syntaxtree.py 47.9 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
from typing import Callable, cast, Iterator, List, AbstractSet, Set, Union, Tuple, Optional, Dict
29

eckhart's avatar
eckhart committed
30
from DHParser.error import Error, ErrorCode, linebreaks, line_col
31
from DHParser.stringview import StringView
32
from DHParser.toolkit import get_config_value, re
di68kap's avatar
di68kap committed
33

34

di68kap's avatar
di68kap committed
35
__all__ = ('WHITESPACE_PTYPE',
36
           'TOKEN_PTYPE',
37
38
           'ZOMBIE_TAG',
           'PLACEHOLDER',
eckhart's avatar
eckhart committed
39
40
41
           'ResultType',
           'StrictResultType',
           'ChildrenType',
42
           'Node',
43
           'serialize',
44
           'FrozenNode',
45
           'tree_sanity_check',
di68kap's avatar
di68kap committed
46
           'RootNode',
47
           'parse_sxpr',
48
           'parse_xml',
eckhart's avatar
eckhart committed
49
           'parse_tree',
50
51
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
52
53


54
55
56
57
58
59
60
61
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


WHITESPACE_PTYPE = ':Whitespace'
62
TOKEN_PTYPE = ':Token'
63

64
ZOMBIE_TAG = "__ZOMBIE__"
65

66
67
68
69
70
71
72
#######################################################################
#
# syntaxtree nodes
#
#######################################################################


73
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
74
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
75
StrictResultType = Union[ChildrenType, StringView, str]
76
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
77
78


79
def flatten_sxpr(sxpr: str, threshold: int = -1) -> str:
80
81
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
82
83
    whitespace.

84
85
86
87
88
89
90
    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
91
    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
92
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
93
94
    '(a (b c))'
    """
95

96
97
98
99
    flat = re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
    if threshold >= 0 and len(flat) > threshold:
        return sxpr.strip()
    return flat
Eckhart Arnold's avatar
Eckhart Arnold committed
100
101


102
def flatten_xml(xml: str) -> str:
103
104
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
105
    i.e. only whitespace within leaf-nodes is preserved.
106
107
    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`.
108
    """
109

110
111
112
113
    # works only with regex
    # return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?<=</\w+>)\s+', '', xml))
    def tag_only(m):
        return m.groupdict()['closing_tag']
114
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
115
116


eckhart's avatar
eckhart committed
117
RX_AMP = re.compile(r'&(?!\w+;)')
118
119


Eckhart Arnold's avatar
Eckhart Arnold committed
120
class Node:  # (collections.abc.Sized): Base class omitted for cython-compatibility
121
122
123
    """
    Represents a node in the concrete or abstract syntax tree.

124
125
    TODO: Add some documentation and doc-tests here...

126
    Attributes:
127
128
        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
129

130
131
132
        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.
133

134
135
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
136

eckhart's avatar
eckhart committed
137
138
139
140
        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
141
        parser (Parser):  The parser which generated this node.
142
143
            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
144
145
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
146
            example by calling ``isinstance(node.parer, ...)``.
147

148
149
150
        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
151
            the length before AST-transformation and will never change
152
            through AST-transformation. READ ONLY!
153

154
155
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
156
            The value of ``pos`` is -1 meaning invalid by default.
157
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
158
            nodes relative to this value.
159
160

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
161
            pos value of the root node should be set to 0 right
162
163
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
164
            Other than that, this value should be considered READ ONLY.
165
166
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
167

168
169
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
170
171
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
172
    """
173

174
    __slots__ = '_result', 'children', '_len', '_pos', 'tag_name', '_xml_attr'
175

176
    def __init__(self, tag_name: str, result: ResultType, leafhint: bool = False) -> None:
177
178
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
179
180
        that generated the node and the parser's result.
        """
eckhart's avatar
eckhart committed
181
        self._pos = -1                  # type: int
182
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
183
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
184
        if leafhint:
eckhart's avatar
eckhart committed
185
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
186
            # self._content = None        # type: Optional[str]
187
            self.children = NoChildren  # type: ChildrenType
188
            self._len = -1              # type: int  # lazy evaluation
189
190
        else:
            self.result = result
191
        self.tag_name = tag_name        # type: str
192

eckhart's avatar
eckhart committed
193
194
    def __deepcopy__(self, memo):
        if self.children:
195
            duplicate = self.__class__(self.tag_name, copy.deepcopy(self.children), False)
eckhart's avatar
eckhart committed
196
        else:
197
            duplicate = self.__class__(self.tag_name, self.result, True)
eckhart's avatar
eckhart committed
198
199
        duplicate._pos = self._pos
        duplicate._len = self._len
eckhart's avatar
eckhart committed
200
        if self.attr_active():
di68kap's avatar
di68kap committed
201
202
            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
203
        return duplicate
Eckhart Arnold's avatar
Eckhart Arnold committed
204

205
    def __str__(self):
206
207
        if isinstance(self, RootNode):
            root = cast(RootNode, self)
208
            errors = root.errors_sorted
209
210
            if errors:
                e_pos = errors[0].pos
Eckhart Arnold's avatar
Eckhart Arnold committed
211
212
                content = self.content
                return content[:e_pos] + \
213
                   ' <<< Error on "%s" | %s >>> ' % \
Eckhart Arnold's avatar
Eckhart Arnold committed
214
                   (content[e_pos - self.pos:], '; '.join(e.message for e in errors))
215
        return self.content
Eckhart Arnold's avatar
Eckhart Arnold committed
216

217
    def __repr__(self):
218
        # mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
219
220
        # name, ptype = (self._tag_name.split(':') + [''])[:2]
        # parg = "MockParser({name}, {ptype})".format(name=name, ptype=ptype)
221
        rarg = str(self) if not self.children else \
222
            "(" + ", ".join(child.__repr__() for child in self.children) + ")"
223
        return "Node(%s, %s)" % (self.tag_name, rarg)
224

Eckhart Arnold's avatar
Eckhart Arnold committed
225

226
    def __len__(self):
227
        if self._len < 0:
di68kap's avatar
di68kap committed
228
            self._len = sum(len(child) for child in self.children) \
229
                if self.children else len(self._result)
230
231
232
233
234
235
236
237
        return self._len


    def __bool__(self):
        # A node that is not None is always True, even if it's empty
        return True


238
239
240
241
242
243
244
245
246
    # can lead to obscure mistakes, where default object comparison behaviour is expected
    # def __eq__(self, other):
    #     """
    #     Equality of nodes: Two nodes are considered as equal, if their tag
    #     name is the same, if their results are equal and if their attributes
    #     and attribute values are the same.
    #     """
    #     return self.tag_name == other.tag_name and self.result == other.result \
    #         and self.compare_attr(other)
247

Eckhart Arnold's avatar
Eckhart Arnold committed
248

249
    def __hash__(self):
250
        return hash(self.tag_name)
251

Eckhart Arnold's avatar
Eckhart Arnold committed
252

253
254
255
    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
256
        an integer or the first child node with the given tag name. Examples::
257

258
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
259
260
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
261
262
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
263
264
265
266

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
267
        Returns:
268
269
            Node: All nodes which have a given tag name.
        """
270
271
272
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
273
            else:
274
275
276
277
278
                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!')
279
280
281
282


    def __contains__(self, tag_name: str) -> bool:
        """
283
        Returns true if a child with the given tag name exists.
284
        Args:
285
286
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
287
288
289
290
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
291
292
293
294
295
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
296
            return False
297
        raise ValueError('Leave node cannot contain other nodes')
298
299


300
    def equals(self, other: 'Node') -> bool:
301
        """
302
303
304
305
306
307
        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`.
308
309
310
311
312
313
314
315
316
317
        """
        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


318
319
320
321
322
323
324
325
326
327
328
329
330
331
    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


332
    def is_anonymous(self) -> bool:
Eckhart Arnold's avatar
Eckhart Arnold committed
333
        return not self.tag_name or self.tag_name[0] == ':'
334

Eckhart Arnold's avatar
Eckhart Arnold committed
335

336
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
337
    def result(self) -> StrictResultType:
338
339
340
341
342
        """
        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.
        """
343
344
        return self._result

345

346
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
347
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
348
        # # made obsolete by static type checking with mypy
Eckhart Arnold's avatar
Eckhart Arnold committed
349
350
351
352
        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
353
354
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
355
        self._len = -1        # lazy evaluation
356
        # self._content = None
357
358
359
360
361
362
363
364
365
        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
366
                self._result = result  # cast(StrictResultType, result)
367

368
369
370
371
372
373
374
375
376
377
378
379
    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]
380
381

    @property
eckhart's avatar
eckhart committed
382
    def content(self) -> str:
383
        """
384
385
386
        Returns content as string. If the node has child-nodes, the
        string content of the child-nodes is recursively read and then
        concatenated.
387
        """
388
389
390
391
392
393
394
395
        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
396
397
        # return "".join(child.content for child in self.children) if self.children \
        #     else str(self._result)
di68kap's avatar
di68kap committed
398
399


400
401
402
403
404
405
    @property
    def structure(self) -> str:
        """
        Return structure (and content) as S-expression on a single line
        without any line breaks.
        """
406
        return flatten_sxpr(self.as_sxpr())
407
408


409
    @property
410
    def pos(self) -> int:
eckhart's avatar
eckhart committed
411
412
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
Eckhart Arnold's avatar
Eckhart Arnold committed
413
            raise AssertionError("Position value not initialized! Use Node.with_pos()")
414
415
        return self._pos

eckhart's avatar
eckhart committed
416

Eckhart Arnold's avatar
Eckhart Arnold committed
417
    def with_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
418
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
419
        Initialize position value. Usually, the parser guard
eckhart's avatar
eckhart committed
420
421
        (`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
422
        when Nodes are created outside the reach of the parser
eckhart's avatar
eckhart committed
423
        guard, their document-position must be assigned manually.
Eckhart Arnold's avatar
Eckhart Arnold committed
424
425
        Position values of the child nodes are assigned recursively, too.
        Returns the node itself for convenience.
eckhart's avatar
eckhart committed
426
        """
427
428
429
430
        # 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
431
432
433
434
435
436
437
438
439
440
        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
441
442
        return self

443

444
445
446
447
448
449
450
451
452
453
454
455
456
    @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
457
458
459
460
461
462
463
464
465
466
467
468
469
    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


470
    def compare_attr(self, other: 'Node') -> bool:
471
        """
472
473
        Returns True, if `self` and `other` have the same attributes with the
        same attribute values.
474
        """
475
476
477
478
479
480
481
482
483
        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
484

Eckhart Arnold's avatar
Eckhart Arnold committed
485

486
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
487
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
488
489
490
491
492
493
494
495
        """
        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.

496
        Args:
497
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
498
            open_fn:   (Node->str) A function that returns an opening
499
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
500
            close_fn:  (Node->str) A function that returns a closeF
501
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
502
            data_fn:   (str->str) A function that filters the data string
503
504
505
506
507
508
                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
509
510
        head = open_fn(self)
        tail = close_fn(self)
511
512

        if not self.result:
513
            return head.rstrip() + tail.lstrip()
514

eckhart's avatar
eckhart committed
515
        tail = tail.lstrip(None if density & 2 else '')
516

517
        inline = inline or inline_fn(self)
518
519
520
521
522
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
523
524
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
525

526
527
        if self.children:
            content = []
528
            for child in self.children:
529
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
530
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
531
                if subtree:
eckhart's avatar
eckhart committed
532
533
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
534
            return head + usetab + (sep + usetab).join(content) + tail
535

eckhart's avatar
eckhart committed
536
        res = self.content
eckhart's avatar
eckhart committed
537
        if not inline and not head:
eckhart's avatar
eckhart committed
538
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
539
            res = res.strip()
540
541
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
542
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
543
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
544
        else:
545
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
546

Eckhart Arnold's avatar
Eckhart Arnold committed
547

548
549
    def as_sxpr(self, src: str = None,
                indentation: int = 2,
550
551
                compact: bool = False,
                flatten_threshold: int = 0) -> str:
552
        """
553
554
        Returns content as S-expression, i.e. in lisp-like form. If this
        method is callad on a RootNode-object,
555

556
        Args:
557
558
559
            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.
560
561
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
562
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
563
                tree structure.
564
565
566
            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.
567
568
        """

eckhart's avatar
eckhart committed
569
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
570
        lbreaks = linebreaks(src) if src else []  # type: List[int]
571
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[RootNode]
572

573
        def opening(node: Node) -> str:
eckhart's avatar
eckhart committed
574
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
575
            txt = [left_bracket, node.tag_name]
576
            # s += " '(pos %i)" % node.add_pos
577
            # txt.append(str(id(node)))  # for debugging
eckhart's avatar
eckhart committed
578
            if node.attr_active():
579
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
580
            if src:
581
582
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
583
584
            if root and id(node) in root.error_nodes:
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in root.get_errors(node)))
585
            return "".join(txt) + '\n'
586

587
        def closing(node: Node) -> str:
eckhart's avatar
eckhart committed
588
589
            """Returns the closing string for the representation of `node`."""
            return right_bracket
590

591
        def pretty(strg: str) -> str:
eckhart's avatar
eckhart committed
592
593
594
595
            """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'\"')
596

597
598
        sxpr = self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
        return flatten_sxpr(sxpr, flatten_threshold)
Eckhart Arnold's avatar
Eckhart Arnold committed
599

eckhart's avatar
eckhart committed
600

601
602
    def as_xml(self, src: str = None,
               indentation: int = 2,
eckhart's avatar
eckhart committed
603
604
605
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
606
607
608
        """
        Returns content as XML-tree.

609
        Args:
610
611
612
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
613
            indentation: The number of whitespaces for indentation
614
615
616
            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
617
                neither the opening tag nor its attr nor the closing tag. This
618
619
620
                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.
621
622
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
623
        """
624
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[RootNode]
625

626
        def opening(node: Node) -> str:
627
628
629
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
630
            txt = ['<', node.tag_name]
eckhart's avatar
eckhart committed
631
            has_reserved_attrs = node.attr_active() \
eckhart's avatar
eckhart committed
632
                and any(r in node.attr for r in {'err', 'line', 'col'})
eckhart's avatar
eckhart committed
633
            if node.attr_active():
634
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
635
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
636
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
637
            if root and id(node) in root.error_nodes and not has_reserved_attrs:
eckhart's avatar
eckhart committed
638
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
eckhart's avatar
eckhart committed
639
                                                 for err in root.get_errors(node)))
640
641
642
            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
643
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
644
645
646
            else:
                ending = ">\n"
            return "".join(txt + [ending])
647

648
        def closing(node: Node):
649
            """Returns the closing string for the representation of `node`."""
650
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
651
                return ''
652
            return ('\n</') + node.tag_name + '>'
653

654
655
656
657
658
659
        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

660
        def inlining(node: Node):
661
662
663
664
            """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
665
            return node.tag_name in inline_tags \
eckhart's avatar
eckhart committed
666
                or (node.attr_active()
eckhart's avatar
eckhart committed
667
                    and node.attr.get('xml:space', 'default') == 'preserve')
668

669
        line_breaks = linebreaks(src) if src else []
670
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
671
                               density=1, inline_fn=inlining)
672

Eckhart Arnold's avatar
Eckhart Arnold committed
673

eckhart's avatar
eckhart committed
674
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
675
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
676
        """
677
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
678

679
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
680
        given `match_function` evaluates to True. The tree is
681
682
683
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
684

685
686
687
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
688
689
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
690
691
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
692
        Yields:
693
            Node: All nodes of the tree for which
694
695
            ``match_function(node)`` returns True
        """
696
        if include_root and match_function(self):
697
            yield self
698
699
700
        child_iterator = reversed(self.children) if reverse else self.children
        for child in child_iterator:
            for node in child.select(match_function, True, reverse):
eckhart's avatar
eckhart committed
701
                yield node
702

Eckhart Arnold's avatar
Eckhart Arnold committed
703

eckhart's avatar
eckhart committed
704
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
eckhart's avatar
eckhart committed
705
                      include_root: bool = False) -> Iterator['Node']:
706
        """
707
708
709
710
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
711

712
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
713
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
714
            ['(X (c "d"))', '(X "F")']
715
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
716
            ['(b "X")', '(X (c "d"))', '(X "F")']
717
            >>> any(tree.select_by_tag('a', False))
718
            False
719
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
720
            ['(a (b "X") (X (c "d")) (e (X "F")))']
721
722
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
723
724

        Args:
725
726
727
728
            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.
729
730
731
        Yields:
            Node: All nodes which have a given tag name.
        """
732
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
733
            tag_names = frozenset({tag_names})
734
        return self.select(lambda node: node.tag_name in tag_names, include_root)
735
736


737
738
739
740
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

741
        This function is mostly just syntactic sugar for
742
743
744
745
746
747
748
749
750
751
        ``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


752
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
753
754
755
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
756
757
758
        return sum(child.tree_size() for child in self.children) + 1


759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
def serialize(node: Node, how: str='default') -> str:
    """
    Serializes the tree starting with `node` either as S-expression, XML
    or in compact form. Possible values for `how` are 'S-expression',
    'XML', 'compact' accordingly, or 'AST', 'CST', 'default' in which case
    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':
777
        return node.as_sxpr(flatten_threshold=get_config_value('flatten_sxpr_threshold'))
778
779
780
781
782
783
784
785
    elif switch == 'xml':
        return node.as_xml()
    elif switch == 'compact':
        return node.as_sxpr(compact=True)
    else:
        raise ValueError('Unknown serialization %s, %s' % (how, switch))


786
class FrozenNode(Node):
787
788
789
790
791
792
793
794
795
796
    """
    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()`.
    """
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813

    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
814
815
816
817
    @property
    def attr(self):
        raise AssertionError("Attributes cannot be accessed on a frozen node")

Eckhart Arnold's avatar
Eckhart Arnold committed
818
    def with_pos(self, pos: int) -> 'Node':
819
820
821
        pass


822
PLACEHOLDER = FrozenNode('__PLACEHOLDER__', '')
di68kap's avatar
di68kap committed
823
824


825
def tree_sanity_check(tree: Node) -> bool:
826
827
828
829
830
831
832
833
834
    """
    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.
    """
835
    node_set = set()  # type: Set[Node]
836
    for node in tree.select(lambda nd: True, include_root=True):
837
        if node in node_set or isinstance(Node, FrozenNode):
838
839
840
841
842
            return False
        node_set.add(node)
    return True


di68kap's avatar
di68kap committed
843
class RootNode(Node):
844
845
846
847
848
849
850
851
852
853
854
855
    """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
856
857
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
858

859
860
861
862
863
864
        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

865
866
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
867
868
869
870
871
872

        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
873
    """
874

eckhart's avatar
eckhart committed
875
    def __init__(self, node: Optional[Node] = None):
876
        super().__init__('__not_yet_ready__', '')
877
        self.errors = []               # type: List[Error]
878
879
        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
880
        self.error_flag = 0
eckhart's avatar
eckhart committed
881
882
        if node is not None:
            self.swallow(node)
883
        # customization for XML-Representation
eckhart's avatar
eckhart committed
884
        self.inline_tags = set()  # type: Set[str]
885
886
        self.omit_tags = set()    # type: Set[str]
        self.empty_tags = set()   # type: Set[str]
di68kap's avatar
di68kap committed
887

888
889
890
891
892
893
894
895
896
897
    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
        duplicate._len = self._len
eckhart's avatar
eckhart committed
898
        if self.attr_active():
di68kap's avatar
di68kap committed
899
            duplicate.attr.update(copy.deepcopy(self._xml_attr))
900
            # duplicate._xml_attr = copy.deepcopy(self._xml_attr)  # this is blocked by cython
901
        duplicate.errors = copy.copy(self.errors)
902
903
        duplicate.error_nodes = copy.copy(self.error_nodes)
        duplicate.error_positions = copy.deepcopy(self.error_positions)
904
905
906
907
        duplicate.error_flag = self.error_flag
        duplicate.inline_tags = self.inline_tags
        duplicate.omit_tags = self.omit_tags
        duplicate.empty_tags = self.empty_tags
908
        duplicate.tag_name = self.tag_name
909
910
911
        return duplicate


912
    def swallow(self, node: Node) -> 'RootNode':
913
914
        """
        Put `self` in the place of `node` by copying all its data.
915
916
917
918
919
920
921
922
923
        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
924
925
926
927
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
928
        self.tag_name = node.tag_name
eckhart's avatar
eckhart committed
929
        if node.attr_active():
930
            self._xml_attr = node._xml_attr
931
        # self._content = node._content
932
933
        if id(node) in self.error_nodes:
            self.error_nodes[id(self)] = self.error_nodes[id(node)]
934
        return self
di68kap's avatar
di68kap committed
935

936
    def add_error(self, node: Optional[Node], error: Error) -> 'RootNode':
937
938
939
        """
        Adds an Error object to the tree, locating it at a specific node.
        """
940
941
        if not node:
            node =