transform.py 32.9 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
# transform.py - transformation functions for converting the
#                concrete into the abstract syntax tree
#
# Copyright 2016  by Eckhart Arnold (arnold@badw.de)
#                 Bavarian Academy of Sciences an Humanities (badw.de)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied.  See the License for the specific language governing
# permissions and limitations under the License.
18
19


20
21
22
"""
Module ``transform`` contains the functions for transforming the
concrete syntax tree (CST) into an abstract syntax tree (AST).
23

24
25
26
As these functions are very generic, they can in principle be
used for any kind of tree transformations, not necessarily only
for CST -> AST transformations.
27
28
"""

29

30
31
32
import inspect
from functools import partial, reduce, singledispatch

33
from DHParser.error import Error
eckhart's avatar
eckhart committed
34
from DHParser.syntaxtree import Node, WHITESPACE_PTYPE, TOKEN_PTYPE, MockParser, ZOMBIE_NODE
35
36
from DHParser.toolkit import expand_table, smart_list, re, typing
from typing import AbstractSet, Any, ByteString, Callable, cast, Container, Dict, \
eckhart's avatar
eckhart committed
37
    List, Sequence, Union, Text
38

39
40
__all__ = ('TransformationDict',
           'TransformationProc',
41
           'TransformationFunc',
42
43
44
           'ConditionFunc',
           'KeyFunc',
           'transformation_factory',
45
46
47
           'key_parser_name',
           'key_tag_name',
           'traverse',
48
           'is_named',
49
           'replace_by_single_child',
Eckhart Arnold's avatar
Eckhart Arnold committed
50
           'reduce_single_child',
51
           'replace_or_reduce',
52
53
           'replace_parser',
           'collapse',
54
           'merge_children',
55
           'replace_content',
56
           'replace_content_by',
57
           'apply_if',
eckhart's avatar
eckhart committed
58
           'apply_unless',
59
           'traverse_locally',
60
           'is_anonymous',
61
62
63
64
           'is_whitespace',
           'is_empty',
           'is_expendable',
           'is_token',
65
           'is_one_of',
66
           'has_content',
di68kap's avatar
di68kap committed
67
           'has_parent',
68
69
70
71
72
73
74
75
           'lstrip',
           'rstrip',
           'strip',
           'keep_children',
           'keep_children_if',
           'keep_tokens',
           'keep_nodes',
           'keep_content',
76
           'remove_children_if',
eckhart's avatar
eckhart committed
77
           'remove_nodes',
78
79
80
81
82
           'remove_content',
           'remove_first',
           'remove_last',
           'remove_whitespace',
           'remove_empty',
di68kap's avatar
di68kap committed
83
           'remove_anonymous_empty',
84
85
           'remove_anonymous_expendables',
           'remove_anonymous_tokens',
86
87
           'remove_expendables',
           'remove_brackets',
88
89
           'remove_infix_operator',
           'remove_single_child',
90
91
92
93
           'remove_tokens',
           'flatten',
           'forbid',
           'require',
94
           'assert_content',
95
96
           'error_on',
           'warn_on',
eckhart's avatar
eckhart committed
97
           'assert_has_children')
98
99


100
TransformationProc = Callable[[List[Node]], None]
Eckhart Arnold's avatar
Eckhart Arnold committed
101
TransformationDict = Dict[str, Sequence[Callable]]
102
TransformationFunc = Union[Callable[[Node], Any], partial]
Eckhart Arnold's avatar
Eckhart Arnold committed
103
ProcessingTableType = Dict[str, Union[Sequence[Callable], TransformationDict]]
104
105
ConditionFunc = Callable  # Callable[[List[Node]], bool]
KeyFunc = Callable[[Node], str]
eckhart's avatar
eckhart committed
106
CriteriaType = Union[int, str, Callable]
107
108


109
def transformation_factory(t1=None, t2=None, t3=None, t4=None, t5=None):
110
    """Creates factory functions from transformation-functions that
111
    dispatch on the first parameter after the context parameter.
112
113

    Decorating a transformation-function that has more than merely the
Eckhart Arnold's avatar
Eckhart Arnold committed
114
    ``context``-parameter with ``transformation_factory`` creates a
115
    function with the same name, which returns a partial-function that
116
    takes just the context-parameter.
117
118
119
120
121
122
123
124
125

    Additionally, there is some some syntactic sugar for
    transformation-functions that receive a collection as their second
    parameter and do not have any further parameters. In this case a
    list of parameters passed to the factory function will be converted
    into a collection.

    Main benefit is readability of processing tables.

126
127
    Usage::

eckhart's avatar
eckhart committed
128
        @transformation_factory(AbstractSet[str])
129
        def remove_tokens(context, tokens):
130
            ...
131
132
133

    or, alternatively::

134
        @transformation_factory
135
        def remove_tokens(context, tokens: AbstractSet[str]):
136
137
            ...

138
139
    Example::

140
        trans_table = { 'expression': remove_tokens('+', '-') }
141
142
143

    instead of::

144
        trans_table = { 'expression': partial(remove_tokens, tokens={'+', '-'}) }
145
146

    Parameters:
147
        t1:  type of the second argument of the transformation function,
148
149
            only necessary if the transformation functions' parameter list
            does not have type annotations.
150
151
152
153
154
155
156
    """

    def decorator(f):
        sig = inspect.signature(f)
        params = list(sig.parameters.values())[1:]
        if len(params) == 0:
            return f  # '@transformer' not needed w/o free parameters
157
        assert t1 or params[0].annotation != params[0].empty, \
158
            "No type information on second parameter found! Please, use type " \
eckhart's avatar
eckhart committed
159
            "annotation or provide the type information via transformer-decorator."
160
        p1type = t1 or params[0].annotation
161
        f = singledispatch(f)
eckhart's avatar
eckhart committed
162
163
        try:
            if len(params) == 1 and issubclass(p1type, Container) \
eckhart's avatar
eckhart committed
164
                    and not (issubclass(p1type, Text) or issubclass(p1type, ByteString)):
eckhart's avatar
eckhart committed
165
166
167
168
169
170
171
172
173
                def gen_special(*args):
                    c = set(args) if issubclass(p1type, AbstractSet) else \
                        list(args) if issubclass(p1type, Sequence) else args
                    d = {params[0].name: c}
                    return partial(f, **d)

                f.register(p1type.__args__[0], gen_special)
        except AttributeError:
            pass  # Union Type does not allow subclassing, but is not needed here
174
175
176
177
178
179

        def gen_partial(*args, **kwargs):
            d = {p.name: arg for p, arg in zip(params, args)}
            d.update(kwargs)
            return partial(f, **d)

180
181
182
183
184
        for t in (p1type, t2, t3, t4, t5):
            if t:
                f.register(t, gen_partial)
            else:
                break
185
186
        return f

187
    if isinstance(t1, type(lambda: 1)):
188
189
190
        # Provide for the case that transformation_factory has been
        # written as plain decorator and not as a function call that
        # returns the decorator proper.
191
192
        func = t1
        t1 = None
193
194
195
196
197
        return decorator(func)
    else:
        return decorator


198
def key_parser_name(node: Node) -> str:
199
200
201
    return node.parser.name


202
def key_tag_name(node: Node) -> str:
203
204
205
    return node.tag_name


206
def traverse(root_node: Node,
Eckhart Arnold's avatar
Eckhart Arnold committed
207
             processing_table: ProcessingTableType,
208
209
210
             key_func: KeyFunc=key_tag_name) -> None:
    """
    Traverses the snytax tree starting with the given ``node`` depth
211
    first and applies the sequences of callback-functions registered
212
    in the ``processing_table``-dictionary.
213
214
215
216
217
218
219
220
221

    The most important use case is the transformation of a concrete
    syntax tree into an abstract tree (AST). But it is also imaginable
    to employ tree-traversal for the semantic analysis of the AST.

    In order to assign sequences of callback-functions to nodes, a
    dictionary ("processing table") is used. The keys usually represent
    tag names, but any other key function is possible. There exist
    three special keys:
222
223
224
225
226

    - '+': always called (before any other processing function)
    - '*': called for those nodes for which no (other) processing
      function appears in the table
    - '~': always called (after any other processing function)
227
228
229
230
231

    Args:
        root_node (Node): The root-node of the syntax tree to be traversed
        processing_table (dict): node key -> sequence of functions that
            will be applied to matching nodes in order. This dictionary
232
233
            is interpreted as a ``compact_table``. See
            :func:`expand_table` or :func:`EBNFCompiler.EBNFTransTable`
234
235
236
        key_func (function): A mapping key_func(node) -> keystr. The default
            key_func yields node.parser.name.

237
238
    Example::

239
        table = { "term": [replace_by_single_child, flatten],
240
                  "factor, flowmarker, retrieveop": replace_by_single_child }
241
        traverse(node, table)
242

243
    """
244
245
246
    # Is this optimazation really needed?
    if '__cache__' in processing_table:
        # assume that processing table has already been expanded
eckhart's avatar
eckhart committed
247
        table = processing_table               # type: ProcessingTableType
eckhart's avatar
eckhart committed
248
        cache = cast(TransformationDict, processing_table['__cache__'])  # type: TransformationDict
249
    else:
250
251
        # normalize processing_table entries by turning single values
        # into lists with a single value
252
253
        table = {name: cast(Sequence[Callable], smart_list(call))
                 for name, call in list(processing_table.items())}
254
        table = expand_table(table)
eckhart's avatar
eckhart committed
255
256
        cache = cast(TransformationDict,
                     table.setdefault('__cache__', cast(TransformationDict, dict())))
257
258
        # change processing table in place, so its already expanded and cache filled next time
        processing_table.clear()
259
260
261
262
263
264
        processing_table.update(table)

    # assert '__cache__' in processing_table
    # # Code without optimization
    # table = {name: smart_list(call) for name, call in list(processing_table.items())}
    # table = expand_table(table)
Eckhart Arnold's avatar
Eckhart Arnold committed
265
    # cache = {}  # type: Dict[str, List[Callable]]
266

267
    def traverse_recursive(context):
eckhart's avatar
eckhart committed
268
        nonlocal cache
269
        node = context[-1]
270
        if node.children:
eckhart's avatar
eckhart committed
271
            context.append(ZOMBIE_NODE)
272
            for child in node.result:
eckhart's avatar
eckhart committed
273
                context[-1] = child
274
                traverse_recursive(context)  # depth first
eckhart's avatar
eckhart committed
275
            context.pop()
276
277

        key = key_func(node)
278
279
280
        try:
            sequence = cache[key]
        except KeyError:
281
282
283
            sequence = table.get('+', []) \
                + table.get(key, table.get('*', [])) \
                + table.get('~', [])
284
285
286
287
288
289
290
            # '+' always called (before any other processing function)
            # '*' called for those nodes for which no (other) processing function
            #     appears in the table
            # '~' always called (after any other processing function)
            cache[key] = sequence

        for call in sequence:
291
            call(context)
292

293
    traverse_recursive([root_node])
294
295
    # assert processing_table['__cache__']

296

297
#######################################################################
298
#
299
300
# meta transformations, i.e. transformations that call other
# transformations
301
#
302
#######################################################################
303
304


eckhart's avatar
eckhart committed
305
@transformation_factory(dict)
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
def traverse_locally(context: List[Node],
                     processing_table: Dict,            # actually: ProcessingTableType
                     key_func: Callable=key_tag_name):  # actually: KeyFunc
    """Transforms the syntax tree starting from the last node in the context
    according to the given processing table. The purpose of this function is
    to apply certain transformations locally, i.e. only for those nodes that
    have the last node in the context as their parent node.
    """
    traverse(context[-1], processing_table, key_func)


# @transformation_factory(List[Callable])
# def apply_to_child(context: List[Node], transformations: List[Callable], condition: Callable):
#     """Applies a list of transformations to those children that meet a specifc condition."""
#     node = context[-1]
#     for child in node.children:
#         context.append(child)
#         if condition(context):
#             for transform in transformations:
#                 transform(context)
#         context.pop()


@transformation_factory(Callable)
def apply_if(context: List[Node], transformation: Callable, condition: Callable):
    """Applies a transformation only if a certain condition is met."""
    if condition(context):
        transformation(context)


eckhart's avatar
eckhart committed
336
337
338
339
340
341
342
@transformation_factory(Callable)
def apply_unless(context: List[Node], transformation: Callable, condition: Callable):
    """Applies a transformation if a certain condition is *not* met."""
    if not condition(context):
        transformation(context)


343
344
345
346
347
348
349
350
351
352
353
354
355
356
#######################################################################
#
# conditionals that determine whether the context (or the last node in
# the context for that matter) fulfill a specific condition.
# ---------------------------------------------------------------------
#
# The context of a node is understood as a list of all parent nodes
# leading up to and including the node itself. If represented as list,
# the last element of the list is the node itself.
#
#######################################################################


def is_single_child(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
357
    """Returns ``True`` if the current node does not have any siblings."""
358
359
360
361
    return len(context[-2].children) == 1


def is_named(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
362
    """Returns ``True`` if the current node's parser is a named parser."""
363
364
365
366
    return bool(context[-1].parser.name)


def is_anonymous(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
367
    """Returns ``True`` if the current node's parser is an anonymous parser."""
368
369
370
371
372
373
374
375
376
377
    return not context[-1].parser.name


def is_whitespace(context: List[Node]) -> bool:
    """Removes whitespace and comments defined with the
    ``@comment``-directive."""
    return context[-1].parser.ptype == WHITESPACE_PTYPE


def is_empty(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
378
    """Returns ``True`` if the current node's content is empty."""
379
380
381
382
    return not context[-1].result


def is_expendable(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
383
384
    """Returns ``True`` if the current node either is a node containing
    whitespace or an empty node."""
385
386
387
    return is_empty(context) or is_whitespace(context)


eckhart's avatar
eckhart committed
388
@transformation_factory  # (AbstractSet[str])
389
390
391
392
def is_token(context: List[Node], tokens: AbstractSet[str] = frozenset()) -> bool:
    """Checks whether the last node in the context has `ptype == TOKEN_PTYPE`
    and it's content matches one of the given tokens. Leading and trailing
    whitespace-tokens will be ignored. In case an empty set of tokens is passed,
eckhart's avatar
eckhart committed
393
    any token is a match.
394
395
    """
    def stripped(nd: Node) -> str:
396
        """Removes leading and trailing whitespace-nodes from content."""
397
398
399
400
401
        # assert node.parser.ptype == TOKEN_PTYPE
        if nd.children:
            i, k = 0, len(nd.children)
            while i < len(nd.children) and nd.children[i].parser.ptype == WHITESPACE_PTYPE:
                i += 1
402
            while k > 0 and nd.children[k - 1].parser.ptype == WHITESPACE_PTYPE:
403
404
405
406
                k -= 1
            return "".join(child.content for child in node.children[i:k])
        return nd.content
    node = context[-1]
407
    return node.parser.ptype == TOKEN_PTYPE and (not tokens or stripped(node) in tokens)
408
409


eckhart's avatar
eckhart committed
410
@transformation_factory
411
412
413
414
415
def is_one_of(context: List[Node], tag_name_set: AbstractSet[str]) -> bool:
    """Returns true, if the node's tag_name is one of the given tag names."""
    return context[-1].tag_name in tag_name_set


eckhart's avatar
eckhart committed
416
@transformation_factory
417
def has_content(context: List[Node], regexp: str) -> bool:
418
419
420
421
422
423
424
425
    """
    Checks a node's content against a regular expression.

    In contrast to ``re.match`` the regular expression must match the complete
    string and not just the beginning of the string to succeed!
    """
    if not regexp.endswith('$'):
        regexp += "$"
426
427
428
    return bool(re.match(regexp, context[-1].content))


eckhart's avatar
eckhart committed
429
@transformation_factory
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
def has_parent(context: List[Node], tag_name_set: AbstractSet[str]) -> bool:
    """Checks whether a node with one of the given tag names appears somewhere
     in the context before the last node in the context."""
    for i in range(2, len(context)):
        if context[-i].tag_name in tag_name_set:
            return True
    return False


#######################################################################
#
# utility functions (private)
#
#######################################################################


def _replace_by(node: Node, child: Node):
447
448
449
450
    if not child.parser.name:
        child.parser = MockParser(node.parser.name, child.parser.ptype)
        # parser names must not be overwritten, else: child.parser.name = node.parser.name
    node.parser = child.parser
eckhart's avatar
eckhart committed
451
    node.errors.extend(child.errors)
452
453
454
    node.result = child.result


455
def _reduce_child(node: Node, child: Node):
eckhart's avatar
eckhart committed
456
    node.errors.extend(child.errors)
457
458
459
    node.result = child.result


eckhart's avatar
eckhart committed
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
# def _pick_child(context: List[Node], criteria: CriteriaType):
#     """Returns the first child that meets the criteria."""
#     if isinstance(criteria, int):
#         try:
#             return context[-1].children[criteria]
#         except IndexError:
#             return None
#     elif isinstance(criteria, str):
#         for child in context[-1].children:
#             if child.tag_name == criteria:
#                 return child
#         return None
#     else:  # assume criteria has type ConditionFunc
#         for child in context[-1].children:
#             context.append(child)
#             evaluation = criteria(context)
#             context.pop()
#             if evaluation:
#                 return child
#         return None
480
481


482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
#######################################################################
#
# rearranging transformations
#
# - tree may be rearranged (e.g.flattened)
# - nodes that are not leaves may be dropped
# - order is preserved
# - leave content is preserved (though not necessarily the leaves
#   themselves)
#
#######################################################################


# @transformation_factory(int, str, Callable)
# def replace_by_child(context: List[Node], criteria: CriteriaType=is_single_child):
#     """
#     Replaces a node by the first of its immediate descendants
#     that meets the `criteria`. The criteria can either be the
#     index of the child (counting from zero), or the tag name or
#     a boolean-valued function on the context of the child.
#     If no child matching the criteria is found, the node will
#     not be replaced.
#     With the default value for `criteria` the same semantics is
#     the same that of `replace_by_single_child`.
#     """
#     child = _pick_child(context, criteria)
#     if child:
#         _replace_by(context[-1], child)
#
#
# @transformation_factory(int, str, Callable)
# def content_from_child(context: List[Node], criteria: CriteriaType = is_single_child):
#     """
#     Reduces a node, by transferring the result of the first of its
#     immediate descendants that meets the `criteria` to this node,
#     but keeping this node's parser entry. The criteria can either
#     be the index of the child (counting from zero), or the tag
#     name or a boolean-valued function on the context of the child.
#     If no child matching the criteria is found, the node will
#     not be replaced.
#     With the default value for `criteria` this has the same semantics
#     as `content_from_single_child`.
#     """
#     child = _pick_child(context, criteria)
#     if child:
#         _reduce_child(context[-1], child)
528
529


530
531
def replace_by_single_child(context: List[Node]):
    """
532
533
534
    Removes single branch node, replacing it by its immediate descendant.
    Replacement only takes place, if the last node in the context has
    exactly one child.
535
536
537
    """
    node = context[-1]
    if len(node.children) == 1:
538
        _replace_by(node, node.children[0])
539
540


Eckhart Arnold's avatar
Eckhart Arnold committed
541
def reduce_single_child(context: List[Node]):
542
    """
543
    Reduces a single branch node by transferring the result of its
544
    immediate descendant to this node, but keeping this node's parser entry.
545
546
    Reduction only takes place if the last node in the context has
    exactly one child.
547
548
549
    """
    node = context[-1]
    if len(node.children) == 1:
550
        _reduce_child(node, node.children[0])
551
552
553


@transformation_factory(Callable)
554
555
556
def replace_or_reduce(context: List[Node], condition: Callable=is_named):
    """
    Replaces node by a single child, if condition is met on child,
557
558
    otherwise (i.e. if the child is anonymous) reduces the child.
    """
559
    node = context[-1]
560
    if len(node.children) == 1:
di68kap's avatar
di68kap committed
561
        child = node.children[0]
562
        if condition(context):
563
            _replace_by(node, child)
564
        else:
565
            _reduce_child(node, child)
566
567
568


@transformation_factory
569
570
571
def replace_parser(context: List[Node], name: str):
    """
    Replaces the parser of a Node with a mock parser with the given
572
573
574
    name.

    Parameters:
575
        context: the context where the parser shall be replaced
eckhart's avatar
eckhart committed
576
        name: "NAME:PTYPE" of the surrogate. The ptype is optional
577
    """
578
    node = context[-1]
579
580
581
582
583
    name, ptype = (name.split(':') + [''])[:2]
    node.parser = MockParser(name, ptype)


@transformation_factory(Callable)
584
585
def flatten(context: List[Node], condition: Callable=is_anonymous, recursive: bool=True):
    """
586
    Flattens all children, that fulfil the given ``condition``
587
588
589
    (default: all unnamed children). Flattening means that wherever a
    node has child nodes, the child nodes are inserted in place of the
    node.
590
591

    If the parameter ``recursive`` is ``True`` the same will recursively be
592
593
594
    done with the child-nodes, first. In other words, all leaves of
    this node and its child nodes are collected in-order as direct
    children of this node.
595
596
597
598

    Applying flatten recursively will result in these kinds of
    structural transformation::

di68kap's avatar
di68kap committed
599
        (1 (+ 2) (+ 3))    ->   (1 + 2 + 3)
600
601
        (1 (+ (2 + (3))))  ->   (1 + 2 + 3)
    """
602
    node = context[-1]
603
    if node.children:
Eckhart Arnold's avatar
Eckhart Arnold committed
604
        new_result = []     # type: List[Node]
eckhart's avatar
eckhart committed
605
        context.append(ZOMBIE_NODE)
606
        for child in node.children:
eckhart's avatar
eckhart committed
607
            context[-1] = child
608
            if child.children and condition(context):
609
                if recursive:
610
                    flatten(context, condition, recursive)
611
612
613
                new_result.extend(child.children)
            else:
                new_result.append(child)
eckhart's avatar
eckhart committed
614
        context.pop()
615
616
617
        node.result = tuple(new_result)


618
619
620
def collapse(context: List[Node]):
    """
    Collapses all sub-nodes of a node by replacing them with the
621
    string representation of the node.
622
    """
623
    node = context[-1]
624
    node.result = node.content
625
626


627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
# @transformation_factory
# def collect_leaves(context: List[Node], whitespace: str=''):
#     """
#     Collects all leave nodes dropping any intermediary nodes.
#     Optionally adds whitespace between the nodes.
#     """
#     assert context[-1].children
#     node = context[-1]
#     leaves_iterator = node.select(lambda nd: not nd.children, include_root=False)
#     if whitespace:
#         mock_ws_parser = MockParser('', WHITESPACE_PTYPE)
#         result = []
#         for leave in leaves_iterator:
#             result.append(leave)
#             result.append(Node(mock_ws_parser, whitespace, leafhint=True))
#         result.pop()
#         node.result = tuple(result)
#     else:
#         node.result = (nd for nd in leaves_iterator)


648
@transformation_factory
649
650
def merge_children(context: List[Node], tag_names: List[str]):
    """
651
652
653
    Joins all children next to each other and with particular tag-names
    into a single child node with a mock-parser with the name of the
    first tag-name in the list.
654
    """
Eckhart Arnold's avatar
Eckhart Arnold committed
655
    node = context[-1]
656
    result = []
657
    name, ptype = ('', tag_names[0]) if tag_names[0][:1] == ':' else (tag_names[0], '')
658
    if node.children:
659
        i = 0
660
661
662
663
664
665
666
667
668
669
670
        L = len(node.children)
        while i < L:
            while i < L and not node.children[i].tag_name in tag_names:
                result.append(node.children[i])
                i += 1
            k = i + 1
            while (k < L and node.children[k].tag_name in tag_names
                   and bool(node.children[i].children) == bool(node.children[k].children)):
                k += 1
            if i < L:
                result.append(Node(MockParser(name, ptype),
Eckhart Arnold's avatar
Eckhart Arnold committed
671
672
                                   reduce(lambda a, b: a + b,
                                          (node.children for node in node.children[i:k]))))
673
674
675
676
            i = k
        node.result = tuple(result)


eckhart's avatar
eckhart committed
677
@transformation_factory(Callable)
678
def replace_content(context: List[Node], func: Callable):  # Callable[[Node], ResultType]
di68kap's avatar
di68kap committed
679
    """Replaces the content of the node. ``func`` takes the node's result
680
681
    as an argument an returns the mapped result.
    """
682
    node = context[-1]
683
684
685
    node.result = func(node.result)


686
687
688
689
690
691
692
693
@transformation_factory(str)
def replace_content_by(context: List[Node], content: str):  # Callable[[Node], ResultType]
    """Replaces the content of the node with the given text content.
    """
    node = context[-1]
    node.result = content


694
695
696
697
698
699
700
701
702
#######################################################################
#
# destructive transformations:
#
# - leaves may be dropped (e.g. if deemed irrelevant)
# - errors of dropped leaves will be lost
# - no promise that order will be preserved
#
#######################################################################
703
704


705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
@transformation_factory(Callable)
def lstrip(context: List[Node], condition: Callable = is_expendable):
    """Recursively removes all leading child-nodes that fulfill a given condition."""
    node = context[-1]
    i = 1
    while i > 0 and node.children:
        lstrip(context + [node.children[0]], condition)
        i, L = 0, len(node.children)
        while i < L and condition(context + [node.children[i]]):
            i += 1
        if i > 0:
            node.result = node.children[i:]


@transformation_factory(Callable)
def rstrip(context: List[Node], condition: Callable = is_expendable):
    """Recursively removes all leading nodes that fulfill a given condition."""
    node = context[-1]
    i, L = 0, len(node.children)
    while i < L and node.children:
        rstrip(context + [node.children[-1]], condition)
        L = len(node.children)
        i = L
        while i > 0 and condition(context + [node.children[i-1]]):
            i -= 1
        if i < L:
            node.result = node.children[:i]


@transformation_factory(Callable)
eckhart's avatar
eckhart committed
735
def strip(context: List[Node], condition: Callable = is_expendable):
736
737
738
739
740
    """Removes leading and trailing child-nodes that fulfill a given condition."""
    lstrip(context, condition)
    rstrip(context, condition)


741
@transformation_factory(slice)
742
def keep_children(context: List[Node], section: slice = slice(None)):
743
    """Keeps only child-nodes which fall into a slice of the result field."""
744
    node = context[-1]
745
    if node.children:
746
        node.result = node.children[section]
747
748
749


@transformation_factory(Callable)
750
751
752
753
754
755
756
757
def keep_children_if(context: List[Node], condition: Callable):
    """Removes all children for which `condition()` returns `True`."""
    node = context[-1]
    if node.children:
        node.result = tuple(c for c in node.children if condition(context + [c]))


@transformation_factory
eckhart's avatar
eckhart committed
758
def keep_tokens(context: List[Node], tokens: AbstractSet[str]=frozenset()):
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
    """Removes any among a particular set of tokens from the immediate
    descendants of a node. If ``tokens`` is the empty set, all tokens
    are removed."""
    keep_children_if(context, partial(is_token, tokens=tokens))


@transformation_factory
def keep_nodes(context: List[Node], tag_names: AbstractSet[str]):
    """Removes children by tag name."""
    keep_children_if(context, partial(is_one_of, tag_name_set=tag_names))


@transformation_factory
def keep_content(context: List[Node], regexp: str):
    """Removes children depending on their string value."""
    keep_children_if(context, partial(has_content, regexp=regexp))


@transformation_factory(Callable)
def remove_children_if(context: List[Node], condition: Callable):
779
780
781
782
    """Removes all children for which `condition()` returns `True`."""
    node = context[-1]
    if node.children:
        node.result = tuple(c for c in node.children if not condition(context + [c]))
eckhart's avatar
eckhart committed
783
    pass
784

eckhart's avatar
eckhart committed
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
# @transformation_factory(Callable)
# def remove_children(context: List[Node],
#                     condition: Callable = TRUE_CONDITION,
#                     section: slice = slice(None)):
#     """Removes all nodes from a slice of the result field if the function
#     `condition(child_node)` evaluates to `True`."""
#     node = context[-1]
#     if node.children:
#         c = node.children
#         N = len(c)
#         rng = range(*section.indices(N))
#         node.result = tuple(c[i] for i in range(N)
#                             if i not in rng or not condition(context + [c[i]]))
#         # selection = []
#         # for i in range(N):
#         #     context.append(c[i])
#         #     if not i in rng or not condition(context):
#         #         selection.append(c[i])
#         #     context.pop()
#         # if len(selection) != c:
#         #     node.result = tuple(selection)
806
807


808
809
remove_whitespace = remove_children_if(is_whitespace)
# partial(remove_children_if, condition=is_whitespace)
810
remove_empty = remove_children_if(is_empty)
di68kap's avatar
di68kap committed
811
remove_anonymous_empty = remove_children_if(lambda ctx: is_empty(ctx) and is_anonymous(ctx))
812
813
814
815
remove_expendables = remove_children_if(is_expendable)
# partial(remove_children_if, condition=is_expendable)
remove_anonymous_expendables = remove_children_if(lambda ctx: is_anonymous(ctx)
                                                  and is_expendable(ctx))
816
remove_anonymous_tokens = remove_children_if(lambda ctx: is_token(ctx) and is_anonymous(ctx))
817
818
819
remove_first = apply_if(keep_children(slice(1, None)), lambda ctx: len(ctx[-1].children) > 1)
remove_last = apply_if(keep_children(slice(None, -1)), lambda ctx: len(ctx[-1].children) > 1)
remove_brackets = apply_if(keep_children(slice(1, -1)), lambda ctx: len(ctx[-1].children) >= 2)
820
remove_infix_operator = keep_children(slice(0, None, 2))
821
remove_single_child = apply_if(keep_children(slice(0)), lambda ctx: len(ctx[-1].children) == 1)
822
823


eckhart's avatar
eckhart committed
824
825
@transformation_factory(AbstractSet[str])
def remove_tokens(context: List[Node], tokens: AbstractSet[str]=frozenset()):
826
    """Removes any among a particular set of tokens from the immediate
827
828
    descendants of a node. If ``tokens`` is the empty set, all tokens
    are removed."""
829
    remove_children_if(context, partial(is_token, tokens=tokens))
830
831
832


@transformation_factory
eckhart's avatar
eckhart committed
833
def remove_nodes(context: List[Node], tag_names: AbstractSet[str]):
Eckhart Arnold's avatar
Eckhart Arnold committed
834
    """Removes children by tag name."""
835
    remove_children_if(context, partial(is_one_of, tag_name_set=tag_names))
836
837
838


@transformation_factory
839
def remove_content(context: List[Node], regexp: str):
840
    """Removes children depending on their string value."""
841
    remove_children_if(context, partial(has_content, regexp=regexp))
842
843
844
845


########################################################################
#
846
# AST semantic validation functions (EXPERIMENTAL!!!)
847
848
849
#
########################################################################

850
@transformation_factory(Callable)
851
852
853
854
def error_on(context: List[Node], condition: Callable, error_msg: str = ''):
    """
    Checks for `condition`; adds an error message if condition is not met.
    """
855
    node = context[-1]
856
    if not condition(context):
857
858
859
860
861
862
        if error_msg:
            node.add_error(error_msg % node.tag_name if error_msg.find("%s") > 0 else error_msg)
        else:
            cond_name = condition.__name__ if hasattr(condition, '__name__') \
                        else condition.__class__.__name__ if hasattr(condition, '__class__') \
                        else '<unknown>'
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
            node.add_error("transform.error_on: Failed to meet condition " + cond_name)


@transformation_factory(Callable)
def warn_on(context: List[Node], condition: Callable, warning: str = ''):
    """
    Checks for `condition`; adds an warning message if condition is not met.
    """
    node = context[-1]
    if not condition(context):
        if warning:
            node.add_error(warning % node.tag_name if warning.find("%s") > 0 else warning,
                           Error.WARNING)
        else:
            cond_name = condition.__name__ if hasattr(condition, '__name__') \
                        else condition.__class__.__name__ if hasattr(condition, '__class__') \
                        else '<unknown>'
            node.add_error("transform.warn_on: Failed to meet condition " + cond_name,
                           Error.WARNING)
882
883


884
assert_has_children = error_on(lambda nd: nd.children, 'Element "%s" has no children')
885
886
887


@transformation_factory
888
def assert_content(context: List[Node], regexp: str):
889
    node = context[-1]
890
    if not has_content(context, regexp):
eckhart's avatar
eckhart committed
891
        context[0].new_error(node, 'Element "%s" violates %s on %s' %
eckhart's avatar
eckhart committed
892
                             (node.parser.name, str(regexp), node.content))
893

894
895

@transformation_factory
896
def require(context: List[Node], child_tags: AbstractSet[str]):
897
    node = context[-1]
898
899
    for child in node.children:
        if child.tag_name not in child_tags:
eckhart's avatar
eckhart committed
900
            context[0].new_error(node, 'Element "%s" is not allowed inside "%s".' %
eckhart's avatar
eckhart committed
901
                                 (child.parser.name, node.parser.name))
902
903
904


@transformation_factory
905
def forbid(context: List[Node], child_tags: AbstractSet[str]):
906
    node = context[-1]
907
908
    for child in node.children:
        if child.tag_name in child_tags:
eckhart's avatar
eckhart committed
909
            context[0].new_error(node, 'Element "%s" cannot be nested inside "%s".' %
eckhart's avatar
eckhart committed
910
                                 (child.parser.name, node.parser.name))
911