syntaxtree.py 41.3 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

Eckhart Arnold's avatar
Eckhart Arnold committed
27
import collections.abc
28
from collections import OrderedDict
eckhart's avatar
eckhart committed
29
import copy
30

eckhart's avatar
eckhart committed
31
from DHParser.error import Error, ErrorCode, linebreaks, line_col
32
from DHParser.stringview import StringView
eckhart's avatar
eckhart committed
33
from DHParser.toolkit import re, typing
eckhart's avatar
eckhart committed
34
from typing import Callable, cast, Iterator, List, AbstractSet, Set, Dict, Union, Tuple, Optional
di68kap's avatar
di68kap committed
35

36

37
38
39
40
41
42
__all__ = ('ParserBase',
           'WHITESPACE_PTYPE',
           'TOKEN_PTYPE',
           'MockParser',
           'ZombieParser',
           'ZOMBIE_PARSER',
eckhart's avatar
eckhart committed
43
           'ZOMBIE_NODE',
44
           'Node',
di68kap's avatar
di68kap committed
45
           'RootNode',
eckhart's avatar
eckhart committed
46
           'ZOMBIE_ROOTNODE',
47
           'parse_sxpr',
48
49
50
           'parse_xml',
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
51
52


53
54
55
56
57
58
59
60
61
62
63
64
65
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


class ParserBase:
    """
    ParserBase is the base class for all real and mock parser classes.
    It is defined here, because Node objects require a parser object
    for instantiation.
    """
66

67
    __slots__ = 'name', 'ptype'
Eckhart Arnold's avatar
Eckhart Arnold committed
68

69
70
71
    def __init__(self,):  # , pbases=frozenset()):
        self.name = ''  # type: str
        self.ptype = ':' + self.__class__.__name__  # type: str
72
73

    def __repr__(self):
eckhart's avatar
eckhart committed
74
        return self.name + self.ptype
75
76
77
78

    def __str__(self):
        return self.name + (' = ' if self.name else '') + repr(self)

79
80
81
    def __call__(self, text: StringView) -> Tuple[Optional['Node'], StringView]:
        return None, text

82
83
    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
84
        """Returns the parser's name if it has a name and repr()"""
85
86
        return self.name if self.name else repr(self)

87
    def reset(self):
88
        """Resets any parser variables. (Should be overridden.)"""
89
90
        pass

Eckhart Arnold's avatar
Eckhart Arnold committed
91
92
    @property
    def grammar(self) -> Optional['Grammar']:
93
94
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
95
96
97
        return None

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
98
        """Applies the function `func` to the parser. Returns False, if
99
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
100
        otherwise."""
101
102
        return False

103
104

WHITESPACE_PTYPE = ':Whitespace'
105
TOKEN_PTYPE = ':Token'
106
107
108
109
110
111


class MockParser(ParserBase):
    """
    MockParser objects can be used to reconstruct syntax trees from a
    serialized form like S-expressions or XML. Mock objects can mimic
eckhart's avatar
eckhart committed
112
    different parser types by assigning them a `ptype` on initialization.
113
114
115
116
117

    Mock objects should not be used for anything other than
    syntax tree (re-)construction. In all other cases where a parser
    object substitute is needed, chose the singleton ZOMBIE_PARSER.
    """
118

Eckhart Arnold's avatar
Eckhart Arnold committed
119
120
    __slots__ = ()

121
122
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
123
124
        super().__init__()
        self.name = name
di68kap's avatar
di68kap committed
125
126
        if ptype:
            self.ptype = ptype  # or ':' + self.__class__.__name__
127
128
129
130
131
132
133
134
135
136
137
138


class ZombieParser(MockParser):
    """
    Serves as a substitute for a Parser instance.

    ``ZombieParser`` is the class of the singelton object
    ``ZOMBIE_PARSER``. The  ``ZOMBIE_PARSER`` has a name and can be
    called, but it never matches. It serves as a substitute where only
    these (or one of these properties) is needed, but no real Parser-
    object is instantiated.
    """
139

140
    alive = False
Eckhart Arnold's avatar
Eckhart Arnold committed
141
    __slots__ = ()
142
143

    def __init__(self):
144
        super(ZombieParser, self).__init__()
145
146
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
147
        self.name = "__ZOMBIE__"
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
        self.__class__.alive = True

    def __copy__(self):
        return self

    def __deepcopy__(self, memo):
        return self

    def __call__(self, text):
        """Better call Saul ;-)"""
        return None, text


ZOMBIE_PARSER = ZombieParser()


#######################################################################
#
# syntaxtree nodes
#
#######################################################################


171
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
172
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
173
StrictResultType = Union[ChildrenType, StringView, str]
174
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
175
176


Eckhart Arnold's avatar
Eckhart Arnold committed
177
def flatten_sxpr(sxpr: str) -> str:
178
179
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
180
181
182
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
183
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
184
185
    '(a (b c))'
    """
186

187
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
188
189


190
def flatten_xml(xml: str) -> str:
191
192
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
193
    i.e. only whitespace within leaf-nodes is preserved.
194
195
    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`.
196
    """
197

198
199
200
201
    # 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']
202
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
203
204


eckhart's avatar
eckhart committed
205
RX_AMP = re.compile(r'&(?!\w+;)')
206
207


Eckhart Arnold's avatar
Eckhart Arnold committed
208
class Node(collections.abc.Sized):
209
210
211
    """
    Represents a node in the concrete or abstract syntax tree.

212
213
    TODO: Add some documentation and doc-tests here...

214
    Attributes:
215
216
        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
217

218
219
220
        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.
221

222
223
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
224

eckhart's avatar
eckhart committed
225
226
227
228
        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
229
        parser (Parser):  The parser which generated this node.
230
231
            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
232
233
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
234
            example by calling ``isinstance(node.parer, ...)``.
235

236
237
238
        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
239
            the length before AST-transformation and will never change
240
            through AST-transformation. READ ONLY!
241

242
243
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
244
            The value of ``pos`` is -1 meaning invalid by default.
245
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
246
            nodes relative to this value.
247
248

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
249
            pos value of the root node should be set to 0 right
250
251
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
252
            Other than that, this value should be considered READ ONLY.
253
254
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
255

256
257
        errors (list):  A list of all errors that occured on this node.

258
259
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
260
261
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
262
    """
263

Eckhart Arnold's avatar
Eckhart Arnold committed
264
    __slots__ = '_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content'
265

266
267
268
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
269
270
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
271
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
272
        self._pos = -1                  # type: int
273
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
274
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
275
        if leafhint:
eckhart's avatar
eckhart committed
276
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
eckhart's avatar
eckhart committed
277
            self._content = None        # type: Optional[str]
278
            self.children = NoChildren  # type: ChildrenType
279
            self._len = -1              # type: int  # lazy evaluation
280
281
        else:
            self.result = result
282
283
        self.parser = parser or ZOMBIE_PARSER

eckhart's avatar
eckhart committed
284
285
286
287
288
289
290
291
292
293
294
    def __deepcopy__(self, memo):
        if self.children:
            duplicate = self.__class__(self.parser, copy.deepcopy(self.children), False)
        else:
            duplicate = self.__class__(self.parser, self.result, True)
        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
295

296
    def __str__(self):
297
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
298
        if self.errors:
299
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
300
                   (s, '; '.join(e.message for e in self.errors))
301
        return s
302

Eckhart Arnold's avatar
Eckhart Arnold committed
303

304
305
306
307
    def __repr__(self):
        mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
        parg = "MockParser({name}, {ptype})".format(**mpargs)
        rarg = str(self) if not self.children else \
eckhart's avatar
eckhart committed
308
            "(" + ", ".join(repr(child) for child in self.children) + ")"
309
        return "Node(%s, %s)" % (parg, rarg)
310

Eckhart Arnold's avatar
Eckhart Arnold committed
311

312
    def __len__(self):
313
        if self._len < 0:
di68kap's avatar
di68kap committed
314
            self._len = sum(len(child) for child in self.children) \
315
                if self.children else len(self._result)
316
317
318
319
320
321
322
323
        return self._len


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


324
    def __eq__(self, other):
325
326
327
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
328
329
330
331

        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!
332
        """
333
        return self.tag_name == other.tag_name and self.result == other.result
334

Eckhart Arnold's avatar
Eckhart Arnold committed
335

336
    def __hash__(self):
337
        return hash(self.tag_name)
338

Eckhart Arnold's avatar
Eckhart Arnold committed
339

340
341
342
    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
343
        an integer or the first child node with the given tag name. Examples::
344

345
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
346
347
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
348
349
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
350
351
352
353

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
354
        Returns:
355
356
            Node: All nodes which have a given tag name.
        """
357
358
359
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
360
            else:
361
362
363
364
365
                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!')
366
367
368
369


    def __contains__(self, tag_name: str) -> bool:
        """
370
        Returns true if a child with the given tag name exists.
371
        Args:
372
373
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
374
375
376
377
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
378
379
380
381
382
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
383
            return False
384
        raise ValueError('Leave node cannot contain other nodes')
385
386


387
388
389
390
391
392
393
394
395
396
397
398
399
400
    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


401
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
402
    def tag_name(self) -> str:
403
404
405
406
407
408
        """
        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`.
        """
409
        return self.parser.name or self.parser.ptype
410

Eckhart Arnold's avatar
Eckhart Arnold committed
411

412
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
413
    def result(self) -> StrictResultType:
414
415
416
417
418
        """
        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.
        """
419
420
        return self._result

421

422
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
423
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
424
        # # made obsolete by static type checking with mypy
425
426
427
        # 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
428
429
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
430
431
        self._len = -1        # lazy evaluation
        self._content = None
432
433
434
435
436
437
438
439
440
        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
441
                self._result = result  # cast(StrictResultType, result)
442

443
444

    @property
eckhart's avatar
eckhart committed
445
    def content(self) -> str:
446
        """
di68kap's avatar
di68kap committed
447
448
449
        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.
450
        """
451
452
453
454
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
455
                # self._content = self._result
456
457
458
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
459
460


di68kap's avatar
di68kap committed
461
462
463
464
465
    @content.setter
    def content(self, content: str):
        self.result = content


466
467
468
469
470
471
472
473
474
    @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))


475
    @property
476
    def pos(self) -> int:
eckhart's avatar
eckhart committed
477
478
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
479
            raise AssertionError("Position value not initialized!")
480
481
        return self._pos

eckhart's avatar
eckhart committed
482

eckhart's avatar
eckhart committed
483
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
484
485
486
487
488
489
490
491
492
        """
        (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
493
494
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
495
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
496
        offset = self.pos
497
        for child in self.children:
eckhart's avatar
eckhart committed
498
            child.init_pos(offset)
eckhart's avatar
eckhart committed
499
500
501
            offset = child.pos + len(child)
        return self

502

503
    @property
504
    def attr(self):
505
        """
506
        Returns a dictionary of XML-attr attached to the node.
507
        """
508
509
510
511
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
512

513
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
514
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
515
516
517
518
519
520
521
522
        """
        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.

523
        Args:
524
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
525
            open_fn:   (Node->str) A function that returns an opening
526
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
527
            close_fn:  (Node->str) A function that returns a closeF
528
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
529
            data_fn:   (str->str) A function that filters the data string
530
531
532
533
534
535
                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
536
537
        head = open_fn(self)
        tail = close_fn(self)
538
539

        if not self.result:
540
            return head.rstrip() + tail.lstrip()
541

eckhart's avatar
eckhart committed
542
        tail = tail.lstrip(None if density & 2 else '')
543

544
        inline = inline or inline_fn(self)
545
546
547
548
549
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
550
551
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
552

553
554
        if self.children:
            content = []
555
            for child in self.children:
556
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
557
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
558
                if subtree:
eckhart's avatar
eckhart committed
559
560
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
561
            return head + usetab + (sep + usetab).join(content) + tail
562

eckhart's avatar
eckhart committed
563
        res = self.content
eckhart's avatar
eckhart committed
564
        if not inline and not head:
eckhart's avatar
eckhart committed
565
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
566
            res = res.strip()
567
568
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
569
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
570
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
571
        else:
572
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
573

Eckhart Arnold's avatar
Eckhart Arnold committed
574

575
576
577
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
578
                compact: bool = False) -> str:
579
580
581
        """
        Returns content as S-expression, i.e. in lisp-like form.

582
        Args:
583
584
585
            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.
586
587
588
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
589
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
590
                tree structure.
591
592
        """

eckhart's avatar
eckhart committed
593
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
594
        lbreaks = linebreaks(src) if src else []  # type: List[int]
595

596
        def opening(node) -> str:
eckhart's avatar
eckhart committed
597
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
598
            txt = [left_bracket, node.tag_name]
599
            # s += " '(pos %i)" % node.add_pos
600
            if hasattr(node, '_xml_attr'):
601
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
602
            if src:
603
604
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
605
            if showerrors and node.errors:
606
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
607
            return "".join(txt) + '\n'
608

eckhart's avatar
eckhart committed
609
610
611
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
612

eckhart's avatar
eckhart committed
613
614
615
616
617
        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'\"')
618

619
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
620

eckhart's avatar
eckhart committed
621

622
623
624
    def as_xml(self, src: str = None,
               showerrors: bool = True,
               indentation: int = 2,
eckhart's avatar
eckhart committed
625
626
627
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
628
629
630
        """
        Returns content as XML-tree.

631
        Args:
632
633
634
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
635
636
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
637
638
639
            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
640
                neither the opening tag nor its attr nor the closing tag. This
641
642
643
                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.
644
645
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
646
647
        """

648
        def opening(node) -> str:
649
650
651
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
652
            txt = ['<', node.tag_name]
653
            has_reserved_attrs = hasattr(node, '_xml_attr') \
eckhart's avatar
eckhart committed
654
                and any(r in node.attr for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
655
            if hasattr(node, '_xml_attr'):
656
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
657
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
658
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
659
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
660
661
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
662
663
664
            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
665
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
666
667
668
            else:
                ending = ">\n"
            return "".join(txt + [ending])
669
670

        def closing(node):
671
            """Returns the closing string for the representation of `node`."""
672
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
673
                return ''
674
            return ('\n</') + node.tag_name + '>'
675

676
677
678
679
680
681
682
        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


683
684
685
686
687
        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
688
689
690
            return node.tag_name in inline_tags \
                or (hasattr(node, '_xml_attr')
                    and node.attr.get('xml:space', 'default') == 'preserve')
691

692
        line_breaks = linebreaks(src) if src else []
693
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
694
                               density=1, inline_fn=inlining)
695

Eckhart Arnold's avatar
Eckhart Arnold committed
696

eckhart's avatar
eckhart committed
697
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
698
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
699
        """
700
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
701

702
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
703
        given `match_function` evaluates to True. The tree is
704
705
706
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
707

708
709
710
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
711
712
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
713
714
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
715
        Yields:
716
            Node: All nodes of the tree for which
717
718
            ``match_function(node)`` returns True
        """
719
        if include_root and match_function(self):
720
            yield self
721
722
723
        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
724
                yield node
725

Eckhart Arnold's avatar
Eckhart Arnold committed
726

eckhart's avatar
eckhart committed
727
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
eckhart's avatar
eckhart committed
728
                      include_root: bool = False) -> Iterator['Node']:
729
        """
730
731
732
733
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
734

735
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
736
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
737
            ['(X (c "d"))', '(X "F")']
738
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
739
            ['(b "X")', '(X (c "d"))', '(X "F")']
740
            >>> any(tree.select_by_tag('a', False))
741
            False
742
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
743
            ['(a (b "X") (X (c "d")) (e (X "F")))']
744
745
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
746
747

        Args:
748
749
750
751
            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.
752
753
754
        Yields:
            Node: All nodes which have a given tag name.
        """
755
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
756
            tag_names = frozenset({tag_names})
757
        return self.select(lambda node: node.tag_name in tag_names, include_root)
758
759


760
761
762
763
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

764
        This function is mostly just syntactic sugar for
765
766
767
768
769
770
771
772
773
774
        ``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


775
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
776
777
778
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
779
780
781
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
782
783
784
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


di68kap's avatar
di68kap committed
785
class RootNode(Node):
786
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
787

Eckhart Arnold's avatar
Eckhart Arnold committed
788
        all_errors (list):  A list of all errors that have occured so far during
789
790
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
791

792
793
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
794
    """
795

eckhart's avatar
eckhart committed
796
    def __init__(self, node: Optional[Node] = None):
di68kap's avatar
di68kap committed
797
        super().__init__(ZOMBIE_PARSER, '')
eckhart's avatar
eckhart committed
798
        self.all_errors = []  # type: List[Error]
di68kap's avatar
di68kap committed
799
        self.error_flag = 0
eckhart's avatar
eckhart committed
800
801
        if node is not None:
            self.swallow(node)
802
        # customization for XML-Representation
eckhart's avatar
eckhart committed
803
804
805
        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
806

807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
    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
        duplicate.parser = self.parser
        return duplicate


829
    def swallow(self, node: Node) -> 'RootNode':
830
831
        """
        Put `self` in the place of `node` by copying all its data.
832
833
834
835
836
837
838
839
840
        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
841
842
843
844
845
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
846
847
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
848
        self._content = node._content
849
        return self
di68kap's avatar
di68kap committed
850

eckhart's avatar
eckhart committed
851
    def add_error(self, node: Node, error: Error) -> 'RootNode':
852
853
854
        """
        Adds an Error object to the tree, locating it at a specific node.
        """
di68kap's avatar
di68kap committed
855
856
857
858
859
        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
860
    def new_error(self,
861
                  node: Node,
di68kap's avatar
di68kap committed
862
                  message: str,
eckhart's avatar
eckhart committed
863
                  code: ErrorCode = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
864
        """
eckhart's avatar
eckhart committed
865
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
866
867
868
869
870
        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
871
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
872
873
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
874

eckhart's avatar
eckhart committed
875
    def collect_errors(self) -> List[Error]:
876
877
        """
        Returns the list of errors, ordered bv their position.
di68kap's avatar
di68kap committed
878
        """
879
        self.all_errors.sort(key=lambda e: e.pos)
Eckhart Arnold's avatar
Eckhart Arnold committed
880
        return self.all_errors
di68kap's avatar
di68kap committed
881

882
    def customized_XML(self):
883
884
        """
        Returns a customized XML representation of the tree.
885
        See the docstring of `Node.as_xml()` for an explanation of the
886
887
        customizations.
        """
eckhart's avatar
eckhart committed
888
        return self.as_xml(inline_tags=self.inline_tags,
889
890
891
                           omit_tags=self.omit_tags,
                           empty_tags=self.empty_tags)

di68kap's avatar
di68kap committed
892

eckhart's avatar
eckhart committed
893
894
ZOMBIE_ROOTNODE = RootNode()

di68kap's avatar
di68kap committed
895
896
897
898
899
#######################################################################
#
# S-expression- and XML-parsers
#
#######################################################################
eckhart's avatar
eckhart committed
900
901


eckhart's avatar
eckhart committed
902
def parse_sxpr(sxpr: Union[str, StringView]) -> Node:
903
    """
904
905
906
907
908
    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.
909
910

    Example:
911
    >>> parse_sxpr("(a (b c))").as_sxpr()
912
    '(a\\n  (b\\n    "c"\\n  )\\n)'
913
    """
914

eckhart's avatar
eckhart committed
915
    sxpr = StringView(sxpr).strip() if isinstance(sxpr, str) else sxpr.strip()
eckhart's avatar
eckhart committed
916
    mock_parsers = dict()  # type: Dict[str, MockParser]