
56 | 56 | |
|
57 | 57 | AUTOCALL_PAREN = ( |
|
58 | 58 | ["/f 1 2 3\n"], (1, 0), |
|
59 | 59 | ['f(1, 2, 3)\n'] |
|
60 | 60 | ) |
|
61 | 61 | |
|
62 | 62 | SIMPLE_HELP = ( |
|
63 | 63 | ["foo?\n"], (1, 0), |
|
64 | 64 | ["get_ipython().run_line_magic('pinfo', 'foo')\n"] |
|
65 | 65 | ) |
|
66 | 66 | |
|
67 | 67 | DETAILED_HELP = ( |
|
68 | 68 | ["foo??\n"], (1, 0), |
|
69 | 69 | ["get_ipython().run_line_magic('pinfo2', 'foo')\n"] |
|
70 | 70 | ) |
|
71 | 71 | |
|
72 | 72 | MAGIC_HELP = ( |
|
73 | 73 | ["%foo?\n"], (1, 0), |
|
74 | 74 | ["get_ipython().run_line_magic('pinfo', '%foo')\n"] |
|
75 | 75 | ) |
|
76 | 76 | |
|
77 | 77 | HELP_IN_EXPR = ( |
|
78 | 78 | ["a = b + c?\n"], (1, 0), |
|
79 | 79 | ["get_ipython().set_next_input('a = b + c');" |
|
80 | 80 | "get_ipython().run_line_magic('pinfo', 'c')\n"] |
|
81 | 81 | ) |
|
82 | 82 | |
|
83 | 83 | HELP_CONTINUED_LINE = ("""\ |
|
84 | 84 | a = \\ |
|
85 | 85 | zip? |
|
86 | 86 | """.splitlines(keepends=True), (1, 0), |
|
87 | 87 | [r"get_ipython().set_next_input('a = \\\nzip');get_ipython().run_line_magic('pinfo', 'zip')" + "\n"] |
|
88 | 88 | ) |
|
89 | 89 | |
|
90 | 90 | HELP_MULTILINE = ("""\ |
|
91 | 91 | (a, |
|
92 | 92 | b) = zip? |
|
93 | 93 | """.splitlines(keepends=True), (1, 0), |
|
94 | 94 | [r"get_ipython().set_next_input('(a,\nb) = zip');get_ipython().run_line_magic('pinfo', 'zip')" + "\n"] |
|
95 | 95 | ) |
|
96 | 96 | |
|
97 | 97 | def check_find(transformer, case, match=True): |
|
98 | 98 | sample, expected_start, _ = case |
|
99 | 99 | tbl = make_tokens_by_line(sample) |
|
100 | 100 | res = transformer.find(tbl) |
|
101 | 101 | if match: |
|
102 | 102 | # start_line is stored 0-indexed, expected values are 1-indexed |
|
103 | 103 | nt.assert_equal((res.start_line+1, res.start_col), expected_start) |
|
104 | 104 | return res |
|
105 | 105 | else: |
|
106 | 106 | nt.assert_is(res, None) |
|
107 | 107 | |
|
108 | 108 | def check_transform(transformer_cls, case): |
|
109 | 109 | lines, start, expected = case |
|
110 | 110 | transformer = transformer_cls(start) |
|
111 | 111 | nt.assert_equal(transformer.transform(lines), expected) |
|
112 | 112 | |
|
113 | 113 | def test_continued_line(): |
|
114 | 114 | lines = MULTILINE_MAGIC_ASSIGN[0] |
|
115 | 115 | nt.assert_equal(ipt2.find_end_of_continued_line(lines, 1), 2) |
|
116 | 116 | |
|
117 | 117 | nt.assert_equal(ipt2.assemble_continued_line(lines, (1, 5), 2), "foo bar") |
|
118 | 118 | |
|
119 | 119 | def test_find_assign_magic(): |
|
120 | 120 | check_find(ipt2.MagicAssign, MULTILINE_MAGIC_ASSIGN) |
|
121 | 121 | check_find(ipt2.MagicAssign, MULTILINE_SYSTEM_ASSIGN, match=False) |
|
122 | 122 | |
|
123 | 123 | def test_transform_assign_magic(): |
|
124 | 124 | check_transform(ipt2.MagicAssign, MULTILINE_MAGIC_ASSIGN) |
|
125 | 125 | |
|
126 | 126 | def test_find_assign_system(): |
|
127 | 127 | check_find(ipt2.SystemAssign, MULTILINE_SYSTEM_ASSIGN) |
|
128 | 128 | check_find(ipt2.SystemAssign, (["a = !ls\n"], (1, 5), None)) |
|
129 | 129 | check_find(ipt2.SystemAssign, (["a=!ls\n"], (1, 2), None)) |
|
130 | 130 | check_find(ipt2.SystemAssign, MULTILINE_MAGIC_ASSIGN, match=False) |
|
131 | 131 | |
|
132 | 132 | def test_transform_assign_system(): |
|
133 | 133 | check_transform(ipt2.SystemAssign, MULTILINE_SYSTEM_ASSIGN) |
|
134 | 134 | |
|
135 | 135 | def test_find_magic_escape(): |
|
136 | 136 | check_find(ipt2.EscapedCommand, MULTILINE_MAGIC) |
|
137 | 137 | check_find(ipt2.EscapedCommand, INDENTED_MAGIC) |
|
138 | 138 | check_find(ipt2.EscapedCommand, MULTILINE_MAGIC_ASSIGN, match=False) |
|
139 | 139 | |
|
140 | 140 | def test_transform_magic_escape(): |
|
141 | 141 | check_transform(ipt2.EscapedCommand, MULTILINE_MAGIC) |
|
142 | 142 | check_transform(ipt2.EscapedCommand, INDENTED_MAGIC) |
|
143 | 143 | |
|
144 | 144 | def test_find_autocalls(): |
|
145 | 145 | for case in [AUTOCALL_QUOTE, AUTOCALL_QUOTE2, AUTOCALL_PAREN]: |
|
146 | 146 | print("Testing %r" % case[0]) |
|
147 | 147 | check_find(ipt2.EscapedCommand, case) |
|
148 | 148 | |
|
149 | 149 | def test_transform_autocall(): |
|
150 | 150 | for case in [AUTOCALL_QUOTE, AUTOCALL_QUOTE2, AUTOCALL_PAREN]: |
|
151 | 151 | print("Testing %r" % case[0]) |
|
152 | 152 | check_transform(ipt2.EscapedCommand, case) |
|
153 | 153 | |
|
154 | 154 | def test_find_help(): |
|
155 | 155 | for case in [SIMPLE_HELP, DETAILED_HELP, MAGIC_HELP, HELP_IN_EXPR]: |
|
156 | 156 | check_find(ipt2.HelpEnd, case) |
|
157 | 157 | |
|
158 | 158 | tf = check_find(ipt2.HelpEnd, HELP_CONTINUED_LINE) |
|
159 | 159 | nt.assert_equal(tf.q_line, 1) |
|
160 | 160 | nt.assert_equal(tf.q_col, 3) |
|
161 | 161 | |
|
162 | 162 | tf = check_find(ipt2.HelpEnd, HELP_MULTILINE) |
|
163 | 163 | nt.assert_equal(tf.q_line, 1) |
|
164 | 164 | nt.assert_equal(tf.q_col, 8) |
|
165 | 165 | |
|
166 | 166 | # ? in a comment does not trigger help |
|
167 | 167 | check_find(ipt2.HelpEnd, (["foo # bar?\n"], None, None), match=False) |
|
168 | 168 | # Nor in a string |
|
169 | 169 | check_find(ipt2.HelpEnd, (["foo = '''bar?\n"], None, None), match=False) |
|
170 | 170 | |
|
171 | 171 | def test_transform_help(): |
|
172 | 172 | tf = ipt2.HelpEnd((1, 0), (1, 9)) |
|
173 | 173 | nt.assert_equal(tf.transform(HELP_IN_EXPR[0]), HELP_IN_EXPR[2]) |
|
174 | 174 | |
|
175 | 175 | tf = ipt2.HelpEnd((1, 0), (2, 3)) |
|
176 | 176 | nt.assert_equal(tf.transform(HELP_CONTINUED_LINE[0]), HELP_CONTINUED_LINE[2]) |
|
177 | 177 | |
|
178 | 178 | tf = ipt2.HelpEnd((1, 0), (2, 8)) |
|
179 | 179 | nt.assert_equal(tf.transform(HELP_MULTILINE[0]), HELP_MULTILINE[2]) |
|
180 | ||
|
181 | def test_check_complete(): | |
|
182 | tm = ipt2.TransformerManager() | |
|
183 | nt.assert_equal(tm.check_complete("a = 1"), ('complete', None)) | |
|
184 | nt.assert_equal(tm.check_complete("for a in range(5):"), ('incomplete', 4)) | |
|
185 | nt.assert_equal(tm.check_complete("raise = 2"), ('invalid', None)) | |
|
186 | nt.assert_equal(tm.check_complete("a = [1,\n2,"), ('incomplete', 0)) | |
|
187 | nt.assert_equal(tm.check_complete("a = '''\n hi"), ('incomplete', 3)) | |
|
188 | nt.assert_equal(tm.check_complete("def a():\n x=1\n global x"), ('invalid', None)) |
General Comments 0
You need to be logged in to leave comments.
Login now