|
| 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 | +""" |
| 27 | +from _ast import * |
| 28 | + |
| 29 | + |
| 30 | +def parse(source, filename='<unknown>', mode='exec'): |
| 31 | + """ |
| 32 | + Parse the source into an AST node. |
| 33 | + Equivalent to compile(source, filename, mode, PyCF_ONLY_AST). |
| 34 | + """ |
| 35 | + return compile(source, filename, mode, PyCF_ONLY_AST) |
| 36 | + |
| 37 | + |
| 38 | +def literal_eval(node_or_string): |
| 39 | + """ |
| 40 | + Safely evaluate an expression node or a string containing a Python |
| 41 | + expression. The string or node provided may only consist of the following |
| 42 | + Python literal structures: strings, bytes, numbers, tuples, lists, dicts, |
| 43 | + sets, booleans, and None. |
| 44 | + """ |
| 45 | + if isinstance(node_or_string, str): |
| 46 | + node_or_string = parse(node_or_string, mode='eval') |
| 47 | + if isinstance(node_or_string, Expression): |
| 48 | + node_or_string = node_or_string.body |
| 49 | + def _convert_num(node): |
| 50 | + if isinstance(node, Constant): |
| 51 | + if isinstance(node.value, (int, float, complex)): |
| 52 | + return node.value |
| 53 | + elif isinstance(node, Num): |
| 54 | + return node.n |
| 55 | + raise ValueError('malformed node or string: ' + repr(node)) |
| 56 | + def _convert_signed_num(node): |
| 57 | + if isinstance(node, UnaryOp) and isinstance(node.op, (UAdd, USub)): |
| 58 | + operand = _convert_num(node.operand) |
| 59 | + if isinstance(node.op, UAdd): |
| 60 | + return + operand |
| 61 | + else: |
| 62 | + return - operand |
| 63 | + return _convert_num(node) |
| 64 | + def _convert(node): |
| 65 | + if isinstance(node, Constant): |
| 66 | + return node.value |
| 67 | + elif isinstance(node, (Str, Bytes)): |
| 68 | + return node.s |
| 69 | + elif isinstance(node, Num): |
| 70 | + return node.n |
| 71 | + elif isinstance(node, Tuple): |
| 72 | + return tuple(map(_convert, node.elts)) |
| 73 | + elif isinstance(node, List): |
| 74 | + return list(map(_convert, node.elts)) |
| 75 | + elif isinstance(node, Set): |
| 76 | + return set(map(_convert, node.elts)) |
| 77 | + elif isinstance(node, Dict): |
| 78 | + return dict(zip(map(_convert, node.keys), |
| 79 | + map(_convert, node.values))) |
| 80 | + elif isinstance(node, NameConstant): |
| 81 | + return node.value |
| 82 | + elif isinstance(node, BinOp) and isinstance(node.op, (Add, Sub)): |
| 83 | + left = _convert_signed_num(node.left) |
| 84 | + right = _convert_num(node.right) |
| 85 | + if isinstance(left, (int, float)) and isinstance(right, complex): |
| 86 | + if isinstance(node.op, Add): |
| 87 | + return left + right |
| 88 | + else: |
| 89 | + return left - right |
| 90 | + return _convert_signed_num(node) |
| 91 | + return _convert(node_or_string) |
| 92 | + |
| 93 | + |
| 94 | +def dump(node, annotate_fields=True, include_attributes=False): |
| 95 | + """ |
| 96 | + Return a formatted dump of the tree in *node*. This is mainly useful for |
| 97 | + debugging purposes. The returned string will show the names and the values |
| 98 | + for fields. This makes the code impossible to evaluate, so if evaluation is |
| 99 | + wanted *annotate_fields* must be set to False. Attributes such as line |
| 100 | + numbers and column offsets are not dumped by default. If this is wanted, |
| 101 | + *include_attributes* can be set to True. |
| 102 | + """ |
| 103 | + def _format(node): |
| 104 | + if isinstance(node, AST): |
| 105 | + fields = [(a, _format(b)) for a, b in iter_fields(node)] |
| 106 | + rv = '%s(%s' % (node.__class__.__name__, ', '.join( |
| 107 | + ('%s=%s' % field for field in fields) |
| 108 | + if annotate_fields else |
| 109 | + (b for a, b in fields) |
| 110 | + )) |
| 111 | + if include_attributes and node._attributes: |
| 112 | + rv += fields and ', ' or ' ' |
| 113 | + rv += ', '.join('%s=%s' % (a, _format(getattr(node, a))) |
| 114 | + for a in node._attributes) |
| 115 | + return rv + ')' |
| 116 | + elif isinstance(node, list): |
| 117 | + return '[%s]' % ', '.join(_format(x) for x in node) |
| 118 | + return repr(node) |
| 119 | + if not isinstance(node, AST): |
| 120 | + raise TypeError('expected AST, got %r' % node.__class__.__name__) |
| 121 | + return _format(node) |
| 122 | + |
| 123 | + |
| 124 | +def copy_location(new_node, old_node): |
| 125 | + """ |
| 126 | + Copy source location (`lineno` and `col_offset` attributes) from |
| 127 | + *old_node* to *new_node* if possible, and return *new_node*. |
| 128 | + """ |
| 129 | + for attr in 'lineno', 'col_offset': |
| 130 | + if attr in old_node._attributes and attr in new_node._attributes \ |
| 131 | + and hasattr(old_node, attr): |
| 132 | + setattr(new_node, attr, getattr(old_node, attr)) |
| 133 | + return new_node |
| 134 | + |
| 135 | + |
| 136 | +def fix_missing_locations(node): |
| 137 | + """ |
| 138 | + When you compile a node tree with compile(), the compiler expects lineno and |
| 139 | + col_offset attributes for every node that supports them. This is rather |
| 140 | + tedious to fill in for generated nodes, so this helper adds these attributes |
| 141 | + recursively where not already set, by setting them to the values of the |
| 142 | + parent node. It works recursively starting at *node*. |
| 143 | + """ |
| 144 | + def _fix(node, lineno, col_offset): |
| 145 | + if 'lineno' in node._attributes: |
| 146 | + if not hasattr(node, 'lineno'): |
| 147 | + node.lineno = lineno |
| 148 | + else: |
| 149 | + lineno = node.lineno |
| 150 | + if 'col_offset' in node._attributes: |
| 151 | + if not hasattr(node, 'col_offset'): |
| 152 | + node.col_offset = col_offset |
| 153 | + else: |
| 154 | + col_offset = node.col_offset |
| 155 | + for child in iter_child_nodes(node): |
| 156 | + _fix(child, lineno, col_offset) |
| 157 | + _fix(node, 1, 0) |
| 158 | + return node |
| 159 | + |
| 160 | + |
| 161 | +def increment_lineno(node, n=1): |
| 162 | + """ |
| 163 | + Increment the line number of each node in the tree starting at *node* by *n*. |
| 164 | + This is useful to "move code" to a different location in a file. |
| 165 | + """ |
| 166 | + for child in walk(node): |
| 167 | + if 'lineno' in child._attributes: |
| 168 | + child.lineno = getattr(child, 'lineno', 0) + n |
| 169 | + return node |
| 170 | + |
| 171 | + |
| 172 | +def iter_fields(node): |
| 173 | + """ |
| 174 | + Yield a tuple of ``(fieldname, value)`` for each field in ``node._fields`` |
| 175 | + that is present on *node*. |
| 176 | + """ |
| 177 | + for field in node._fields: |
| 178 | + try: |
| 179 | + yield field, getattr(node, field) |
| 180 | + except AttributeError: |
| 181 | + pass |
| 182 | + |
| 183 | + |
| 184 | +def iter_child_nodes(node): |
| 185 | + """ |
| 186 | + Yield all direct child nodes of *node*, that is, all fields that are nodes |
| 187 | + and all items of fields that are lists of nodes. |
| 188 | + """ |
| 189 | + for name, field in iter_fields(node): |
| 190 | + if isinstance(field, AST): |
| 191 | + yield field |
| 192 | + elif isinstance(field, list): |
| 193 | + for item in field: |
| 194 | + if isinstance(item, AST): |
| 195 | + yield item |
| 196 | + |
| 197 | + |
| 198 | +def get_docstring(node, clean=True): |
| 199 | + """ |
| 200 | + Return the docstring for the given node or None if no docstring can |
| 201 | + be found. If the node provided does not have docstrings a TypeError |
| 202 | + will be raised. |
| 203 | +
|
| 204 | + If *clean* is `True`, all tabs are expanded to spaces and any whitespace |
| 205 | + that can be uniformly removed from the second line onwards is removed. |
| 206 | + """ |
| 207 | + if not isinstance(node, (AsyncFunctionDef, FunctionDef, ClassDef, Module)): |
| 208 | + raise TypeError("%r can't have docstrings" % node.__class__.__name__) |
| 209 | + if not(node.body and isinstance(node.body[0], Expr)): |
| 210 | + return None |
| 211 | + node = node.body[0].value |
| 212 | + if isinstance(node, Str): |
| 213 | + text = node.s |
| 214 | + elif isinstance(node, Constant) and isinstance(node.value, str): |
| 215 | + text = node.value |
| 216 | + else: |
| 217 | + return None |
| 218 | + if clean: |
| 219 | + import inspect |
| 220 | + text = inspect.cleandoc(text) |
| 221 | + return text |
| 222 | + |
| 223 | + |
| 224 | +def walk(node): |
| 225 | + """ |
| 226 | + Recursively yield all descendant nodes in the tree starting at *node* |
| 227 | + (including *node* itself), in no specified order. This is useful if you |
| 228 | + only want to modify nodes in place and don't care about the context. |
| 229 | + """ |
| 230 | + from collections import deque |
| 231 | + todo = deque([node]) |
| 232 | + while todo: |
| 233 | + node = todo.popleft() |
| 234 | + todo.extend(iter_child_nodes(node)) |
| 235 | + yield node |
| 236 | + |
| 237 | + |
| 238 | +class NodeVisitor(object): |
| 239 | + """ |
| 240 | + A node visitor base class that walks the abstract syntax tree and calls a |
| 241 | + visitor function for every node found. This function may return a value |
| 242 | + which is forwarded by the `visit` method. |
| 243 | +
|
| 244 | + This class is meant to be subclassed, with the subclass adding visitor |
| 245 | + methods. |
| 246 | +
|
| 247 | + Per default the visitor functions for the nodes are ``'visit_'`` + |
| 248 | + class name of the node. So a `TryFinally` node visit function would |
| 249 | + be `visit_TryFinally`. This behavior can be changed by overriding |
| 250 | + the `visit` method. If no visitor function exists for a node |
| 251 | + (return value `None`) the `generic_visit` visitor is used instead. |
| 252 | +
|
| 253 | + Don't use the `NodeVisitor` if you want to apply changes to nodes during |
| 254 | + traversing. For this a special visitor exists (`NodeTransformer`) that |
| 255 | + allows modifications. |
| 256 | + """ |
| 257 | + |
| 258 | + def visit(self, node): |
| 259 | + """Visit a node.""" |
| 260 | + method = 'visit_' + node.__class__.__name__ |
| 261 | + visitor = getattr(self, method, self.generic_visit) |
| 262 | + return visitor(node) |
| 263 | + |
| 264 | + def generic_visit(self, node): |
| 265 | + """Called if no explicit visitor function exists for a node.""" |
| 266 | + for field, value in iter_fields(node): |
| 267 | + if isinstance(value, list): |
| 268 | + for item in value: |
| 269 | + if isinstance(item, AST): |
| 270 | + self.visit(item) |
| 271 | + elif isinstance(value, AST): |
| 272 | + self.visit(value) |
| 273 | + |
| 274 | + |
| 275 | +class NodeTransformer(NodeVisitor): |
| 276 | + """ |
| 277 | + A :class:`NodeVisitor` subclass that walks the abstract syntax tree and |
| 278 | + allows modification of nodes. |
| 279 | +
|
| 280 | + The `NodeTransformer` will walk the AST and use the return value of the |
| 281 | + visitor methods to replace or remove the old node. If the return value of |
| 282 | + the visitor method is ``None``, the node will be removed from its location, |
| 283 | + otherwise it is replaced with the return value. The return value may be the |
| 284 | + original node in which case no replacement takes place. |
| 285 | +
|
| 286 | + Here is an example transformer that rewrites all occurrences of name lookups |
| 287 | + (``foo``) to ``data['foo']``:: |
| 288 | +
|
| 289 | + class RewriteName(NodeTransformer): |
| 290 | +
|
| 291 | + def visit_Name(self, node): |
| 292 | + return copy_location(Subscript( |
| 293 | + value=Name(id='data', ctx=Load()), |
| 294 | + slice=Index(value=Str(s=node.id)), |
| 295 | + ctx=node.ctx |
| 296 | + ), node) |
| 297 | +
|
| 298 | + Keep in mind that if the node you're operating on has child nodes you must |
| 299 | + either transform the child nodes yourself or call the :meth:`generic_visit` |
| 300 | + method for the node first. |
| 301 | +
|
| 302 | + For nodes that were part of a collection of statements (that applies to all |
| 303 | + statement nodes), the visitor may also return a list of nodes rather than |
| 304 | + just a single node. |
| 305 | +
|
| 306 | + Usually you use the transformer like this:: |
| 307 | +
|
| 308 | + node = YourTransformer().visit(node) |
| 309 | + """ |
| 310 | + |
| 311 | + def generic_visit(self, node): |
| 312 | + for field, old_value in iter_fields(node): |
| 313 | + if isinstance(old_value, list): |
| 314 | + new_values = [] |
| 315 | + for value in old_value: |
| 316 | + if isinstance(value, AST): |
| 317 | + value = self.visit(value) |
| 318 | + if value is None: |
| 319 | + continue |
| 320 | + elif not isinstance(value, AST): |
| 321 | + new_values.extend(value) |
| 322 | + continue |
| 323 | + new_values.append(value) |
| 324 | + old_value[:] = new_values |
| 325 | + elif isinstance(old_value, AST): |
| 326 | + new_node = self.visit(old_value) |
| 327 | + if new_node is None: |
| 328 | + delattr(node, field) |
| 329 | + else: |
| 330 | + setattr(node, field, new_node) |
| 331 | + return node |
0 commit comments