| 1 | n/a | """ |
|---|
| 2 | n/a | ast |
|---|
| 3 | n/a | ~~~ |
|---|
| 4 | n/a | |
|---|
| 5 | n/a | The `ast` module helps Python applications to process trees of the Python |
|---|
| 6 | n/a | abstract syntax grammar. The abstract syntax itself might change with |
|---|
| 7 | n/a | each Python release; this module helps to find out programmatically what |
|---|
| 8 | n/a | the current grammar looks like and allows modifications of it. |
|---|
| 9 | n/a | |
|---|
| 10 | n/a | An abstract syntax tree can be generated by passing `ast.PyCF_ONLY_AST` as |
|---|
| 11 | n/a | a flag to the `compile()` builtin function or by using the `parse()` |
|---|
| 12 | n/a | function from this module. The result will be a tree of objects whose |
|---|
| 13 | n/a | classes all inherit from `ast.AST`. |
|---|
| 14 | n/a | |
|---|
| 15 | n/a | A modified abstract syntax tree can be compiled into a Python code object |
|---|
| 16 | n/a | using the built-in `compile()` function. |
|---|
| 17 | n/a | |
|---|
| 18 | n/a | Additionally various helper functions are provided that make working with |
|---|
| 19 | n/a | the trees simpler. The main intention of the helper functions and this |
|---|
| 20 | n/a | module in general is to provide an easy to use interface for libraries |
|---|
| 21 | n/a | that work tightly with the python syntax (template engines for example). |
|---|
| 22 | n/a | |
|---|
| 23 | n/a | |
|---|
| 24 | n/a | :copyright: Copyright 2008 by Armin Ronacher. |
|---|
| 25 | n/a | :license: Python License. |
|---|
| 26 | n/a | """ |
|---|
| 27 | n/a | from _ast import * |
|---|
| 28 | n/a | |
|---|
| 29 | n/a | |
|---|
| 30 | n/a | def parse(source, filename='<unknown>', mode='exec'): |
|---|
| 31 | n/a | """ |
|---|
| 32 | n/a | Parse the source into an AST node. |
|---|
| 33 | n/a | Equivalent to compile(source, filename, mode, PyCF_ONLY_AST). |
|---|
| 34 | n/a | """ |
|---|
| 35 | n/a | return compile(source, filename, mode, PyCF_ONLY_AST) |
|---|
| 36 | n/a | |
|---|
| 37 | n/a | |
|---|
| 38 | n/a | _NUM_TYPES = (int, float, complex) |
|---|
| 39 | n/a | |
|---|
| 40 | n/a | def literal_eval(node_or_string): |
|---|
| 41 | n/a | """ |
|---|
| 42 | n/a | Safely evaluate an expression node or a string containing a Python |
|---|
| 43 | n/a | expression. The string or node provided may only consist of the following |
|---|
| 44 | n/a | Python literal structures: strings, bytes, numbers, tuples, lists, dicts, |
|---|
| 45 | n/a | sets, booleans, and None. |
|---|
| 46 | n/a | """ |
|---|
| 47 | n/a | if isinstance(node_or_string, str): |
|---|
| 48 | n/a | node_or_string = parse(node_or_string, mode='eval') |
|---|
| 49 | n/a | if isinstance(node_or_string, Expression): |
|---|
| 50 | n/a | node_or_string = node_or_string.body |
|---|
| 51 | n/a | def _convert(node): |
|---|
| 52 | n/a | if isinstance(node, Constant): |
|---|
| 53 | n/a | return node.value |
|---|
| 54 | n/a | elif isinstance(node, (Str, Bytes)): |
|---|
| 55 | n/a | return node.s |
|---|
| 56 | n/a | elif isinstance(node, Num): |
|---|
| 57 | n/a | return node.n |
|---|
| 58 | n/a | elif isinstance(node, Tuple): |
|---|
| 59 | n/a | return tuple(map(_convert, node.elts)) |
|---|
| 60 | n/a | elif isinstance(node, List): |
|---|
| 61 | n/a | return list(map(_convert, node.elts)) |
|---|
| 62 | n/a | elif isinstance(node, Set): |
|---|
| 63 | n/a | return set(map(_convert, node.elts)) |
|---|
| 64 | n/a | elif isinstance(node, Dict): |
|---|
| 65 | n/a | return dict((_convert(k), _convert(v)) for k, v |
|---|
| 66 | n/a | in zip(node.keys, node.values)) |
|---|
| 67 | n/a | elif isinstance(node, NameConstant): |
|---|
| 68 | n/a | return node.value |
|---|
| 69 | n/a | elif isinstance(node, UnaryOp) and isinstance(node.op, (UAdd, USub)): |
|---|
| 70 | n/a | operand = _convert(node.operand) |
|---|
| 71 | n/a | if isinstance(operand, _NUM_TYPES): |
|---|
| 72 | n/a | if isinstance(node.op, UAdd): |
|---|
| 73 | n/a | return + operand |
|---|
| 74 | n/a | else: |
|---|
| 75 | n/a | return - operand |
|---|
| 76 | n/a | elif isinstance(node, BinOp) and isinstance(node.op, (Add, Sub)): |
|---|
| 77 | n/a | left = _convert(node.left) |
|---|
| 78 | n/a | right = _convert(node.right) |
|---|
| 79 | n/a | if isinstance(left, _NUM_TYPES) and isinstance(right, _NUM_TYPES): |
|---|
| 80 | n/a | if isinstance(node.op, Add): |
|---|
| 81 | n/a | return left + right |
|---|
| 82 | n/a | else: |
|---|
| 83 | n/a | return left - right |
|---|
| 84 | n/a | raise ValueError('malformed node or string: ' + repr(node)) |
|---|
| 85 | n/a | return _convert(node_or_string) |
|---|
| 86 | n/a | |
|---|
| 87 | n/a | |
|---|
| 88 | n/a | def dump(node, annotate_fields=True, include_attributes=False): |
|---|
| 89 | n/a | """ |
|---|
| 90 | n/a | Return a formatted dump of the tree in *node*. This is mainly useful for |
|---|
| 91 | n/a | debugging purposes. The returned string will show the names and the values |
|---|
| 92 | n/a | for fields. This makes the code impossible to evaluate, so if evaluation is |
|---|
| 93 | n/a | wanted *annotate_fields* must be set to False. Attributes such as line |
|---|
| 94 | n/a | numbers and column offsets are not dumped by default. If this is wanted, |
|---|
| 95 | n/a | *include_attributes* can be set to True. |
|---|
| 96 | n/a | """ |
|---|
| 97 | n/a | def _format(node): |
|---|
| 98 | n/a | if isinstance(node, AST): |
|---|
| 99 | n/a | fields = [(a, _format(b)) for a, b in iter_fields(node)] |
|---|
| 100 | n/a | rv = '%s(%s' % (node.__class__.__name__, ', '.join( |
|---|
| 101 | n/a | ('%s=%s' % field for field in fields) |
|---|
| 102 | n/a | if annotate_fields else |
|---|
| 103 | n/a | (b for a, b in fields) |
|---|
| 104 | n/a | )) |
|---|
| 105 | n/a | if include_attributes and node._attributes: |
|---|
| 106 | n/a | rv += fields and ', ' or ' ' |
|---|
| 107 | n/a | rv += ', '.join('%s=%s' % (a, _format(getattr(node, a))) |
|---|
| 108 | n/a | for a in node._attributes) |
|---|
| 109 | n/a | return rv + ')' |
|---|
| 110 | n/a | elif isinstance(node, list): |
|---|
| 111 | n/a | return '[%s]' % ', '.join(_format(x) for x in node) |
|---|
| 112 | n/a | return repr(node) |
|---|
| 113 | n/a | if not isinstance(node, AST): |
|---|
| 114 | n/a | raise TypeError('expected AST, got %r' % node.__class__.__name__) |
|---|
| 115 | n/a | return _format(node) |
|---|
| 116 | n/a | |
|---|
| 117 | n/a | |
|---|
| 118 | n/a | def copy_location(new_node, old_node): |
|---|
| 119 | n/a | """ |
|---|
| 120 | n/a | Copy source location (`lineno` and `col_offset` attributes) from |
|---|
| 121 | n/a | *old_node* to *new_node* if possible, and return *new_node*. |
|---|
| 122 | n/a | """ |
|---|
| 123 | n/a | for attr in 'lineno', 'col_offset': |
|---|
| 124 | n/a | if attr in old_node._attributes and attr in new_node._attributes \ |
|---|
| 125 | n/a | and hasattr(old_node, attr): |
|---|
| 126 | n/a | setattr(new_node, attr, getattr(old_node, attr)) |
|---|
| 127 | n/a | return new_node |
|---|
| 128 | n/a | |
|---|
| 129 | n/a | |
|---|
| 130 | n/a | def fix_missing_locations(node): |
|---|
| 131 | n/a | """ |
|---|
| 132 | n/a | When you compile a node tree with compile(), the compiler expects lineno and |
|---|
| 133 | n/a | col_offset attributes for every node that supports them. This is rather |
|---|
| 134 | n/a | tedious to fill in for generated nodes, so this helper adds these attributes |
|---|
| 135 | n/a | recursively where not already set, by setting them to the values of the |
|---|
| 136 | n/a | parent node. It works recursively starting at *node*. |
|---|
| 137 | n/a | """ |
|---|
| 138 | n/a | def _fix(node, lineno, col_offset): |
|---|
| 139 | n/a | if 'lineno' in node._attributes: |
|---|
| 140 | n/a | if not hasattr(node, 'lineno'): |
|---|
| 141 | n/a | node.lineno = lineno |
|---|
| 142 | n/a | else: |
|---|
| 143 | n/a | lineno = node.lineno |
|---|
| 144 | n/a | if 'col_offset' in node._attributes: |
|---|
| 145 | n/a | if not hasattr(node, 'col_offset'): |
|---|
| 146 | n/a | node.col_offset = col_offset |
|---|
| 147 | n/a | else: |
|---|
| 148 | n/a | col_offset = node.col_offset |
|---|
| 149 | n/a | for child in iter_child_nodes(node): |
|---|
| 150 | n/a | _fix(child, lineno, col_offset) |
|---|
| 151 | n/a | _fix(node, 1, 0) |
|---|
| 152 | n/a | return node |
|---|
| 153 | n/a | |
|---|
| 154 | n/a | |
|---|
| 155 | n/a | def increment_lineno(node, n=1): |
|---|
| 156 | n/a | """ |
|---|
| 157 | n/a | Increment the line number of each node in the tree starting at *node* by *n*. |
|---|
| 158 | n/a | This is useful to "move code" to a different location in a file. |
|---|
| 159 | n/a | """ |
|---|
| 160 | n/a | for child in walk(node): |
|---|
| 161 | n/a | if 'lineno' in child._attributes: |
|---|
| 162 | n/a | child.lineno = getattr(child, 'lineno', 0) + n |
|---|
| 163 | n/a | return node |
|---|
| 164 | n/a | |
|---|
| 165 | n/a | |
|---|
| 166 | n/a | def iter_fields(node): |
|---|
| 167 | n/a | """ |
|---|
| 168 | n/a | Yield a tuple of ``(fieldname, value)`` for each field in ``node._fields`` |
|---|
| 169 | n/a | that is present on *node*. |
|---|
| 170 | n/a | """ |
|---|
| 171 | n/a | for field in node._fields: |
|---|
| 172 | n/a | try: |
|---|
| 173 | n/a | yield field, getattr(node, field) |
|---|
| 174 | n/a | except AttributeError: |
|---|
| 175 | n/a | pass |
|---|
| 176 | n/a | |
|---|
| 177 | n/a | |
|---|
| 178 | n/a | def iter_child_nodes(node): |
|---|
| 179 | n/a | """ |
|---|
| 180 | n/a | Yield all direct child nodes of *node*, that is, all fields that are nodes |
|---|
| 181 | n/a | and all items of fields that are lists of nodes. |
|---|
| 182 | n/a | """ |
|---|
| 183 | n/a | for name, field in iter_fields(node): |
|---|
| 184 | n/a | if isinstance(field, AST): |
|---|
| 185 | n/a | yield field |
|---|
| 186 | n/a | elif isinstance(field, list): |
|---|
| 187 | n/a | for item in field: |
|---|
| 188 | n/a | if isinstance(item, AST): |
|---|
| 189 | n/a | yield item |
|---|
| 190 | n/a | |
|---|
| 191 | n/a | |
|---|
| 192 | n/a | def get_docstring(node, clean=True): |
|---|
| 193 | n/a | """ |
|---|
| 194 | n/a | Return the docstring for the given node or None if no docstring can |
|---|
| 195 | n/a | be found. If the node provided does not have docstrings a TypeError |
|---|
| 196 | n/a | will be raised. |
|---|
| 197 | n/a | """ |
|---|
| 198 | n/a | if not isinstance(node, (AsyncFunctionDef, FunctionDef, ClassDef, Module)): |
|---|
| 199 | n/a | raise TypeError("%r can't have docstrings" % node.__class__.__name__) |
|---|
| 200 | n/a | if not(node.body and isinstance(node.body[0], Expr)): |
|---|
| 201 | n/a | return |
|---|
| 202 | n/a | node = node.body[0].value |
|---|
| 203 | n/a | if isinstance(node, Str): |
|---|
| 204 | n/a | text = node.s |
|---|
| 205 | n/a | elif isinstance(node, Constant) and isinstance(node.value, str): |
|---|
| 206 | n/a | text = node.value |
|---|
| 207 | n/a | else: |
|---|
| 208 | n/a | return |
|---|
| 209 | n/a | if clean: |
|---|
| 210 | n/a | import inspect |
|---|
| 211 | n/a | text = inspect.cleandoc(text) |
|---|
| 212 | n/a | return text |
|---|
| 213 | n/a | |
|---|
| 214 | n/a | |
|---|
| 215 | n/a | def walk(node): |
|---|
| 216 | n/a | """ |
|---|
| 217 | n/a | Recursively yield all descendant nodes in the tree starting at *node* |
|---|
| 218 | n/a | (including *node* itself), in no specified order. This is useful if you |
|---|
| 219 | n/a | only want to modify nodes in place and don't care about the context. |
|---|
| 220 | n/a | """ |
|---|
| 221 | n/a | from collections import deque |
|---|
| 222 | n/a | todo = deque([node]) |
|---|
| 223 | n/a | while todo: |
|---|
| 224 | n/a | node = todo.popleft() |
|---|
| 225 | n/a | todo.extend(iter_child_nodes(node)) |
|---|
| 226 | n/a | yield node |
|---|
| 227 | n/a | |
|---|
| 228 | n/a | |
|---|
| 229 | n/a | class NodeVisitor(object): |
|---|
| 230 | n/a | """ |
|---|
| 231 | n/a | A node visitor base class that walks the abstract syntax tree and calls a |
|---|
| 232 | n/a | visitor function for every node found. This function may return a value |
|---|
| 233 | n/a | which is forwarded by the `visit` method. |
|---|
| 234 | n/a | |
|---|
| 235 | n/a | This class is meant to be subclassed, with the subclass adding visitor |
|---|
| 236 | n/a | methods. |
|---|
| 237 | n/a | |
|---|
| 238 | n/a | Per default the visitor functions for the nodes are ``'visit_'`` + |
|---|
| 239 | n/a | class name of the node. So a `TryFinally` node visit function would |
|---|
| 240 | n/a | be `visit_TryFinally`. This behavior can be changed by overriding |
|---|
| 241 | n/a | the `visit` method. If no visitor function exists for a node |
|---|
| 242 | n/a | (return value `None`) the `generic_visit` visitor is used instead. |
|---|
| 243 | n/a | |
|---|
| 244 | n/a | Don't use the `NodeVisitor` if you want to apply changes to nodes during |
|---|
| 245 | n/a | traversing. For this a special visitor exists (`NodeTransformer`) that |
|---|
| 246 | n/a | allows modifications. |
|---|
| 247 | n/a | """ |
|---|
| 248 | n/a | |
|---|
| 249 | n/a | def visit(self, node): |
|---|
| 250 | n/a | """Visit a node.""" |
|---|
| 251 | n/a | method = 'visit_' + node.__class__.__name__ |
|---|
| 252 | n/a | visitor = getattr(self, method, self.generic_visit) |
|---|
| 253 | n/a | return visitor(node) |
|---|
| 254 | n/a | |
|---|
| 255 | n/a | def generic_visit(self, node): |
|---|
| 256 | n/a | """Called if no explicit visitor function exists for a node.""" |
|---|
| 257 | n/a | for field, value in iter_fields(node): |
|---|
| 258 | n/a | if isinstance(value, list): |
|---|
| 259 | n/a | for item in value: |
|---|
| 260 | n/a | if isinstance(item, AST): |
|---|
| 261 | n/a | self.visit(item) |
|---|
| 262 | n/a | elif isinstance(value, AST): |
|---|
| 263 | n/a | self.visit(value) |
|---|
| 264 | n/a | |
|---|
| 265 | n/a | |
|---|
| 266 | n/a | class NodeTransformer(NodeVisitor): |
|---|
| 267 | n/a | """ |
|---|
| 268 | n/a | A :class:`NodeVisitor` subclass that walks the abstract syntax tree and |
|---|
| 269 | n/a | allows modification of nodes. |
|---|
| 270 | n/a | |
|---|
| 271 | n/a | The `NodeTransformer` will walk the AST and use the return value of the |
|---|
| 272 | n/a | visitor methods to replace or remove the old node. If the return value of |
|---|
| 273 | n/a | the visitor method is ``None``, the node will be removed from its location, |
|---|
| 274 | n/a | otherwise it is replaced with the return value. The return value may be the |
|---|
| 275 | n/a | original node in which case no replacement takes place. |
|---|
| 276 | n/a | |
|---|
| 277 | n/a | Here is an example transformer that rewrites all occurrences of name lookups |
|---|
| 278 | n/a | (``foo``) to ``data['foo']``:: |
|---|
| 279 | n/a | |
|---|
| 280 | n/a | class RewriteName(NodeTransformer): |
|---|
| 281 | n/a | |
|---|
| 282 | n/a | def visit_Name(self, node): |
|---|
| 283 | n/a | return copy_location(Subscript( |
|---|
| 284 | n/a | value=Name(id='data', ctx=Load()), |
|---|
| 285 | n/a | slice=Index(value=Str(s=node.id)), |
|---|
| 286 | n/a | ctx=node.ctx |
|---|
| 287 | n/a | ), node) |
|---|
| 288 | n/a | |
|---|
| 289 | n/a | Keep in mind that if the node you're operating on has child nodes you must |
|---|
| 290 | n/a | either transform the child nodes yourself or call the :meth:`generic_visit` |
|---|
| 291 | n/a | method for the node first. |
|---|
| 292 | n/a | |
|---|
| 293 | n/a | For nodes that were part of a collection of statements (that applies to all |
|---|
| 294 | n/a | statement nodes), the visitor may also return a list of nodes rather than |
|---|
| 295 | n/a | just a single node. |
|---|
| 296 | n/a | |
|---|
| 297 | n/a | Usually you use the transformer like this:: |
|---|
| 298 | n/a | |
|---|
| 299 | n/a | node = YourTransformer().visit(node) |
|---|
| 300 | n/a | """ |
|---|
| 301 | n/a | |
|---|
| 302 | n/a | def generic_visit(self, node): |
|---|
| 303 | n/a | for field, old_value in iter_fields(node): |
|---|
| 304 | n/a | if isinstance(old_value, list): |
|---|
| 305 | n/a | new_values = [] |
|---|
| 306 | n/a | for value in old_value: |
|---|
| 307 | n/a | if isinstance(value, AST): |
|---|
| 308 | n/a | value = self.visit(value) |
|---|
| 309 | n/a | if value is None: |
|---|
| 310 | n/a | continue |
|---|
| 311 | n/a | elif not isinstance(value, AST): |
|---|
| 312 | n/a | new_values.extend(value) |
|---|
| 313 | n/a | continue |
|---|
| 314 | n/a | new_values.append(value) |
|---|
| 315 | n/a | old_value[:] = new_values |
|---|
| 316 | n/a | elif isinstance(old_value, AST): |
|---|
| 317 | n/a | new_node = self.visit(old_value) |
|---|
| 318 | n/a | if new_node is None: |
|---|
| 319 | n/a | delattr(node, field) |
|---|
| 320 | n/a | else: |
|---|
| 321 | n/a | setattr(node, field, new_node) |
|---|
| 322 | n/a | return node |
|---|