2.12.2021, 9:00 - 11:00: Due to updates GitLab may be unavailable for some minutes between 09:00 and 11:00.

syntaxtree.py 39.3 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
29

30
from DHParser.error import Error, linebreaks, line_col
31
from DHParser.stringview import StringView
eckhart's avatar
eckhart committed
32
from DHParser.toolkit import re, typing
eckhart's avatar
eckhart committed
33
from typing import Callable, cast, Iterator, List, AbstractSet, Set, Union, Tuple, Optional
di68kap's avatar
di68kap committed
34

35

36
37
38
39
40
41
__all__ = ('ParserBase',
           'WHITESPACE_PTYPE',
           'TOKEN_PTYPE',
           'MockParser',
           'ZombieParser',
           'ZOMBIE_PARSER',
eckhart's avatar
eckhart committed
42
           'ZOMBIE_NODE',
43
           'Node',
di68kap's avatar
di68kap committed
44
           '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
59
60
61
62
63
#######################################################################
#
# 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.
    """
64

65
    __slots__ = 'name', 'ptype'
Eckhart Arnold's avatar
Eckhart Arnold committed
66

67
68
69
    def __init__(self,):  # , pbases=frozenset()):
        self.name = ''  # type: str
        self.ptype = ':' + self.__class__.__name__  # type: str
70
71

    def __repr__(self):
di68kap's avatar
di68kap committed
72
         return self.name + self.ptype
73
74
75
76

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

77
78
79
    def __call__(self, text: StringView) -> Tuple[Optional['Node'], StringView]:
        return None, text

80
81
    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
82
        """Returns the parser's name if it has a name and repr()"""
83
84
        return self.name if self.name else repr(self)

85
    def reset(self):
86
        """Resets any parser variables. (Should be overridden.)"""
87
88
        pass

eckhart's avatar
eckhart committed
89
    def grammar(self) -> Optional[object]:
90
91
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
92
93
94
        return None

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
95
        """Applies the function `func` to the parser. Returns False, if
96
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
97
        otherwise."""
98
99
        return False

100
101

WHITESPACE_PTYPE = ':Whitespace'
102
TOKEN_PTYPE = ':Token'
103
104
105
106
107
108


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
109
    different parser types by assigning them a `ptype` on initialization.
110
111
112
113
114

    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.
    """
115

Eckhart Arnold's avatar
Eckhart Arnold committed
116
117
    __slots__ = ()

118
119
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
120
121
        super().__init__()
        self.name = name
di68kap's avatar
di68kap committed
122
123
        if ptype:
            self.ptype = ptype  # or ':' + self.__class__.__name__
124
125
126
127
128
129
130
131
132
133
134
135


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.
    """
136

137
    alive = False
Eckhart Arnold's avatar
Eckhart Arnold committed
138
    __slots__ = ()
139
140

    def __init__(self):
141
        super(ZombieParser, self).__init__()
142
143
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
144
        self.name = "__ZOMBIE__"
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
        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
#
#######################################################################


168
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
169
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
170
StrictResultType = Union[ChildrenType, StringView, str]
171
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
172
173


Eckhart Arnold's avatar
Eckhart Arnold committed
174
def flatten_sxpr(sxpr: str) -> str:
175
176
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
177
178
179
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
180
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
181
182
    '(a (b c))'
    """
183

184
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
185
186


187
def flatten_xml(xml: str) -> str:
188
189
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
190
    i.e. only whitespace within leaf-nodes is preserved.
191
192
    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`.
193
    """
194

195
196
197
198
    # 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']
199
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
200
201


202
203
204
RX_AMP = re.compile('&(?!\w+;)')


Eckhart Arnold's avatar
Eckhart Arnold committed
205
class Node(collections.abc.Sized):
206
207
208
    """
    Represents a node in the concrete or abstract syntax tree.

209
210
    TODO: Add some documentation and doc-tests here...

211
    Attributes:
212
213
        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
214

215
216
217
        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.
218

219
220
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
221

eckhart's avatar
eckhart committed
222
223
224
225
        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
226
        parser (Parser):  The parser which generated this node.
227
228
            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
229
230
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
231
            example by calling ``isinstance(node.parer, ...)``.
232

233
234
235
        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
236
            the length before AST-transformation and will never change
237
            through AST-transformation. READ ONLY!
238

239
240
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
241
            The value of ``pos`` is -1 meaning invalid by default.
242
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
243
            nodes relative to this value.
244
245

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
246
            pos value of the root node should be set to 0 right
247
248
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
249
            Other than that, this value should be considered READ ONLY.
250
251
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
252

253
254
        errors (list):  A list of all errors that occured on this node.

255
256
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
257
258
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
259
    """
260

Eckhart Arnold's avatar
Eckhart Arnold committed
261
    __slots__ = '_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content'
262

263
264
265
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
266
267
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
268
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
269
        self._pos = -1                  # type: int
270
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
271
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
272
        if leafhint:
eckhart's avatar
eckhart committed
273
            self._result = result       # type: StrictResultType
eckhart's avatar
eckhart committed
274
            self._content = None        # type: Optional[str]
275
            self.children = NoChildren  # type: ChildrenType
276
            self._len = -1              # type: int  # lazy evaluation
277
278
        else:
            self.result = result
279
280
        self.parser = parser or ZOMBIE_PARSER

Eckhart Arnold's avatar
Eckhart Arnold committed
281

282
    def __str__(self):
283
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
284
        if self.errors:
285
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
286
                   (s, '; '.join(e.message for e in self.errors))
287
        return s
288

Eckhart Arnold's avatar
Eckhart Arnold committed
289

290
291
292
293
294
295
    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 \
               "(" + ", ".join(repr(child) for child in self.children) + ")"
        return "Node(%s, %s)" % (parg, rarg)
296

Eckhart Arnold's avatar
Eckhart Arnold committed
297

298
    def __len__(self):
299
        if self._len < 0:
di68kap's avatar
di68kap committed
300
            self._len = sum(len(child) for child in self.children) \
301
                if self.children else len(self._result)
302
303
304
305
306
307
308
309
        return self._len


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


310
    def __eq__(self, other):
311
312
313
314
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
        """
315
        return self.tag_name == other.tag_name and self.result == other.result
316

Eckhart Arnold's avatar
Eckhart Arnold committed
317

318
    def __hash__(self):
319
        return hash(self.tag_name)
320

Eckhart Arnold's avatar
Eckhart Arnold committed
321

322
323
324
    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
325
        an integer or the first child node with the given tag name. Examples::
326

327
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
328
329
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
330
331
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
332
333
334
335

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
336
        Returns:
337
338
            Node: All nodes which have a given tag name.
        """
339
340
341
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
342
            else:
343
344
345
346
347
                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!')
348
349
350
351


    def __contains__(self, tag_name: str) -> bool:
        """
352
        Returns true if a child with the given tag name exists.
353
        Args:
354
355
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
356
357
358
359
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
360
361
362
363
364
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
365
            return False
366
        raise ValueError('Leave node cannot contain other nodes')
367
368


369
370
371
372
373
374
375
376
377
378
379
380
381
382
    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


383
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
384
    def tag_name(self) -> str:
385
386
387
388
389
390
        """
        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`.
        """
391
        return self.parser.name or self.parser.ptype
392

Eckhart Arnold's avatar
Eckhart Arnold committed
393

394
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
395
    def result(self) -> StrictResultType:
396
397
398
399
400
        """
        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.
        """
401
402
        return self._result

403

404
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
405
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
406
        # # made obsolete by static type checking with mypy
407
408
409
        # 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
410
411
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
412
413
        self._len = -1        # lazy evaluation
        self._content = None
414
415
416
417
418
419
420
421
422
        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
423
                self._result = result
424

425
426

    @property
eckhart's avatar
eckhart committed
427
    def content(self) -> str:
428
        """
di68kap's avatar
di68kap committed
429
430
431
        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.
432
        """
433
434
435
436
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
437
                # self._content = self._result
438
439
440
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
441
442


di68kap's avatar
di68kap committed
443
444
445
446
447
    @content.setter
    def content(self, content: str):
        self.result = content


448
449
450
451
452
453
454
455
456
    @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))


457
    @property
458
    def pos(self) -> int:
eckhart's avatar
eckhart committed
459
460
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
461
            raise AssertionError("Position value not initialized!")
462
463
        return self._pos

eckhart's avatar
eckhart committed
464

eckhart's avatar
eckhart committed
465
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
466
467
468
469
470
471
472
473
474
        """
        (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
475
476
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
477
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
478
        offset = self.pos
479
        for child in self.children:
eckhart's avatar
eckhart committed
480
            child.init_pos(offset)
eckhart's avatar
eckhart committed
481
482
483
            offset = child.pos + len(child)
        return self

484

485
    @property
486
    def attr(self):
487
        """
488
        Returns a dictionary of XML-attr attached to the node.
489
        """
490
491
492
493
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
494

495
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
496
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
497
498
499
500
501
502
503
504
        """
        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.

505
        Args:
506
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
507
            open_fn:   (Node->str) A function that returns an opening
508
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
509
            close_fn:  (Node->str) A function that returns a closeF
510
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
511
            data_fn:   (str->str) A function that filters the data string
512
513
514
515
516
517
                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
518
519
        head = open_fn(self)
        tail = close_fn(self)
520
521

        if not self.result:
522
            return head.rstrip() + tail.lstrip()
523

eckhart's avatar
eckhart committed
524
        tail = tail.lstrip(None if density & 2 else '')
525

526
        inline = inline or inline_fn(self)
527
528
529
530
531
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
532
533
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
534

535
536
        if self.children:
            content = []
537
            for child in self.children:
538
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
539
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
540
541
542
                if subtree:
                    subtree = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in subtree))
543
            return head + usetab + (sep + usetab).join(content) + tail
544

545
        res = self.content  # cast(str, self.result)  # safe, because if there are no children, result is a string
eckhart's avatar
eckhart committed
546
        if not inline and not head:
eckhart's avatar
eckhart committed
547
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
548
            res = res.strip()
549
550
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
551
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
552
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
553
        else:
554
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
555

Eckhart Arnold's avatar
Eckhart Arnold committed
556

557
558
559
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
560
                compact: bool = False) -> str:
561
562
563
        """
        Returns content as S-expression, i.e. in lisp-like form.

564
        Args:
565
566
567
            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.
568
569
570
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
571
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
572
                tree structure.
573
574
        """

eckhart's avatar
eckhart committed
575
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
576
        lbreaks = linebreaks(src) if src else []  # type: List[int]
577

578
        def opening(node) -> str:
eckhart's avatar
eckhart committed
579
            """Returns the opening string for the representation of `node`."""
580
            txt = [left_bracket,  node.tag_name]
581
            # s += " '(pos %i)" % node.add_pos
582
            if hasattr(node, '_xml_attr'):
583
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
584
            if src:
585
586
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
587
            if showerrors and node.errors:
588
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
589
            return "".join(txt) + '\n'
590

eckhart's avatar
eckhart committed
591
592
593
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
594

eckhart's avatar
eckhart committed
595
596
597
598
599
        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'\"')
600

601
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
602

eckhart's avatar
eckhart committed
603

604
605
606
607
608
609
    def as_xml(self, src: str = None,
               showerrors: bool = True,
               indentation: int = 2,
               inline_tags: Set[str]=set(),
               omit_tags: Set[str]=set(),
               empty_tags: Set[str]=set()) -> str:
610
611
612
        """
        Returns content as XML-tree.

613
        Args:
614
615
616
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
617
618
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
619
620
621
            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
622
                neither the opening tag nor its attr nor the closing tag. This
623
624
625
                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.
626
627
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
628
629
        """

630
        def opening(node) -> str:
631
632
633
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
634
            txt = ['<', node.tag_name]
635
            has_reserved_attrs = hasattr(node, '_xml_attr') \
636
                and any (r in node.attr for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
637
            if hasattr(node, '_xml_attr'):
638
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
639
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
640
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
641
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
642
643
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
644
645
646
            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
647
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
648
649
650
            else:
                ending = ">\n"
            return "".join(txt + [ending])
651
652

        def closing(node):
653
            """Returns the closing string for the representation of `node`."""
654
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
655
                return ''
656
            return ('\n</') + node.tag_name + '>'
657

658
659
660
661
662
663
664
        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


665
666
667
668
669
        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.
            """
670
            return node.tag_name in inline_tags or (hasattr(node, '_xml_attr') \
671
                                                    and node.attr.get('xml:space', 'default') == 'preserve')
672

673
        line_breaks = linebreaks(src) if src else []
674
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
675
                               density=1, inline_fn=inlining)
676

Eckhart Arnold's avatar
Eckhart Arnold committed
677

678
679
    def select(self, match_function: Callable, include_root: bool=False, reverse: bool=False) \
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
680
        """
681
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
682

683
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
684
        given `match_function` evaluates to True. The tree is
685
686
687
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
688

689
690
691
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
692
693
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
694
695
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
696
        Yields:
697
            Node: All nodes of the tree for which
698
699
            ``match_function(node)`` returns True
        """
700
        if include_root and match_function(self):
701
            yield self
702
703
704
        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
705
                yield node
706

Eckhart Arnold's avatar
Eckhart Arnold committed
707

eckhart's avatar
eckhart committed
708
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
709
                      include_root: bool=False) -> Iterator['Node']:
710
        """
711
712
713
714
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
715

716
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
717
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
718
            ['(X (c "d"))', '(X "F")']
719
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
720
            ['(b "X")', '(X (c "d"))', '(X "F")']
721
            >>> any(tree.select_by_tag('a', False))
722
            False
723
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
724
            ['(a (b "X") (X (c "d")) (e (X "F")))']
725
726
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
727
728

        Args:
729
730
731
732
            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.
733
734
735
        Yields:
            Node: All nodes which have a given tag name.
        """
736
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
737
            tag_names = frozenset({tag_names})
738
        return self.select(lambda node: node.tag_name in tag_names, include_root)
739
740


741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

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


756
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
757
758
759
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
760
761
762
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
763
764
765
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


di68kap's avatar
di68kap committed
766
class RootNode(Node):
767
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
768

Eckhart Arnold's avatar
Eckhart Arnold committed
769
        all_errors (list):  A list of all errors that have occured so far during
770
771
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
772

773
774
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
775
    """
776

eckhart's avatar
eckhart committed
777
    def __init__(self, node: Optional[Node] = None) -> 'RootNode':
di68kap's avatar
di68kap committed
778
        super().__init__(ZOMBIE_PARSER, '')
779
        self.all_errors = []
di68kap's avatar
di68kap committed
780
        self.error_flag = 0
eckhart's avatar
eckhart committed
781
782
        if node is not None:
            self.swallow(node)
783
784
785
786
        # customization for XML-Representation
        self.inline_tags = set()
        self.omit_tags = set()
        self.empty_tags = set()
di68kap's avatar
di68kap committed
787

788
    def swallow(self, node: Node) -> 'RootNode':
789
790
        """
        Put `self` in the place of `node` by copying all its data.
791
792
793
794
795
796
797
798
799
        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
800
801
802
803
804
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
805
806
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
807
        self._content = node._content
808
        return self
di68kap's avatar
di68kap committed
809

eckhart's avatar
eckhart committed
810
    def add_error(self, node: Node, error: Error) -> 'RootNode':
811
812
813
        """
        Adds an Error object to the tree, locating it at a specific node.
        """
di68kap's avatar
di68kap committed
814
815
816
817
818
        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
819
    def new_error(self,
820
                  node: Node,
di68kap's avatar
di68kap committed
821
                  message: str,
di68kap's avatar
di68kap committed
822
                  code: int = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
823
        """
eckhart's avatar
eckhart committed
824
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
825
826
827
828
829
        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
830
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
831
832
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
833

eckhart's avatar
eckhart committed
834
    def collect_errors(self) -> List[Error]:
835
836
        """
        Returns the list of errors, ordered bv their position.
di68kap's avatar
di68kap committed
837
        """
838
        self.all_errors.sort(key=lambda e: e.pos)
Eckhart Arnold's avatar
Eckhart Arnold committed
839
        return self.all_errors
di68kap's avatar
di68kap committed
840

841
    def customized_XML(self):
842
843
        """
        Returns a customized XML representation of the tree.
844
        See the docstring of `Node.as_xml()` for an explanation of the
845
846
        customizations.
        """
847
848
849
850
        return self.as_xml(inline_tags = self.inline_tags,
                           omit_tags=self.omit_tags,
                           empty_tags=self.empty_tags)

di68kap's avatar
di68kap committed
851

di68kap's avatar
di68kap committed
852
853
854
855
856
#######################################################################
#
# S-expression- and XML-parsers
#
#######################################################################
eckhart's avatar
eckhart committed
857
858


859
def parse_sxpr(sxpr: str) -> Node:
860
    """
861
862
863
864
865
    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.
866
867

    Example:
868
    >>> parse_sxpr("(a (b c))").as_sxpr()
869
    '(a\\n  (b\\n    "c"\\n  )\\n)'
870
    """
871

872
873
    sxpr = StringView(sxpr).strip()
    mock_parsers = dict()
874

875
    def next_block(s: StringView):
eckhart's avatar
eckhart committed
876
877
        """Generator that yields all characters until the next closing bracket
        that does not match an opening bracket matched earlier within the same
878
879
        package.
        """
880
        s = s.strip()
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
        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)
900

901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
    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', ''))
        end = match.end() - sxpr.begin
        tagname = sxpr[:end]
        name, class_name = (tagname.split(':') + [''])[:2]
        sxpr = sxpr[end:].strip()
        attributes = OrderedDict()
        if sxpr[0] == '(':
            result = tuple(inner_parser(block) for block in next_block(sxpr))
        else:
            lines = []
            while sxpr and sxpr[0:1] != ')':
920
                # parse attr
921
922
923
924
925
926
927
928
929
930
931
932
                while sxpr[:2] == "`(":
                    i = sxpr.find('"')
                    k = sxpr.find(')')
                    # read very special attribute pos
                    if sxpr[2:5] == "pos" and 0 < i < k:
                        pos = int(sxpr[5:k].strip().split(' ')[0])
                    # ignore very special attribute err
                    elif sxpr[2:5] == "err" and 0 <= sxpr.find('`', 5) < k:
                        m = sxpr.find('(', 5)
                        while m >= 0 and m < k:
                            m = sxpr.find('(', k)
                            k = max(k, sxpr.find(')', max(m, 0)))
933
                    # read attr