syntaxtree.py 38.9 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

eckhart's avatar
eckhart committed
29
from DHParser.error import Error, ErrorCode, linebreaks, line_col
30
from DHParser.stringview import StringView
di68kap's avatar
di68kap committed
31
32
from DHParser.toolkit import re
from typing import Callable, cast, Iterator, List, AbstractSet, Set, Union, Tuple, Optional
di68kap's avatar
di68kap committed
33

34

di68kap's avatar
di68kap committed
35
__all__ = ('WHITESPACE_PTYPE',
36
           'TOKEN_PTYPE',
37
           'ZOMBIE',
eckhart's avatar
eckhart committed
38
           'ZOMBIE_NODE',
eckhart's avatar
eckhart committed
39
40
41
           'ResultType',
           'StrictResultType',
           'ChildrenType',
42
           'Node',
di68kap's avatar
di68kap committed
43
           'RootNode',
eckhart's avatar
eckhart committed
44
           'ZOMBIE_ROOTNODE',
45
           'parse_sxpr',
46
47
48
           'parse_xml',
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
49
50


51
52
53
54
55
56
57
58
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


WHITESPACE_PTYPE = ':Whitespace'
59
TOKEN_PTYPE = ':Token'
60

61
62
ZOMBIE = "__ZOMBIE__"

63
64
65
66
67
68
69
#######################################################################
#
# syntaxtree nodes
#
#######################################################################


70
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
71
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
72
StrictResultType = Union[ChildrenType, StringView, str]
73
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
74
75


Eckhart Arnold's avatar
Eckhart Arnold committed
76
def flatten_sxpr(sxpr: str) -> str:
77
78
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
79
80
81
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
82
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
83
84
    '(a (b c))'
    """
85

86
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
87
88


89
def flatten_xml(xml: str) -> str:
90
91
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
92
    i.e. only whitespace within leaf-nodes is preserved.
93
94
    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`.
95
    """
96

97
98
99
100
    # 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']
101
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
102
103


eckhart's avatar
eckhart committed
104
RX_AMP = re.compile(r'&(?!\w+;)')
105
106


Eckhart Arnold's avatar
Eckhart Arnold committed
107
class Node:  # (collections.abc.Sized): Base class omitted for cython-compatibility
108
109
110
    """
    Represents a node in the concrete or abstract syntax tree.

111
112
    TODO: Add some documentation and doc-tests here...

113
    Attributes:
114
115
        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
116

117
118
119
        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.
120

121
122
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
123

eckhart's avatar
eckhart committed
124
125
126
127
        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
128
        parser (Parser):  The parser which generated this node.
129
130
            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
131
132
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
133
            example by calling ``isinstance(node.parer, ...)``.
134

135
136
137
        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
138
            the length before AST-transformation and will never change
139
            through AST-transformation. READ ONLY!
140

141
142
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
143
            The value of ``pos`` is -1 meaning invalid by default.
144
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
145
            nodes relative to this value.
146
147

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
148
            pos value of the root node should be set to 0 right
149
150
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
151
            Other than that, this value should be considered READ ONLY.
152
153
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
154

155
156
        errors (list):  A list of all errors that occured on this node.

157
158
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
159
160
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
161
    """
162

163
    __slots__ = '_result', 'children', '_len', '_pos', '_tag_name', 'errors', '_xml_attr', '_content'
164

165
    def __init__(self, tag_name: Optional[str], result: ResultType, leafhint: bool = False) -> None:
166
167
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
168
169
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
170
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
171
        self._pos = -1                  # type: int
172
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
173
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
174
        if leafhint:
eckhart's avatar
eckhart committed
175
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
eckhart's avatar
eckhart committed
176
            self._content = None        # type: Optional[str]
177
            self.children = NoChildren  # type: ChildrenType
178
            self._len = -1              # type: int  # lazy evaluation
179
180
        else:
            self.result = result
181
182
        assert tag_name is None or isinstance(tag_name, str)   # TODO: Delete this line
        self._tag_name = tag_name if tag_name else ZOMBIE
183
184
185
186
        # if parser is None:
        #     self._tag_name = ZOMBIE
        # else:
        #     self._tag_name = parser.name or parser.ptype
187

eckhart's avatar
eckhart committed
188
189
    def __deepcopy__(self, memo):
        if self.children:
190
            duplicate = self.__class__(self._tag_name, copy.deepcopy(self.children), False)
eckhart's avatar
eckhart committed
191
        else:
192
            duplicate = self.__class__(self._tag_name, self.result, True)
eckhart's avatar
eckhart committed
193
194
195
196
197
198
        duplicate.errors = copy.deepcopy(self.errors) if self.errors else []
        duplicate._pos = self._pos
        duplicate._len = self._len
        if hasattr(self, '_xml_attr'):
            duplicate._xml_attr = copy.deepcopy(self._xml_attr)
        return duplicate
Eckhart Arnold's avatar
Eckhart Arnold committed
199

200
    def __str__(self):
201
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
202
        if self.errors:
203
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
204
                   (s, '; '.join(e.message for e in self.errors))
205
        return s
206

Eckhart Arnold's avatar
Eckhart Arnold committed
207

208
    def __repr__(self):
209
        # mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
210
211
        # name, ptype = (self._tag_name.split(':') + [''])[:2]
        # parg = "MockParser({name}, {ptype})".format(name=name, ptype=ptype)
212
        rarg = str(self) if not self.children else \
eckhart's avatar
eckhart committed
213
            "(" + ", ".join(repr(child) for child in self.children) + ")"
214
        return "Node(%s, %s)" % (self._tag_name, rarg)
215

Eckhart Arnold's avatar
Eckhart Arnold committed
216

217
    def __len__(self):
218
        if self._len < 0:
di68kap's avatar
di68kap committed
219
            self._len = sum(len(child) for child in self.children) \
220
                if self.children else len(self._result)
221
222
223
224
225
226
227
228
        return self._len


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


229
    def __eq__(self, other):
230
231
232
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
233
234
235
236

        Note: It is not required that two nodes have the same errors attached.
        In case you need to check for error equality as well, compare a
        serialization that includes error messages, as_sxpr() will do!
237
        """
238
        return self.tag_name == other.tag_name and self.result == other.result
239

Eckhart Arnold's avatar
Eckhart Arnold committed
240

241
    def __hash__(self):
242
        return hash(self.tag_name)
243

Eckhart Arnold's avatar
Eckhart Arnold committed
244

245
246
247
    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
248
        an integer or the first child node with the given tag name. Examples::
249

250
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
251
252
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
253
254
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
255
256
257
258

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
259
        Returns:
260
261
            Node: All nodes which have a given tag name.
        """
262
263
264
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
265
            else:
266
267
268
269
270
                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!')
271
272
273
274


    def __contains__(self, tag_name: str) -> bool:
        """
275
        Returns true if a child with the given tag name exists.
276
        Args:
277
278
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
279
280
281
282
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
283
284
285
286
287
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
288
            return False
289
        raise ValueError('Leave node cannot contain other nodes')
290
291


292
293
294
295
296
297
298
299
300
301
302
303
304
305
    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


306
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
307
    def tag_name(self) -> str:
308
309
310
311
312
313
        """
        Returns the tage name of Node, i.e. the name for XML or
        S-expression representation. By default the tag name is the
        name of the node's parser or, if the node's parser is unnamed, the
        node's parser's `ptype`.
        """
314
315
316
317
318
319
320
321
322
323
324
325
        return self._tag_name

    @tag_name.setter
    def tag_name(self, tag_name: str):
        assert tag_name
        self._tag_name = tag_name


    def is_anonymous(self):
        return self._tag_name[0] == ':'


326
327
328
329
    # @property
    # def parser(self) -> MockParser:
    #     name, ptype = (self.tag_name.split(':') + [''])[:2]
    #     return MockParser(name, ':' + ptype)
330

Eckhart Arnold's avatar
Eckhart Arnold committed
331

332
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
333
    def result(self) -> StrictResultType:
334
335
336
337
338
        """
        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.
        """
339
340
        return self._result

341

342
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
343
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
344
        # # made obsolete by static type checking with mypy
345
346
347
        # assert ((isinstance(result, tuple) and all(isinstance(child, Node) for child in result))
        #         or isinstance(result, Node)
        #         or isinstance(result, str)), str(result)
Eckhart Arnold's avatar
Eckhart Arnold committed
348
349
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
350
351
        self._len = -1        # lazy evaluation
        self._content = None
352
353
354
355
356
357
358
359
360
        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
361
                self._result = result  # cast(StrictResultType, result)
362

363
364

    @property
eckhart's avatar
eckhart committed
365
    def content(self) -> str:
366
        """
di68kap's avatar
di68kap committed
367
368
369
        Returns content as string, omitting error messages. If the node has
        child-nodes, the string content of the child-nodes is recursively read
        and then concatenated.
370
        """
371
372
373
374
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
375
                # self._content = self._result
376
377
378
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
379
380


di68kap's avatar
di68kap committed
381
382
383
384
385
    @content.setter
    def content(self, content: str):
        self.result = content


386
387
388
389
390
391
392
393
394
    @property
    def structure(self) -> str:
        """
        Return structure (and content) as S-expression on a single line
        without any line breaks.
        """
        return flatten_sxpr(self.as_sxpr(showerrors=False))


395
    @property
396
    def pos(self) -> int:
eckhart's avatar
eckhart committed
397
398
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
399
            raise AssertionError("Position value not initialized!")
400
401
        return self._pos

eckhart's avatar
eckhart committed
402

eckhart's avatar
eckhart committed
403
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
404
405
406
407
408
409
410
411
412
        """
        (Re-)initialize position value. Usually, the parser guard
        (`parsers.add_parser_guard()`) takes care of assigning the
        position in the document to newly created nodes. However,
        where Nodes are created outside the reach of the parser
        guard, their document-position must be assigned manually.
        This function recursively reassigns the position values
        of the child nodes, too.
        """
eckhart's avatar
eckhart committed
413
414
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
415
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
416
        offset = self.pos
417
        for child in self.children:
eckhart's avatar
eckhart committed
418
            child.init_pos(offset)
eckhart's avatar
eckhart committed
419
420
421
            offset = child.pos + len(child)
        return self

422

423
    @property
424
    def attr(self):
425
        """
426
        Returns a dictionary of XML-attr attached to the node.
427
        """
428
429
430
431
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
432

433
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
434
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
435
436
437
438
439
440
441
442
        """
        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.

443
        Args:
444
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
445
            open_fn:   (Node->str) A function that returns an opening
446
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
447
            close_fn:  (Node->str) A function that returns a closeF
448
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
449
            data_fn:   (str->str) A function that filters the data string
450
451
452
453
454
455
                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
456
457
        head = open_fn(self)
        tail = close_fn(self)
458
459

        if not self.result:
460
            return head.rstrip() + tail.lstrip()
461

eckhart's avatar
eckhart committed
462
        tail = tail.lstrip(None if density & 2 else '')
463

464
        inline = inline or inline_fn(self)
465
466
467
468
469
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
470
471
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
472

473
474
        if self.children:
            content = []
475
            for child in self.children:
476
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
477
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
478
                if subtree:
eckhart's avatar
eckhart committed
479
480
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
481
            return head + usetab + (sep + usetab).join(content) + tail
482

eckhart's avatar
eckhart committed
483
        res = self.content
eckhart's avatar
eckhart committed
484
        if not inline and not head:
eckhart's avatar
eckhart committed
485
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
486
            res = res.strip()
487
488
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
489
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
490
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
491
        else:
492
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
493

Eckhart Arnold's avatar
Eckhart Arnold committed
494

495
496
497
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
498
                compact: bool = False) -> str:
499
500
501
        """
        Returns content as S-expression, i.e. in lisp-like form.

502
        Args:
503
504
505
            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.
506
507
508
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
509
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
510
                tree structure.
511
512
        """

eckhart's avatar
eckhart committed
513
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
514
        lbreaks = linebreaks(src) if src else []  # type: List[int]
515

516
        def opening(node) -> str:
eckhart's avatar
eckhart committed
517
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
518
            txt = [left_bracket, node.tag_name]
519
            # s += " '(pos %i)" % node.add_pos
520
            if hasattr(node, '_xml_attr'):
521
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
522
            if src:
523
524
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
525
            if showerrors and node.errors:
526
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
527
            return "".join(txt) + '\n'
528

eckhart's avatar
eckhart committed
529
530
531
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
532

eckhart's avatar
eckhart committed
533
534
535
536
537
        def pretty(strg):
            """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'\"')
538

539
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
540

eckhart's avatar
eckhart committed
541

542
543
544
    def as_xml(self, src: str = None,
               showerrors: bool = True,
               indentation: int = 2,
eckhart's avatar
eckhart committed
545
546
547
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
548
549
550
        """
        Returns content as XML-tree.

551
        Args:
552
553
554
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
555
556
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
557
558
559
            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
560
                neither the opening tag nor its attr nor the closing tag. This
561
562
563
                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.
564
565
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
566
567
        """

568
        def opening(node) -> str:
569
570
571
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
572
            txt = ['<', node.tag_name]
573
            has_reserved_attrs = hasattr(node, '_xml_attr') \
eckhart's avatar
eckhart committed
574
                and any(r in node.attr for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
575
            if hasattr(node, '_xml_attr'):
576
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
577
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
578
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
579
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
580
581
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
582
583
584
            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
585
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
586
587
588
            else:
                ending = ">\n"
            return "".join(txt + [ending])
589
590

        def closing(node):
591
            """Returns the closing string for the representation of `node`."""
592
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
593
                return ''
594
            return ('\n</') + node.tag_name + '>'
595

596
597
598
599
600
601
602
        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


603
604
605
606
607
        def inlining(node):
            """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
608
609
610
            return node.tag_name in inline_tags \
                or (hasattr(node, '_xml_attr')
                    and node.attr.get('xml:space', 'default') == 'preserve')
611

612
        line_breaks = linebreaks(src) if src else []
613
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
614
                               density=1, inline_fn=inlining)
615

Eckhart Arnold's avatar
Eckhart Arnold committed
616

eckhart's avatar
eckhart committed
617
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
618
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
619
        """
620
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
621

622
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
623
        given `match_function` evaluates to True. The tree is
624
625
626
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
627

628
629
630
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
631
632
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
633
634
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
635
        Yields:
636
            Node: All nodes of the tree for which
637
638
            ``match_function(node)`` returns True
        """
639
        if include_root and match_function(self):
640
            yield self
641
642
643
        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
644
                yield node
645

Eckhart Arnold's avatar
Eckhart Arnold committed
646

eckhart's avatar
eckhart committed
647
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
eckhart's avatar
eckhart committed
648
                      include_root: bool = False) -> Iterator['Node']:
649
        """
650
651
652
653
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
654

655
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
656
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
657
            ['(X (c "d"))', '(X "F")']
658
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
659
            ['(b "X")', '(X (c "d"))', '(X "F")']
660
            >>> any(tree.select_by_tag('a', False))
661
            False
662
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
663
            ['(a (b "X") (X (c "d")) (e (X "F")))']
664
665
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
666
667

        Args:
668
669
670
671
            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.
672
673
674
        Yields:
            Node: All nodes which have a given tag name.
        """
675
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
676
            tag_names = frozenset({tag_names})
677
        return self.select(lambda node: node.tag_name in tag_names, include_root)
678
679


680
681
682
683
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

684
        This function is mostly just syntactic sugar for
685
686
687
688
689
690
691
692
693
694
        ``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


695
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
696
697
698
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
699
700
701
        return sum(child.tree_size() for child in self.children) + 1


702
ZOMBIE_NODE = Node(ZOMBIE, '')
di68kap's avatar
di68kap committed
703
704


di68kap's avatar
di68kap committed
705
class RootNode(Node):
706
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
707

Eckhart Arnold's avatar
Eckhart Arnold committed
708
        all_errors (list):  A list of all errors that have occured so far during
709
710
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
711

712
713
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
714
    """
715

eckhart's avatar
eckhart committed
716
    def __init__(self, node: Optional[Node] = None):
717
        super().__init__(ZOMBIE, '')
eckhart's avatar
eckhart committed
718
        self.all_errors = []  # type: List[Error]
di68kap's avatar
di68kap committed
719
        self.error_flag = 0
eckhart's avatar
eckhart committed
720
721
        if node is not None:
            self.swallow(node)
722
        # customization for XML-Representation
eckhart's avatar
eckhart committed
723
724
725
        self.inline_tags = set()  # type: Set[str]
        self.omit_tags = set()  # type: Set[str]
        self.empty_tags = set()  # type: Set[str]
di68kap's avatar
di68kap committed
726

727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
    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.errors = copy.deepcopy(self.errors) if self.errors else []
        duplicate._pos = self._pos
        duplicate._len = self._len
        if hasattr(self, '_xml_attr'):
            duplicate._xml_attr = copy.deepcopy(self._xml_attr)
        duplicate.all_errors = copy.deepcopy(self.all_errors)
        duplicate.error_flag = self.error_flag
        duplicate.inline_tags = self.inline_tags
        duplicate.omit_tags = self.omit_tags
        duplicate.empty_tags = self.empty_tags
745
        duplicate.tag_name = self.tag_name
746
747
748
        return duplicate


749
    def swallow(self, node: Node) -> 'RootNode':
750
751
        """
        Put `self` in the place of `node` by copying all its data.
752
753
754
755
756
757
758
759
760
        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
761
762
763
764
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
765
        self.tag_name = node.tag_name
766
767
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
768
        self._content = node._content
769
        return self
di68kap's avatar
di68kap committed
770

eckhart's avatar
eckhart committed
771
    def add_error(self, node: Node, error: Error) -> 'RootNode':
772
773
774
        """
        Adds an Error object to the tree, locating it at a specific node.
        """
di68kap's avatar
di68kap committed
775
776
777
778
779
        self.all_errors.append(error)
        self.error_flag = max(self.error_flag, error.code)
        node.errors.append(error)
        return self

eckhart's avatar
eckhart committed
780
    def new_error(self,
781
                  node: Node,
di68kap's avatar
di68kap committed
782
                  message: str,
eckhart's avatar
eckhart committed
783
                  code: ErrorCode = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
784
        """
eckhart's avatar
eckhart committed
785
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
786
787
788
789
790
        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
791
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
792
793
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
794

eckhart's avatar
eckhart committed
795
    def collect_errors(self) -> List[Error]:
796
797
        """
        Returns the list of errors, ordered bv their position.
di68kap's avatar
di68kap committed
798
        """
799
        self.all_errors.sort(key=lambda e: e.pos)
Eckhart Arnold's avatar
Eckhart Arnold committed
800
        return self.all_errors
di68kap's avatar
di68kap committed
801

802
    def customized_XML(self):
803
804
        """
        Returns a customized XML representation of the tree.
805
        See the docstring of `Node.as_xml()` for an explanation of the
806
807
        customizations.
        """
eckhart's avatar
eckhart committed
808
        return self.as_xml(inline_tags=self.inline_tags,
809
810
811
                           omit_tags=self.omit_tags,
                           empty_tags=self.empty_tags)

di68kap's avatar
di68kap committed
812

eckhart's avatar
eckhart committed
813
814
ZOMBIE_ROOTNODE = RootNode()

di68kap's avatar
di68kap committed
815
816
817
818
819
#######################################################################
#
# S-expression- and XML-parsers
#
#######################################################################
eckhart's avatar
eckhart committed
820
821


eckhart's avatar
eckhart committed
822
def parse_sxpr(sxpr: Union[str, StringView]) -> Node:
823
    """
824
825
826
827
828
    Generates a tree of nodes from an S-expression.

    This can - among other things - be used for deserialization of trees that
    have been serialized with `Node.as_sxpr()` or as a convenient way to
    generate test data.
829
830

    Example:
831
    >>> parse_sxpr("(a (b c))").as_sxpr()
832
    '(a\\n  (b\\n    "c"\\n  )\\n)'
833
    """
834

eckhart's avatar
eckhart committed
835
    sxpr = StringView(sxpr).strip() if isinstance(sxpr, str) else sxpr.strip()
836
    # mock_parsers = dict()  # type: Dict[StringView, MockParser]
837

838
    def next_block(s: StringView):
eckhart's avatar
eckhart committed
839
840
        """Generator that yields all characters until the next closing bracket
        that does not match an opening bracket matched earlier within the same
841
842
        package.
        """
843
        s = s.strip()
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
        try:
            while s[0] != ')':
                if s[0] != '(':
                    raise ValueError('"(" expected, not ' + s[:10])
                # assert s[0] == '(', s
                level = 1
                k = 1
                while level > 0:
                    if s[k] == '(':
                        level += 1
                    elif s[k] == ')':
                        level -= 1
                    k += 1
                yield s[:k]
                s = s[k:].strip()
        except IndexError:
            errmsg = ('Malformed S-expression. Unprocessed part: "%s"' % s) if s \
                else 'Malformed S-expression. Closing bracket(s) ")" missing.'
            raise AssertionError(errmsg)
863

864
865
866
867
868
869
870
871
872
    def inner_parser(sxpr: StringView) -> Node:
        if sxpr[0] != '(':
            raise ValueError('"(" expected, not ' + sxpr[:10])
        # assert sxpr[0] == '(', sxpr
        sxpr = sxpr[1:].strip()
        match = sxpr.match(re.compile(r'[\w:]+'))
        if match is None:
            raise AssertionError('Malformed S-expression Node-tagname or identifier expected, '
                                 'not "%s"' % sxpr[:40].replace('\n', ''))
di68kap's avatar
di68kap committed
873
        end = sxpr.index(match.end())
874
875
876
        tagname = sxpr[:end]
        name, class_name = (tagname.split(':') + [''])[:2]
        sxpr = sxpr[end:].strip()
eckhart's avatar
eckhart committed
877
        attributes = OrderedDict()  # type: OrderedDict[str, str]
878
        if sxpr[0] == '(':
eckhart's avatar
eckhart committed
879
            result = tuple(inner_parser(block) for block in next_block(sxpr))  # type: ResultType
880
881
882
        else:
            lines = []
            while sxpr and sxpr[0:1] != ')':
883
                # parse attr
884
885
886
887
888
                while sxpr[:2] == "`(":
                    i = sxpr.find('"')
                    k = sxpr.find(')')
                    # read very special attribute pos
                    if sxpr[2:5] == "pos" and 0 < i < k:
eckhart's avatar
eckhart committed
889
890
                        # pos = int(sxpr[5:k].strip().split(' ')[0])
                        pass
891
892
893
                    # ignore very special attribute err
                    elif sxpr[2:5] == "err" and 0 <= sxpr.find('`', 5) < k:
                        m = sxpr.find('(', 5)
eckhart's avatar
eckhart committed
894
                        while 0 <= m < k:
895
896
                            m = sxpr.find('(', k)
                            k = max(k, sxpr.find(')', max(m, 0)))
897
                    # read attr
898
899
900
901
                    else:
                        attr = sxpr[2:i].strip()