transform.py 33.9 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
import collections.abc
31
32
33
import inspect
from functools import partial, reduce, singledispatch

34
from DHParser.error import Error
eckhart's avatar
eckhart committed
35
from DHParser.syntaxtree import Node, WHITESPACE_PTYPE, TOKEN_PTYPE, MockParser, ZOMBIE_NODE
36
37
from DHParser.toolkit import expand_table, smart_list, re, typing
from typing import AbstractSet, Any, ByteString, Callable, cast, Container, Dict, \
38
    Tuple, List, Sequence, Union, Text, GenericMeta
39

40
41
__all__ = ('TransformationDict',
           'TransformationProc',
42
           'TransformationFunc',
43
44
45
           'ConditionFunc',
           'KeyFunc',
           'transformation_factory',
46
47
48
           'key_parser_name',
           'key_tag_name',
           'traverse',
49
           'is_named',
50
           'replace_by_single_child',
Eckhart Arnold's avatar
Eckhart Arnold committed
51
           'reduce_single_child',
52
           'replace_or_reduce',
53
54
           'replace_parser',
           'collapse',
55
           'merge_children',
56
           'replace_content',
57
           'replace_content_by',
58
           'apply_if',
eckhart's avatar
eckhart committed
59
           'apply_unless',
60
           'traverse_locally',
61
           'is_anonymous',
62
63
64
65
           'is_whitespace',
           'is_empty',
           'is_expendable',
           'is_token',
66
           'is_one_of',
67
           'has_content',
di68kap's avatar
di68kap committed
68
           'has_parent',
69
70
71
72
73
74
75
76
           'lstrip',
           'rstrip',
           'strip',
           'keep_children',
           'keep_children_if',
           'keep_tokens',
           'keep_nodes',
           'keep_content',
77
           'remove_children_if',
eckhart's avatar
eckhart committed
78
           'remove_nodes',
79
80
81
82
83
           'remove_content',
           'remove_first',
           'remove_last',
           'remove_whitespace',
           'remove_empty',
di68kap's avatar
di68kap committed
84
           'remove_anonymous_empty',
85
86
           'remove_anonymous_expendables',
           'remove_anonymous_tokens',
87
88
           'remove_expendables',
           'remove_brackets',
89
90
           'remove_infix_operator',
           'remove_single_child',
91
92
93
94
           'remove_tokens',
           'flatten',
           'forbid',
           'require',
95
           'assert_content',
96
97
           'error_on',
           'warn_on',
eckhart's avatar
eckhart committed
98
           'assert_has_children')
99
100


101
TransformationProc = Callable[[List[Node]], None]
Eckhart Arnold's avatar
Eckhart Arnold committed
102
TransformationDict = Dict[str, Sequence[Callable]]
103
TransformationFunc = Union[Callable[[Node], Any], partial]
Eckhart Arnold's avatar
Eckhart Arnold committed
104
ProcessingTableType = Dict[str, Union[Sequence[Callable], TransformationDict]]
105
106
ConditionFunc = Callable  # Callable[[List[Node]], bool]
KeyFunc = Callable[[Node], str]
eckhart's avatar
eckhart committed
107
CriteriaType = Union[int, str, Callable]
108
109


110
def transformation_factory(t1=None, t2=None, t3=None, t4=None, t5=None):
111
    """Creates factory functions from transformation-functions that
112
    dispatch on the first parameter after the context parameter.
113
114

    Decorating a transformation-function that has more than merely the
Eckhart Arnold's avatar
Eckhart Arnold committed
115
    ``context``-parameter with ``transformation_factory`` creates a
116
    function with the same name, which returns a partial-function that
117
    takes just the context-parameter.
118
119
120
121
122
123
124
125
126

    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.

127
128
    Usage::

eckhart's avatar
eckhart committed
129
        @transformation_factory(AbstractSet[str])
130
        def remove_tokens(context, tokens):
131
            ...
132
133
134

    or, alternatively::

135
        @transformation_factory
136
        def remove_tokens(context, tokens: AbstractSet[str]):
137
138
            ...

139
140
    Example::

141
        trans_table = { 'expression': remove_tokens('+', '-') }
142
143
144

    instead of::

145
        trans_table = { 'expression': partial(remove_tokens, tokens={'+', '-'}) }
146
147

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

153
154
155
156
157
158
159
160
161
162
163
164
165
166
    def type_guard(t):
        """Raises an error if type `t` is a generic type or could be mistaken
        for the type of the canonical first parameter "List[Node] of
        transformation functions. Returns `t`."""
        if isinstance(t, GenericMeta):
            raise TypeError("Generic Type %s not permitted\n in transformation_factory "
                            "decorator. Use the equivalent non-generic type instead!"
                            % str(t))
        if issubclass(List[Node], t):
            raise TypeError("Sequence type %s not permitted\nin transformation_factory "
                            "decorator, because it could be mistaken for a base class "
                            "of List[Node].\nTry 'tuple' instead!" % str(t))
        return t

167
    def decorator(f):
168
        nonlocal t1
169
170
171
172
        sig = inspect.signature(f)
        params = list(sig.parameters.values())[1:]
        if len(params) == 0:
            return f  # '@transformer' not needed w/o free parameters
173
        assert t1 or params[0].annotation != params[0].empty, \
174
            "No type information on second parameter found! Please, use type " \
eckhart's avatar
eckhart committed
175
            "annotation or provide the type information via transformer-decorator."
176
        f = singledispatch(f)
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
        p1type = params[0].annotation
        if t1 is None:
            t1 = type_guard(p1type)
        elif issubclass(p1type, type_guard(t1)):
            try:
                if len(params) == 1 and issubclass(p1type, Container) \
                        and not (issubclass(p1type, Text) or issubclass(p1type, ByteString)):
                    def gen_special(*args):
                        c = set(args) if issubclass(p1type, AbstractSet) else \
                            tuple(args) if issubclass(p1type, Sequence) else args
                        d = {params[0].name: c}
                        return partial(f, **d)

                    f.register(type_guard(p1type.__args__[0]), gen_special)
            except AttributeError:
                pass  # Union Type does not allow subclassing, but is not needed here
        else:
            raise TypeError("Annotated type %s is not a subclass of decorated type %s !"
                            % (str(p1type), str(t1)))
196
197
198
199
200
201

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

202
        for t in (t1, t2, t3, t4, t5):
203
            if t:
204
                f.register(type_guard(t), gen_partial)
205
206
            else:
                break
207
208
        return f

209
    if isinstance(t1, type(lambda: 1)):
210
211
212
        # Provide for the case that transformation_factory has been
        # written as plain decorator and not as a function call that
        # returns the decorator proper.
213
214
        func = t1
        t1 = None
215
216
217
218
219
        return decorator(func)
    else:
        return decorator


220
def key_parser_name(node: Node) -> str:
221
222
223
    return node.parser.name


224
def key_tag_name(node: Node) -> str:
225
226
227
    return node.tag_name


228
def traverse(root_node: Node,
Eckhart Arnold's avatar
Eckhart Arnold committed
229
             processing_table: ProcessingTableType,
230
231
232
             key_func: KeyFunc=key_tag_name) -> None:
    """
    Traverses the snytax tree starting with the given ``node`` depth
233
    first and applies the sequences of callback-functions registered
234
    in the ``processing_table``-dictionary.
235
236
237
238
239
240
241
242
243

    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:
244
245
246
247
248

    - '+': 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)
249
250
251
252
253

    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
254
255
            is interpreted as a ``compact_table``. See
            :func:`expand_table` or :func:`EBNFCompiler.EBNFTransTable`
256
257
258
        key_func (function): A mapping key_func(node) -> keystr. The default
            key_func yields node.parser.name.

259
260
    Example::

261
        table = { "term": [replace_by_single_child, flatten],
262
                  "factor, flowmarker, retrieveop": replace_by_single_child }
263
        traverse(node, table)
264

265
    """
266
267
268
    # Is this optimazation really needed?
    if '__cache__' in processing_table:
        # assume that processing table has already been expanded
eckhart's avatar
eckhart committed
269
        table = processing_table               # type: ProcessingTableType
eckhart's avatar
eckhart committed
270
        cache = cast(TransformationDict, processing_table['__cache__'])  # type: TransformationDict
271
    else:
272
273
        # normalize processing_table entries by turning single values
        # into lists with a single value
274
275
        table = {name: cast(Sequence[Callable], smart_list(call))
                 for name, call in list(processing_table.items())}
276
        table = expand_table(table)
eckhart's avatar
eckhart committed
277
278
        cache = cast(TransformationDict,
                     table.setdefault('__cache__', cast(TransformationDict, dict())))
279
280
        # change processing table in place, so its already expanded and cache filled next time
        processing_table.clear()
281
282
283
284
285
286
        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
287
    # cache = {}  # type: Dict[str, List[Callable]]
288

289
    def traverse_recursive(context):
eckhart's avatar
eckhart committed
290
        nonlocal cache
291
        node = context[-1]
292
        if node.children:
eckhart's avatar
eckhart committed
293
            context.append(ZOMBIE_NODE)
294
            for child in node.result:
eckhart's avatar
eckhart committed
295
                context[-1] = child
296
                traverse_recursive(context)  # depth first
eckhart's avatar
eckhart committed
297
            context.pop()
298
299

        key = key_func(node)
300
301
302
        try:
            sequence = cache[key]
        except KeyError:
303
304
305
            sequence = table.get('+', []) \
                + table.get(key, table.get('*', [])) \
                + table.get('~', [])
306
307
308
309
310
311
312
            # '+' 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:
313
            call(context)
314

315
    traverse_recursive([root_node])
316
317
    # assert processing_table['__cache__']

318

319
#######################################################################
320
#
321
322
# meta transformations, i.e. transformations that call other
# transformations
323
#
324
#######################################################################
325
326


eckhart's avatar
eckhart committed
327
@transformation_factory(dict)
328
329
330
331
332
333
334
335
336
337
338
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)


339
@transformation_factory(collections.abc.Callable)
340
341
342
343
344
345
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)


346
@transformation_factory(collections.abc.Callable)
eckhart's avatar
eckhart committed
347
348
349
350
351
352
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)


353
354
355
356
357
358
359
360
361
362
363
364
365
366
#######################################################################
#
# 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
367
    """Returns ``True`` if the current node does not have any siblings."""
368
369
370
371
    return len(context[-2].children) == 1


def is_named(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
372
    """Returns ``True`` if the current node's parser is a named parser."""
373
374
375
376
    return bool(context[-1].parser.name)


def is_anonymous(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
377
    """Returns ``True`` if the current node's parser is an anonymous parser."""
378
379
380
381
382
383
384
385
386
387
    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
388
    """Returns ``True`` if the current node's content is empty."""
389
390
391
392
    return not context[-1].result


def is_expendable(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
393
394
    """Returns ``True`` if the current node either is a node containing
    whitespace or an empty node."""
395
396
397
    return is_empty(context) or is_whitespace(context)


398
@transformation_factory(collections.abc.Set)
399
400
401
402
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
403
    any token is a match.
404
405
    """
    def stripped(nd: Node) -> str:
406
        """Removes leading and trailing whitespace-nodes from content."""
407
408
409
410
411
        # 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
412
            while k > 0 and nd.children[k - 1].parser.ptype == WHITESPACE_PTYPE:
413
414
415
416
                k -= 1
            return "".join(child.content for child in node.children[i:k])
        return nd.content
    node = context[-1]
417
    return node.parser.ptype == TOKEN_PTYPE and (not tokens or stripped(node) in tokens)
418
419


420
@transformation_factory(collections.abc.Set)
421
422
423
424
425
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
426
@transformation_factory
427
def has_content(context: List[Node], regexp: str) -> bool:
428
429
430
431
432
433
434
435
    """
    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('$'):
        regexp += "$"
436
437
438
    return bool(re.match(regexp, context[-1].content))


439
@transformation_factory(collections.abc.Set)
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
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):
457
458
459
460
    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
461
    node.errors.extend(child.errors)
462
463
464
    node.result = child.result


465
def _reduce_child(node: Node, child: Node):
eckhart's avatar
eckhart committed
466
    node.errors.extend(child.errors)
467
468
469
    node.result = child.result


eckhart's avatar
eckhart committed
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
# 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
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
528
529
530
531
532
533
534
535
536
537
#######################################################################
#
# 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)
538
539


540
541
def replace_by_single_child(context: List[Node]):
    """
542
543
544
    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.
545
546
547
    """
    node = context[-1]
    if len(node.children) == 1:
548
        _replace_by(node, node.children[0])
549
550


Eckhart Arnold's avatar
Eckhart Arnold committed
551
def reduce_single_child(context: List[Node]):
552
    """
553
    Reduces a single branch node by transferring the result of its
554
    immediate descendant to this node, but keeping this node's parser entry.
555
556
    Reduction only takes place if the last node in the context has
    exactly one child.
557
558
559
    """
    node = context[-1]
    if len(node.children) == 1:
560
        _reduce_child(node, node.children[0])
561
562


563
@transformation_factory(collections.abc.Callable)
564
565
566
def replace_or_reduce(context: List[Node], condition: Callable=is_named):
    """
    Replaces node by a single child, if condition is met on child,
567
568
    otherwise (i.e. if the child is anonymous) reduces the child.
    """
569
    node = context[-1]
570
    if len(node.children) == 1:
di68kap's avatar
di68kap committed
571
        child = node.children[0]
572
        if condition(context):
573
            _replace_by(node, child)
574
        else:
575
            _reduce_child(node, child)
576
577
578


@transformation_factory
579
580
581
def replace_parser(context: List[Node], name: str):
    """
    Replaces the parser of a Node with a mock parser with the given
582
583
584
    name.

    Parameters:
585
        context: the context where the parser shall be replaced
eckhart's avatar
eckhart committed
586
        name: "NAME:PTYPE" of the surrogate. The ptype is optional
587
    """
588
    node = context[-1]
589
590
591
592
    name, ptype = (name.split(':') + [''])[:2]
    node.parser = MockParser(name, ptype)


593
@transformation_factory(collections.abc.Callable)
594
595
def flatten(context: List[Node], condition: Callable=is_anonymous, recursive: bool=True):
    """
596
    Flattens all children, that fulfil the given ``condition``
597
598
599
    (default: all unnamed children). Flattening means that wherever a
    node has child nodes, the child nodes are inserted in place of the
    node.
600
601

    If the parameter ``recursive`` is ``True`` the same will recursively be
602
603
604
    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.
605
606
607
608

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

di68kap's avatar
di68kap committed
609
        (1 (+ 2) (+ 3))    ->   (1 + 2 + 3)
610
611
        (1 (+ (2 + (3))))  ->   (1 + 2 + 3)
    """
612
    node = context[-1]
613
    if node.children:
Eckhart Arnold's avatar
Eckhart Arnold committed
614
        new_result = []     # type: List[Node]
eckhart's avatar
eckhart committed
615
        context.append(ZOMBIE_NODE)
616
        for child in node.children:
eckhart's avatar
eckhart committed
617
            context[-1] = child
618
            if child.children and condition(context):
619
                if recursive:
620
                    flatten(context, condition, recursive)
621
622
623
                new_result.extend(child.children)
            else:
                new_result.append(child)
eckhart's avatar
eckhart committed
624
        context.pop()
625
626
627
        node.result = tuple(new_result)


628
629
630
def collapse(context: List[Node]):
    """
    Collapses all sub-nodes of a node by replacing them with the
631
    string representation of the node.
632
    """
633
    node = context[-1]
634
    node.result = node.content
635
636


637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
# @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)


658
659
@transformation_factory(tuple)
def merge_children(context: List[Node], tag_names: Tuple[str]):
660
    """
661
662
663
    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.
664
    """
Eckhart Arnold's avatar
Eckhart Arnold committed
665
    node = context[-1]
666
    result = []
667
    name, ptype = ('', tag_names[0]) if tag_names[0][:1] == ':' else (tag_names[0], '')
668
    if node.children:
669
        i = 0
670
671
672
673
674
675
676
677
678
679
680
        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
681
682
                                   reduce(lambda a, b: a + b,
                                          (node.children for node in node.children[i:k]))))
683
684
685
686
            i = k
        node.result = tuple(result)


687
@transformation_factory(collections.abc.Callable)
688
def replace_content(context: List[Node], func: Callable):  # Callable[[Node], ResultType]
di68kap's avatar
di68kap committed
689
    """Replaces the content of the node. ``func`` takes the node's result
690
691
    as an argument an returns the mapped result.
    """
692
    node = context[-1]
693
694
695
    node.result = func(node.result)


696
@transformation_factory  # (str)
697
698
699
700
701
702
703
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


704
705
706
707
708
709
710
711
712
#######################################################################
#
# 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
#
#######################################################################
713
714


715
@transformation_factory(collections.abc.Callable)
716
717
718
719
720
721
722
723
724
725
726
727
728
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:]


729
@transformation_factory(collections.abc.Callable)
730
731
732
733
734
735
736
737
738
739
740
741
742
743
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]


744
@transformation_factory(collections.abc.Callable)
eckhart's avatar
eckhart committed
745
def strip(context: List[Node], condition: Callable = is_expendable):
746
747
748
749
750
    """Removes leading and trailing child-nodes that fulfill a given condition."""
    lstrip(context, condition)
    rstrip(context, condition)


751
@transformation_factory  # (slice)
752
def keep_children(context: List[Node], section: slice = slice(None)):
753
    """Keeps only child-nodes which fall into a slice of the result field."""
754
    node = context[-1]
755
    if node.children:
756
        node.result = node.children[section]
757
758


759
@transformation_factory(collections.abc.Callable)
760
761
762
763
764
765
766
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]))


767
@transformation_factory(collections.abc.Set)
eckhart's avatar
eckhart committed
768
def keep_tokens(context: List[Node], tokens: AbstractSet[str]=frozenset()):
769
770
771
772
773
774
    """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))


775
@transformation_factory(collections.abc.Set)
776
777
778
779
780
781
782
783
784
785
786
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))


787
@transformation_factory(collections.abc.Callable)
788
def remove_children_if(context: List[Node], condition: Callable):
789
790
791
792
    """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
793
    pass
794

eckhart's avatar
eckhart committed
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
# @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)
816
817


818
819
remove_whitespace = remove_children_if(is_whitespace)
# partial(remove_children_if, condition=is_whitespace)
820
remove_empty = remove_children_if(is_empty)
di68kap's avatar
di68kap committed
821
remove_anonymous_empty = remove_children_if(lambda ctx: is_empty(ctx) and is_anonymous(ctx))
822
823
824
825
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))
826
remove_anonymous_tokens = remove_children_if(lambda ctx: is_token(ctx) and is_anonymous(ctx))
827
828
829
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)
830
remove_infix_operator = keep_children(slice(0, None, 2))
831
remove_single_child = apply_if(keep_children(slice(0)), lambda ctx: len(ctx[-1].children) == 1)
832
833


834
@transformation_factory(collections.abc.Set)
eckhart's avatar
eckhart committed
835
def remove_tokens(context: List[Node], tokens: AbstractSet[str]=frozenset()):
836
    """Removes any among a particular set of tokens from the immediate
837
838
    descendants of a node. If ``tokens`` is the empty set, all tokens
    are removed."""
839
    remove_children_if(context, partial(is_token, tokens=tokens))
840
841


842
@transformation_factory(collections.abc.Set)
eckhart's avatar
eckhart committed
843
def remove_nodes(context: List[Node], tag_names: AbstractSet[str]):
Eckhart Arnold's avatar
Eckhart Arnold committed
844
    """Removes children by tag name."""
845
    remove_children_if(context, partial(is_one_of, tag_name_set=tag_names))
846
847
848


@transformation_factory
849
def remove_content(context: List[Node], regexp: str):
850
    """Removes children depending on their string value."""
851
    remove_children_if(context, partial(has_content, regexp=regexp))
852
853
854
855


########################################################################
#
856
# AST semantic validation functions (EXPERIMENTAL!!!)
857
858
859
#
########################################################################

860
@transformation_factory(collections.abc.Callable)
861
862
863
864
def error_on(context: List[Node], condition: Callable, error_msg: str = ''):
    """
    Checks for `condition`; adds an error message if condition is not met.
    """
865
    node = context[-1]
866
    if not condition(context):
867
868
869
870
871
872
        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>'
873
874
875
            node.add_error("transform.error_on: Failed to meet condition " + cond_name)


876
@transformation_factory(collections.abc.Callable)
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
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)
892
893


894
assert_has_children = error_on(lambda nd: nd.children, 'Element "%s" has no children')
895
896
897


@transformation_factory
898
def assert_content(context: List[Node], regexp: str):
899
    node = context[-1]
900
    if not has_content(context, regexp):
eckhart's avatar
eckhart committed
901
        context[0].new_error(node, 'Element "%s" violates %s on %s' %
eckhart's avatar
eckhart committed
902
                             (node.parser.name, str(regexp), node.content))
903

904

905
@transformation_factory(collections.abc.Set)
906
def require(context: List[Node], child_tags: AbstractSet[str]):
907
    node = context[-1]
908
909
    for child in node.children:
        if child.tag_name not in child_tags:
eckhart's avatar
eckhart committed
910
            context[0].new_error(node, 'Element "%s" is not allowed inside "%s".' %
eckhart's avatar
eckhart committed
911
                                 (child.parser.name, node.parser.name))
912
913


914
@transformation_factory(collections.abc.Set)
915
def forbid(context: List[Node], child_tags: AbstractSet[str]):
916
    node = context[-1]
917
918
    for child in node.children:
        if child.tag_name in child_tags:
eckhart's avatar
eckhart committed
919
            context[0].new_error(node, 'Element "%s" cannot be nested inside "%s".' %
eckhart's avatar
eckhart committed
920
                                 (child.parser.name, node.parser.name))
921