Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 1 | """ |
| 2 | ast |
| 3 | ~~~ |
| 4 | |
| 5 | The `ast` module helps Python applications to process trees of the Python |
| 6 | abstract syntax grammar. The abstract syntax itself might change with |
| 7 | each Python release; this module helps to find out programmatically what |
| 8 | the current grammar looks like and allows modifications of it. |
| 9 | |
| 10 | An abstract syntax tree can be generated by passing `ast.PyCF_ONLY_AST` as |
| 11 | a flag to the `compile()` builtin function or by using the `parse()` |
| 12 | function from this module. The result will be a tree of objects whose |
| 13 | classes all inherit from `ast.AST`. |
| 14 | |
| 15 | A modified abstract syntax tree can be compiled into a Python code object |
| 16 | using the built-in `compile()` function. |
| 17 | |
| 18 | Additionally various helper functions are provided that make working with |
| 19 | the trees simpler. The main intention of the helper functions and this |
| 20 | module in general is to provide an easy to use interface for libraries |
| 21 | that work tightly with the python syntax (template engines for example). |
| 22 | |
| 23 | |
| 24 | :copyright: Copyright 2008 by Armin Ronacher. |
| 25 | :license: Python License. |
| 26 | """ |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 27 | import sys |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 28 | from _ast import * |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 29 | from contextlib import contextmanager, nullcontext |
| 30 | from enum import IntEnum, auto |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 31 | |
| 32 | |
| 33 | def parse(source, filename='<unknown>', mode='exec', *, |
| 34 | type_comments=False, feature_version=None): |
| 35 | """ |
| 36 | Parse the source into an AST node. |
| 37 | Equivalent to compile(source, filename, mode, PyCF_ONLY_AST). |
| 38 | Pass type_comments=True to get back type comments where the syntax allows. |
| 39 | """ |
| 40 | flags = PyCF_ONLY_AST |
| 41 | if type_comments: |
| 42 | flags |= PyCF_TYPE_COMMENTS |
| 43 | if isinstance(feature_version, tuple): |
| 44 | major, minor = feature_version # Should be a 2-tuple. |
| 45 | assert major == 3 |
| 46 | feature_version = minor |
| 47 | elif feature_version is None: |
| 48 | feature_version = -1 |
| 49 | # Else it should be an int giving the minor version for 3.x. |
| 50 | return compile(source, filename, mode, flags, |
| 51 | _feature_version=feature_version) |
| 52 | |
| 53 | |
| 54 | def literal_eval(node_or_string): |
| 55 | """ |
| 56 | Safely evaluate an expression node or a string containing a Python |
| 57 | expression. The string or node provided may only consist of the following |
| 58 | Python literal structures: strings, bytes, numbers, tuples, lists, dicts, |
| 59 | sets, booleans, and None. |
| 60 | """ |
| 61 | if isinstance(node_or_string, str): |
Yi Kong | 7119932 | 2022-08-30 15:53:45 +0800 | [diff] [blame] | 62 | node_or_string = parse(node_or_string.lstrip(" \t"), mode='eval') |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 63 | if isinstance(node_or_string, Expression): |
| 64 | node_or_string = node_or_string.body |
Haibo Huang | f5f93a7 | 2020-10-19 15:43:42 -0700 | [diff] [blame] | 65 | def _raise_malformed_node(node): |
Yi Kong | 7119932 | 2022-08-30 15:53:45 +0800 | [diff] [blame] | 66 | msg = "malformed node or string" |
| 67 | if lno := getattr(node, 'lineno', None): |
| 68 | msg += f' on line {lno}' |
| 69 | raise ValueError(msg + f': {node!r}') |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 70 | def _convert_num(node): |
Haibo Huang | f5f93a7 | 2020-10-19 15:43:42 -0700 | [diff] [blame] | 71 | if not isinstance(node, Constant) or type(node.value) not in (int, float, complex): |
| 72 | _raise_malformed_node(node) |
| 73 | return node.value |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 74 | def _convert_signed_num(node): |
| 75 | if isinstance(node, UnaryOp) and isinstance(node.op, (UAdd, USub)): |
| 76 | operand = _convert_num(node.operand) |
| 77 | if isinstance(node.op, UAdd): |
| 78 | return + operand |
| 79 | else: |
| 80 | return - operand |
| 81 | return _convert_num(node) |
| 82 | def _convert(node): |
| 83 | if isinstance(node, Constant): |
| 84 | return node.value |
| 85 | elif isinstance(node, Tuple): |
| 86 | return tuple(map(_convert, node.elts)) |
| 87 | elif isinstance(node, List): |
| 88 | return list(map(_convert, node.elts)) |
| 89 | elif isinstance(node, Set): |
| 90 | return set(map(_convert, node.elts)) |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 91 | elif (isinstance(node, Call) and isinstance(node.func, Name) and |
| 92 | node.func.id == 'set' and node.args == node.keywords == []): |
| 93 | return set() |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 94 | elif isinstance(node, Dict): |
Haibo Huang | f5f93a7 | 2020-10-19 15:43:42 -0700 | [diff] [blame] | 95 | if len(node.keys) != len(node.values): |
| 96 | _raise_malformed_node(node) |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 97 | return dict(zip(map(_convert, node.keys), |
| 98 | map(_convert, node.values))) |
| 99 | elif isinstance(node, BinOp) and isinstance(node.op, (Add, Sub)): |
| 100 | left = _convert_signed_num(node.left) |
| 101 | right = _convert_num(node.right) |
| 102 | if isinstance(left, (int, float)) and isinstance(right, complex): |
| 103 | if isinstance(node.op, Add): |
| 104 | return left + right |
| 105 | else: |
| 106 | return left - right |
| 107 | return _convert_signed_num(node) |
| 108 | return _convert(node_or_string) |
| 109 | |
| 110 | |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 111 | def dump(node, annotate_fields=True, include_attributes=False, *, indent=None): |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 112 | """ |
| 113 | Return a formatted dump of the tree in node. This is mainly useful for |
| 114 | debugging purposes. If annotate_fields is true (by default), |
| 115 | the returned string will show the names and the values for fields. |
| 116 | If annotate_fields is false, the result string will be more compact by |
| 117 | omitting unambiguous field names. Attributes such as line |
| 118 | numbers and column offsets are not dumped by default. If this is wanted, |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 119 | include_attributes can be set to true. If indent is a non-negative |
| 120 | integer or string, then the tree will be pretty-printed with that indent |
| 121 | level. None (the default) selects the single line representation. |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 122 | """ |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 123 | def _format(node, level=0): |
| 124 | if indent is not None: |
| 125 | level += 1 |
| 126 | prefix = '\n' + indent * level |
| 127 | sep = ',\n' + indent * level |
| 128 | else: |
| 129 | prefix = '' |
| 130 | sep = ', ' |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 131 | if isinstance(node, AST): |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 132 | cls = type(node) |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 133 | args = [] |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 134 | allsimple = True |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 135 | keywords = annotate_fields |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 136 | for name in node._fields: |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 137 | try: |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 138 | value = getattr(node, name) |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 139 | except AttributeError: |
| 140 | keywords = True |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 141 | continue |
| 142 | if value is None and getattr(cls, name, ...) is None: |
| 143 | keywords = True |
| 144 | continue |
| 145 | value, simple = _format(value, level) |
| 146 | allsimple = allsimple and simple |
| 147 | if keywords: |
| 148 | args.append('%s=%s' % (name, value)) |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 149 | else: |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 150 | args.append(value) |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 151 | if include_attributes and node._attributes: |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 152 | for name in node._attributes: |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 153 | try: |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 154 | value = getattr(node, name) |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 155 | except AttributeError: |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 156 | continue |
| 157 | if value is None and getattr(cls, name, ...) is None: |
| 158 | continue |
| 159 | value, simple = _format(value, level) |
| 160 | allsimple = allsimple and simple |
| 161 | args.append('%s=%s' % (name, value)) |
| 162 | if allsimple and len(args) <= 3: |
| 163 | return '%s(%s)' % (node.__class__.__name__, ', '.join(args)), not args |
| 164 | return '%s(%s%s)' % (node.__class__.__name__, prefix, sep.join(args)), False |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 165 | elif isinstance(node, list): |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 166 | if not node: |
| 167 | return '[]', True |
| 168 | return '[%s%s]' % (prefix, sep.join(_format(x, level)[0] for x in node)), False |
| 169 | return repr(node), True |
| 170 | |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 171 | if not isinstance(node, AST): |
| 172 | raise TypeError('expected AST, got %r' % node.__class__.__name__) |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 173 | if indent is not None and not isinstance(indent, str): |
| 174 | indent = ' ' * indent |
| 175 | return _format(node)[0] |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 176 | |
| 177 | |
| 178 | def copy_location(new_node, old_node): |
| 179 | """ |
| 180 | Copy source location (`lineno`, `col_offset`, `end_lineno`, and `end_col_offset` |
| 181 | attributes) from *old_node* to *new_node* if possible, and return *new_node*. |
| 182 | """ |
| 183 | for attr in 'lineno', 'col_offset', 'end_lineno', 'end_col_offset': |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 184 | if attr in old_node._attributes and attr in new_node._attributes: |
| 185 | value = getattr(old_node, attr, None) |
| 186 | # end_lineno and end_col_offset are optional attributes, and they |
| 187 | # should be copied whether the value is None or not. |
| 188 | if value is not None or ( |
| 189 | hasattr(old_node, attr) and attr.startswith("end_") |
| 190 | ): |
| 191 | setattr(new_node, attr, value) |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 192 | return new_node |
| 193 | |
| 194 | |
| 195 | def fix_missing_locations(node): |
| 196 | """ |
| 197 | When you compile a node tree with compile(), the compiler expects lineno and |
| 198 | col_offset attributes for every node that supports them. This is rather |
| 199 | tedious to fill in for generated nodes, so this helper adds these attributes |
| 200 | recursively where not already set, by setting them to the values of the |
| 201 | parent node. It works recursively starting at *node*. |
| 202 | """ |
| 203 | def _fix(node, lineno, col_offset, end_lineno, end_col_offset): |
| 204 | if 'lineno' in node._attributes: |
| 205 | if not hasattr(node, 'lineno'): |
| 206 | node.lineno = lineno |
| 207 | else: |
| 208 | lineno = node.lineno |
| 209 | if 'end_lineno' in node._attributes: |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 210 | if getattr(node, 'end_lineno', None) is None: |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 211 | node.end_lineno = end_lineno |
| 212 | else: |
| 213 | end_lineno = node.end_lineno |
| 214 | if 'col_offset' in node._attributes: |
| 215 | if not hasattr(node, 'col_offset'): |
| 216 | node.col_offset = col_offset |
| 217 | else: |
| 218 | col_offset = node.col_offset |
| 219 | if 'end_col_offset' in node._attributes: |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 220 | if getattr(node, 'end_col_offset', None) is None: |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 221 | node.end_col_offset = end_col_offset |
| 222 | else: |
| 223 | end_col_offset = node.end_col_offset |
| 224 | for child in iter_child_nodes(node): |
| 225 | _fix(child, lineno, col_offset, end_lineno, end_col_offset) |
| 226 | _fix(node, 1, 0, 1, 0) |
| 227 | return node |
| 228 | |
| 229 | |
| 230 | def increment_lineno(node, n=1): |
| 231 | """ |
| 232 | Increment the line number and end line number of each node in the tree |
| 233 | starting at *node* by *n*. This is useful to "move code" to a different |
| 234 | location in a file. |
| 235 | """ |
| 236 | for child in walk(node): |
| 237 | if 'lineno' in child._attributes: |
| 238 | child.lineno = getattr(child, 'lineno', 0) + n |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 239 | if ( |
| 240 | "end_lineno" in child._attributes |
| 241 | and (end_lineno := getattr(child, "end_lineno", 0)) is not None |
| 242 | ): |
| 243 | child.end_lineno = end_lineno + n |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 244 | return node |
| 245 | |
| 246 | |
| 247 | def iter_fields(node): |
| 248 | """ |
| 249 | Yield a tuple of ``(fieldname, value)`` for each field in ``node._fields`` |
| 250 | that is present on *node*. |
| 251 | """ |
| 252 | for field in node._fields: |
| 253 | try: |
| 254 | yield field, getattr(node, field) |
| 255 | except AttributeError: |
| 256 | pass |
| 257 | |
| 258 | |
| 259 | def iter_child_nodes(node): |
| 260 | """ |
| 261 | Yield all direct child nodes of *node*, that is, all fields that are nodes |
| 262 | and all items of fields that are lists of nodes. |
| 263 | """ |
| 264 | for name, field in iter_fields(node): |
| 265 | if isinstance(field, AST): |
| 266 | yield field |
| 267 | elif isinstance(field, list): |
| 268 | for item in field: |
| 269 | if isinstance(item, AST): |
| 270 | yield item |
| 271 | |
| 272 | |
| 273 | def get_docstring(node, clean=True): |
| 274 | """ |
| 275 | Return the docstring for the given node or None if no docstring can |
| 276 | be found. If the node provided does not have docstrings a TypeError |
| 277 | will be raised. |
| 278 | |
| 279 | If *clean* is `True`, all tabs are expanded to spaces and any whitespace |
| 280 | that can be uniformly removed from the second line onwards is removed. |
| 281 | """ |
| 282 | if not isinstance(node, (AsyncFunctionDef, FunctionDef, ClassDef, Module)): |
| 283 | raise TypeError("%r can't have docstrings" % node.__class__.__name__) |
| 284 | if not(node.body and isinstance(node.body[0], Expr)): |
| 285 | return None |
| 286 | node = node.body[0].value |
| 287 | if isinstance(node, Str): |
| 288 | text = node.s |
| 289 | elif isinstance(node, Constant) and isinstance(node.value, str): |
| 290 | text = node.value |
| 291 | else: |
| 292 | return None |
| 293 | if clean: |
| 294 | import inspect |
| 295 | text = inspect.cleandoc(text) |
| 296 | return text |
| 297 | |
| 298 | |
| 299 | def _splitlines_no_ff(source): |
| 300 | """Split a string into lines ignoring form feed and other chars. |
| 301 | |
| 302 | This mimics how the Python parser splits source code. |
| 303 | """ |
| 304 | idx = 0 |
| 305 | lines = [] |
| 306 | next_line = '' |
| 307 | while idx < len(source): |
| 308 | c = source[idx] |
| 309 | next_line += c |
| 310 | idx += 1 |
| 311 | # Keep \r\n together |
| 312 | if c == '\r' and idx < len(source) and source[idx] == '\n': |
| 313 | next_line += '\n' |
| 314 | idx += 1 |
| 315 | if c in '\r\n': |
| 316 | lines.append(next_line) |
| 317 | next_line = '' |
| 318 | |
| 319 | if next_line: |
| 320 | lines.append(next_line) |
| 321 | return lines |
| 322 | |
| 323 | |
| 324 | def _pad_whitespace(source): |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 325 | r"""Replace all chars except '\f\t' in a line with spaces.""" |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 326 | result = '' |
| 327 | for c in source: |
| 328 | if c in '\f\t': |
| 329 | result += c |
| 330 | else: |
| 331 | result += ' ' |
| 332 | return result |
| 333 | |
| 334 | |
| 335 | def get_source_segment(source, node, *, padded=False): |
| 336 | """Get source code segment of the *source* that generated *node*. |
| 337 | |
| 338 | If some location information (`lineno`, `end_lineno`, `col_offset`, |
| 339 | or `end_col_offset`) is missing, return None. |
| 340 | |
| 341 | If *padded* is `True`, the first line of a multi-line statement will |
| 342 | be padded with spaces to match its original position. |
| 343 | """ |
| 344 | try: |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 345 | if node.end_lineno is None or node.end_col_offset is None: |
| 346 | return None |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 347 | lineno = node.lineno - 1 |
| 348 | end_lineno = node.end_lineno - 1 |
| 349 | col_offset = node.col_offset |
| 350 | end_col_offset = node.end_col_offset |
| 351 | except AttributeError: |
| 352 | return None |
| 353 | |
| 354 | lines = _splitlines_no_ff(source) |
| 355 | if end_lineno == lineno: |
| 356 | return lines[lineno].encode()[col_offset:end_col_offset].decode() |
| 357 | |
| 358 | if padded: |
| 359 | padding = _pad_whitespace(lines[lineno].encode()[:col_offset].decode()) |
| 360 | else: |
| 361 | padding = '' |
| 362 | |
| 363 | first = padding + lines[lineno].encode()[col_offset:].decode() |
| 364 | last = lines[end_lineno].encode()[:end_col_offset].decode() |
| 365 | lines = lines[lineno+1:end_lineno] |
| 366 | |
| 367 | lines.insert(0, first) |
| 368 | lines.append(last) |
| 369 | return ''.join(lines) |
| 370 | |
| 371 | |
| 372 | def walk(node): |
| 373 | """ |
| 374 | Recursively yield all descendant nodes in the tree starting at *node* |
| 375 | (including *node* itself), in no specified order. This is useful if you |
| 376 | only want to modify nodes in place and don't care about the context. |
| 377 | """ |
| 378 | from collections import deque |
| 379 | todo = deque([node]) |
| 380 | while todo: |
| 381 | node = todo.popleft() |
| 382 | todo.extend(iter_child_nodes(node)) |
| 383 | yield node |
| 384 | |
| 385 | |
| 386 | class NodeVisitor(object): |
| 387 | """ |
| 388 | A node visitor base class that walks the abstract syntax tree and calls a |
| 389 | visitor function for every node found. This function may return a value |
| 390 | which is forwarded by the `visit` method. |
| 391 | |
| 392 | This class is meant to be subclassed, with the subclass adding visitor |
| 393 | methods. |
| 394 | |
| 395 | Per default the visitor functions for the nodes are ``'visit_'`` + |
| 396 | class name of the node. So a `TryFinally` node visit function would |
| 397 | be `visit_TryFinally`. This behavior can be changed by overriding |
| 398 | the `visit` method. If no visitor function exists for a node |
| 399 | (return value `None`) the `generic_visit` visitor is used instead. |
| 400 | |
| 401 | Don't use the `NodeVisitor` if you want to apply changes to nodes during |
| 402 | traversing. For this a special visitor exists (`NodeTransformer`) that |
| 403 | allows modifications. |
| 404 | """ |
| 405 | |
| 406 | def visit(self, node): |
| 407 | """Visit a node.""" |
| 408 | method = 'visit_' + node.__class__.__name__ |
| 409 | visitor = getattr(self, method, self.generic_visit) |
| 410 | return visitor(node) |
| 411 | |
| 412 | def generic_visit(self, node): |
| 413 | """Called if no explicit visitor function exists for a node.""" |
| 414 | for field, value in iter_fields(node): |
| 415 | if isinstance(value, list): |
| 416 | for item in value: |
| 417 | if isinstance(item, AST): |
| 418 | self.visit(item) |
| 419 | elif isinstance(value, AST): |
| 420 | self.visit(value) |
| 421 | |
| 422 | def visit_Constant(self, node): |
| 423 | value = node.value |
| 424 | type_name = _const_node_type_names.get(type(value)) |
| 425 | if type_name is None: |
| 426 | for cls, name in _const_node_type_names.items(): |
| 427 | if isinstance(value, cls): |
| 428 | type_name = name |
| 429 | break |
| 430 | if type_name is not None: |
| 431 | method = 'visit_' + type_name |
| 432 | try: |
| 433 | visitor = getattr(self, method) |
| 434 | except AttributeError: |
| 435 | pass |
| 436 | else: |
| 437 | import warnings |
| 438 | warnings.warn(f"{method} is deprecated; add visit_Constant", |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 439 | DeprecationWarning, 2) |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 440 | return visitor(node) |
| 441 | return self.generic_visit(node) |
| 442 | |
| 443 | |
| 444 | class NodeTransformer(NodeVisitor): |
| 445 | """ |
| 446 | A :class:`NodeVisitor` subclass that walks the abstract syntax tree and |
| 447 | allows modification of nodes. |
| 448 | |
| 449 | The `NodeTransformer` will walk the AST and use the return value of the |
| 450 | visitor methods to replace or remove the old node. If the return value of |
| 451 | the visitor method is ``None``, the node will be removed from its location, |
| 452 | otherwise it is replaced with the return value. The return value may be the |
| 453 | original node in which case no replacement takes place. |
| 454 | |
| 455 | Here is an example transformer that rewrites all occurrences of name lookups |
| 456 | (``foo``) to ``data['foo']``:: |
| 457 | |
| 458 | class RewriteName(NodeTransformer): |
| 459 | |
| 460 | def visit_Name(self, node): |
Haibo Huang | f5f93a7 | 2020-10-19 15:43:42 -0700 | [diff] [blame] | 461 | return Subscript( |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 462 | value=Name(id='data', ctx=Load()), |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 463 | slice=Constant(value=node.id), |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 464 | ctx=node.ctx |
Haibo Huang | f5f93a7 | 2020-10-19 15:43:42 -0700 | [diff] [blame] | 465 | ) |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 466 | |
| 467 | Keep in mind that if the node you're operating on has child nodes you must |
| 468 | either transform the child nodes yourself or call the :meth:`generic_visit` |
| 469 | method for the node first. |
| 470 | |
| 471 | For nodes that were part of a collection of statements (that applies to all |
| 472 | statement nodes), the visitor may also return a list of nodes rather than |
| 473 | just a single node. |
| 474 | |
| 475 | Usually you use the transformer like this:: |
| 476 | |
| 477 | node = YourTransformer().visit(node) |
| 478 | """ |
| 479 | |
| 480 | def generic_visit(self, node): |
| 481 | for field, old_value in iter_fields(node): |
| 482 | if isinstance(old_value, list): |
| 483 | new_values = [] |
| 484 | for value in old_value: |
| 485 | if isinstance(value, AST): |
| 486 | value = self.visit(value) |
| 487 | if value is None: |
| 488 | continue |
| 489 | elif not isinstance(value, AST): |
| 490 | new_values.extend(value) |
| 491 | continue |
| 492 | new_values.append(value) |
| 493 | old_value[:] = new_values |
| 494 | elif isinstance(old_value, AST): |
| 495 | new_node = self.visit(old_value) |
| 496 | if new_node is None: |
| 497 | delattr(node, field) |
| 498 | else: |
| 499 | setattr(node, field, new_node) |
| 500 | return node |
| 501 | |
| 502 | |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 503 | # If the ast module is loaded more than once, only add deprecated methods once |
| 504 | if not hasattr(Constant, 'n'): |
| 505 | # The following code is for backward compatibility. |
| 506 | # It will be removed in future. |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 507 | |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 508 | def _getter(self): |
| 509 | """Deprecated. Use value instead.""" |
| 510 | return self.value |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 511 | |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 512 | def _setter(self, value): |
| 513 | self.value = value |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 514 | |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 515 | Constant.n = property(_getter, _setter) |
| 516 | Constant.s = property(_getter, _setter) |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 517 | |
| 518 | class _ABC(type): |
| 519 | |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 520 | def __init__(cls, *args): |
| 521 | cls.__doc__ = """Deprecated AST node class. Use ast.Constant instead""" |
| 522 | |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 523 | def __instancecheck__(cls, inst): |
| 524 | if not isinstance(inst, Constant): |
| 525 | return False |
| 526 | if cls in _const_types: |
| 527 | try: |
| 528 | value = inst.value |
| 529 | except AttributeError: |
| 530 | return False |
| 531 | else: |
| 532 | return ( |
| 533 | isinstance(value, _const_types[cls]) and |
| 534 | not isinstance(value, _const_types_not.get(cls, ())) |
| 535 | ) |
| 536 | return type.__instancecheck__(cls, inst) |
| 537 | |
| 538 | def _new(cls, *args, **kwargs): |
Haibo Huang | f5f93a7 | 2020-10-19 15:43:42 -0700 | [diff] [blame] | 539 | for key in kwargs: |
| 540 | if key not in cls._fields: |
| 541 | # arbitrary keyword arguments are accepted |
| 542 | continue |
| 543 | pos = cls._fields.index(key) |
| 544 | if pos < len(args): |
| 545 | raise TypeError(f"{cls.__name__} got multiple values for argument {key!r}") |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 546 | if cls in _const_types: |
| 547 | return Constant(*args, **kwargs) |
| 548 | return Constant.__new__(cls, *args, **kwargs) |
| 549 | |
| 550 | class Num(Constant, metaclass=_ABC): |
| 551 | _fields = ('n',) |
| 552 | __new__ = _new |
| 553 | |
| 554 | class Str(Constant, metaclass=_ABC): |
| 555 | _fields = ('s',) |
| 556 | __new__ = _new |
| 557 | |
| 558 | class Bytes(Constant, metaclass=_ABC): |
| 559 | _fields = ('s',) |
| 560 | __new__ = _new |
| 561 | |
| 562 | class NameConstant(Constant, metaclass=_ABC): |
| 563 | __new__ = _new |
| 564 | |
| 565 | class Ellipsis(Constant, metaclass=_ABC): |
| 566 | _fields = () |
| 567 | |
| 568 | def __new__(cls, *args, **kwargs): |
| 569 | if cls is Ellipsis: |
| 570 | return Constant(..., *args, **kwargs) |
| 571 | return Constant.__new__(cls, *args, **kwargs) |
| 572 | |
| 573 | _const_types = { |
| 574 | Num: (int, float, complex), |
| 575 | Str: (str,), |
| 576 | Bytes: (bytes,), |
| 577 | NameConstant: (type(None), bool), |
| 578 | Ellipsis: (type(...),), |
| 579 | } |
| 580 | _const_types_not = { |
| 581 | Num: (bool,), |
| 582 | } |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 583 | |
Haibo Huang | d883030 | 2020-03-03 10:09:46 -0800 | [diff] [blame] | 584 | _const_node_type_names = { |
| 585 | bool: 'NameConstant', # should be before int |
| 586 | type(None): 'NameConstant', |
| 587 | int: 'Num', |
| 588 | float: 'Num', |
| 589 | complex: 'Num', |
| 590 | str: 'Str', |
| 591 | bytes: 'Bytes', |
| 592 | type(...): 'Ellipsis', |
| 593 | } |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 594 | |
| 595 | class slice(AST): |
| 596 | """Deprecated AST node class.""" |
| 597 | |
| 598 | class Index(slice): |
| 599 | """Deprecated AST node class. Use the index value directly instead.""" |
| 600 | def __new__(cls, value, **kwargs): |
| 601 | return value |
| 602 | |
| 603 | class ExtSlice(slice): |
| 604 | """Deprecated AST node class. Use ast.Tuple instead.""" |
| 605 | def __new__(cls, dims=(), **kwargs): |
| 606 | return Tuple(list(dims), Load(), **kwargs) |
| 607 | |
| 608 | # If the ast module is loaded more than once, only add deprecated methods once |
| 609 | if not hasattr(Tuple, 'dims'): |
| 610 | # The following code is for backward compatibility. |
| 611 | # It will be removed in future. |
| 612 | |
| 613 | def _dims_getter(self): |
| 614 | """Deprecated. Use elts instead.""" |
| 615 | return self.elts |
| 616 | |
| 617 | def _dims_setter(self, value): |
| 618 | self.elts = value |
| 619 | |
| 620 | Tuple.dims = property(_dims_getter, _dims_setter) |
| 621 | |
| 622 | class Suite(mod): |
| 623 | """Deprecated AST node class. Unused in Python 3.""" |
| 624 | |
| 625 | class AugLoad(expr_context): |
| 626 | """Deprecated AST node class. Unused in Python 3.""" |
| 627 | |
| 628 | class AugStore(expr_context): |
| 629 | """Deprecated AST node class. Unused in Python 3.""" |
| 630 | |
| 631 | class Param(expr_context): |
| 632 | """Deprecated AST node class. Unused in Python 3.""" |
| 633 | |
| 634 | |
| 635 | # Large float and imaginary literals get turned into infinities in the AST. |
| 636 | # We unparse those infinities to INFSTR. |
| 637 | _INFSTR = "1e" + repr(sys.float_info.max_10_exp + 1) |
| 638 | |
| 639 | class _Precedence(IntEnum): |
| 640 | """Precedence table that originated from python grammar.""" |
| 641 | |
| 642 | TUPLE = auto() |
| 643 | YIELD = auto() # 'yield', 'yield from' |
| 644 | TEST = auto() # 'if'-'else', 'lambda' |
| 645 | OR = auto() # 'or' |
| 646 | AND = auto() # 'and' |
| 647 | NOT = auto() # 'not' |
| 648 | CMP = auto() # '<', '>', '==', '>=', '<=', '!=', |
| 649 | # 'in', 'not in', 'is', 'is not' |
| 650 | EXPR = auto() |
| 651 | BOR = EXPR # '|' |
| 652 | BXOR = auto() # '^' |
| 653 | BAND = auto() # '&' |
| 654 | SHIFT = auto() # '<<', '>>' |
| 655 | ARITH = auto() # '+', '-' |
| 656 | TERM = auto() # '*', '@', '/', '%', '//' |
| 657 | FACTOR = auto() # unary '+', '-', '~' |
| 658 | POWER = auto() # '**' |
| 659 | AWAIT = auto() # 'await' |
| 660 | ATOM = auto() |
| 661 | |
| 662 | def next(self): |
| 663 | try: |
| 664 | return self.__class__(self + 1) |
| 665 | except ValueError: |
| 666 | return self |
| 667 | |
| 668 | |
| 669 | _SINGLE_QUOTES = ("'", '"') |
| 670 | _MULTI_QUOTES = ('"""', "'''") |
| 671 | _ALL_QUOTES = (*_SINGLE_QUOTES, *_MULTI_QUOTES) |
| 672 | |
| 673 | class _Unparser(NodeVisitor): |
| 674 | """Methods in this class recursively traverse an AST and |
| 675 | output source code for the abstract syntax; original formatting |
| 676 | is disregarded.""" |
| 677 | |
| 678 | def __init__(self, *, _avoid_backslashes=False): |
| 679 | self._source = [] |
| 680 | self._buffer = [] |
| 681 | self._precedences = {} |
| 682 | self._type_ignores = {} |
| 683 | self._indent = 0 |
| 684 | self._avoid_backslashes = _avoid_backslashes |
| 685 | |
| 686 | def interleave(self, inter, f, seq): |
| 687 | """Call f on each item in seq, calling inter() in between.""" |
| 688 | seq = iter(seq) |
| 689 | try: |
| 690 | f(next(seq)) |
| 691 | except StopIteration: |
| 692 | pass |
| 693 | else: |
| 694 | for x in seq: |
| 695 | inter() |
| 696 | f(x) |
| 697 | |
| 698 | def items_view(self, traverser, items): |
| 699 | """Traverse and separate the given *items* with a comma and append it to |
| 700 | the buffer. If *items* is a single item sequence, a trailing comma |
| 701 | will be added.""" |
| 702 | if len(items) == 1: |
| 703 | traverser(items[0]) |
| 704 | self.write(",") |
| 705 | else: |
| 706 | self.interleave(lambda: self.write(", "), traverser, items) |
| 707 | |
| 708 | def maybe_newline(self): |
| 709 | """Adds a newline if it isn't the start of generated source""" |
| 710 | if self._source: |
| 711 | self.write("\n") |
| 712 | |
| 713 | def fill(self, text=""): |
| 714 | """Indent a piece of text and append it, according to the current |
| 715 | indentation level""" |
| 716 | self.maybe_newline() |
| 717 | self.write(" " * self._indent + text) |
| 718 | |
| 719 | def write(self, text): |
| 720 | """Append a piece of text""" |
| 721 | self._source.append(text) |
| 722 | |
| 723 | def buffer_writer(self, text): |
| 724 | self._buffer.append(text) |
| 725 | |
| 726 | @property |
| 727 | def buffer(self): |
| 728 | value = "".join(self._buffer) |
| 729 | self._buffer.clear() |
| 730 | return value |
| 731 | |
| 732 | @contextmanager |
| 733 | def block(self, *, extra = None): |
| 734 | """A context manager for preparing the source for blocks. It adds |
| 735 | the character':', increases the indentation on enter and decreases |
| 736 | the indentation on exit. If *extra* is given, it will be directly |
| 737 | appended after the colon character. |
| 738 | """ |
| 739 | self.write(":") |
| 740 | if extra: |
| 741 | self.write(extra) |
| 742 | self._indent += 1 |
| 743 | yield |
| 744 | self._indent -= 1 |
| 745 | |
| 746 | @contextmanager |
| 747 | def delimit(self, start, end): |
| 748 | """A context manager for preparing the source for expressions. It adds |
| 749 | *start* to the buffer and enters, after exit it adds *end*.""" |
| 750 | |
| 751 | self.write(start) |
| 752 | yield |
| 753 | self.write(end) |
| 754 | |
| 755 | def delimit_if(self, start, end, condition): |
| 756 | if condition: |
| 757 | return self.delimit(start, end) |
| 758 | else: |
| 759 | return nullcontext() |
| 760 | |
| 761 | def require_parens(self, precedence, node): |
| 762 | """Shortcut to adding precedence related parens""" |
| 763 | return self.delimit_if("(", ")", self.get_precedence(node) > precedence) |
| 764 | |
| 765 | def get_precedence(self, node): |
| 766 | return self._precedences.get(node, _Precedence.TEST) |
| 767 | |
| 768 | def set_precedence(self, precedence, *nodes): |
| 769 | for node in nodes: |
| 770 | self._precedences[node] = precedence |
| 771 | |
| 772 | def get_raw_docstring(self, node): |
| 773 | """If a docstring node is found in the body of the *node* parameter, |
| 774 | return that docstring node, None otherwise. |
| 775 | |
| 776 | Logic mirrored from ``_PyAST_GetDocString``.""" |
| 777 | if not isinstance( |
| 778 | node, (AsyncFunctionDef, FunctionDef, ClassDef, Module) |
| 779 | ) or len(node.body) < 1: |
| 780 | return None |
| 781 | node = node.body[0] |
| 782 | if not isinstance(node, Expr): |
| 783 | return None |
| 784 | node = node.value |
| 785 | if isinstance(node, Constant) and isinstance(node.value, str): |
| 786 | return node |
| 787 | |
| 788 | def get_type_comment(self, node): |
| 789 | comment = self._type_ignores.get(node.lineno) or node.type_comment |
| 790 | if comment is not None: |
| 791 | return f" # type: {comment}" |
| 792 | |
| 793 | def traverse(self, node): |
| 794 | if isinstance(node, list): |
| 795 | for item in node: |
| 796 | self.traverse(item) |
| 797 | else: |
| 798 | super().visit(node) |
| 799 | |
Yi Kong | 7119932 | 2022-08-30 15:53:45 +0800 | [diff] [blame] | 800 | # Note: as visit() resets the output text, do NOT rely on |
| 801 | # NodeVisitor.generic_visit to handle any nodes (as it calls back in to |
| 802 | # the subclass visit() method, which resets self._source to an empty list) |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 803 | def visit(self, node): |
| 804 | """Outputs a source code string that, if converted back to an ast |
| 805 | (using ast.parse) will generate an AST equivalent to *node*""" |
| 806 | self._source = [] |
| 807 | self.traverse(node) |
| 808 | return "".join(self._source) |
| 809 | |
| 810 | def _write_docstring_and_traverse_body(self, node): |
| 811 | if (docstring := self.get_raw_docstring(node)): |
| 812 | self._write_docstring(docstring) |
| 813 | self.traverse(node.body[1:]) |
| 814 | else: |
| 815 | self.traverse(node.body) |
| 816 | |
| 817 | def visit_Module(self, node): |
| 818 | self._type_ignores = { |
| 819 | ignore.lineno: f"ignore{ignore.tag}" |
| 820 | for ignore in node.type_ignores |
| 821 | } |
| 822 | self._write_docstring_and_traverse_body(node) |
| 823 | self._type_ignores.clear() |
| 824 | |
| 825 | def visit_FunctionType(self, node): |
| 826 | with self.delimit("(", ")"): |
| 827 | self.interleave( |
| 828 | lambda: self.write(", "), self.traverse, node.argtypes |
| 829 | ) |
| 830 | |
| 831 | self.write(" -> ") |
| 832 | self.traverse(node.returns) |
| 833 | |
| 834 | def visit_Expr(self, node): |
| 835 | self.fill() |
| 836 | self.set_precedence(_Precedence.YIELD, node.value) |
| 837 | self.traverse(node.value) |
| 838 | |
| 839 | def visit_NamedExpr(self, node): |
| 840 | with self.require_parens(_Precedence.TUPLE, node): |
| 841 | self.set_precedence(_Precedence.ATOM, node.target, node.value) |
| 842 | self.traverse(node.target) |
| 843 | self.write(" := ") |
| 844 | self.traverse(node.value) |
| 845 | |
| 846 | def visit_Import(self, node): |
| 847 | self.fill("import ") |
| 848 | self.interleave(lambda: self.write(", "), self.traverse, node.names) |
| 849 | |
| 850 | def visit_ImportFrom(self, node): |
| 851 | self.fill("from ") |
| 852 | self.write("." * node.level) |
| 853 | if node.module: |
| 854 | self.write(node.module) |
| 855 | self.write(" import ") |
| 856 | self.interleave(lambda: self.write(", "), self.traverse, node.names) |
| 857 | |
| 858 | def visit_Assign(self, node): |
| 859 | self.fill() |
| 860 | for target in node.targets: |
| 861 | self.traverse(target) |
| 862 | self.write(" = ") |
| 863 | self.traverse(node.value) |
| 864 | if type_comment := self.get_type_comment(node): |
| 865 | self.write(type_comment) |
| 866 | |
| 867 | def visit_AugAssign(self, node): |
| 868 | self.fill() |
| 869 | self.traverse(node.target) |
| 870 | self.write(" " + self.binop[node.op.__class__.__name__] + "= ") |
| 871 | self.traverse(node.value) |
| 872 | |
| 873 | def visit_AnnAssign(self, node): |
| 874 | self.fill() |
| 875 | with self.delimit_if("(", ")", not node.simple and isinstance(node.target, Name)): |
| 876 | self.traverse(node.target) |
| 877 | self.write(": ") |
| 878 | self.traverse(node.annotation) |
| 879 | if node.value: |
| 880 | self.write(" = ") |
| 881 | self.traverse(node.value) |
| 882 | |
| 883 | def visit_Return(self, node): |
| 884 | self.fill("return") |
| 885 | if node.value: |
| 886 | self.write(" ") |
| 887 | self.traverse(node.value) |
| 888 | |
| 889 | def visit_Pass(self, node): |
| 890 | self.fill("pass") |
| 891 | |
| 892 | def visit_Break(self, node): |
| 893 | self.fill("break") |
| 894 | |
| 895 | def visit_Continue(self, node): |
| 896 | self.fill("continue") |
| 897 | |
| 898 | def visit_Delete(self, node): |
| 899 | self.fill("del ") |
| 900 | self.interleave(lambda: self.write(", "), self.traverse, node.targets) |
| 901 | |
| 902 | def visit_Assert(self, node): |
| 903 | self.fill("assert ") |
| 904 | self.traverse(node.test) |
| 905 | if node.msg: |
| 906 | self.write(", ") |
| 907 | self.traverse(node.msg) |
| 908 | |
| 909 | def visit_Global(self, node): |
| 910 | self.fill("global ") |
| 911 | self.interleave(lambda: self.write(", "), self.write, node.names) |
| 912 | |
| 913 | def visit_Nonlocal(self, node): |
| 914 | self.fill("nonlocal ") |
| 915 | self.interleave(lambda: self.write(", "), self.write, node.names) |
| 916 | |
| 917 | def visit_Await(self, node): |
| 918 | with self.require_parens(_Precedence.AWAIT, node): |
| 919 | self.write("await") |
| 920 | if node.value: |
| 921 | self.write(" ") |
| 922 | self.set_precedence(_Precedence.ATOM, node.value) |
| 923 | self.traverse(node.value) |
| 924 | |
| 925 | def visit_Yield(self, node): |
| 926 | with self.require_parens(_Precedence.YIELD, node): |
| 927 | self.write("yield") |
| 928 | if node.value: |
| 929 | self.write(" ") |
| 930 | self.set_precedence(_Precedence.ATOM, node.value) |
| 931 | self.traverse(node.value) |
| 932 | |
| 933 | def visit_YieldFrom(self, node): |
| 934 | with self.require_parens(_Precedence.YIELD, node): |
| 935 | self.write("yield from ") |
| 936 | if not node.value: |
| 937 | raise ValueError("Node can't be used without a value attribute.") |
| 938 | self.set_precedence(_Precedence.ATOM, node.value) |
| 939 | self.traverse(node.value) |
| 940 | |
| 941 | def visit_Raise(self, node): |
| 942 | self.fill("raise") |
| 943 | if not node.exc: |
| 944 | if node.cause: |
| 945 | raise ValueError(f"Node can't use cause without an exception.") |
| 946 | return |
| 947 | self.write(" ") |
| 948 | self.traverse(node.exc) |
| 949 | if node.cause: |
| 950 | self.write(" from ") |
| 951 | self.traverse(node.cause) |
| 952 | |
| 953 | def visit_Try(self, node): |
| 954 | self.fill("try") |
| 955 | with self.block(): |
| 956 | self.traverse(node.body) |
| 957 | for ex in node.handlers: |
| 958 | self.traverse(ex) |
| 959 | if node.orelse: |
| 960 | self.fill("else") |
| 961 | with self.block(): |
| 962 | self.traverse(node.orelse) |
| 963 | if node.finalbody: |
| 964 | self.fill("finally") |
| 965 | with self.block(): |
| 966 | self.traverse(node.finalbody) |
| 967 | |
| 968 | def visit_ExceptHandler(self, node): |
| 969 | self.fill("except") |
| 970 | if node.type: |
| 971 | self.write(" ") |
| 972 | self.traverse(node.type) |
| 973 | if node.name: |
| 974 | self.write(" as ") |
| 975 | self.write(node.name) |
| 976 | with self.block(): |
| 977 | self.traverse(node.body) |
| 978 | |
| 979 | def visit_ClassDef(self, node): |
| 980 | self.maybe_newline() |
| 981 | for deco in node.decorator_list: |
| 982 | self.fill("@") |
| 983 | self.traverse(deco) |
| 984 | self.fill("class " + node.name) |
| 985 | with self.delimit_if("(", ")", condition = node.bases or node.keywords): |
| 986 | comma = False |
| 987 | for e in node.bases: |
| 988 | if comma: |
| 989 | self.write(", ") |
| 990 | else: |
| 991 | comma = True |
| 992 | self.traverse(e) |
| 993 | for e in node.keywords: |
| 994 | if comma: |
| 995 | self.write(", ") |
| 996 | else: |
| 997 | comma = True |
| 998 | self.traverse(e) |
| 999 | |
| 1000 | with self.block(): |
| 1001 | self._write_docstring_and_traverse_body(node) |
| 1002 | |
| 1003 | def visit_FunctionDef(self, node): |
| 1004 | self._function_helper(node, "def") |
| 1005 | |
| 1006 | def visit_AsyncFunctionDef(self, node): |
| 1007 | self._function_helper(node, "async def") |
| 1008 | |
| 1009 | def _function_helper(self, node, fill_suffix): |
| 1010 | self.maybe_newline() |
| 1011 | for deco in node.decorator_list: |
| 1012 | self.fill("@") |
| 1013 | self.traverse(deco) |
| 1014 | def_str = fill_suffix + " " + node.name |
| 1015 | self.fill(def_str) |
| 1016 | with self.delimit("(", ")"): |
| 1017 | self.traverse(node.args) |
| 1018 | if node.returns: |
| 1019 | self.write(" -> ") |
| 1020 | self.traverse(node.returns) |
| 1021 | with self.block(extra=self.get_type_comment(node)): |
| 1022 | self._write_docstring_and_traverse_body(node) |
| 1023 | |
| 1024 | def visit_For(self, node): |
| 1025 | self._for_helper("for ", node) |
| 1026 | |
| 1027 | def visit_AsyncFor(self, node): |
| 1028 | self._for_helper("async for ", node) |
| 1029 | |
| 1030 | def _for_helper(self, fill, node): |
| 1031 | self.fill(fill) |
| 1032 | self.traverse(node.target) |
| 1033 | self.write(" in ") |
| 1034 | self.traverse(node.iter) |
| 1035 | with self.block(extra=self.get_type_comment(node)): |
| 1036 | self.traverse(node.body) |
| 1037 | if node.orelse: |
| 1038 | self.fill("else") |
| 1039 | with self.block(): |
| 1040 | self.traverse(node.orelse) |
| 1041 | |
| 1042 | def visit_If(self, node): |
| 1043 | self.fill("if ") |
| 1044 | self.traverse(node.test) |
| 1045 | with self.block(): |
| 1046 | self.traverse(node.body) |
| 1047 | # collapse nested ifs into equivalent elifs. |
| 1048 | while node.orelse and len(node.orelse) == 1 and isinstance(node.orelse[0], If): |
| 1049 | node = node.orelse[0] |
| 1050 | self.fill("elif ") |
| 1051 | self.traverse(node.test) |
| 1052 | with self.block(): |
| 1053 | self.traverse(node.body) |
| 1054 | # final else |
| 1055 | if node.orelse: |
| 1056 | self.fill("else") |
| 1057 | with self.block(): |
| 1058 | self.traverse(node.orelse) |
| 1059 | |
| 1060 | def visit_While(self, node): |
| 1061 | self.fill("while ") |
| 1062 | self.traverse(node.test) |
| 1063 | with self.block(): |
| 1064 | self.traverse(node.body) |
| 1065 | if node.orelse: |
| 1066 | self.fill("else") |
| 1067 | with self.block(): |
| 1068 | self.traverse(node.orelse) |
| 1069 | |
| 1070 | def visit_With(self, node): |
| 1071 | self.fill("with ") |
| 1072 | self.interleave(lambda: self.write(", "), self.traverse, node.items) |
| 1073 | with self.block(extra=self.get_type_comment(node)): |
| 1074 | self.traverse(node.body) |
| 1075 | |
| 1076 | def visit_AsyncWith(self, node): |
| 1077 | self.fill("async with ") |
| 1078 | self.interleave(lambda: self.write(", "), self.traverse, node.items) |
| 1079 | with self.block(extra=self.get_type_comment(node)): |
| 1080 | self.traverse(node.body) |
| 1081 | |
| 1082 | def _str_literal_helper( |
| 1083 | self, string, *, quote_types=_ALL_QUOTES, escape_special_whitespace=False |
| 1084 | ): |
| 1085 | """Helper for writing string literals, minimizing escapes. |
| 1086 | Returns the tuple (string literal to write, possible quote types). |
| 1087 | """ |
| 1088 | def escape_char(c): |
| 1089 | # \n and \t are non-printable, but we only escape them if |
| 1090 | # escape_special_whitespace is True |
| 1091 | if not escape_special_whitespace and c in "\n\t": |
| 1092 | return c |
| 1093 | # Always escape backslashes and other non-printable characters |
| 1094 | if c == "\\" or not c.isprintable(): |
| 1095 | return c.encode("unicode_escape").decode("ascii") |
| 1096 | return c |
| 1097 | |
| 1098 | escaped_string = "".join(map(escape_char, string)) |
| 1099 | possible_quotes = quote_types |
| 1100 | if "\n" in escaped_string: |
| 1101 | possible_quotes = [q for q in possible_quotes if q in _MULTI_QUOTES] |
| 1102 | possible_quotes = [q for q in possible_quotes if q not in escaped_string] |
| 1103 | if not possible_quotes: |
| 1104 | # If there aren't any possible_quotes, fallback to using repr |
| 1105 | # on the original string. Try to use a quote from quote_types, |
| 1106 | # e.g., so that we use triple quotes for docstrings. |
| 1107 | string = repr(string) |
| 1108 | quote = next((q for q in quote_types if string[0] in q), string[0]) |
| 1109 | return string[1:-1], [quote] |
| 1110 | if escaped_string: |
| 1111 | # Sort so that we prefer '''"''' over """\"""" |
| 1112 | possible_quotes.sort(key=lambda q: q[0] == escaped_string[-1]) |
| 1113 | # If we're using triple quotes and we'd need to escape a final |
| 1114 | # quote, escape it |
| 1115 | if possible_quotes[0][0] == escaped_string[-1]: |
| 1116 | assert len(possible_quotes[0]) == 3 |
| 1117 | escaped_string = escaped_string[:-1] + "\\" + escaped_string[-1] |
| 1118 | return escaped_string, possible_quotes |
| 1119 | |
| 1120 | def _write_str_avoiding_backslashes(self, string, *, quote_types=_ALL_QUOTES): |
| 1121 | """Write string literal value with a best effort attempt to avoid backslashes.""" |
| 1122 | string, quote_types = self._str_literal_helper(string, quote_types=quote_types) |
| 1123 | quote_type = quote_types[0] |
| 1124 | self.write(f"{quote_type}{string}{quote_type}") |
| 1125 | |
| 1126 | def visit_JoinedStr(self, node): |
| 1127 | self.write("f") |
| 1128 | if self._avoid_backslashes: |
| 1129 | self._fstring_JoinedStr(node, self.buffer_writer) |
| 1130 | self._write_str_avoiding_backslashes(self.buffer) |
| 1131 | return |
| 1132 | |
| 1133 | # If we don't need to avoid backslashes globally (i.e., we only need |
| 1134 | # to avoid them inside FormattedValues), it's cosmetically preferred |
| 1135 | # to use escaped whitespace. That is, it's preferred to use backslashes |
| 1136 | # for cases like: f"{x}\n". To accomplish this, we keep track of what |
| 1137 | # in our buffer corresponds to FormattedValues and what corresponds to |
| 1138 | # Constant parts of the f-string, and allow escapes accordingly. |
| 1139 | buffer = [] |
| 1140 | for value in node.values: |
| 1141 | meth = getattr(self, "_fstring_" + type(value).__name__) |
| 1142 | meth(value, self.buffer_writer) |
| 1143 | buffer.append((self.buffer, isinstance(value, Constant))) |
| 1144 | new_buffer = [] |
| 1145 | quote_types = _ALL_QUOTES |
| 1146 | for value, is_constant in buffer: |
| 1147 | # Repeatedly narrow down the list of possible quote_types |
| 1148 | value, quote_types = self._str_literal_helper( |
| 1149 | value, quote_types=quote_types, |
| 1150 | escape_special_whitespace=is_constant |
| 1151 | ) |
| 1152 | new_buffer.append(value) |
| 1153 | value = "".join(new_buffer) |
| 1154 | quote_type = quote_types[0] |
| 1155 | self.write(f"{quote_type}{value}{quote_type}") |
| 1156 | |
| 1157 | def visit_FormattedValue(self, node): |
| 1158 | self.write("f") |
| 1159 | self._fstring_FormattedValue(node, self.buffer_writer) |
| 1160 | self._write_str_avoiding_backslashes(self.buffer) |
| 1161 | |
| 1162 | def _fstring_JoinedStr(self, node, write): |
| 1163 | for value in node.values: |
| 1164 | meth = getattr(self, "_fstring_" + type(value).__name__) |
| 1165 | meth(value, write) |
| 1166 | |
| 1167 | def _fstring_Constant(self, node, write): |
| 1168 | if not isinstance(node.value, str): |
| 1169 | raise ValueError("Constants inside JoinedStr should be a string.") |
| 1170 | value = node.value.replace("{", "{{").replace("}", "}}") |
| 1171 | write(value) |
| 1172 | |
| 1173 | def _fstring_FormattedValue(self, node, write): |
| 1174 | write("{") |
| 1175 | unparser = type(self)(_avoid_backslashes=True) |
| 1176 | unparser.set_precedence(_Precedence.TEST.next(), node.value) |
| 1177 | expr = unparser.visit(node.value) |
| 1178 | if expr.startswith("{"): |
| 1179 | write(" ") # Separate pair of opening brackets as "{ {" |
| 1180 | if "\\" in expr: |
| 1181 | raise ValueError("Unable to avoid backslash in f-string expression part") |
| 1182 | write(expr) |
| 1183 | if node.conversion != -1: |
| 1184 | conversion = chr(node.conversion) |
| 1185 | if conversion not in "sra": |
| 1186 | raise ValueError("Unknown f-string conversion.") |
| 1187 | write(f"!{conversion}") |
| 1188 | if node.format_spec: |
| 1189 | write(":") |
| 1190 | meth = getattr(self, "_fstring_" + type(node.format_spec).__name__) |
| 1191 | meth(node.format_spec, write) |
| 1192 | write("}") |
| 1193 | |
| 1194 | def visit_Name(self, node): |
| 1195 | self.write(node.id) |
| 1196 | |
| 1197 | def _write_docstring(self, node): |
| 1198 | self.fill() |
| 1199 | if node.kind == "u": |
| 1200 | self.write("u") |
| 1201 | self._write_str_avoiding_backslashes(node.value, quote_types=_MULTI_QUOTES) |
| 1202 | |
| 1203 | def _write_constant(self, value): |
| 1204 | if isinstance(value, (float, complex)): |
Yi Kong | 7119932 | 2022-08-30 15:53:45 +0800 | [diff] [blame] | 1205 | # Substitute overflowing decimal literal for AST infinities, |
| 1206 | # and inf - inf for NaNs. |
| 1207 | self.write( |
| 1208 | repr(value) |
| 1209 | .replace("inf", _INFSTR) |
| 1210 | .replace("nan", f"({_INFSTR}-{_INFSTR})") |
| 1211 | ) |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 1212 | elif self._avoid_backslashes and isinstance(value, str): |
| 1213 | self._write_str_avoiding_backslashes(value) |
| 1214 | else: |
| 1215 | self.write(repr(value)) |
| 1216 | |
| 1217 | def visit_Constant(self, node): |
| 1218 | value = node.value |
| 1219 | if isinstance(value, tuple): |
| 1220 | with self.delimit("(", ")"): |
| 1221 | self.items_view(self._write_constant, value) |
| 1222 | elif value is ...: |
| 1223 | self.write("...") |
| 1224 | else: |
| 1225 | if node.kind == "u": |
| 1226 | self.write("u") |
| 1227 | self._write_constant(node.value) |
| 1228 | |
| 1229 | def visit_List(self, node): |
| 1230 | with self.delimit("[", "]"): |
| 1231 | self.interleave(lambda: self.write(", "), self.traverse, node.elts) |
| 1232 | |
| 1233 | def visit_ListComp(self, node): |
| 1234 | with self.delimit("[", "]"): |
| 1235 | self.traverse(node.elt) |
| 1236 | for gen in node.generators: |
| 1237 | self.traverse(gen) |
| 1238 | |
| 1239 | def visit_GeneratorExp(self, node): |
| 1240 | with self.delimit("(", ")"): |
| 1241 | self.traverse(node.elt) |
| 1242 | for gen in node.generators: |
| 1243 | self.traverse(gen) |
| 1244 | |
| 1245 | def visit_SetComp(self, node): |
| 1246 | with self.delimit("{", "}"): |
| 1247 | self.traverse(node.elt) |
| 1248 | for gen in node.generators: |
| 1249 | self.traverse(gen) |
| 1250 | |
| 1251 | def visit_DictComp(self, node): |
| 1252 | with self.delimit("{", "}"): |
| 1253 | self.traverse(node.key) |
| 1254 | self.write(": ") |
| 1255 | self.traverse(node.value) |
| 1256 | for gen in node.generators: |
| 1257 | self.traverse(gen) |
| 1258 | |
| 1259 | def visit_comprehension(self, node): |
| 1260 | if node.is_async: |
| 1261 | self.write(" async for ") |
| 1262 | else: |
| 1263 | self.write(" for ") |
| 1264 | self.set_precedence(_Precedence.TUPLE, node.target) |
| 1265 | self.traverse(node.target) |
| 1266 | self.write(" in ") |
| 1267 | self.set_precedence(_Precedence.TEST.next(), node.iter, *node.ifs) |
| 1268 | self.traverse(node.iter) |
| 1269 | for if_clause in node.ifs: |
| 1270 | self.write(" if ") |
| 1271 | self.traverse(if_clause) |
| 1272 | |
| 1273 | def visit_IfExp(self, node): |
| 1274 | with self.require_parens(_Precedence.TEST, node): |
| 1275 | self.set_precedence(_Precedence.TEST.next(), node.body, node.test) |
| 1276 | self.traverse(node.body) |
| 1277 | self.write(" if ") |
| 1278 | self.traverse(node.test) |
| 1279 | self.write(" else ") |
| 1280 | self.set_precedence(_Precedence.TEST, node.orelse) |
| 1281 | self.traverse(node.orelse) |
| 1282 | |
| 1283 | def visit_Set(self, node): |
Yi Kong | 7119932 | 2022-08-30 15:53:45 +0800 | [diff] [blame] | 1284 | if node.elts: |
| 1285 | with self.delimit("{", "}"): |
| 1286 | self.interleave(lambda: self.write(", "), self.traverse, node.elts) |
| 1287 | else: |
| 1288 | # `{}` would be interpreted as a dictionary literal, and |
| 1289 | # `set` might be shadowed. Thus: |
| 1290 | self.write('{*()}') |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 1291 | |
| 1292 | def visit_Dict(self, node): |
| 1293 | def write_key_value_pair(k, v): |
| 1294 | self.traverse(k) |
| 1295 | self.write(": ") |
| 1296 | self.traverse(v) |
| 1297 | |
| 1298 | def write_item(item): |
| 1299 | k, v = item |
| 1300 | if k is None: |
| 1301 | # for dictionary unpacking operator in dicts {**{'y': 2}} |
| 1302 | # see PEP 448 for details |
| 1303 | self.write("**") |
| 1304 | self.set_precedence(_Precedence.EXPR, v) |
| 1305 | self.traverse(v) |
| 1306 | else: |
| 1307 | write_key_value_pair(k, v) |
| 1308 | |
| 1309 | with self.delimit("{", "}"): |
| 1310 | self.interleave( |
| 1311 | lambda: self.write(", "), write_item, zip(node.keys, node.values) |
| 1312 | ) |
| 1313 | |
| 1314 | def visit_Tuple(self, node): |
| 1315 | with self.delimit("(", ")"): |
| 1316 | self.items_view(self.traverse, node.elts) |
| 1317 | |
| 1318 | unop = {"Invert": "~", "Not": "not", "UAdd": "+", "USub": "-"} |
| 1319 | unop_precedence = { |
| 1320 | "not": _Precedence.NOT, |
| 1321 | "~": _Precedence.FACTOR, |
| 1322 | "+": _Precedence.FACTOR, |
| 1323 | "-": _Precedence.FACTOR, |
| 1324 | } |
| 1325 | |
| 1326 | def visit_UnaryOp(self, node): |
| 1327 | operator = self.unop[node.op.__class__.__name__] |
| 1328 | operator_precedence = self.unop_precedence[operator] |
| 1329 | with self.require_parens(operator_precedence, node): |
| 1330 | self.write(operator) |
| 1331 | # factor prefixes (+, -, ~) shouldn't be seperated |
| 1332 | # from the value they belong, (e.g: +1 instead of + 1) |
| 1333 | if operator_precedence is not _Precedence.FACTOR: |
| 1334 | self.write(" ") |
| 1335 | self.set_precedence(operator_precedence, node.operand) |
| 1336 | self.traverse(node.operand) |
| 1337 | |
| 1338 | binop = { |
| 1339 | "Add": "+", |
| 1340 | "Sub": "-", |
| 1341 | "Mult": "*", |
| 1342 | "MatMult": "@", |
| 1343 | "Div": "/", |
| 1344 | "Mod": "%", |
| 1345 | "LShift": "<<", |
| 1346 | "RShift": ">>", |
| 1347 | "BitOr": "|", |
| 1348 | "BitXor": "^", |
| 1349 | "BitAnd": "&", |
| 1350 | "FloorDiv": "//", |
| 1351 | "Pow": "**", |
| 1352 | } |
| 1353 | |
| 1354 | binop_precedence = { |
| 1355 | "+": _Precedence.ARITH, |
| 1356 | "-": _Precedence.ARITH, |
| 1357 | "*": _Precedence.TERM, |
| 1358 | "@": _Precedence.TERM, |
| 1359 | "/": _Precedence.TERM, |
| 1360 | "%": _Precedence.TERM, |
| 1361 | "<<": _Precedence.SHIFT, |
| 1362 | ">>": _Precedence.SHIFT, |
| 1363 | "|": _Precedence.BOR, |
| 1364 | "^": _Precedence.BXOR, |
| 1365 | "&": _Precedence.BAND, |
| 1366 | "//": _Precedence.TERM, |
| 1367 | "**": _Precedence.POWER, |
| 1368 | } |
| 1369 | |
| 1370 | binop_rassoc = frozenset(("**",)) |
| 1371 | def visit_BinOp(self, node): |
| 1372 | operator = self.binop[node.op.__class__.__name__] |
| 1373 | operator_precedence = self.binop_precedence[operator] |
| 1374 | with self.require_parens(operator_precedence, node): |
| 1375 | if operator in self.binop_rassoc: |
| 1376 | left_precedence = operator_precedence.next() |
| 1377 | right_precedence = operator_precedence |
| 1378 | else: |
| 1379 | left_precedence = operator_precedence |
| 1380 | right_precedence = operator_precedence.next() |
| 1381 | |
| 1382 | self.set_precedence(left_precedence, node.left) |
| 1383 | self.traverse(node.left) |
| 1384 | self.write(f" {operator} ") |
| 1385 | self.set_precedence(right_precedence, node.right) |
| 1386 | self.traverse(node.right) |
| 1387 | |
| 1388 | cmpops = { |
| 1389 | "Eq": "==", |
| 1390 | "NotEq": "!=", |
| 1391 | "Lt": "<", |
| 1392 | "LtE": "<=", |
| 1393 | "Gt": ">", |
| 1394 | "GtE": ">=", |
| 1395 | "Is": "is", |
| 1396 | "IsNot": "is not", |
| 1397 | "In": "in", |
| 1398 | "NotIn": "not in", |
| 1399 | } |
| 1400 | |
| 1401 | def visit_Compare(self, node): |
| 1402 | with self.require_parens(_Precedence.CMP, node): |
| 1403 | self.set_precedence(_Precedence.CMP.next(), node.left, *node.comparators) |
| 1404 | self.traverse(node.left) |
| 1405 | for o, e in zip(node.ops, node.comparators): |
| 1406 | self.write(" " + self.cmpops[o.__class__.__name__] + " ") |
| 1407 | self.traverse(e) |
| 1408 | |
| 1409 | boolops = {"And": "and", "Or": "or"} |
| 1410 | boolop_precedence = {"and": _Precedence.AND, "or": _Precedence.OR} |
| 1411 | |
| 1412 | def visit_BoolOp(self, node): |
| 1413 | operator = self.boolops[node.op.__class__.__name__] |
| 1414 | operator_precedence = self.boolop_precedence[operator] |
| 1415 | |
| 1416 | def increasing_level_traverse(node): |
| 1417 | nonlocal operator_precedence |
| 1418 | operator_precedence = operator_precedence.next() |
| 1419 | self.set_precedence(operator_precedence, node) |
| 1420 | self.traverse(node) |
| 1421 | |
| 1422 | with self.require_parens(operator_precedence, node): |
| 1423 | s = f" {operator} " |
| 1424 | self.interleave(lambda: self.write(s), increasing_level_traverse, node.values) |
| 1425 | |
| 1426 | def visit_Attribute(self, node): |
| 1427 | self.set_precedence(_Precedence.ATOM, node.value) |
| 1428 | self.traverse(node.value) |
| 1429 | # Special case: 3.__abs__() is a syntax error, so if node.value |
| 1430 | # is an integer literal then we need to either parenthesize |
| 1431 | # it or add an extra space to get 3 .__abs__(). |
| 1432 | if isinstance(node.value, Constant) and isinstance(node.value.value, int): |
| 1433 | self.write(" ") |
| 1434 | self.write(".") |
| 1435 | self.write(node.attr) |
| 1436 | |
| 1437 | def visit_Call(self, node): |
| 1438 | self.set_precedence(_Precedence.ATOM, node.func) |
| 1439 | self.traverse(node.func) |
| 1440 | with self.delimit("(", ")"): |
| 1441 | comma = False |
| 1442 | for e in node.args: |
| 1443 | if comma: |
| 1444 | self.write(", ") |
| 1445 | else: |
| 1446 | comma = True |
| 1447 | self.traverse(e) |
| 1448 | for e in node.keywords: |
| 1449 | if comma: |
| 1450 | self.write(", ") |
| 1451 | else: |
| 1452 | comma = True |
| 1453 | self.traverse(e) |
| 1454 | |
| 1455 | def visit_Subscript(self, node): |
| 1456 | def is_simple_tuple(slice_value): |
Yi Kong | 7119932 | 2022-08-30 15:53:45 +0800 | [diff] [blame] | 1457 | # when unparsing a non-empty tuple, the parentheses can be safely |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 1458 | # omitted if there aren't any elements that explicitly requires |
Yi Kong | 7119932 | 2022-08-30 15:53:45 +0800 | [diff] [blame] | 1459 | # parentheses (such as starred expressions). |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 1460 | return ( |
| 1461 | isinstance(slice_value, Tuple) |
| 1462 | and slice_value.elts |
| 1463 | and not any(isinstance(elt, Starred) for elt in slice_value.elts) |
| 1464 | ) |
| 1465 | |
| 1466 | self.set_precedence(_Precedence.ATOM, node.value) |
| 1467 | self.traverse(node.value) |
| 1468 | with self.delimit("[", "]"): |
| 1469 | if is_simple_tuple(node.slice): |
| 1470 | self.items_view(self.traverse, node.slice.elts) |
| 1471 | else: |
| 1472 | self.traverse(node.slice) |
| 1473 | |
| 1474 | def visit_Starred(self, node): |
| 1475 | self.write("*") |
| 1476 | self.set_precedence(_Precedence.EXPR, node.value) |
| 1477 | self.traverse(node.value) |
| 1478 | |
| 1479 | def visit_Ellipsis(self, node): |
| 1480 | self.write("...") |
| 1481 | |
| 1482 | def visit_Slice(self, node): |
| 1483 | if node.lower: |
| 1484 | self.traverse(node.lower) |
| 1485 | self.write(":") |
| 1486 | if node.upper: |
| 1487 | self.traverse(node.upper) |
| 1488 | if node.step: |
| 1489 | self.write(":") |
| 1490 | self.traverse(node.step) |
| 1491 | |
Yi Kong | 7119932 | 2022-08-30 15:53:45 +0800 | [diff] [blame] | 1492 | def visit_Match(self, node): |
| 1493 | self.fill("match ") |
| 1494 | self.traverse(node.subject) |
| 1495 | with self.block(): |
| 1496 | for case in node.cases: |
| 1497 | self.traverse(case) |
| 1498 | |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 1499 | def visit_arg(self, node): |
| 1500 | self.write(node.arg) |
| 1501 | if node.annotation: |
| 1502 | self.write(": ") |
| 1503 | self.traverse(node.annotation) |
| 1504 | |
| 1505 | def visit_arguments(self, node): |
| 1506 | first = True |
| 1507 | # normal arguments |
| 1508 | all_args = node.posonlyargs + node.args |
| 1509 | defaults = [None] * (len(all_args) - len(node.defaults)) + node.defaults |
| 1510 | for index, elements in enumerate(zip(all_args, defaults), 1): |
| 1511 | a, d = elements |
| 1512 | if first: |
| 1513 | first = False |
| 1514 | else: |
| 1515 | self.write(", ") |
| 1516 | self.traverse(a) |
| 1517 | if d: |
| 1518 | self.write("=") |
| 1519 | self.traverse(d) |
| 1520 | if index == len(node.posonlyargs): |
| 1521 | self.write(", /") |
| 1522 | |
| 1523 | # varargs, or bare '*' if no varargs but keyword-only arguments present |
| 1524 | if node.vararg or node.kwonlyargs: |
| 1525 | if first: |
| 1526 | first = False |
| 1527 | else: |
| 1528 | self.write(", ") |
| 1529 | self.write("*") |
| 1530 | if node.vararg: |
| 1531 | self.write(node.vararg.arg) |
| 1532 | if node.vararg.annotation: |
| 1533 | self.write(": ") |
| 1534 | self.traverse(node.vararg.annotation) |
| 1535 | |
| 1536 | # keyword-only arguments |
| 1537 | if node.kwonlyargs: |
| 1538 | for a, d in zip(node.kwonlyargs, node.kw_defaults): |
| 1539 | self.write(", ") |
| 1540 | self.traverse(a) |
| 1541 | if d: |
| 1542 | self.write("=") |
| 1543 | self.traverse(d) |
| 1544 | |
| 1545 | # kwargs |
| 1546 | if node.kwarg: |
| 1547 | if first: |
| 1548 | first = False |
| 1549 | else: |
| 1550 | self.write(", ") |
| 1551 | self.write("**" + node.kwarg.arg) |
| 1552 | if node.kwarg.annotation: |
| 1553 | self.write(": ") |
| 1554 | self.traverse(node.kwarg.annotation) |
| 1555 | |
| 1556 | def visit_keyword(self, node): |
| 1557 | if node.arg is None: |
| 1558 | self.write("**") |
| 1559 | else: |
| 1560 | self.write(node.arg) |
| 1561 | self.write("=") |
| 1562 | self.traverse(node.value) |
| 1563 | |
| 1564 | def visit_Lambda(self, node): |
| 1565 | with self.require_parens(_Precedence.TEST, node): |
| 1566 | self.write("lambda ") |
| 1567 | self.traverse(node.args) |
| 1568 | self.write(": ") |
| 1569 | self.set_precedence(_Precedence.TEST, node.body) |
| 1570 | self.traverse(node.body) |
| 1571 | |
| 1572 | def visit_alias(self, node): |
| 1573 | self.write(node.name) |
| 1574 | if node.asname: |
| 1575 | self.write(" as " + node.asname) |
| 1576 | |
| 1577 | def visit_withitem(self, node): |
| 1578 | self.traverse(node.context_expr) |
| 1579 | if node.optional_vars: |
| 1580 | self.write(" as ") |
| 1581 | self.traverse(node.optional_vars) |
| 1582 | |
Yi Kong | 7119932 | 2022-08-30 15:53:45 +0800 | [diff] [blame] | 1583 | def visit_match_case(self, node): |
| 1584 | self.fill("case ") |
| 1585 | self.traverse(node.pattern) |
| 1586 | if node.guard: |
| 1587 | self.write(" if ") |
| 1588 | self.traverse(node.guard) |
| 1589 | with self.block(): |
| 1590 | self.traverse(node.body) |
| 1591 | |
| 1592 | def visit_MatchValue(self, node): |
| 1593 | self.traverse(node.value) |
| 1594 | |
| 1595 | def visit_MatchSingleton(self, node): |
| 1596 | self._write_constant(node.value) |
| 1597 | |
| 1598 | def visit_MatchSequence(self, node): |
| 1599 | with self.delimit("[", "]"): |
| 1600 | self.interleave( |
| 1601 | lambda: self.write(", "), self.traverse, node.patterns |
| 1602 | ) |
| 1603 | |
| 1604 | def visit_MatchStar(self, node): |
| 1605 | name = node.name |
| 1606 | if name is None: |
| 1607 | name = "_" |
| 1608 | self.write(f"*{name}") |
| 1609 | |
| 1610 | def visit_MatchMapping(self, node): |
| 1611 | def write_key_pattern_pair(pair): |
| 1612 | k, p = pair |
| 1613 | self.traverse(k) |
| 1614 | self.write(": ") |
| 1615 | self.traverse(p) |
| 1616 | |
| 1617 | with self.delimit("{", "}"): |
| 1618 | keys = node.keys |
| 1619 | self.interleave( |
| 1620 | lambda: self.write(", "), |
| 1621 | write_key_pattern_pair, |
| 1622 | zip(keys, node.patterns, strict=True), |
| 1623 | ) |
| 1624 | rest = node.rest |
| 1625 | if rest is not None: |
| 1626 | if keys: |
| 1627 | self.write(", ") |
| 1628 | self.write(f"**{rest}") |
| 1629 | |
| 1630 | def visit_MatchClass(self, node): |
| 1631 | self.set_precedence(_Precedence.ATOM, node.cls) |
| 1632 | self.traverse(node.cls) |
| 1633 | with self.delimit("(", ")"): |
| 1634 | patterns = node.patterns |
| 1635 | self.interleave( |
| 1636 | lambda: self.write(", "), self.traverse, patterns |
| 1637 | ) |
| 1638 | attrs = node.kwd_attrs |
| 1639 | if attrs: |
| 1640 | def write_attr_pattern(pair): |
| 1641 | attr, pattern = pair |
| 1642 | self.write(f"{attr}=") |
| 1643 | self.traverse(pattern) |
| 1644 | |
| 1645 | if patterns: |
| 1646 | self.write(", ") |
| 1647 | self.interleave( |
| 1648 | lambda: self.write(", "), |
| 1649 | write_attr_pattern, |
| 1650 | zip(attrs, node.kwd_patterns, strict=True), |
| 1651 | ) |
| 1652 | |
| 1653 | def visit_MatchAs(self, node): |
| 1654 | name = node.name |
| 1655 | pattern = node.pattern |
| 1656 | if name is None: |
| 1657 | self.write("_") |
| 1658 | elif pattern is None: |
| 1659 | self.write(node.name) |
| 1660 | else: |
| 1661 | with self.require_parens(_Precedence.TEST, node): |
| 1662 | self.set_precedence(_Precedence.BOR, node.pattern) |
| 1663 | self.traverse(node.pattern) |
| 1664 | self.write(f" as {node.name}") |
| 1665 | |
| 1666 | def visit_MatchOr(self, node): |
| 1667 | with self.require_parens(_Precedence.BOR, node): |
| 1668 | self.set_precedence(_Precedence.BOR.next(), *node.patterns) |
| 1669 | self.interleave(lambda: self.write(" | "), self.traverse, node.patterns) |
| 1670 | |
Haibo Huang | 5eba2b4 | 2021-01-22 11:22:02 -0800 | [diff] [blame] | 1671 | def unparse(ast_obj): |
| 1672 | unparser = _Unparser() |
| 1673 | return unparser.visit(ast_obj) |
| 1674 | |
| 1675 | |
| 1676 | def main(): |
| 1677 | import argparse |
| 1678 | |
| 1679 | parser = argparse.ArgumentParser(prog='python -m ast') |
| 1680 | parser.add_argument('infile', type=argparse.FileType(mode='rb'), nargs='?', |
| 1681 | default='-', |
| 1682 | help='the file to parse; defaults to stdin') |
| 1683 | parser.add_argument('-m', '--mode', default='exec', |
| 1684 | choices=('exec', 'single', 'eval', 'func_type'), |
| 1685 | help='specify what kind of code must be parsed') |
| 1686 | parser.add_argument('--no-type-comments', default=True, action='store_false', |
| 1687 | help="don't add information about type comments") |
| 1688 | parser.add_argument('-a', '--include-attributes', action='store_true', |
| 1689 | help='include attributes such as line numbers and ' |
| 1690 | 'column offsets') |
| 1691 | parser.add_argument('-i', '--indent', type=int, default=3, |
| 1692 | help='indentation of nodes (number of spaces)') |
| 1693 | args = parser.parse_args() |
| 1694 | |
| 1695 | with args.infile as infile: |
| 1696 | source = infile.read() |
| 1697 | tree = parse(source, args.infile.name, args.mode, type_comments=args.no_type_comments) |
| 1698 | print(dump(tree, include_attributes=args.include_attributes, indent=args.indent)) |
| 1699 | |
| 1700 | if __name__ == '__main__': |
| 1701 | main() |