syntaxtree.py 49.1 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
# syntaxtree.py - syntax tree classes for DHParser
#
# 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.
17
18
19


"""
20
21
Module ``syntaxtree`` defines the ``Node``-class for syntax trees as well
as an abstract base class for parser-objects. The latter is defined
eckhart's avatar
eckhart committed
22
here, because node-objects refer to parser-objects. All concrete
23
24
25
parser classes are defined in the ``parse`` module.
"""

26
from collections import OrderedDict
eckhart's avatar
eckhart committed
27
import copy
28
import json
29
from typing import Callable, cast, Iterator, List, AbstractSet, Set, Union, Tuple, Optional, Dict
30

eckhart's avatar
eckhart committed
31
from DHParser.error import Error, ErrorCode, linebreaks, line_col
32
from DHParser.stringview import StringView
33
from DHParser.toolkit import get_config_value, re
di68kap's avatar
di68kap committed
34

35

di68kap's avatar
di68kap committed
36
__all__ = ('WHITESPACE_PTYPE',
37
           'TOKEN_PTYPE',
38
39
           'ZOMBIE_TAG',
           'PLACEHOLDER',
eckhart's avatar
eckhart committed
40
41
42
           'ResultType',
           'StrictResultType',
           'ChildrenType',
43
           'Node',
44
           'serialize',
45
           'FrozenNode',
46
           'tree_sanity_check',
di68kap's avatar
di68kap committed
47
           'RootNode',
48
           'parse_sxpr',
49
           'parse_xml',
eckhart's avatar
eckhart committed
50
           'parse_tree',
51
52
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
53
54


55
56
57
58
59
60
61
62
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


WHITESPACE_PTYPE = ':Whitespace'
63
TOKEN_PTYPE = ':Token'
64

65
ZOMBIE_TAG = "__ZOMBIE__"
66

67
68
69
70
71
72
73
#######################################################################
#
# syntaxtree nodes
#
#######################################################################


eckhart's avatar
eckhart committed
74
75
76
77
RX_IS_SXPR = re.compile(r'\s*\(')
RX_IS_XML = re.compile(r'\s*<')


78
def flatten_sxpr(sxpr: str, threshold: int = -1) -> str:
79
80
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
81
82
    whitespace.

83
84
85
86
87
88
89
    The ``threshold`` value is a maximum number of
    characters allowed in the flattened expression. If this number
    is exceeded the the unflattened S-expression is returned. A
    negative number means that the S-expression will always be
    flattened. Zero or (any postive integer <= 3) essentially means
    that the expression will not be flattened.

Eckhart Arnold's avatar
Eckhart Arnold committed
90
    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
91
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
92
93
    '(a (b c))'
    """
eckhart's avatar
eckhart committed
94
95
96
    assert RX_IS_SXPR.match(sxpr)
    if threshold == 0:
        return sxpr
97
    flat = re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
eckhart's avatar
eckhart committed
98
    if len(flat) > threshold >= 0:
99
100
        return sxpr.strip()
    return flat
Eckhart Arnold's avatar
Eckhart Arnold committed
101
102


103
def flatten_xml(xml: str) -> str:
104
105
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
106
    i.e. only whitespace within leaf-nodes is preserved.
107
108
    A more precise alternative to `flatten_xml` is to use Node.as_xml()
    ans passing a set containing the top level tag to parameter `inline_tags`.
109
    """
110
111
    # works only with regex
    # return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?<=</\w+>)\s+', '', xml))
eckhart's avatar
eckhart committed
112
    assert RX_IS_XML.match(xml)
113
114
    def tag_only(m):
        return m.groupdict()['closing_tag']
115
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
116
117


118
119
120
121
122
ChildrenType = Tuple['Node', ...]
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
StrictResultType = Union[ChildrenType, StringView, str]
ResultType = Union[ChildrenType, 'Node', StringView, str, None]

eckhart's avatar
eckhart committed
123
RX_AMP = re.compile(r'&(?!\w+;)')
124
125


Eckhart Arnold's avatar
Eckhart Arnold committed
126
class Node:  # (collections.abc.Sized): Base class omitted for cython-compatibility
127
128
129
    """
    Represents a node in the concrete or abstract syntax tree.

130
131
    TODO: Add some documentation and doc-tests here...

132
    Attributes:
133
134
        tag_name (str):  The name of the node, which is either its
            parser's name or, if that is empty, the parser's class name
135

136
137
138
        result (str or tuple):  The result of the parser which
            generated this node, which can be either a string or a
            tuple of child nodes.
139

140
141
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
142

eckhart's avatar
eckhart committed
143
144
145
146
        content (str):  Yields the contents of the tree as string. The
            difference to ``str(node)`` is that ``node.content`` does
            not add the error messages to the returned string.

Eckhart Arnold's avatar
Eckhart Arnold committed
147
        parser (Parser):  The parser which generated this node.
148
149
            WARNING: In case you use mock syntax trees for testing or
            parser replacement during the AST-transformation: DO NOT
Eckhart Arnold's avatar
Eckhart Arnold committed
150
151
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
152
            example by calling ``isinstance(node.parer, ...)``.
153

154
155
156
        len (int):  The full length of the node's string result if the
            node is a leaf node or, otherwise, the concatenated string
            result's of its descendants. The figure always represents
157
            the length before AST-transformation and will never change
158
            through AST-transformation. READ ONLY!
159

160
161
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
162
            The value of ``pos`` is -1 meaning invalid by default.
163
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
164
            nodes relative to this value.
165
166

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
167
            pos value of the root node should be set to 0 right
168
169
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
170
            Other than that, this value should be considered READ ONLY.
171
172
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
173

174
175
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
176
177
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
178
    """
179

180
    __slots__ = '_result', 'children', '_pos', 'tag_name', '_xml_attr'
181

182
    def __init__(self, tag_name: str, result: ResultType, leafhint: bool = False) -> None:
183
184
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
185
186
        that generated the node and the parser's result.
        """
eckhart's avatar
eckhart committed
187
        self._pos = -1                  # type: int
188
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
189
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
190
        if leafhint:
eckhart's avatar
eckhart committed
191
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
192
193
194
            self.children = NoChildren  # type: ChildrenType
        else:
            self.result = result
195
        self.tag_name = tag_name        # type: str
196

eckhart's avatar
eckhart committed
197
198
    def __deepcopy__(self, memo):
        if self.children:
199
            duplicate = self.__class__(self.tag_name, copy.deepcopy(self.children), False)
eckhart's avatar
eckhart committed
200
        else:
201
            duplicate = self.__class__(self.tag_name, self.result, True)
eckhart's avatar
eckhart committed
202
        duplicate._pos = self._pos
eckhart's avatar
eckhart committed
203
        if self.attr_active():
di68kap's avatar
di68kap committed
204
205
            duplicate.attr.update(copy.deepcopy(self._xml_attr))
            # duplicate._xml_attr = copy.deepcopy(self._xml_attr)  # this is not cython compatible
eckhart's avatar
eckhart committed
206
        return duplicate
Eckhart Arnold's avatar
Eckhart Arnold committed
207

208
    def __str__(self):
209
210
        if isinstance(self, RootNode):
            root = cast(RootNode, self)
211
            errors = root.errors_sorted
212
213
            if errors:
                e_pos = errors[0].pos
Eckhart Arnold's avatar
Eckhart Arnold committed
214
215
                content = self.content
                return content[:e_pos] + \
216
                   ' <<< Error on "%s" | %s >>> ' % \
Eckhart Arnold's avatar
Eckhart Arnold committed
217
                   (content[e_pos - self.pos:], '; '.join(e.message for e in errors))
218
        return self.content
Eckhart Arnold's avatar
Eckhart Arnold committed
219

220
    def __repr__(self):
221
        # mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
222
223
        # name, ptype = (self._tag_name.split(':') + [''])[:2]
        # parg = "MockParser({name}, {ptype})".format(name=name, ptype=ptype)
224
        rarg = str(self) if not self.children else \
225
            "(" + ", ".join(child.__repr__() for child in self.children) + ")"
226
        return "Node(%s, %s)" % (self.tag_name, rarg)
227

Eckhart Arnold's avatar
Eckhart Arnold committed
228

229
    def __len__(self):
230
231
        return (sum(len(child) for child in self.children)
                if self.children else len(self._result))
232
233
234


    def __bool__(self):
235
236
237
238
        """Returns the bool value of a node, which is always True. The reason
        for this is that a boolean test on a variable that can contain a node
        or None will only yield `False` in case of None.
        """
239
240
241
        return True


242
    def __hash__(self):
243
        return hash(self.tag_name)
244

Eckhart Arnold's avatar
Eckhart Arnold committed
245

246
247
248
    def __getitem__(self, index_or_tagname: Union[int, str]) -> Union['Node', Iterator['Node']]:
        """
        Returns the child node with the given index if ``index_or_tagname`` is
249
        an integer or the first child node with the given tag name. Examples::
250

251
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
252
253
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
254
255
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
256
257
258
259

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
260
        Returns:
261
262
            Node: All nodes which have a given tag name.
        """
263
264
265
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
266
            else:
267
268
269
270
271
                for child in self.children:
                    if child.tag_name == index_or_tagname:
                        return child
                raise KeyError(index_or_tagname)
        raise ValueError('Leave nodes have no children that can be indexed!')
272
273
274
275


    def __contains__(self, tag_name: str) -> bool:
        """
276
        Returns true if a child with the given tag name exists.
277
        Args:
278
279
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
280
281
282
283
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
284
285
286
287
288
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
289
            return False
290
        raise ValueError('Leave node cannot contain other nodes')
291
292


293
    def equals(self, other: 'Node') -> bool:
294
        """
295
296
297
298
299
300
        Equality of value: Two nodes are considered as having the same value,
        if their tag name is the same, if their results are equal and
        if their attributes and attribute values are the same.

        Returns True, if the tree originating in node `self` is equal by
        value to the tree originating in node `other`.
301
302
303
304
305
306
307
308
309
310
        """
        if self.tag_name == other.tag_name and self.compare_attr(other):
            if self.children:
                return (len(self.children) == len(other.children)
                        and all(a.equals(b) for a, b in zip(self.children, other.children)))
            else:
                return self.result == other.result
        return False


311
312
313
314
315
316
317
318
319
320
321
322
323
324
    def get(self, index_or_tagname: Union[int, str],
            surrogate: Union['Node', Iterator['Node']]) -> Union['Node', Iterator['Node']]:
        """Returns the child node with the given index if ``index_or_tagname``
        is an integer or the first child node with the given tag name. If no
        child with the given index or tag_name exists, the ``surrogate`` is
        returned instead. This mimics the behaviour of Python's dictionary's
        get-method.
        """
        try:
            return self[index_or_tagname]
        except KeyError:
            return surrogate


325
    def is_anonymous(self) -> bool:
Eckhart Arnold's avatar
Eckhart Arnold committed
326
        return not self.tag_name or self.tag_name[0] == ':'
327

Eckhart Arnold's avatar
Eckhart Arnold committed
328

329
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
330
    def result(self) -> StrictResultType:
331
332
333
334
335
        """
        Returns the result from the parser that created the node.
        Error messages are not included in the result. Use `self.content()`
        if the result plus any error messages is needed.
        """
336
337
        return self._result

338

339
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
340
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
341
        # # made obsolete by static type checking with mypy
Eckhart Arnold's avatar
Eckhart Arnold committed
342
343
344
345
        assert ((isinstance(result, tuple) and all(isinstance(child, Node) for child in result))
                or isinstance(result, Node)
                or isinstance(result, str)
                or isinstance(result, StringView)), "%s (%s)" % (str(result), str(type(result)))
Eckhart Arnold's avatar
Eckhart Arnold committed
346
347
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
348
        # self._content = None
349
350
351
352
353
354
355
356
357
        if isinstance(result, Node):
            self.children = (result,)
            self._result = self.children
        else:
            if isinstance(result, tuple):
                self.children = result
                self._result = result or ''
            else:
                self.children = NoChildren
eckhart's avatar
eckhart committed
358
                self._result = result  # cast(StrictResultType, result)
359

360

361
362
363
364
365
366
367
368
369
370
371
372
    def _content(self) -> List[str]:
        """
        Returns string content as list of string fragments
        that are gathered from all child nodes in order.
        """
        if self.children:
            fragments = []
            for child in self.children:
                fragments.extend(child._content())
            return fragments
        self._result = str(self._result)
        return [self._result]
373

374

375
    @property
eckhart's avatar
eckhart committed
376
    def content(self) -> str:
377
        """
378
379
380
        Returns content as string. If the node has child-nodes, the
        string content of the child-nodes is recursively read and then
        concatenated.
381
        """
382
383
384
385
386
387
388
389
        if self.children:
            fragments = []
            for child in self.children:
                fragments.extend(child._content())
            return ''.join(fragments)
        self._result = str(self._result)
        return self._result
        # unoptimized
Eckhart Arnold's avatar
Eckhart Arnold committed
390
391
        # return "".join(child.content for child in self.children) if self.children \
        #     else str(self._result)
di68kap's avatar
di68kap committed
392
393


394
    @property
395
    def pos(self) -> int:
eckhart's avatar
eckhart committed
396
397
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
Eckhart Arnold's avatar
Eckhart Arnold committed
398
            raise AssertionError("Position value not initialized! Use Node.with_pos()")
399
400
        return self._pos

eckhart's avatar
eckhart committed
401

Eckhart Arnold's avatar
Eckhart Arnold committed
402
    def with_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
403
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
404
        Initialize position value. Usually, the parser guard
eckhart's avatar
eckhart committed
405
406
        (`parsers.add_parser_guard()`) takes care of assigning the
        position in the document to newly created nodes. However,
Eckhart Arnold's avatar
Eckhart Arnold committed
407
        when Nodes are created outside the reach of the parser
eckhart's avatar
eckhart committed
408
        guard, their document-position must be assigned manually.
Eckhart Arnold's avatar
Eckhart Arnold committed
409
410
        Position values of the child nodes are assigned recursively, too.
        Returns the node itself for convenience.
eckhart's avatar
eckhart committed
411
        """
412
413
414
415
        # condition self.pos == pos cannot be assumed when tokens or whitespace
        # are dropped early!
        # assert self._pos < 0 or self.pos == pos, ("pos mismatch %i != %i at Node: %s"
        #                                           % (self._pos, pos, repr(self)))
Eckhart Arnold's avatar
Eckhart Arnold committed
416
417
418
419
420
421
422
423
424
425
        if pos != self._pos >= 0:
            raise AssertionError("Position value cannot be reassigned to a different value!")
        if self._pos < 0:
            self._pos = pos
            # recursively adjust pos-values of all children
            offset = self.pos
            for child in self.children:
                if child._pos < 0:
                    child.with_pos(offset)
                offset = child.pos + len(child)
eckhart's avatar
eckhart committed
426
427
        return self

428

429
430
431
432
433
434
435
436
437
438
439
440
441
    @property
    def attr(self):
        """
        Returns a dictionary of XML-attr attached to the node.
        """
        try:
            if self._xml_attr is None:          # cython compatibility
                self._xml_attr = OrderedDict()
        except AttributeError:
            self._xml_attr = OrderedDict()
        return self._xml_attr


eckhart's avatar
eckhart committed
442
443
444
445
446
447
448
449
450
451
452
453
454
    def attr_active(self) -> bool:
        """
        Returns True, if XML-Attributes of this node have ever been set
        or queried, even if unsuccessfully.
        """
        try:
            if self._xml_attr is not None:
                return True
        except AttributeError:
            pass
        return False


455
    def compare_attr(self, other: 'Node') -> bool:
456
        """
457
458
        Returns True, if `self` and `other` have the same attributes with the
        same attribute values.
459
        """
460
461
462
463
464
465
466
467
468
        if self.attr_active():
            if other.attr_active():
                return self.attr == other.attr
            return len(self.attr) == 0
            # self has empty dictionary and other has no attributes
        elif other.attr_active():
            return len(other.attr) == 0
            # other has empty attribute dictionary and self as no attributes
        return True  # neither self nor other have any attributes
469

Eckhart Arnold's avatar
Eckhart Arnold committed
470

471
472
473
474
475
476
477
478
479
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
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
            -> Iterator['Node']:
        """
        Finds nodes in the tree that fulfill a given criterion.

        `select` is a generator that yields all nodes for which the
        given `match_function` evaluates to True. The tree is
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.

        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
        Yields:
            Node: All nodes of the tree for which
            ``match_function(node)`` returns True
        """
        if include_root and match_function(self):
            yield self
        child_iterator = reversed(self.children) if reverse else self.children
        for child in child_iterator:
            if match_function(child):
                yield child
            yield from child.select(match_function, False, reverse)
        # The above variant is slightly faster
        # for child in child_iterator:
        #     yield from child.select(match_function, True, reverse)


    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
                      include_root: bool = False) -> Iterator['Node']:
        """
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::

            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
            ['(X (c "d"))', '(X "F")']
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
            ['(b "X")', '(X (c "d"))', '(X "F")']
            >>> any(tree.select_by_tag('a', False))
            False
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
            ['(a (b "X") (X (c "d")) (e (X "F")))']
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'

        Args:
            tag_name(set): A tag name or set of tag names that is being
                searched for
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
        Yields:
            Node: All nodes which have a given tag name.
        """
        if isinstance(tag_names, str):
            tag_names = frozenset({tag_names})
        return self.select(lambda node: node.tag_name in tag_names, include_root)


    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

        This function is mostly just syntactic sugar for
        ``next(node.select_by_tag(tag_names, False))``. However, rather than
        raising a StopIterationError if no descendant with the given tag-name
        exists, it returns None.
        """
        try:
            return next(self.select_by_tag(tag_names, False))
        except StopIteration:
            return None


    def tree_size(self) -> int:
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
        return sum(child.tree_size() for child in self.children) + 1

    #
    # serialization methods
    #

563
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
564
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
565
566
567
568
569
570
571
572
        """
        Generates a tree representation of this node and its children
        in string from.

        The kind ot tree-representation that is determined by several
        function parameters. This could be an XML-representation or a
        lisp-like S-expression.

573
        Args:
574
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
575
            open_fn:   (Node->str) A function that returns an opening
576
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
577
            close_fn:  (Node->str) A function that returns a closeF
578
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
579
            data_fn:   (str->str) A function that filters the data string
580
581
582
583
584
585
                before printing, e.g. to add quotation marks

        Returns (str):
            A string that contains a (serialized) tree representation
            of the node and its children.
        """
eckhart's avatar
eckhart committed
586
587
        head = open_fn(self)
        tail = close_fn(self)
588
589

        if not self.result:
590
            return head.rstrip() + tail.lstrip()
591

eckhart's avatar
eckhart committed
592
        tail = tail.lstrip(None if density & 2 else '')
593

594
        inline = inline or inline_fn(self)
595
596
597
598
599
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
600
601
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
602

603
604
        if self.children:
            content = []
605
            for child in self.children:
606
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
607
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
608
                if subtree:
eckhart's avatar
eckhart committed
609
610
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
611
            return head + usetab + (sep + usetab).join(content) + tail
612

eckhart's avatar
eckhart committed
613
        res = self.content
eckhart's avatar
eckhart committed
614
        if not inline and not head:
eckhart's avatar
eckhart committed
615
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
616
            res = res.strip()
617
618
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
619
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
620
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
621
        else:
622
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
623

Eckhart Arnold's avatar
Eckhart Arnold committed
624

625
626
    def as_sxpr(self, src: str = None,
                indentation: int = 2,
627
                compact: bool = False,
eckhart's avatar
eckhart committed
628
                flatten_threshold: int = 92) -> str:
629
        """
630
631
        Returns content as S-expression, i.e. in lisp-like form. If this
        method is callad on a RootNode-object,
632

633
        Args:
634
635
636
            src:  The source text or `None`. In case the source text is
                given the position of the element in the text will be
                reported as line and column.
637
638
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
639
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
640
                tree structure.
641
642
643
            flatten_threshold:  Return the S-expression in flattened form if
                the flattened expression does not exceed the threshold length.
                A negative number means that it will always be flattened.
644
645
        """

eckhart's avatar
eckhart committed
646
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
647
        lbreaks = linebreaks(src) if src else []  # type: List[int]
648
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[RootNode]
649

650
        def opening(node: Node) -> str:
eckhart's avatar
eckhart committed
651
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
652
            txt = [left_bracket, node.tag_name]
653
            # s += " '(pos %i)" % node.add_pos
654
            # txt.append(str(id(node)))  # for debugging
eckhart's avatar
eckhart committed
655
            if node.attr_active():
656
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
657
            if src:
658
659
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
660
661
            if root and id(node) in root.error_nodes:
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in root.get_errors(node)))
662
            return "".join(txt) + '\n'
663

664
        def closing(node: Node) -> str:
eckhart's avatar
eckhart committed
665
666
            """Returns the closing string for the representation of `node`."""
            return right_bracket
667

668
        def pretty(strg: str) -> str:
eckhart's avatar
eckhart committed
669
670
671
672
            """Encloses `strg` with the right kind of quotation marks."""
            return '"%s"' % strg if strg.find('"') < 0 \
                else "'%s'" % strg if strg.find("'") < 0 \
                else '"%s"' % strg.replace('"', r'\"')
673

674
        sxpr = self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
eckhart's avatar
eckhart committed
675
        return sxpr if compact else flatten_sxpr(sxpr, flatten_threshold)
Eckhart Arnold's avatar
Eckhart Arnold committed
676

eckhart's avatar
eckhart committed
677

678
679
    def as_xml(self, src: str = None,
               indentation: int = 2,
eckhart's avatar
eckhart committed
680
681
682
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
683
684
685
        """
        Returns content as XML-tree.

686
        Args:
687
688
689
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
690
            indentation: The number of whitespaces for indentation
691
692
693
            inline_tags:  A set of tag names, the content of which will always be written
                on a single line, unless it contains explicit line feeds ('\n').
            omit_tags:  A set of tags from which only the content will be printed, but
694
                neither the opening tag nor its attr nor the closing tag. This
695
696
697
                allows producing a mix of plain text and child tags in the output,
                which otherwise is not supported by the Node object, because it
                requires its content to be either a tuple of children or string content.
698
699
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
700
        """
701
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[RootNode]
702

703
        def opening(node: Node) -> str:
704
705
706
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
707
            txt = ['<', node.tag_name]
eckhart's avatar
eckhart committed
708
            has_reserved_attrs = node.attr_active() \
eckhart's avatar
eckhart committed
709
                and any(r in node.attr for r in {'err', 'line', 'col'})
eckhart's avatar
eckhart committed
710
            if node.attr_active():
711
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
712
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
713
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
714
            if root and id(node) in root.error_nodes and not has_reserved_attrs:
eckhart's avatar
eckhart committed
715
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
eckhart's avatar
eckhart committed
716
                                                 for err in root.get_errors(node)))
717
718
719
            if node.tag_name in empty_tags:
                assert not node.result, ("Node %s with content %s is not an empty element!" %
                                         (node.tag_name, str(node)))
eckhart's avatar
eckhart committed
720
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
721
722
723
            else:
                ending = ">\n"
            return "".join(txt + [ending])
724

725
        def closing(node: Node):
726
            """Returns the closing string for the representation of `node`."""
727
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
728
                return ''
729
            return ('\n</') + node.tag_name + '>'
730

731
732
733
734
735
736
        def sanitizer(content: str) -> str:
            """Substitute "&", "<", ">" in XML-content by the respective entities."""
            content = RX_AMP.sub('&amp;', content)
            content = content.replace('<', '&lt;').replace('>', '&gt;')
            return content

737
        def inlining(node: Node):
738
739
740
741
            """Returns True, if `node`'s tag name is contained in `inline_tags`,
            thereby signalling that the children of this node shall not be
            printed on several lines to avoid unwanted gaps in the output.
            """
eckhart's avatar
eckhart committed
742
            return node.tag_name in inline_tags \
eckhart's avatar
eckhart committed
743
                or (node.attr_active()
eckhart's avatar
eckhart committed
744
                    and node.attr.get('xml:space', 'default') == 'preserve')
745

746
        line_breaks = linebreaks(src) if src else []
747
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
748
                               density=1, inline_fn=inlining)
749

Eckhart Arnold's avatar
Eckhart Arnold committed
750

751
    def to_json_obj(self) -> Dict:
752
        """Seralize a node or tree as json-object"""
Eckhart Arnold's avatar
Eckhart Arnold committed
753
754
755
        data = [self.tag_name,
                [child.to_json_obj() for child in self.children]
                if self.children else str(self._result)]
756
757
758
759
760
761
        has_attr = self.attr_active()
        if self._pos >= 0 or has_attr:
            data.append(self._pos)
        if has_attr:
            data.append(dict(self._xml_attr))
        return { '__class__': 'DHParser.Node', 'data': data }
762
763


764
765
    @staticmethod
    def from_json_obj(json_obj: Dict) -> 'Node':
766
767
768
769
770
771
        """Convert a json object representing a node (or tree) back into a
        Node object. Raises a ValueError, if `json_obj` does not represent
        a node."""
        if json_obj.get('__class__', '') != 'DHParser.Node':
            raise ValueError('JSON object: ' + str(json_obj) +
                             ' does not represent a Node object.')
772
        tag_name, result, pos, attr = (json_obj['data'] + [-1, None])[:4]
773
        if isinstance(result, str):
774
775
776
777
778
779
780
            leafhint = True
        else:
            leafhint = False
            result = tuple(Node.from_json_obj(child) for child in result)
        node = Node(tag_name, result, leafhint)
        node._pos = pos
        if attr:
781
            node.attr.update(attr)
782
        return node
783

784
785
786
787
    def as_json(self, indent: Optional[int] = 2, ensure_ascii=False) -> str:
        return json.dumps(self.to_json_obj(), indent=indent, ensure_ascii=ensure_ascii,
                          separators=(', ', ': ') if indent is not None else (',', ':'))

788

789
790
def serialize(node: Node, how: str='default') -> str:
    """
791
    Serializes the tree starting with `node` either as S-expression, XML, JSON,
792
    or in compact form. Possible values for `how` are 'S-expression',
793
    'XML', 'JSON', 'compact' accordingly, or 'AST', 'CST', 'default' in which case
794
795
796
797
798
799
800
801
802
803
804
805
806
    the value of respective configuration variable determines the
    serialization format. (See module `configuration.py`.)
    """
    switch = how.lower()

    if switch == 'ast':
        switch = get_config_value('ast_serialization').lower()
    elif switch == 'cst':
        switch = get_config_value('cst_serialization').lower()
    elif switch == 'default':
        switch = get_config_value('default_serialization').lower()

    if switch == 's-expression':
807
        return node.as_sxpr(flatten_threshold=get_config_value('flatten_sxpr_threshold'))
808
809
    elif switch == 'xml':
        return node.as_xml()
810
811
    elif switch == 'json':
        return node.as_json()
812
813
814
815
816
817
    elif switch == 'compact':
        return node.as_sxpr(compact=True)
    else:
        raise ValueError('Unknown serialization %s, %s' % (how, switch))


818
class FrozenNode(Node):
819
820
821
822
823
824
825
826
827
828
    """
    FrozenNode is an immutable kind of Node, i.e. it must not be changed
    after initialization. The purpose is mainly to allow certain kinds of
    optimization, like not having to instantiate empty nodes (because they
    are always the same and will be dropped while parsing, anyway).

    Frozen nodes must be used only temporarily during parsing or
    tree-transformation and should not occur in the product of the
    transformation any more. This can be verified with `tree_sanity_check()`.
    """
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845

    def __init__(self, tag_name: str, result: ResultType) -> None:
        if isinstance(result, str) or isinstance(result, StringView):
            result = str(result)
        else:
            raise TypeError('FrozenNode only accepts string as results. '
                            '(Only leaf-nodes can be frozen nodes.)')
        super(FrozenNode, self).__init__(tag_name, result, True)

    @property
    def result(self) -> StrictResultType:
        return self._result

    @result.setter
    def result(self, result: ResultType):
        raise TypeError('FrozenNode does not allow re-assignment of results.')

eckhart's avatar
eckhart committed
846
847
848
849
    @property
    def attr(self):
        raise AssertionError("Attributes cannot be accessed on a frozen node")

Eckhart Arnold's avatar
Eckhart Arnold committed
850
    def with_pos(self, pos: int) -> 'Node':
851
852
853
        pass


854
PLACEHOLDER = FrozenNode('__PLACEHOLDER__', '')
di68kap's avatar
di68kap committed
855
856


857
def tree_sanity_check(tree: Node) -> bool:
858
859
860
861
862
863
864
865
866
    """
    Sanity check for syntax trees: One and the same node must never appear
    twice in the syntax tree. Frozen Nodes (EMTPY_NODE, PLACEHOLDER)
    should only exist temporarily and must have been dropped or eliminated
    before any kind of tree generation (i.e. parsing) or transformation
    is finished.
    :param tree: the root of the tree to be checked
    :return: True, if the tree is `sane`, False otherwise.
    """
867
    node_set = set()  # type: Set[Node]
868
    for node in tree.select(lambda nd: True, include_root=True):
869
        if node in node_set or isinstance(Node, FrozenNode):
870
871
872
873
874
            return False
        node_set.add(node)
    return True


di68kap's avatar
di68kap committed
875
class RootNode(Node):
876
877
878
879
880
881
882
883
884
885
886
887
    """The root node for the syntax tree is a special kind of node that keeps
    and manages global properties of the tree as a whole. These are first and
    foremost the list off errors that occurred during tree generation
    (i.e. parsing) or any transformation of the tree. Other properties concern
    the customization of the XML-serialization.

    The root node can be instantiated before the tree is fully parsed. This is
    necessary, because the root node is needed for managing error messages
    during the parsing process, already. In order to connect the root node to
    the tree, when parsing is finished, the swallow()-method must be called.

        errors (list):  A list of all errors that have occurred so far during
888
889
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
890

891
892
893
894
895
896
        error_nodes (dict): A mapping of node-ids to a list of errors that
                occurred on the node with the respective id.

        error_positions (dict): A mapping of locations to a set of ids of
                nodes that contain an error at that particular location

897
898
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
899
900
901
902
903
904

        inline_tags (set of strings): see `Node.as_xml()` for an explanation.

        omit_tags (set of strings): see `Node.as_xml()` for an explanation.

        empty_tags (set oif strings): see `Node.as_xml()` for an explanation.
di68kap's avatar
di68kap committed
905
    """
906

eckhart's avatar
eckhart committed
907
    def __init__(self, node: Optional[Node] = None):
908
        super().__init__('__not_yet_ready__', '')
909
        self.errors = []               # type: List[Error]
910
911
        self.error_nodes = dict()      # type: Dict[int, List[Error]]  # id(node) -> error list
        self.error_positions = dict()  # type: Dict[int, Set[int]]  # pos -> set of id(node)
di68kap's avatar
di68kap committed
912
        self.error_flag = 0
eckhart's avatar
eckhart committed
913
914
        if node is not None:
            self.swallow(node)
915
        # customization for XML-Representation
eckhart's avatar
eckhart committed
916
        self.inline_tags = set()  # type: Set[str]
917
918
        self.omit_tags = set()    # type: Set[str]
        self.empty_tags = set()   # type: Set[str]
di68kap's avatar
di68kap committed
919

920
921
922
923
924
925
926
927
928
    def __deepcopy__(self, memodict={}):
        duplicate = self.__class__(None)
        if self.children:
            duplicate.children = copy.deepcopy(self.children)
            duplicate._result = duplicate.children
        else:
            duplicate.children = NoChildren
            duplicate._result = self._result
        duplicate._pos = self._pos
eckhart's avatar
eckhart committed
929
        if self.attr_active():
di68kap's avatar
di68kap committed
930
            duplicate.attr.update(copy.deepcopy(self._xml_attr))
931
            # duplicate._xml_attr = copy.deepcopy(self._xml_attr)  # this is blocked by cython
932
        duplicate.errors = copy.copy(self.errors)
933
934
        duplicate.error_nodes = copy.copy(self.error_nodes)
        duplicate.error_positions = copy.deepcopy(self.error_positions)
935
936
937
938
        duplicate.error_flag = self.error_flag
        duplicate.inline_tags = self.inline_tags
        duplicate.omit_tags = self.omit_tags
        duplicate.empty_tags = self.empty_tags
939
        duplicate.tag_name = self.tag_name
940
941
942
        return duplicate


943
    def swallow(self, node: Node) -> 'RootNode':
944
945
        """
        Put `self` in the place of `node` by copying all its data.
946
947
948
949
950
951
952
953
954
        Returns self.

        This is done by the parse.Grammar object after
        parsing has finished, so that the Grammar object always
        returns a syntax tree rooted in a RootNode object.

        It is possible to add errors to a RootNode object, before it
        has actually swallowed the root of the syntax tree.
        """
di68kap's avatar
di68kap committed
955
956
957
        self._result = node._result
        self.children = node.children
        self._pos = node._pos
958
        self.tag_name = node.tag_name
eckhart's avatar
eckhart committed
959
        if node.attr_active():
960
            self._xml_attr = node._xml_attr
961
        # self._content = node._content
962
963
        if id(node) in self.error_nodes:
            self.error_nodes[id(self)] = self.error_nodes[id(node)]
964
        return self
di68kap's avatar
di68kap committed
965

966
    def add_error(self, node: Optional[Node], error: Error) -> 'RootNode':
967
968
969
        """
        Adds an Error object to the tree, locating it at a specific node.
        """
970
971
        if not node:
            node = Node(ZOMBIE_TAG, '').with_pos(error.pos)
972
        assert node.pos == error.pos or isinstance(node, FrozenNode)
973
        self.error_nodes.setdefault(id(node), []).append(error)
974
        self.error_positions.setdefault(error.pos, set()).add(id(node))
975
976
        self.errors.append(error)
        self.error_flag = max(self.error_flag, error.code)
di68kap's avatar
di68kap committed
977
978
        return self

eckhart's avatar
eckhart committed
979
    def new_error(self,
980
                  node: Node,
di68kap's avatar
di68kap committed
981
                  message: str,
eckhart's avatar
eckhart committed
982
                  code: ErrorCode = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
983
        """
eckhart's avatar
eckhart committed
984
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
985
986
987
988
989
        Parameters:
            pos(int):     The position of the error in the source text
            message(str): A string with the error message.abs
            code(int):    An error code to identify the kind of error
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
990
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
991
992
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
993

Eckhart Arnold's avatar