transform.py 35.4 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, Generic
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
    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`."""
157
158
159
160
161
        # 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 hasattr(t, '__parameters__') and t.__parameters:
162
163
164
            raise TypeError("Generic Type %s not permitted\n in transformation_factory "
                            "decorator. Use the equivalent non-generic type instead!"
                            % str(t))
165
        if t == List[Node]:  #  issubclass(List[Node], t)
166
167
            raise TypeError("Sequence type %s not permitted\nin transformation_factory "
                            "decorator, because it could be mistaken for a base class "
168
169
170
                            "of List[Node]\nwhich is the type of the canonical first "
                            "argument of transformation functions. Try 'tuple' instead!"
                            % str(t))
171
172
        return t

173
    def decorator(f):
174
        nonlocal t1
175
176
177
178
        sig = inspect.signature(f)
        params = list(sig.parameters.values())[1:]
        if len(params) == 0:
            return f  # '@transformer' not needed w/o free parameters
179
        assert t1 or params[0].annotation != params[0].empty, \
180
            "No type information on second parameter found! Please, use type " \
eckhart's avatar
eckhart committed
181
            "annotation or provide the type information via transformer-decorator."
182
        f = singledispatch(f)
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
        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)))
202
203
204
205
206
207

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

208
        for t in (t1, t2, t3, t4, t5):
209
            if t:
210
                f.register(type_guard(t), gen_partial)
211
212
            else:
                break
213
214
        return f

215
    if isinstance(t1, type(lambda: 1)):
216
217
218
        # Provide for the case that transformation_factory has been
        # written as plain decorator and not as a function call that
        # returns the decorator proper.
219
220
        func = t1
        t1 = None
221
222
223
224
225
        return decorator(func)
    else:
        return decorator


226
def key_parser_name(node: Node) -> str:
227
228
229
    return node.parser.name


230
def key_tag_name(node: Node) -> str:
231
232
233
    return node.tag_name


234
def traverse(root_node: Node,
Eckhart Arnold's avatar
Eckhart Arnold committed
235
             processing_table: ProcessingTableType,
236
237
238
             key_func: KeyFunc=key_tag_name) -> None:
    """
    Traverses the snytax tree starting with the given ``node`` depth
239
    first and applies the sequences of callback-functions registered
240
    in the ``processing_table``-dictionary.
241
242
243
244
245
246
247
248
249

    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:
250
251
252
253
254

    - '+': 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)
255
256
257
258
259

    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
260
261
            is interpreted as a ``compact_table``. See
            :func:`expand_table` or :func:`EBNFCompiler.EBNFTransTable`
262
263
264
        key_func (function): A mapping key_func(node) -> keystr. The default
            key_func yields node.parser.name.

265
266
    Example::

267
        table = { "term": [replace_by_single_child, flatten],
268
                  "factor, flowmarker, retrieveop": replace_by_single_child }
269
        traverse(node, table)
270

271
    """
272
273
274
    # Is this optimazation really needed?
    if '__cache__' in processing_table:
        # assume that processing table has already been expanded
eckhart's avatar
eckhart committed
275
        table = processing_table               # type: ProcessingTableType
eckhart's avatar
eckhart committed
276
        cache = cast(TransformationDict, processing_table['__cache__'])  # type: TransformationDict
277
    else:
278
279
        # normalize processing_table entries by turning single values
        # into lists with a single value
280
281
        table = {name: cast(Sequence[Callable], smart_list(call))
                 for name, call in list(processing_table.items())}
282
        table = expand_table(table)
eckhart's avatar
eckhart committed
283
284
        cache = cast(TransformationDict,
                     table.setdefault('__cache__', cast(TransformationDict, dict())))
285
286
        # change processing table in place, so its already expanded and cache filled next time
        processing_table.clear()
287
288
289
290
291
292
        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
293
    # cache = {}  # type: Dict[str, List[Callable]]
294

295
    def traverse_recursive(context):
eckhart's avatar
eckhart committed
296
        nonlocal cache
297
        node = context[-1]
298
        if node.children:
eckhart's avatar
eckhart committed
299
            context.append(ZOMBIE_NODE)
di68kap's avatar
di68kap committed
300
            for child in node.children:
eckhart's avatar
eckhart committed
301
                context[-1] = child
302
                traverse_recursive(context)  # depth first
eckhart's avatar
eckhart committed
303
            context.pop()
304
305

        key = key_func(node)
306
307
308
        try:
            sequence = cache[key]
        except KeyError:
309
310
311
            sequence = table.get('+', []) \
                + table.get(key, table.get('*', [])) \
                + table.get('~', [])
312
313
314
315
316
317
318
            # '+' 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:
319
            call(context)
320

321
    traverse_recursive([root_node])
322
323
    # assert processing_table['__cache__']

324

325
#######################################################################
326
#
327
328
# meta transformations, i.e. transformations that call other
# transformations
329
#
330
#######################################################################
331
332


eckhart's avatar
eckhart committed
333
@transformation_factory(dict)
334
335
336
337
338
339
340
341
342
343
344
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)


345
@transformation_factory(collections.abc.Callable)
346
347
348
349
350
351
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)


352
@transformation_factory(collections.abc.Callable)
eckhart's avatar
eckhart committed
353
354
355
356
357
358
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)


359
360
361
362
363
364
365
366
367
368
369
370
371
372
#######################################################################
#
# 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
373
    """Returns ``True`` if the current node does not have any siblings."""
374
375
376
377
    return len(context[-2].children) == 1


def is_named(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
378
    """Returns ``True`` if the current node's parser is a named parser."""
379
380
381
382
    return bool(context[-1].parser.name)


def is_anonymous(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
383
    """Returns ``True`` if the current node's parser is an anonymous parser."""
384
385
386
387
388
389
390
391
392
393
    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
394
    """Returns ``True`` if the current node's content is empty."""
395
396
397
398
    return not context[-1].result


def is_expendable(context: List[Node]) -> bool:
eckhart's avatar
eckhart committed
399
400
    """Returns ``True`` if the current node either is a node containing
    whitespace or an empty node."""
401
402
403
    return is_empty(context) or is_whitespace(context)


404
@transformation_factory(collections.abc.Set)
405
406
407
408
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
409
    any token is a match.
410
411
    """
    def stripped(nd: Node) -> str:
412
        """Removes leading and trailing whitespace-nodes from content."""
413
414
415
416
417
        # 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
418
            while k > 0 and nd.children[k - 1].parser.ptype == WHITESPACE_PTYPE:
419
420
421
422
                k -= 1
            return "".join(child.content for child in node.children[i:k])
        return nd.content
    node = context[-1]
423
    return node.parser.ptype == TOKEN_PTYPE and (not tokens or stripped(node) in tokens)
424
425


426
@transformation_factory(collections.abc.Set)
427
428
429
430
431
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
432
@transformation_factory
433
def has_content(context: List[Node], regexp: str) -> bool:
434
435
436
437
438
439
440
441
    """
    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('$'):
        regexp += "$"
442
443
444
    return bool(re.match(regexp, context[-1].content))


445
@transformation_factory(collections.abc.Set)
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
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):
463
464
465
466
    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 Arnold's avatar
Eckhart Arnold committed
467
    # node.errors.extend(child.errors)
468
    node.result = child.result
469
470
    if hasattr(child, '_xml_attr'):
        node.attributes.update(child.attributes)
471
472


473
def _reduce_child(node: Node, child: Node):
Eckhart Arnold's avatar
Eckhart Arnold committed
474
    # node.errors.extend(child.errors)
475
    node.result = child.result
476
477
    if hasattr(child, '_xml_attr'):
        node.attributes.update(child.attributes)
478
479


eckhart's avatar
eckhart committed
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
# 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
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
538
539
540
541
542
543
544
545
546
547
#######################################################################
#
# 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)
548
549


550
551
def replace_by_single_child(context: List[Node]):
    """
552
553
554
    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.
555
556
557
    """
    node = context[-1]
    if len(node.children) == 1:
558
        _replace_by(node, node.children[0])
559
560


Eckhart Arnold's avatar
Eckhart Arnold committed
561
def reduce_single_child(context: List[Node]):
562
    """
563
    Reduces a single branch node by transferring the result of its
564
    immediate descendant to this node, but keeping this node's parser entry.
565
566
    Reduction only takes place if the last node in the context has
    exactly one child.
567
568
569
    """
    node = context[-1]
    if len(node.children) == 1:
570
        _reduce_child(node, node.children[0])
571
572


573
@transformation_factory(collections.abc.Callable)
574
575
576
def replace_or_reduce(context: List[Node], condition: Callable=is_named):
    """
    Replaces node by a single child, if condition is met on child,
577
578
    otherwise (i.e. if the child is anonymous) reduces the child.
    """
579
    node = context[-1]
580
    if len(node.children) == 1:
di68kap's avatar
di68kap committed
581
        child = node.children[0]
582
        if condition(context):
583
            _replace_by(node, child)
584
        else:
585
            _reduce_child(node, child)
586
587
588


@transformation_factory
589
590
591
def replace_parser(context: List[Node], name: str):
    """
    Replaces the parser of a Node with a mock parser with the given
592
593
594
    name.

    Parameters:
595
        context: the context where the parser shall be replaced
eckhart's avatar
eckhart committed
596
        name: "NAME:PTYPE" of the surrogate. The ptype is optional
597
    """
598
    node = context[-1]
599
    name, ptype = (name.split(':') + [''])[:2]
di68kap's avatar
di68kap committed
600
    node.parser = MockParser(name, ':' + ptype)
601
602


603
@transformation_factory(collections.abc.Callable)
604
605
def flatten(context: List[Node], condition: Callable=is_anonymous, recursive: bool=True):
    """
606
    Flattens all children, that fulfil the given ``condition``
607
608
609
    (default: all unnamed children). Flattening means that wherever a
    node has child nodes, the child nodes are inserted in place of the
    node.
610
611

    If the parameter ``recursive`` is ``True`` the same will recursively be
612
613
614
    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.
615
616
617
618

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

di68kap's avatar
di68kap committed
619
        (1 (+ 2) (+ 3))    ->   (1 + 2 + 3)
620
621
        (1 (+ (2 + (3))))  ->   (1 + 2 + 3)
    """
622
    node = context[-1]
623
    if node.children:
Eckhart Arnold's avatar
Eckhart Arnold committed
624
        new_result = []     # type: List[Node]
eckhart's avatar
eckhart committed
625
        context.append(ZOMBIE_NODE)
626
        for child in node.children:
eckhart's avatar
eckhart committed
627
            context[-1] = child
628
            if child.children and condition(context):
629
                if recursive:
630
                    flatten(context, condition, recursive)
631
632
633
                new_result.extend(child.children)
            else:
                new_result.append(child)
eckhart's avatar
eckhart committed
634
        context.pop()
635
636
637
        node.result = tuple(new_result)


638
639
640
def collapse(context: List[Node]):
    """
    Collapses all sub-nodes of a node by replacing them with the
641
    string representation of the node.
642
    """
643
    node = context[-1]
644
    node.result = node.content
645
646


647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
# @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)


668
669
@transformation_factory(tuple)
def merge_children(context: List[Node], tag_names: Tuple[str]):
670
    """
671
672
673
    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.
674
    """
Eckhart Arnold's avatar
Eckhart Arnold committed
675
    node = context[-1]
676
    result = []
677
    name, ptype = ('', tag_names[0]) if tag_names[0][:1] == ':' else (tag_names[0], '')
678
    if node.children:
679
        i = 0
680
681
682
683
684
685
686
687
688
689
690
        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
691
692
                                   reduce(lambda a, b: a + b,
                                          (node.children for node in node.children[i:k]))))
693
694
695
696
            i = k
        node.result = tuple(result)


697
@transformation_factory(collections.abc.Callable)
698
def replace_content(context: List[Node], func: Callable):  # Callable[[Node], ResultType]
di68kap's avatar
di68kap committed
699
    """Replaces the content of the node. ``func`` takes the node's result
700
701
    as an argument an returns the mapped result.
    """
702
    node = context[-1]
703
704
705
    node.result = func(node.result)


706
@transformation_factory  # (str)
707
708
709
710
711
712
713
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


714
715
716
717
718
719
720
721
722
#######################################################################
#
# 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
#
#######################################################################
723
724


725
@transformation_factory(collections.abc.Callable)
726
727
728
729
730
731
732
733
734
735
736
737
738
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:]


739
@transformation_factory(collections.abc.Callable)
740
741
742
743
744
745
746
747
748
749
750
751
752
753
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]


754
@transformation_factory(collections.abc.Callable)
eckhart's avatar
eckhart committed
755
def strip(context: List[Node], condition: Callable = is_expendable):
756
757
758
759
760
    """Removes leading and trailing child-nodes that fulfill a given condition."""
    lstrip(context, condition)
    rstrip(context, condition)


761
@transformation_factory  # (slice)
762
def keep_children(context: List[Node], section: slice = slice(None)):
763
    """Keeps only child-nodes which fall into a slice of the result field."""
764
    node = context[-1]
765
    if node.children:
766
        node.result = node.children[section]
767
768


769
@transformation_factory(collections.abc.Callable)
770
771
772
773
774
775
776
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]))


777
@transformation_factory(collections.abc.Set)
eckhart's avatar
eckhart committed
778
def keep_tokens(context: List[Node], tokens: AbstractSet[str]=frozenset()):
779
780
781
782
783
784
    """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))


785
@transformation_factory(collections.abc.Set)
786
787
788
789
790
791
792
793
794
795
796
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))


797
@transformation_factory(collections.abc.Callable)
798
def remove_children_if(context: List[Node], condition: Callable):
799
800
801
802
    """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
803
    pass
804

eckhart's avatar
eckhart committed
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
# @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)
826
827


828
829
remove_whitespace = remove_children_if(is_whitespace)
# partial(remove_children_if, condition=is_whitespace)
830
remove_empty = remove_children_if(is_empty)
di68kap's avatar
di68kap committed
831
remove_anonymous_empty = remove_children_if(lambda ctx: is_empty(ctx) and is_anonymous(ctx))
832
833
834
835
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))
836
remove_anonymous_tokens = remove_children_if(lambda ctx: is_token(ctx) and is_anonymous(ctx))
837
838
839
# 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)
840
remove_infix_operator = keep_children(slice(0, None, 2))
841
remove_single_child = apply_if(keep_children(slice(0)), lambda ctx: len(ctx[-1].children) == 1)
842
843


844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
def remove_first(context: List[Node]):
    """Removes the first non-whitespace child."""
    node = context[-1]
    if node.children:
        for i, child in enumerate(node.children):
            if child.parser.ptype != WHITESPACE_PTYPE:
                break
        else:
            return
        node.result = node.children[:i] + node.children[i+1:]


def remove_last(context: List[Node]):
    """Removes the last non-whitespace child."""
    node = context[-1]
    if node.children:
        for i, child in enumerate(reversed(node.children)):
            if child.parser.ptype != WHITESPACE_PTYPE:
                break
        else:
            return
        i = len(node.children) - i - 1
        node.result = node.children[:i] + node.children[i+1:]


def remove_brackets(context: List[Node]):
    """Removes the first and the last non-whitespace child."""
    remove_first(context)
    remove_last(context)


875
@transformation_factory(collections.abc.Set)
eckhart's avatar
eckhart committed
876
def remove_tokens(context: List[Node], tokens: AbstractSet[str]=frozenset()):
877
    """Removes any among a particular set of tokens from the immediate
878
879
    descendants of a node. If ``tokens`` is the empty set, all tokens
    are removed."""
880
    remove_children_if(context, partial(is_token, tokens=tokens))
881
882


883
@transformation_factory(collections.abc.Set)
eckhart's avatar
eckhart committed
884
def remove_nodes(context: List[Node], tag_names: AbstractSet[str]):
Eckhart Arnold's avatar
Eckhart Arnold committed
885
    """Removes children by tag name."""
886
    remove_children_if(context, partial(is_one_of, tag_name_set=tag_names))
887
888
889


@transformation_factory
890
def remove_content(context: List[Node], regexp: str):
891
    """Removes children depending on their string value."""
892
    remove_children_if(context, partial(has_content, regexp=regexp))
893
894
895
896


########################################################################
#
897
# AST semantic validation functions (EXPERIMENTAL!!!)
898
899
900
#
########################################################################

901
@transformation_factory(collections.abc.Callable)
902
903
904
905
def error_on(context: List[Node], condition: Callable, error_msg: str = ''):
    """
    Checks for `condition`; adds an error message if condition is not met.
    """
906
    node = context[-1]
907
    if not condition(context):
908
909
910
911
912
913
        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>'
914
915
916
            node.add_error("transform.error_on: Failed to meet condition " + cond_name)


917
@transformation_factory(collections.abc.Callable)
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
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)
933
934


935
assert_has_children = error_on(lambda nd: nd.children, 'Element "%s" has no children')
936
937
938


@transformation_factory
939
def assert_content(context: List[Node], regexp: str):
940
    node = context[-1]
941
    if not has_content(context, regexp):
eckhart's avatar
eckhart committed
942
        context[0].new_error(node, 'Element "%s" violates %s on %s' %
eckhart's avatar
eckhart committed
943
                             (node.parser.name, str(regexp), node.content))
944

945

946
@transformation_factory(collections.abc.Set)
947
def require(context: List[Node], child_tags: AbstractSet[str]):
948
    node = context[-1]
949
950
    for child in node.children:
        if child.tag_name not in child_tags:
eckhart's avatar
eckhart committed
951
            context[0].new_error(node, 'Element "%s" is not allowed inside "%s".' %
eckhart's avatar
eckhart committed
952
                                 (child.parser.name, node.parser.name))
953
954


955
@transformation_factory(collections.abc.Set)
956
def forbid(context: List[Node], child_tags: AbstractSet[str]):
957
    node = context[-1]
958
959
    for child in node.children:
        if child.tag_name in child_tags:
eckhart's avatar
eckhart committed
960
            context[0].new_error(node, 'Element "%s" cannot be nested inside "%s".' %
eckhart's avatar
eckhart committed
961
                                 (child.parser.name, node.parser.name))