]>
Commit | Line | Data |
---|---|---|
1 | """ | |
2 | This is the syntax tree for Python 3 syntaxes. The classes represent | |
3 | syntax elements like functions and imports. | |
4 | ||
5 | All of the nodes can be traced back to the `Python grammar file | |
6 | <https://docs.python.org/3/reference/grammar.html>`_. If you want to know how | |
7 | a tree is structured, just analyse that file (for each Python version it's a | |
8 | bit different). | |
9 | ||
10 | There's a lot of logic here that makes it easier for Jedi (and other libraries) | |
11 | to deal with a Python syntax tree. | |
12 | ||
13 | By using :py:meth:`parso.tree.NodeOrLeaf.get_code` on a module, you can get | |
14 | back the 1-to-1 representation of the input given to the parser. This is | |
15 | important if you want to refactor a parser tree. | |
16 | ||
17 | >>> from parso import parse | |
18 | >>> parser = parse('import os') | |
19 | >>> module = parser.get_root_node() | |
20 | >>> module | |
21 | <Module: @1-1> | |
22 | ||
23 | Any subclasses of :class:`Scope`, including :class:`Module` has an attribute | |
24 | :attr:`iter_imports <Scope.iter_imports>`: | |
25 | ||
26 | >>> list(module.iter_imports()) | |
27 | [<ImportName: import os@1,0>] | |
28 | ||
29 | Changes to the Python Grammar | |
30 | ----------------------------- | |
31 | ||
32 | A few things have changed when looking at Python grammar files: | |
33 | ||
34 | - :class:`Param` does not exist in Python grammar files. It is essentially a | |
35 | part of a ``parameters`` node. |parso| splits it up to make it easier to | |
36 | analyse parameters. However this just makes it easier to deal with the syntax | |
37 | tree, it doesn't actually change the valid syntax. | |
38 | - A few nodes like `lambdef` and `lambdef_nocond` have been merged in the | |
39 | syntax tree to make it easier to do deal with them. | |
40 | ||
41 | Parser Tree Classes | |
42 | ------------------- | |
43 | """ | |
44 | ||
45 | import re | |
46 | try: | |
47 | from collections.abc import Mapping | |
48 | except ImportError: | |
49 | from collections import Mapping | |
50 | from typing import Tuple | |
51 | ||
52 | from parso.tree import Node, BaseNode, Leaf, ErrorNode, ErrorLeaf, search_ancestor # noqa | |
53 | from parso.python.prefix import split_prefix | |
54 | from parso.utils import split_lines | |
55 | ||
56 | _FLOW_CONTAINERS = set(['if_stmt', 'while_stmt', 'for_stmt', 'try_stmt', | |
57 | 'with_stmt', 'async_stmt', 'suite']) | |
58 | _RETURN_STMT_CONTAINERS = set(['suite', 'simple_stmt']) | _FLOW_CONTAINERS | |
59 | ||
60 | _FUNC_CONTAINERS = set( | |
61 | ['suite', 'simple_stmt', 'decorated', 'async_funcdef'] | |
62 | ) | _FLOW_CONTAINERS | |
63 | ||
64 | _GET_DEFINITION_TYPES = set([ | |
65 | 'expr_stmt', 'sync_comp_for', 'with_stmt', 'for_stmt', 'import_name', | |
66 | 'import_from', 'param', 'del_stmt', 'namedexpr_test', | |
67 | ]) | |
68 | _IMPORTS = set(['import_name', 'import_from']) | |
69 | ||
70 | ||
71 | class DocstringMixin: | |
72 | __slots__ = () | |
73 | ||
74 | def get_doc_node(self): | |
75 | """ | |
76 | Returns the string leaf of a docstring. e.g. ``r'''foo'''``. | |
77 | """ | |
78 | if self.type == 'file_input': | |
79 | node = self.children[0] | |
80 | elif self.type in ('funcdef', 'classdef'): | |
81 | node = self.children[self.children.index(':') + 1] | |
82 | if node.type == 'suite': # Normally a suite | |
83 | node = node.children[1] # -> NEWLINE stmt | |
84 | else: # ExprStmt | |
85 | simple_stmt = self.parent | |
86 | c = simple_stmt.parent.children | |
87 | index = c.index(simple_stmt) | |
88 | if not index: | |
89 | return None | |
90 | node = c[index - 1] | |
91 | ||
92 | if node.type == 'simple_stmt': | |
93 | node = node.children[0] | |
94 | if node.type == 'string': | |
95 | return node | |
96 | return None | |
97 | ||
98 | ||
99 | class PythonMixin: | |
100 | """ | |
101 | Some Python specific utilities. | |
102 | """ | |
103 | __slots__ = () | |
104 | ||
105 | def get_name_of_position(self, position): | |
106 | """ | |
107 | Given a (line, column) tuple, returns a :py:class:`Name` or ``None`` if | |
108 | there is no name at that position. | |
109 | """ | |
110 | for c in self.children: | |
111 | if isinstance(c, Leaf): | |
112 | if c.type == 'name' and c.start_pos <= position <= c.end_pos: | |
113 | return c | |
114 | else: | |
115 | result = c.get_name_of_position(position) | |
116 | if result is not None: | |
117 | return result | |
118 | return None | |
119 | ||
120 | ||
121 | class PythonLeaf(PythonMixin, Leaf): | |
122 | __slots__ = () | |
123 | ||
124 | def _split_prefix(self): | |
125 | return split_prefix(self, self.get_start_pos_of_prefix()) | |
126 | ||
127 | def get_start_pos_of_prefix(self): | |
128 | """ | |
129 | Basically calls :py:meth:`parso.tree.NodeOrLeaf.get_start_pos_of_prefix`. | |
130 | """ | |
131 | # TODO it is really ugly that we have to override it. Maybe change | |
132 | # indent error leafs somehow? No idea how, though. | |
133 | previous_leaf = self.get_previous_leaf() | |
134 | if previous_leaf is not None and previous_leaf.type == 'error_leaf' \ | |
135 | and previous_leaf.token_type in ('INDENT', 'DEDENT', 'ERROR_DEDENT'): | |
136 | previous_leaf = previous_leaf.get_previous_leaf() | |
137 | ||
138 | if previous_leaf is None: # It's the first leaf. | |
139 | lines = split_lines(self.prefix) | |
140 | # + 1 is needed because split_lines always returns at least ['']. | |
141 | return self.line - len(lines) + 1, 0 # It's the first leaf. | |
142 | return previous_leaf.end_pos | |
143 | ||
144 | ||
145 | class _LeafWithoutNewlines(PythonLeaf): | |
146 | """ | |
147 | Simply here to optimize performance. | |
148 | """ | |
149 | __slots__ = () | |
150 | ||
151 | @property | |
152 | def end_pos(self) -> Tuple[int, int]: | |
153 | return self.line, self.column + len(self.value) | |
154 | ||
155 | ||
156 | # Python base classes | |
157 | class PythonBaseNode(PythonMixin, BaseNode): | |
158 | __slots__ = () | |
159 | ||
160 | ||
161 | class PythonNode(PythonMixin, Node): | |
162 | __slots__ = () | |
163 | ||
164 | ||
165 | class PythonErrorNode(PythonMixin, ErrorNode): | |
166 | __slots__ = () | |
167 | ||
168 | ||
169 | class PythonErrorLeaf(ErrorLeaf, PythonLeaf): | |
170 | __slots__ = () | |
171 | ||
172 | ||
173 | class EndMarker(_LeafWithoutNewlines): | |
174 | __slots__ = () | |
175 | type = 'endmarker' | |
176 | ||
177 | def __repr__(self): | |
178 | return "<%s: prefix=%s end_pos=%s>" % ( | |
179 | type(self).__name__, repr(self.prefix), self.end_pos | |
180 | ) | |
181 | ||
182 | ||
183 | class Newline(PythonLeaf): | |
184 | """Contains NEWLINE and ENDMARKER tokens.""" | |
185 | __slots__ = () | |
186 | type = 'newline' | |
187 | ||
188 | def __repr__(self): | |
189 | return "<%s: %s>" % (type(self).__name__, repr(self.value)) | |
190 | ||
191 | ||
192 | class Name(_LeafWithoutNewlines): | |
193 | """ | |
194 | A string. Sometimes it is important to know if the string belongs to a name | |
195 | or not. | |
196 | """ | |
197 | type = 'name' | |
198 | __slots__ = () | |
199 | ||
200 | def __repr__(self): | |
201 | return "<%s: %s@%s,%s>" % (type(self).__name__, self.value, | |
202 | self.line, self.column) | |
203 | ||
204 | def is_definition(self, include_setitem=False): | |
205 | """ | |
206 | Returns True if the name is being defined. | |
207 | """ | |
208 | return self.get_definition(include_setitem=include_setitem) is not None | |
209 | ||
210 | def get_definition(self, import_name_always=False, include_setitem=False): | |
211 | """ | |
212 | Returns None if there's no definition for a name. | |
213 | ||
214 | :param import_name_always: Specifies if an import name is always a | |
215 | definition. Normally foo in `from foo import bar` is not a | |
216 | definition. | |
217 | """ | |
218 | node = self.parent | |
219 | type_ = node.type | |
220 | ||
221 | if type_ in ('funcdef', 'classdef'): | |
222 | if self == node.name: | |
223 | return node | |
224 | return None | |
225 | ||
226 | if type_ == 'except_clause': | |
227 | if self.get_previous_sibling() == 'as': | |
228 | return node.parent # The try_stmt. | |
229 | return None | |
230 | ||
231 | while node is not None: | |
232 | if node.type == 'suite': | |
233 | return None | |
234 | if node.type in _GET_DEFINITION_TYPES: | |
235 | if self in node.get_defined_names(include_setitem): | |
236 | return node | |
237 | if import_name_always and node.type in _IMPORTS: | |
238 | return node | |
239 | return None | |
240 | node = node.parent | |
241 | return None | |
242 | ||
243 | ||
244 | class Literal(PythonLeaf): | |
245 | __slots__ = () | |
246 | ||
247 | ||
248 | class Number(Literal): | |
249 | type = 'number' | |
250 | __slots__ = () | |
251 | ||
252 | ||
253 | class String(Literal): | |
254 | type = 'string' | |
255 | __slots__ = () | |
256 | ||
257 | @property | |
258 | def string_prefix(self): | |
259 | return re.match(r'\w*(?=[\'"])', self.value).group(0) | |
260 | ||
261 | def _get_payload(self): | |
262 | match = re.search( | |
263 | r'''('{3}|"{3}|'|")(.*)$''', | |
264 | self.value, | |
265 | flags=re.DOTALL | |
266 | ) | |
267 | return match.group(2)[:-len(match.group(1))] | |
268 | ||
269 | ||
270 | class FStringString(PythonLeaf): | |
271 | """ | |
272 | f-strings contain f-string expressions and normal python strings. These are | |
273 | the string parts of f-strings. | |
274 | """ | |
275 | type = 'fstring_string' | |
276 | __slots__ = () | |
277 | ||
278 | ||
279 | class FStringStart(PythonLeaf): | |
280 | """ | |
281 | f-strings contain f-string expressions and normal python strings. These are | |
282 | the string parts of f-strings. | |
283 | """ | |
284 | type = 'fstring_start' | |
285 | __slots__ = () | |
286 | ||
287 | ||
288 | class FStringEnd(PythonLeaf): | |
289 | """ | |
290 | f-strings contain f-string expressions and normal python strings. These are | |
291 | the string parts of f-strings. | |
292 | """ | |
293 | type = 'fstring_end' | |
294 | __slots__ = () | |
295 | ||
296 | ||
297 | class _StringComparisonMixin: | |
298 | def __eq__(self, other): | |
299 | """ | |
300 | Make comparisons with strings easy. | |
301 | Improves the readability of the parser. | |
302 | """ | |
303 | if isinstance(other, str): | |
304 | return self.value == other | |
305 | ||
306 | return self is other | |
307 | ||
308 | def __hash__(self): | |
309 | return hash(self.value) | |
310 | ||
311 | ||
312 | class Operator(_LeafWithoutNewlines, _StringComparisonMixin): | |
313 | type = 'operator' | |
314 | __slots__ = () | |
315 | ||
316 | ||
317 | class Keyword(_LeafWithoutNewlines, _StringComparisonMixin): | |
318 | type = 'keyword' | |
319 | __slots__ = () | |
320 | ||
321 | ||
322 | class Scope(PythonBaseNode, DocstringMixin): | |
323 | """ | |
324 | Super class for the parser tree, which represents the state of a python | |
325 | text file. | |
326 | A Scope is either a function, class or lambda. | |
327 | """ | |
328 | __slots__ = () | |
329 | ||
330 | def __init__(self, children): | |
331 | super().__init__(children) | |
332 | ||
333 | def iter_funcdefs(self): | |
334 | """ | |
335 | Returns a generator of `funcdef` nodes. | |
336 | """ | |
337 | return self._search_in_scope('funcdef') | |
338 | ||
339 | def iter_classdefs(self): | |
340 | """ | |
341 | Returns a generator of `classdef` nodes. | |
342 | """ | |
343 | return self._search_in_scope('classdef') | |
344 | ||
345 | def iter_imports(self): | |
346 | """ | |
347 | Returns a generator of `import_name` and `import_from` nodes. | |
348 | """ | |
349 | return self._search_in_scope('import_name', 'import_from') | |
350 | ||
351 | def _search_in_scope(self, *names): | |
352 | def scan(children): | |
353 | for element in children: | |
354 | if element.type in names: | |
355 | yield element | |
356 | if element.type in _FUNC_CONTAINERS: | |
357 | yield from scan(element.children) | |
358 | ||
359 | return scan(self.children) | |
360 | ||
361 | def get_suite(self): | |
362 | """ | |
363 | Returns the part that is executed by the function. | |
364 | """ | |
365 | return self.children[-1] | |
366 | ||
367 | def __repr__(self): | |
368 | try: | |
369 | name = self.name.value | |
370 | except AttributeError: | |
371 | name = '' | |
372 | ||
373 | return "<%s: %s@%s-%s>" % (type(self).__name__, name, | |
374 | self.start_pos[0], self.end_pos[0]) | |
375 | ||
376 | ||
377 | class Module(Scope): | |
378 | """ | |
379 | The top scope, which is always a module. | |
380 | Depending on the underlying parser this may be a full module or just a part | |
381 | of a module. | |
382 | """ | |
383 | __slots__ = ('_used_names',) | |
384 | type = 'file_input' | |
385 | ||
386 | def __init__(self, children): | |
387 | super().__init__(children) | |
388 | self._used_names = None | |
389 | ||
390 | def _iter_future_import_names(self): | |
391 | """ | |
392 | :return: A list of future import names. | |
393 | :rtype: list of str | |
394 | """ | |
395 | # In Python it's not allowed to use future imports after the first | |
396 | # actual (non-future) statement. However this is not a linter here, | |
397 | # just return all future imports. If people want to scan for issues | |
398 | # they should use the API. | |
399 | for imp in self.iter_imports(): | |
400 | if imp.type == 'import_from' and imp.level == 0: | |
401 | for path in imp.get_paths(): | |
402 | names = [name.value for name in path] | |
403 | if len(names) == 2 and names[0] == '__future__': | |
404 | yield names[1] | |
405 | ||
406 | def get_used_names(self): | |
407 | """ | |
408 | Returns all the :class:`Name` leafs that exist in this module. This | |
409 | includes both definitions and references of names. | |
410 | """ | |
411 | if self._used_names is None: | |
412 | # Don't directly use self._used_names to eliminate a lookup. | |
413 | dct = {} | |
414 | ||
415 | def recurse(node): | |
416 | try: | |
417 | children = node.children | |
418 | except AttributeError: | |
419 | if node.type == 'name': | |
420 | arr = dct.setdefault(node.value, []) | |
421 | arr.append(node) | |
422 | else: | |
423 | for child in children: | |
424 | recurse(child) | |
425 | ||
426 | recurse(self) | |
427 | self._used_names = UsedNamesMapping(dct) | |
428 | return self._used_names | |
429 | ||
430 | ||
431 | class Decorator(PythonBaseNode): | |
432 | type = 'decorator' | |
433 | __slots__ = () | |
434 | ||
435 | ||
436 | class ClassOrFunc(Scope): | |
437 | __slots__ = () | |
438 | ||
439 | @property | |
440 | def name(self): | |
441 | """ | |
442 | Returns the `Name` leaf that defines the function or class name. | |
443 | """ | |
444 | return self.children[1] | |
445 | ||
446 | def get_decorators(self): | |
447 | """ | |
448 | :rtype: list of :class:`Decorator` | |
449 | """ | |
450 | decorated = self.parent | |
451 | if decorated.type == 'async_funcdef': | |
452 | decorated = decorated.parent | |
453 | ||
454 | if decorated.type == 'decorated': | |
455 | if decorated.children[0].type == 'decorators': | |
456 | return decorated.children[0].children | |
457 | else: | |
458 | return decorated.children[:1] | |
459 | else: | |
460 | return [] | |
461 | ||
462 | ||
463 | class Class(ClassOrFunc): | |
464 | """ | |
465 | Used to store the parsed contents of a python class. | |
466 | """ | |
467 | type = 'classdef' | |
468 | __slots__ = () | |
469 | ||
470 | def __init__(self, children): | |
471 | super().__init__(children) | |
472 | ||
473 | def get_super_arglist(self): | |
474 | """ | |
475 | Returns the `arglist` node that defines the super classes. It returns | |
476 | None if there are no arguments. | |
477 | """ | |
478 | if self.children[2] != '(': # Has no parentheses | |
479 | return None | |
480 | else: | |
481 | if self.children[3] == ')': # Empty parentheses | |
482 | return None | |
483 | else: | |
484 | return self.children[3] | |
485 | ||
486 | ||
487 | def _create_params(parent, argslist_list): | |
488 | """ | |
489 | `argslist_list` is a list that can contain an argslist as a first item, but | |
490 | most not. It's basically the items between the parameter brackets (which is | |
491 | at most one item). | |
492 | This function modifies the parser structure. It generates `Param` objects | |
493 | from the normal ast. Those param objects do not exist in a normal ast, but | |
494 | make the evaluation of the ast tree so much easier. | |
495 | You could also say that this function replaces the argslist node with a | |
496 | list of Param objects. | |
497 | """ | |
498 | try: | |
499 | first = argslist_list[0] | |
500 | except IndexError: | |
501 | return [] | |
502 | ||
503 | if first.type in ('name', 'fpdef'): | |
504 | return [Param([first], parent)] | |
505 | elif first == '*': | |
506 | return [first] | |
507 | else: # argslist is a `typedargslist` or a `varargslist`. | |
508 | if first.type == 'tfpdef': | |
509 | children = [first] | |
510 | else: | |
511 | children = first.children | |
512 | new_children = [] | |
513 | start = 0 | |
514 | # Start with offset 1, because the end is higher. | |
515 | for end, child in enumerate(children + [None], 1): | |
516 | if child is None or child == ',': | |
517 | param_children = children[start:end] | |
518 | if param_children: # Could as well be comma and then end. | |
519 | if param_children[0] == '*' \ | |
520 | and (len(param_children) == 1 | |
521 | or param_children[1] == ',') \ | |
522 | or param_children[0] == '/': | |
523 | for p in param_children: | |
524 | p.parent = parent | |
525 | new_children += param_children | |
526 | else: | |
527 | new_children.append(Param(param_children, parent)) | |
528 | start = end | |
529 | return new_children | |
530 | ||
531 | ||
532 | class Function(ClassOrFunc): | |
533 | """ | |
534 | Used to store the parsed contents of a python function. | |
535 | ||
536 | Children:: | |
537 | ||
538 | 0. <Keyword: def> | |
539 | 1. <Name> | |
540 | 2. parameter list (including open-paren and close-paren <Operator>s) | |
541 | 3. or 5. <Operator: :> | |
542 | 4. or 6. Node() representing function body | |
543 | 3. -> (if annotation is also present) | |
544 | 4. annotation (if present) | |
545 | """ | |
546 | type = 'funcdef' | |
547 | ||
548 | def __init__(self, children): | |
549 | super().__init__(children) | |
550 | parameters = self.children[2] # After `def foo` | |
551 | parameters_children = parameters.children[1:-1] | |
552 | # If input parameters list already has Param objects, keep it as is; | |
553 | # otherwise, convert it to a list of Param objects. | |
554 | if not any(isinstance(child, Param) for child in parameters_children): | |
555 | parameters.children[1:-1] = _create_params(parameters, parameters_children) | |
556 | ||
557 | def _get_param_nodes(self): | |
558 | return self.children[2].children | |
559 | ||
560 | def get_params(self): | |
561 | """ | |
562 | Returns a list of `Param()`. | |
563 | """ | |
564 | return [p for p in self._get_param_nodes() if p.type == 'param'] | |
565 | ||
566 | @property | |
567 | def name(self): | |
568 | return self.children[1] # First token after `def` | |
569 | ||
570 | def iter_yield_exprs(self): | |
571 | """ | |
572 | Returns a generator of `yield_expr`. | |
573 | """ | |
574 | def scan(children): | |
575 | for element in children: | |
576 | if element.type in ('classdef', 'funcdef', 'lambdef'): | |
577 | continue | |
578 | ||
579 | try: | |
580 | nested_children = element.children | |
581 | except AttributeError: | |
582 | if element.value == 'yield': | |
583 | if element.parent.type == 'yield_expr': | |
584 | yield element.parent | |
585 | else: | |
586 | yield element | |
587 | else: | |
588 | yield from scan(nested_children) | |
589 | ||
590 | return scan(self.children) | |
591 | ||
592 | def iter_return_stmts(self): | |
593 | """ | |
594 | Returns a generator of `return_stmt`. | |
595 | """ | |
596 | def scan(children): | |
597 | for element in children: | |
598 | if element.type == 'return_stmt' \ | |
599 | or element.type == 'keyword' and element.value == 'return': | |
600 | yield element | |
601 | if element.type in _RETURN_STMT_CONTAINERS: | |
602 | yield from scan(element.children) | |
603 | ||
604 | return scan(self.children) | |
605 | ||
606 | def iter_raise_stmts(self): | |
607 | """ | |
608 | Returns a generator of `raise_stmt`. Includes raise statements inside try-except blocks | |
609 | """ | |
610 | def scan(children): | |
611 | for element in children: | |
612 | if element.type == 'raise_stmt' \ | |
613 | or element.type == 'keyword' and element.value == 'raise': | |
614 | yield element | |
615 | if element.type in _RETURN_STMT_CONTAINERS: | |
616 | yield from scan(element.children) | |
617 | ||
618 | return scan(self.children) | |
619 | ||
620 | def is_generator(self): | |
621 | """ | |
622 | :return bool: Checks if a function is a generator or not. | |
623 | """ | |
624 | return next(self.iter_yield_exprs(), None) is not None | |
625 | ||
626 | @property | |
627 | def annotation(self): | |
628 | """ | |
629 | Returns the test node after `->` or `None` if there is no annotation. | |
630 | """ | |
631 | try: | |
632 | if self.children[3] == "->": | |
633 | return self.children[4] | |
634 | assert self.children[3] == ":" | |
635 | return None | |
636 | except IndexError: | |
637 | return None | |
638 | ||
639 | ||
640 | class Lambda(Function): | |
641 | """ | |
642 | Lambdas are basically trimmed functions, so give it the same interface. | |
643 | ||
644 | Children:: | |
645 | ||
646 | 0. <Keyword: lambda> | |
647 | *. <Param x> for each argument x | |
648 | -2. <Operator: :> | |
649 | -1. Node() representing body | |
650 | """ | |
651 | type = 'lambdef' | |
652 | __slots__ = () | |
653 | ||
654 | def __init__(self, children): | |
655 | # We don't want to call the Function constructor, call its parent. | |
656 | super(Function, self).__init__(children) | |
657 | # Everything between `lambda` and the `:` operator is a parameter. | |
658 | parameters_children = self.children[1:-2] | |
659 | # If input children list already has Param objects, keep it as is; | |
660 | # otherwise, convert it to a list of Param objects. | |
661 | if not any(isinstance(child, Param) for child in parameters_children): | |
662 | self.children[1:-2] = _create_params(self, parameters_children) | |
663 | ||
664 | @property | |
665 | def name(self): | |
666 | """ | |
667 | Raises an AttributeError. Lambdas don't have a defined name. | |
668 | """ | |
669 | raise AttributeError("lambda is not named.") | |
670 | ||
671 | def _get_param_nodes(self): | |
672 | return self.children[1:-2] | |
673 | ||
674 | @property | |
675 | def annotation(self): | |
676 | """ | |
677 | Returns `None`, lambdas don't have annotations. | |
678 | """ | |
679 | return None | |
680 | ||
681 | def __repr__(self): | |
682 | return "<%s@%s>" % (self.__class__.__name__, self.start_pos) | |
683 | ||
684 | ||
685 | class Flow(PythonBaseNode): | |
686 | __slots__ = () | |
687 | ||
688 | ||
689 | class IfStmt(Flow): | |
690 | type = 'if_stmt' | |
691 | __slots__ = () | |
692 | ||
693 | def get_test_nodes(self): | |
694 | """ | |
695 | E.g. returns all the `test` nodes that are named as x, below: | |
696 | ||
697 | if x: | |
698 | pass | |
699 | elif x: | |
700 | pass | |
701 | """ | |
702 | for i, c in enumerate(self.children): | |
703 | if c in ('elif', 'if'): | |
704 | yield self.children[i + 1] | |
705 | ||
706 | def get_corresponding_test_node(self, node): | |
707 | """ | |
708 | Searches for the branch in which the node is and returns the | |
709 | corresponding test node (see function above). However if the node is in | |
710 | the test node itself and not in the suite return None. | |
711 | """ | |
712 | start_pos = node.start_pos | |
713 | for check_node in reversed(list(self.get_test_nodes())): | |
714 | if check_node.start_pos < start_pos: | |
715 | if start_pos < check_node.end_pos: | |
716 | return None | |
717 | # In this case the node is within the check_node itself, | |
718 | # not in the suite | |
719 | else: | |
720 | return check_node | |
721 | ||
722 | def is_node_after_else(self, node): | |
723 | """ | |
724 | Checks if a node is defined after `else`. | |
725 | """ | |
726 | for c in self.children: | |
727 | if c == 'else': | |
728 | if node.start_pos > c.start_pos: | |
729 | return True | |
730 | else: | |
731 | return False | |
732 | ||
733 | ||
734 | class WhileStmt(Flow): | |
735 | type = 'while_stmt' | |
736 | __slots__ = () | |
737 | ||
738 | ||
739 | class ForStmt(Flow): | |
740 | type = 'for_stmt' | |
741 | __slots__ = () | |
742 | ||
743 | def get_testlist(self): | |
744 | """ | |
745 | Returns the input node ``y`` from: ``for x in y:``. | |
746 | """ | |
747 | return self.children[3] | |
748 | ||
749 | def get_defined_names(self, include_setitem=False): | |
750 | return _defined_names(self.children[1], include_setitem) | |
751 | ||
752 | ||
753 | class TryStmt(Flow): | |
754 | type = 'try_stmt' | |
755 | __slots__ = () | |
756 | ||
757 | def get_except_clause_tests(self): | |
758 | """ | |
759 | Returns the ``test`` nodes found in ``except_clause`` nodes. | |
760 | Returns ``[None]`` for except clauses without an exception given. | |
761 | """ | |
762 | for node in self.children: | |
763 | if node.type == 'except_clause': | |
764 | yield node.children[1] | |
765 | elif node == 'except': | |
766 | yield None | |
767 | ||
768 | ||
769 | class WithStmt(Flow): | |
770 | type = 'with_stmt' | |
771 | __slots__ = () | |
772 | ||
773 | def get_defined_names(self, include_setitem=False): | |
774 | """ | |
775 | Returns the a list of `Name` that the with statement defines. The | |
776 | defined names are set after `as`. | |
777 | """ | |
778 | names = [] | |
779 | for with_item in self.children[1:-2:2]: | |
780 | # Check with items for 'as' names. | |
781 | if with_item.type == 'with_item': | |
782 | names += _defined_names(with_item.children[2], include_setitem) | |
783 | return names | |
784 | ||
785 | def get_test_node_from_name(self, name): | |
786 | node = name.search_ancestor("with_item") | |
787 | if node is None: | |
788 | raise ValueError('The name is not actually part of a with statement.') | |
789 | return node.children[0] | |
790 | ||
791 | ||
792 | class Import(PythonBaseNode): | |
793 | __slots__ = () | |
794 | ||
795 | def get_path_for_name(self, name): | |
796 | """ | |
797 | The path is the list of names that leads to the searched name. | |
798 | ||
799 | :return list of Name: | |
800 | """ | |
801 | try: | |
802 | # The name may be an alias. If it is, just map it back to the name. | |
803 | name = self._aliases()[name] | |
804 | except KeyError: | |
805 | pass | |
806 | ||
807 | for path in self.get_paths(): | |
808 | if name in path: | |
809 | return path[:path.index(name) + 1] | |
810 | raise ValueError('Name should be defined in the import itself') | |
811 | ||
812 | def is_nested(self): | |
813 | return False # By default, sub classes may overwrite this behavior | |
814 | ||
815 | def is_star_import(self): | |
816 | return self.children[-1] == '*' | |
817 | ||
818 | ||
819 | class ImportFrom(Import): | |
820 | type = 'import_from' | |
821 | __slots__ = () | |
822 | ||
823 | def get_defined_names(self, include_setitem=False): | |
824 | """ | |
825 | Returns the a list of `Name` that the import defines. The | |
826 | defined names are set after `import` or in case an alias - `as` - is | |
827 | present that name is returned. | |
828 | """ | |
829 | return [alias or name for name, alias in self._as_name_tuples()] | |
830 | ||
831 | def _aliases(self): | |
832 | """Mapping from alias to its corresponding name.""" | |
833 | return dict((alias, name) for name, alias in self._as_name_tuples() | |
834 | if alias is not None) | |
835 | ||
836 | def get_from_names(self): | |
837 | for n in self.children[1:]: | |
838 | if n not in ('.', '...'): | |
839 | break | |
840 | if n.type == 'dotted_name': # from x.y import | |
841 | return n.children[::2] | |
842 | elif n == 'import': # from . import | |
843 | return [] | |
844 | else: # from x import | |
845 | return [n] | |
846 | ||
847 | @property | |
848 | def level(self): | |
849 | """The level parameter of ``__import__``.""" | |
850 | level = 0 | |
851 | for n in self.children[1:]: | |
852 | if n in ('.', '...'): | |
853 | level += len(n.value) | |
854 | else: | |
855 | break | |
856 | return level | |
857 | ||
858 | def _as_name_tuples(self): | |
859 | last = self.children[-1] | |
860 | if last == ')': | |
861 | last = self.children[-2] | |
862 | elif last == '*': | |
863 | return # No names defined directly. | |
864 | ||
865 | if last.type == 'import_as_names': | |
866 | as_names = last.children[::2] | |
867 | else: | |
868 | as_names = [last] | |
869 | for as_name in as_names: | |
870 | if as_name.type == 'name': | |
871 | yield as_name, None | |
872 | else: | |
873 | yield as_name.children[::2] # yields x, y -> ``x as y`` | |
874 | ||
875 | def get_paths(self): | |
876 | """ | |
877 | The import paths defined in an import statement. Typically an array | |
878 | like this: ``[<Name: datetime>, <Name: date>]``. | |
879 | ||
880 | :return list of list of Name: | |
881 | """ | |
882 | dotted = self.get_from_names() | |
883 | ||
884 | if self.children[-1] == '*': | |
885 | return [dotted] | |
886 | return [dotted + [name] for name, alias in self._as_name_tuples()] | |
887 | ||
888 | ||
889 | class ImportName(Import): | |
890 | """For ``import_name`` nodes. Covers normal imports without ``from``.""" | |
891 | type = 'import_name' | |
892 | __slots__ = () | |
893 | ||
894 | def get_defined_names(self, include_setitem=False): | |
895 | """ | |
896 | Returns the a list of `Name` that the import defines. The defined names | |
897 | is always the first name after `import` or in case an alias - `as` - is | |
898 | present that name is returned. | |
899 | """ | |
900 | return [alias or path[0] for path, alias in self._dotted_as_names()] | |
901 | ||
902 | @property | |
903 | def level(self): | |
904 | """The level parameter of ``__import__``.""" | |
905 | return 0 # Obviously 0 for imports without from. | |
906 | ||
907 | def get_paths(self): | |
908 | return [path for path, alias in self._dotted_as_names()] | |
909 | ||
910 | def _dotted_as_names(self): | |
911 | """Generator of (list(path), alias) where alias may be None.""" | |
912 | dotted_as_names = self.children[1] | |
913 | if dotted_as_names.type == 'dotted_as_names': | |
914 | as_names = dotted_as_names.children[::2] | |
915 | else: | |
916 | as_names = [dotted_as_names] | |
917 | ||
918 | for as_name in as_names: | |
919 | if as_name.type == 'dotted_as_name': | |
920 | alias = as_name.children[2] | |
921 | as_name = as_name.children[0] | |
922 | else: | |
923 | alias = None | |
924 | if as_name.type == 'name': | |
925 | yield [as_name], alias | |
926 | else: | |
927 | # dotted_names | |
928 | yield as_name.children[::2], alias | |
929 | ||
930 | def is_nested(self): | |
931 | """ | |
932 | This checks for the special case of nested imports, without aliases and | |
933 | from statement:: | |
934 | ||
935 | import foo.bar | |
936 | """ | |
937 | return bool([1 for path, alias in self._dotted_as_names() | |
938 | if alias is None and len(path) > 1]) | |
939 | ||
940 | def _aliases(self): | |
941 | """ | |
942 | :return list of Name: Returns all the alias | |
943 | """ | |
944 | return dict((alias, path[-1]) for path, alias in self._dotted_as_names() | |
945 | if alias is not None) | |
946 | ||
947 | ||
948 | class KeywordStatement(PythonBaseNode): | |
949 | """ | |
950 | For the following statements: `assert`, `del`, `global`, `nonlocal`, | |
951 | `raise`, `return`, `yield`. | |
952 | ||
953 | `pass`, `continue` and `break` are not in there, because they are just | |
954 | simple keywords and the parser reduces it to a keyword. | |
955 | """ | |
956 | __slots__ = () | |
957 | ||
958 | @property | |
959 | def type(self): | |
960 | """ | |
961 | Keyword statements start with the keyword and end with `_stmt`. You can | |
962 | crosscheck this with the Python grammar. | |
963 | """ | |
964 | return '%s_stmt' % self.keyword | |
965 | ||
966 | @property | |
967 | def keyword(self): | |
968 | return self.children[0].value | |
969 | ||
970 | def get_defined_names(self, include_setitem=False): | |
971 | keyword = self.keyword | |
972 | if keyword == 'del': | |
973 | return _defined_names(self.children[1], include_setitem) | |
974 | if keyword in ('global', 'nonlocal'): | |
975 | return self.children[1::2] | |
976 | return [] | |
977 | ||
978 | ||
979 | class AssertStmt(KeywordStatement): | |
980 | __slots__ = () | |
981 | ||
982 | @property | |
983 | def assertion(self): | |
984 | return self.children[1] | |
985 | ||
986 | ||
987 | class GlobalStmt(KeywordStatement): | |
988 | __slots__ = () | |
989 | ||
990 | def get_global_names(self): | |
991 | return self.children[1::2] | |
992 | ||
993 | ||
994 | class ReturnStmt(KeywordStatement): | |
995 | __slots__ = () | |
996 | ||
997 | ||
998 | class YieldExpr(PythonBaseNode): | |
999 | type = 'yield_expr' | |
1000 | __slots__ = () | |
1001 | ||
1002 | ||
1003 | def _defined_names(current, include_setitem): | |
1004 | """ | |
1005 | A helper function to find the defined names in statements, for loops and | |
1006 | list comprehensions. | |
1007 | """ | |
1008 | names = [] | |
1009 | if current.type in ('testlist_star_expr', 'testlist_comp', 'exprlist', 'testlist'): | |
1010 | for child in current.children[::2]: | |
1011 | names += _defined_names(child, include_setitem) | |
1012 | elif current.type in ('atom', 'star_expr'): | |
1013 | names += _defined_names(current.children[1], include_setitem) | |
1014 | elif current.type in ('power', 'atom_expr'): | |
1015 | if current.children[-2] != '**': # Just if there's no operation | |
1016 | trailer = current.children[-1] | |
1017 | if trailer.children[0] == '.': | |
1018 | names.append(trailer.children[1]) | |
1019 | elif trailer.children[0] == '[' and include_setitem: | |
1020 | for node in current.children[-2::-1]: | |
1021 | if node.type == 'trailer': | |
1022 | names.append(node.children[1]) | |
1023 | break | |
1024 | if node.type == 'name': | |
1025 | names.append(node) | |
1026 | break | |
1027 | else: | |
1028 | names.append(current) | |
1029 | return names | |
1030 | ||
1031 | ||
1032 | class ExprStmt(PythonBaseNode, DocstringMixin): | |
1033 | type = 'expr_stmt' | |
1034 | __slots__ = () | |
1035 | ||
1036 | def get_defined_names(self, include_setitem=False): | |
1037 | """ | |
1038 | Returns a list of `Name` defined before the `=` sign. | |
1039 | """ | |
1040 | names = [] | |
1041 | if self.children[1].type == 'annassign': | |
1042 | names = _defined_names(self.children[0], include_setitem) | |
1043 | return [ | |
1044 | name | |
1045 | for i in range(0, len(self.children) - 2, 2) | |
1046 | if '=' in self.children[i + 1].value | |
1047 | for name in _defined_names(self.children[i], include_setitem) | |
1048 | ] + names | |
1049 | ||
1050 | def get_rhs(self): | |
1051 | """Returns the right-hand-side of the equals.""" | |
1052 | node = self.children[-1] | |
1053 | if node.type == 'annassign': | |
1054 | if len(node.children) == 4: | |
1055 | node = node.children[3] | |
1056 | else: | |
1057 | node = node.children[1] | |
1058 | return node | |
1059 | ||
1060 | def yield_operators(self): | |
1061 | """ | |
1062 | Returns a generator of `+=`, `=`, etc. or None if there is no operation. | |
1063 | """ | |
1064 | first = self.children[1] | |
1065 | if first.type == 'annassign': | |
1066 | if len(first.children) <= 2: | |
1067 | return # No operator is available, it's just PEP 484. | |
1068 | ||
1069 | first = first.children[2] | |
1070 | yield first | |
1071 | ||
1072 | yield from self.children[3::2] | |
1073 | ||
1074 | ||
1075 | class NamedExpr(PythonBaseNode): | |
1076 | type = 'namedexpr_test' | |
1077 | ||
1078 | def get_defined_names(self, include_setitem=False): | |
1079 | return _defined_names(self.children[0], include_setitem) | |
1080 | ||
1081 | ||
1082 | class Param(PythonBaseNode): | |
1083 | """ | |
1084 | It's a helper class that makes business logic with params much easier. The | |
1085 | Python grammar defines no ``param`` node. It defines it in a different way | |
1086 | that is not really suited to working with parameters. | |
1087 | """ | |
1088 | type = 'param' | |
1089 | ||
1090 | def __init__(self, children, parent=None): | |
1091 | super().__init__(children) | |
1092 | self.parent = parent | |
1093 | ||
1094 | @property | |
1095 | def star_count(self): | |
1096 | """ | |
1097 | Is `0` in case of `foo`, `1` in case of `*foo` or `2` in case of | |
1098 | `**foo`. | |
1099 | """ | |
1100 | first = self.children[0] | |
1101 | if first in ('*', '**'): | |
1102 | return len(first.value) | |
1103 | return 0 | |
1104 | ||
1105 | @property | |
1106 | def default(self): | |
1107 | """ | |
1108 | The default is the test node that appears after the `=`. Is `None` in | |
1109 | case no default is present. | |
1110 | """ | |
1111 | has_comma = self.children[-1] == ',' | |
1112 | try: | |
1113 | if self.children[-2 - int(has_comma)] == '=': | |
1114 | return self.children[-1 - int(has_comma)] | |
1115 | except IndexError: | |
1116 | return None | |
1117 | ||
1118 | @property | |
1119 | def annotation(self): | |
1120 | """ | |
1121 | The default is the test node that appears after `:`. Is `None` in case | |
1122 | no annotation is present. | |
1123 | """ | |
1124 | tfpdef = self._tfpdef() | |
1125 | if tfpdef.type == 'tfpdef': | |
1126 | assert tfpdef.children[1] == ":" | |
1127 | assert len(tfpdef.children) == 3 | |
1128 | annotation = tfpdef.children[2] | |
1129 | return annotation | |
1130 | else: | |
1131 | return None | |
1132 | ||
1133 | def _tfpdef(self): | |
1134 | """ | |
1135 | tfpdef: see e.g. grammar36.txt. | |
1136 | """ | |
1137 | offset = int(self.children[0] in ('*', '**')) | |
1138 | return self.children[offset] | |
1139 | ||
1140 | @property | |
1141 | def name(self): | |
1142 | """ | |
1143 | The `Name` leaf of the param. | |
1144 | """ | |
1145 | if self._tfpdef().type == 'tfpdef': | |
1146 | return self._tfpdef().children[0] | |
1147 | else: | |
1148 | return self._tfpdef() | |
1149 | ||
1150 | def get_defined_names(self, include_setitem=False): | |
1151 | return [self.name] | |
1152 | ||
1153 | @property | |
1154 | def position_index(self): | |
1155 | """ | |
1156 | Property for the positional index of a paramter. | |
1157 | """ | |
1158 | index = self.parent.children.index(self) | |
1159 | try: | |
1160 | keyword_only_index = self.parent.children.index('*') | |
1161 | if index > keyword_only_index: | |
1162 | # Skip the ` *, ` | |
1163 | index -= 2 | |
1164 | except ValueError: | |
1165 | pass | |
1166 | try: | |
1167 | keyword_only_index = self.parent.children.index('/') | |
1168 | if index > keyword_only_index: | |
1169 | # Skip the ` /, ` | |
1170 | index -= 2 | |
1171 | except ValueError: | |
1172 | pass | |
1173 | return index - 1 | |
1174 | ||
1175 | def get_parent_function(self): | |
1176 | """ | |
1177 | Returns the function/lambda of a parameter. | |
1178 | """ | |
1179 | return self.search_ancestor('funcdef', 'lambdef') | |
1180 | ||
1181 | def get_code(self, include_prefix=True, include_comma=True): | |
1182 | """ | |
1183 | Like all the other get_code functions, but includes the param | |
1184 | `include_comma`. | |
1185 | ||
1186 | :param include_comma bool: If enabled includes the comma in the string output. | |
1187 | """ | |
1188 | if include_comma: | |
1189 | return super().get_code(include_prefix) | |
1190 | ||
1191 | children = self.children | |
1192 | if children[-1] == ',': | |
1193 | children = children[:-1] | |
1194 | return self._get_code_for_children( | |
1195 | children, | |
1196 | include_prefix=include_prefix | |
1197 | ) | |
1198 | ||
1199 | def __repr__(self): | |
1200 | default = '' if self.default is None else '=%s' % self.default.get_code() | |
1201 | return '<%s: %s>' % (type(self).__name__, str(self._tfpdef()) + default) | |
1202 | ||
1203 | ||
1204 | class SyncCompFor(PythonBaseNode): | |
1205 | type = 'sync_comp_for' | |
1206 | __slots__ = () | |
1207 | ||
1208 | def get_defined_names(self, include_setitem=False): | |
1209 | """ | |
1210 | Returns the a list of `Name` that the comprehension defines. | |
1211 | """ | |
1212 | # allow async for | |
1213 | return _defined_names(self.children[1], include_setitem) | |
1214 | ||
1215 | ||
1216 | # This is simply here so an older Jedi version can work with this new parso | |
1217 | # version. Can be deleted in the next release. | |
1218 | CompFor = SyncCompFor | |
1219 | ||
1220 | ||
1221 | class UsedNamesMapping(Mapping): | |
1222 | """ | |
1223 | This class exists for the sole purpose of creating an immutable dict. | |
1224 | """ | |
1225 | def __init__(self, dct): | |
1226 | self._dict = dct | |
1227 | ||
1228 | def __getitem__(self, key): | |
1229 | return self._dict[key] | |
1230 | ||
1231 | def __len__(self): | |
1232 | return len(self._dict) | |
1233 | ||
1234 | def __iter__(self): | |
1235 | return iter(self._dict) | |
1236 | ||
1237 | def __hash__(self): | |
1238 | return id(self) | |
1239 | ||
1240 | def __eq__(self, other): | |
1241 | # Comparing these dicts does not make sense. | |
1242 | return self is other |