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 38.5 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
    __slots__ = 'name', 'ptype'
Eckhart Arnold's avatar
Eckhart Arnold committed
65

66
67
68
    def __init__(self,):  # , pbases=frozenset()):
        self.name = ''  # type: str
        self.ptype = ':' + self.__class__.__name__  # type: str
69
70

    def __repr__(self):
di68kap's avatar
di68kap committed
71
         return self.name + self.ptype
72
73
74
75

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

76
77
78
    def __call__(self, text: StringView) -> Tuple[Optional['Node'], StringView]:
        return None, text

79
80
81
82
83
84
85
86
87
88
89
    # @property
    # def name(self):
    #     """Returns the name of the parser or the empty string '' for unnamed
    #     parsers."""
    #     return self._name
    #
    # @property
    # def ptype(self) -> str:
    #     """Returns the type of the parser. By default this is the parser's
    #     class name preceded by a colon, e.g. ':ZeroOrMore'."""
    #     return self._ptype
90
91
92

    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
93
        """Returns the parser's name if it has a name and repr()"""
94
95
        return self.name if self.name else repr(self)

96
    def reset(self):
97
        """Resets any parser variables. (Should be overridden.)"""
98
99
        pass

eckhart's avatar
eckhart committed
100
    def grammar(self) -> Optional[object]:
101
102
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
103
104
105
        return None

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
106
        """Applies the function `func` to the parser. Returns False, if
107
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
108
        otherwise."""
109
110
        return False

111
112

WHITESPACE_PTYPE = ':Whitespace'
113
TOKEN_PTYPE = ':Token'
114
115
116
117
118
119


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
120
    different parser types by assigning them a `ptype` on initialization.
121
122
123
124
125

    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.
    """
Eckhart Arnold's avatar
Eckhart Arnold committed
126
127
    __slots__ = ()

128
129
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
130
131
        super().__init__()
        self.name = name
di68kap's avatar
di68kap committed
132
133
        if ptype:
            self.ptype = ptype  # or ':' + self.__class__.__name__
134
135
136
137
138
139
140
141
142
143
144
145
146


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.
    """
    alive = False
Eckhart Arnold's avatar
Eckhart Arnold committed
147
    __slots__ = ()
148
149

    def __init__(self):
150
        super(ZombieParser, self).__init__()
151
152
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
153
        self.name = "__ZOMBIE__"
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
        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
#
#######################################################################


177
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
178
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
179
StrictResultType = Union[ChildrenType, StringView, str]
180
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
181
182


Eckhart Arnold's avatar
Eckhart Arnold committed
183
def flatten_sxpr(sxpr: str) -> str:
184
    """Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
185
186
187
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
188
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
189
190
    '(a (b c))'
    """
191
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
192
193


194
def flatten_xml(xml: str) -> str:
195
    """Returns an XML-tree as a one liner without unnecessary whitespace,
196
197
    i.e. only whitespace within leaf-nodes is preserved.
    """
198
199
200
201
202
    # 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']
    return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?P<closing_tag></\w+>)\s+', tag_only, xml))
203
204


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
    Attributes:
210
211
        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
212

213
214
215
        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.
216

217
218
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
219

eckhart's avatar
eckhart committed
220
221
222
223
        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
224
        parser (Parser):  The parser which generated this node.
225
226
            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
227
228
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
229
            example by calling ``isinstance(node.parer, ...)``.
230

231
232
233
        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
234
            the length before AST-transformation and will never change
235
            through AST-transformation. READ ONLY!
236

237
238
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
239
            The value of ``pos`` is -1 meaning invalid by default.
240
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
241
            nodes relative to this value.
242
243

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
244
            pos value of the root node should be set to 0 right
245
246
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
247
            Other than that, this value should be considered READ ONLY.
248
249
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
250

251
252
        errors (list):  A list of all errors that occured on this node.

253
        attributes (dict): An optional dictionary of XML-attributes. This
254
255
256
            dictionary is created lazily upon first usage. The attributes
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
257
    """
258

Eckhart Arnold's avatar
Eckhart Arnold committed
259
    __slots__ = '_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content'
260

261
262
263
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
264
265
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
266
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
267
        self._pos = -1                  # type: int
di68kap's avatar
di68kap committed
268
269
        # Assignment to self.result initializes the attributes _result, children and _len
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
270
        if leafhint:
eckhart's avatar
eckhart committed
271
            self._result = result       # type: StrictResultType
eckhart's avatar
eckhart committed
272
            self._content = None        # type: Optional[str]
273
            self.children = NoChildren  # type: ChildrenType
274
            self._len = -1              # type: int  # lazy evaluation
275
276
        else:
            self.result = result
277
278
        self.parser = parser or ZOMBIE_PARSER

Eckhart Arnold's avatar
Eckhart Arnold committed
279

280
    def __str__(self):
281
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
282
        if self.errors:
283
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
284
                   (s, '; '.join(e.message for e in self.errors))
285
        return s
286

Eckhart Arnold's avatar
Eckhart Arnold committed
287

288
289
290
291
292
293
    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)
294

Eckhart Arnold's avatar
Eckhart Arnold committed
295

296
    def __len__(self):
297
        if self._len < 0:
di68kap's avatar
di68kap committed
298
            self._len = sum(len(child) for child in self.children) \
299
                if self.children else len(self._result)
300
301
302
303
304
305
306
307
        return self._len


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


308
    def __eq__(self, other):
309
310
311
312
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
        """
313
        return self.tag_name == other.tag_name and self.result == other.result
314

Eckhart Arnold's avatar
Eckhart Arnold committed
315

316
    def __hash__(self):
317
        return hash(self.tag_name)
318

Eckhart Arnold's avatar
Eckhart Arnold committed
319

320
321
322
    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
323
        an integer or the first child node with the given tag name. Examples::
324

325
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
326
327
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
328
329
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
330
331
332
333

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
334
        Returns:
335
336
            Node: All nodes which have a given tag name.
        """
337
338
339
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
340
            else:
341
342
343
344
345
                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!')
346
347
348
349


    def __contains__(self, tag_name: str) -> bool:
        """
350
        Returns true if a child with the given tag name exists.
351
        Args:
352
353
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
354
355
356
357
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
358
359
360
361
362
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
363
            return False
364
        raise ValueError('Leave node cannot contain other nodes')
365
        # generator = self.select_by_tag(tag_name, False)
366
367
368
369
370
        # try:
        #     generator.__next__()
        #     return True
        # except StopIteration:
        #     return False
371
372


373
374
375
376
377
378
379
380
381
382
383
384
385
386
    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


387
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
388
    def tag_name(self) -> str:
389
390
391
392
393
394
        """
        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`.
        """
395
        return self.parser.name or self.parser.ptype
396

Eckhart Arnold's avatar
Eckhart Arnold committed
397

398
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
399
    def result(self) -> StrictResultType:
400
401
402
403
404
        """
        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.
        """
405
406
        return self._result

407

408
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
409
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
410
        # # made obsolete by static type checking with mypy
411
412
413
        # 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
414
415
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
416
417
        self._len = -1        # lazy evaluation
        self._content = None
418
419
420
421
422
423
424
425
426
        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
427
                self._result = result
428

429
430

    @property
eckhart's avatar
eckhart committed
431
    def content(self) -> str:
432
        """
433
        Returns content as string, omitting error messages.
434
        """
435
436
437
438
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
439
                # self._content = self._result
440
441
442
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
443
444
445
446
447
448
449
450
451
452
453


    @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))


454
    @property
455
    def pos(self) -> int:
eckhart's avatar
eckhart committed
456
457
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
458
            raise AssertionError("Position value not initialized!")
459
460
        return self._pos

eckhart's avatar
eckhart committed
461

eckhart's avatar
eckhart committed
462
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
463
464
465
466
467
468
469
470
471
        """
        (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
472
473
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
474
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
475
        offset = self.pos
476
        for child in self.children:
eckhart's avatar
eckhart committed
477
            child.init_pos(offset)
eckhart's avatar
eckhart committed
478
479
480
            offset = child.pos + len(child)
        return self

481

482
483
    @property
    def attributes(self):
484
485
486
        """
        Returns a dictionary of XML-Attributes attached to the Node.
        """
487
488
489
490
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
491

492
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
493
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
494
495
496
497
498
499
500
501
        """
        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.

502
        Args:
503
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
504
            open_fn:   (Node->str) A function that returns an opening
505
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
506
            close_fn:  (Node->str) A function that returns a closeF
507
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
508
            data_fn:   (str->str) A function that filters the data string
509
510
511
512
513
514
                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
515
516
        head = open_fn(self)
        tail = close_fn(self)
517
518

        if not self.result:
519
            return head.rstrip() + tail.lstrip()
520

eckhart's avatar
eckhart committed
521
        tail = tail.lstrip(None if density & 2 else '')
522

523
        inline = inline or inline_fn(self)
524
525
526
527
528
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
529
530
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
531

532
533
        if self.children:
            content = []
534
            for child in self.children:
535
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
536
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
537
538
539
                if subtree:
                    subtree = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in subtree))
540
            return head + usetab + (sep + usetab).join(content) + tail
541

542
        res = cast(str, self.result)  # safe, because if there are no children, result is a string
eckhart's avatar
eckhart committed
543
        if not inline and not head:
eckhart's avatar
eckhart committed
544
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
545
            res = res.strip()
546
547
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
548
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
549
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
550
        else:
551
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
552

Eckhart Arnold's avatar
Eckhart Arnold committed
553

554
555
556
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
557
                compact: bool = False) -> str:
558
559
560
        """
        Returns content as S-expression, i.e. in lisp-like form.

561
        Args:
562
563
564
            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.
565
566
567
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
568
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
569
                tree structure.
570
571
        """

eckhart's avatar
eckhart committed
572
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
573
        lbreaks = linebreaks(src) if src else []  # type: List[int]
574

575
        def opening(node) -> str:
eckhart's avatar
eckhart committed
576
            """Returns the opening string for the representation of `node`."""
577
            txt = [left_bracket,  node.tag_name]
578
            # s += " '(pos %i)" % node.add_pos
579
            if hasattr(node, '_xml_attr'):
580
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attributes.items())
581
            if src:
eckhart's avatar
eckhart committed
582
                txt.append(" `(pos %i %i %i)" % (node.pos, *line_col(lbreaks, node.pos)))
583
            if showerrors and node.errors:
584
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
585
            return "".join(txt) + '\n'
586

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

eckhart's avatar
eckhart committed
591
592
593
594
595
        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'\"')
596

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

eckhart's avatar
eckhart committed
599

600
601
602
603
604
605
    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:
606
607
608
        """
        Returns content as XML-tree.

609
        Args:
610
611
612
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
613
614
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
615
616
617
618
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
                neither the opening tag nor its attributes nor the closing tag. This
                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.
622
623
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
624
625
        """

626
        def opening(node) -> str:
627
628
629
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
630
            txt = ['<', node.tag_name]
631
632
            has_reserved_attrs = hasattr(node, '_xml_attr') \
                and any (r in node.attributes for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
633
            if hasattr(node, '_xml_attr'):
634
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attributes.items())
635
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
636
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
637
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
638
639
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
640
641
642
            if node.tag_name in empty_tags:
                assert not node.result, ("Node %s with content %s is not an empty element!" %
                                         (node.tag_name, str(node)))
eckhart's avatar
eckhart committed
643
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
644
645
646
            else:
                ending = ">\n"
            return "".join(txt + [ending])
647
648

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

654
655
656
657
658
        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.
            """
659
660
            return node.tag_name in inline_tags or (hasattr(node, '_xml_attr') \
                and node.attributes.get('xml:space', 'default') == 'preserve')
661

662
        line_breaks = linebreaks(src) if src else []
663
664
        return self._tree_repr(' ' * indentation, opening, closing,
                               density=1, inline_fn=inlining)
665

Eckhart Arnold's avatar
Eckhart Arnold committed
666

667
    def select(self, match_function: Callable, include_root: bool=False) -> Iterator['Node']:
eckhart's avatar
eckhart committed
668
        """
669
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
670

671
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
672
        given `match_function` evaluates to True. The tree is
673
674
675
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
676

677
678
679
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
680
681
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
682
        Yields:
683
            Node: All nodes of the tree for which
684
685
            ``match_function(node)`` returns True
        """
686
        if include_root and match_function(self):
687
            yield self
eckhart's avatar
eckhart committed
688
689
690
        for child in self.children:
            for node in child.select(match_function, True):
                yield node
691

Eckhart Arnold's avatar
Eckhart Arnold committed
692

eckhart's avatar
eckhart committed
693
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
694
                      include_root: bool=False) -> Iterator['Node']:
695
        """
696
697
698
699
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
700

701
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
702
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
703
            ['(X (c "d"))', '(X "F")']
704
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
705
            ['(b "X")', '(X (c "d"))', '(X "F")']
706
            >>> any(tree.select_by_tag('a', False))
707
            False
708
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
709
            ['(a (b "X") (X (c "d")) (e (X "F")))']
710
711
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
712
713

        Args:
714
715
716
717
            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.
718
719
720
        Yields:
            Node: All nodes which have a given tag name.
        """
721
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
722
            tag_names = frozenset({tag_names})
723
        return self.select(lambda node: node.tag_name in tag_names, include_root)
724
725


726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

        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


741
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
742
743
744
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
745
746
747
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
748
class RootNode(Node):
749
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
750

Eckhart Arnold's avatar
Eckhart Arnold committed
751
        all_errors (list):  A list of all errors that have occured so far during
752
753
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
754

755
756
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
757
    """
eckhart's avatar
eckhart committed
758
    def __init__(self, node: Optional[Node] = None) -> 'RootNode':
di68kap's avatar
di68kap committed
759
        super().__init__(ZOMBIE_PARSER, '')
760
        self.all_errors = []
di68kap's avatar
di68kap committed
761
        self.error_flag = 0
eckhart's avatar
eckhart committed
762
763
        if node is not None:
            self.swallow(node)
764
765
766
767
        # customization for XML-Representation
        self.inline_tags = set()
        self.omit_tags = set()
        self.empty_tags = set()
di68kap's avatar
di68kap committed
768

769
    def swallow(self, node: Node) -> 'RootNode':
770
771
772
773
774
775
776
777
778
779
        """Put `self` in the place of `node` by copying all its data.
        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
780
781
782
783
784
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
785
786
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
787
        self._content = node._content
788
        return self
di68kap's avatar
di68kap committed
789

eckhart's avatar
eckhart committed
790
791
    def add_error(self, node: Node, error: Error) -> 'RootNode':
        """Adds an Error object to the tree, locating it at a specific node."""
di68kap's avatar
di68kap committed
792
793
794
795
796
        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
797
    def new_error(self,
798
                  node: Node,
di68kap's avatar
di68kap committed
799
                  message: str,
di68kap's avatar
di68kap committed
800
                  code: int = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
801
        """
eckhart's avatar
eckhart committed
802
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
803
804
805
806
807
        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
808
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
809
810
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
811

eckhart's avatar
eckhart committed
812
    def collect_errors(self) -> List[Error]:
di68kap's avatar
di68kap committed
813
814
        """Returns the list of errors, ordered bv their position.
        """
815
        self.all_errors.sort(key=lambda e: e.pos)
Eckhart Arnold's avatar
Eckhart Arnold committed
816
        return self.all_errors
di68kap's avatar
di68kap committed
817

818
819
820
821
822
823
824
825
    def customized_XML(self):
        """Returns a customized XML representation of the tree.
        See the docstring of `Node.as_xml()` for an explanation of the
        customizations."""
        return self.as_xml(inline_tags = self.inline_tags,
                           omit_tags=self.omit_tags,
                           empty_tags=self.empty_tags)

di68kap's avatar
di68kap committed
826

eckhart's avatar
eckhart committed
827
828
829
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


830
def parse_sxpr(sxpr: str) -> Node:
831
    """
832
833
834
835
836
    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.
837
838

    Example:
839
    >>> parse_sxpr("(a (b c))").as_sxpr()
840
    '(a\\n  (b\\n    "c"\\n  )\\n)'
841
    """
842
843
    sxpr = StringView(sxpr).strip()
    mock_parsers = dict()
844

845
    def next_block(s: StringView):
eckhart's avatar
eckhart committed
846
847
848
        """Generator that yields all characters until the next closing bracket
        that does not match an opening bracket matched earlier within the same
        package."""
849
        s = s.strip()
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
        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)
869

870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
    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] != ')':
                # parse attributes
                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)))
                    # read attributes
                    else:
                        attr = sxpr[2:i].strip()
                        value = sxpr[i:k].strip()[1:-1]
                        attributes[attr] = value
                    sxpr = sxpr[k+1:].strip()
                # parse content
                for qtmark in ['"""', "'''", '"', "'"]:
                    match = sxpr.match(re.compile(qtmark + r'.*?' + qtmark, re.DOTALL))
                    if match:
                        end = match.end() - sxpr.begin
                        i = len(qtmark)
                        lines.append(str(sxpr[i:end - i]))
                        sxpr = sxpr[end:].strip()
                        break
917
                else:
918
919
920
921
922
923
924
925
926
927
928
929
930
                    match = sxpr.match(re.compile(r'(?:(?!\)).)*', re.DOTALL))
                    end = match.end() - sxpr.begin
                    lines.append(str(sxpr[:end]))
                    sxpr = sxpr[end:]
            result = "\n".join(lines)
        node = Node(mock_parsers.setdefault(tagname, MockParser(name, ':' + class_name)), result)
        if attributes:
            node.attributes.update(attributes)
        return node

    return inner_parser(sxpr)


di68kap's avatar
di68kap committed
931
932
933
RX_WHITESPACE_TAIL = re.compile(r'\s*$')


934
935
936
937
938
def parse_xml(xml: str) -> Node:
    """
    Generates a tree of nodes from a (Pseudo-)XML-source.
    """
    xml = StringView(xml)
939
940
    PlainText = MockParser('', TOKEN_PTYPE)
    mock_parsers = {TOKEN_PTYPE: PlainText}
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960

    def parse_attributes(s: StringView) -> Tuple[StringView, OrderedDict]:
        """Parses a sqeuence of XML-Attributes. Returns the string-slice
        beginning after the end of the attributes."""
        attributes = OrderedDict()
        restart = 0
        for match in s.finditer(re.compile(r'\s*(?P<attr>\w+)\s*=\s*"(?P<value>.*)"\s*')):
            d = match.groupdict()
            attributes[d['attr']] = d['value']
            restart = match.end() - s.begin
        return (s[restart:], attributes)

    def parse_opening_tag(s: StringView) -> Tuple[StringView, str, OrderedDict, bool]:
        """Parses an opening tag. Returns the string segment following the
        the opening tag, the tag name, a dictionary of attributes and
        a flag indicating whether the tag is actually a solitary tag as
        indicated by a slash at the end, i.e. <br/>."""
        match = s.match(re.compile(r'<\s*(?P<tagname>[\w:]+)\s*'))
        assert match
        tagname = match.groupdict()['tagname']
di68kap's avatar
di68kap committed
961
962
        section = s[match.end() - s.begin:]
        s, attributes = parse_attributes(section)
963
964
        i = s.find('>')
        assert i >= 0
965
        return s[i+1:], tagname, attributes, s[i-1] == "/"
966
967

    def parse_closing_tag(s: StringView) -> Tuple[StringView, str]:
968
        """Parses a closing tag and returns the string segment, just after
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
        the closing tag."""
        match = s.match(re.compile(r'</\s*(?P<tagname>[\w:]+)>'))
        assert match
        tagname = match.groupdict()['tagname']
        return s[match.end() - s.begin:], tagname

    def parse_leaf_content(s: StringView) -> Tuple[StringView, str]:
        """Parses a piece of the content of a tag, just until the next opening,
        closing or solitary tag is reached."""
        i = 0
        while s[i] != "<" or s[max(0, i-1)] == "\\":
            i = s.find("<", i)
        return s[i:], s[:i]

    def parse_full_content(s: StringView) -> Tuple[StringView, Node]:
        """Parses the full content of a tag, starting right at the beginning
        of the opening tag and ending right after the closing tag.
        """
        result = []
        s, tagname, attributes, solitary = parse_opening_tag(s)
        name, class_name = (tagname.split(":") + [''])[:2]
        if not solitary:
            while s and not s[:2] == "</":
                s, leaf = parse_leaf_content(s)
di68kap's avatar
di68kap committed
993
                if not leaf.match(RX_WHITESPACE_TAIL):
994
995
996
997
998
999
                    result.append(Node(PlainText, leaf))
                if s[:1] == "<" and s[:2] != "</":
                    s, child = parse_full_content(s)
                    result.append(child)
            s, closing_tagname = parse_closing_tag(s)
            assert tagname == closing_tagname