syntaxtree.py 35.8 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
import copy
30

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

36

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


52
53
54
55
56
57
58
59
60
61
62
63
64
65
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


class ParserBase:
    """
    ParserBase is the base class for all real and mock parser classes.
    It is defined here, because Node objects require a parser object
    for instantiation.
    """
    def __init__(self, name=''):  # , pbases=frozenset()):
66
        self._name = name  # type: str
67
68
69
        self._ptype = ':' + self.__class__.__name__  # type: str

    def __repr__(self):
di68kap's avatar
di68kap committed
70
         return self.name + self.ptype
71
72
73
74

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

75
76
77
    def __call__(self, text: StringView) -> Tuple[Optional['Node'], StringView]:
        return None, text

78
79
    @property
    def name(self):
80
81
        """Returns the name of the parser or the empty string '' for unnamed
        parsers."""
82
83
        return self._name

84
85
    @property
    def ptype(self) -> str:
86
87
        """Returns the type of the parser. By default this is the parser's
        class name preceded by a colon, e.g. ':ZeroOrMore'."""
88
89
90
91
        return self._ptype

    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
92
        """Returns the parser's name if it has a name and repr()"""
93
94
        return self.name if self.name else repr(self)

95
    def reset(self):
96
        """Resets any parser variables. (Should be overridden.)"""
97
98
        pass

eckhart's avatar
eckhart committed
99
    def grammar(self) -> Optional[object]:
100
101
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
102
103
104
        return None

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
105
        """Applies the function `func` to the parser. Returns False, if
106
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
107
        otherwise."""
108
109
        return False

110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126

WHITESPACE_PTYPE = ':Whitespace'
TOKEN_PTYPE = ':Token'


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
    different parser types by assigning them a ptype on initialization.

    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.
    """
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
127
        super().__init__(name)
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
        self._ptype = ptype or ':' + self.__class__.__name__


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

    def __init__(self):
        super(ZombieParser, self).__init__("__ZOMBIE__")
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
        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
#
#######################################################################


170
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
171
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
172
StrictResultType = Union[ChildrenType, StringView, str]
173
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
174
175


Eckhart Arnold's avatar
Eckhart Arnold committed
176
def flatten_sxpr(sxpr: str) -> str:
177
    """Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
178
179
180
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
181
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
182
183
    '(a (b c))'
    """
184
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
185
186


187
188
189
190
191
192
193
def flatten_xml(xml: str) -> str:
    """Returns an XML-tree as a one linter without unnecessary whitespace,
    i.e. only whitespace within leaf-nodes is preserved.
    """
    return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?<=</\w+>)\s+', '', xml))


Eckhart Arnold's avatar
Eckhart Arnold committed
194
class Node(collections.abc.Sized):
195
196
197
    """
    Represents a node in the concrete or abstract syntax tree.

198
    Attributes:
199
200
        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
201

202
203
204
        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.
205

206
207
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
208

eckhart's avatar
eckhart committed
209
210
211
212
        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
213
        parser (Parser):  The parser which generated this node.
214
215
            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
216
217
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
218
            example by calling ``isinstance(node.parer, ...)``.
219

220
221
222
        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
223
            the length before AST-transformation and will never change
224
            through AST-transformation. READ ONLY!
225

226
227
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
228
            The value of ``pos`` is -1 meaning invalid by default.
229
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
230
            nodes relative to this value.
231
232

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
233
            pos value of the root node should be set to 0 right
234
235
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
236
            Other than that, this value should be considered READ ONLY.
237
238
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
239

240
241
        errors (list):  A list of all errors that occured on this node.

242
        attributes (dict): An optional dictionary of XML-attributes. This
243
244
245
            dictionary is created lazily upon first usage. The attributes
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
246
    """
247

eckhart's avatar
eckhart committed
248
    __slots__ = ['_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content']
249

250
251
252
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
253
254
        that generated the node and the parser's result.
        """
eckhart's avatar
eckhart committed
255
        self.errors = []               # type: List[Error]
eckhart's avatar
eckhart committed
256
        self._pos = -1                  # type: int
di68kap's avatar
di68kap committed
257
258
        # 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
259
        if leafhint:
eckhart's avatar
eckhart committed
260
            self._result = result       # type: StrictResultType
eckhart's avatar
eckhart committed
261
            self._content = None        # type: Optional[str]
262
            self.children = NoChildren  # type: ChildrenType
263
            self._len = -1              # type: int  # lazy evaluation
264
265
        else:
            self.result = result
266
267
        self.parser = parser or ZOMBIE_PARSER

Eckhart Arnold's avatar
Eckhart Arnold committed
268

269
    def __str__(self):
270
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
271
        if self.errors:
272
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
273
                   (s, '; '.join(e.message for e in self.errors))
274
        return s
275

Eckhart Arnold's avatar
Eckhart Arnold committed
276

277
278
279
280
281
282
    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)
283

Eckhart Arnold's avatar
Eckhart Arnold committed
284

285
    def __len__(self):
286
        if self._len < 0:
di68kap's avatar
di68kap committed
287
            self._len = sum(len(child) for child in self.children) \
288
                if self.children else len(self._result)
289
290
291
292
293
294
295
296
        return self._len


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


297
    def __eq__(self, other):
298
299
300
301
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
        """
302
        return self.tag_name == other.tag_name and self.result == other.result
303

Eckhart Arnold's avatar
Eckhart Arnold committed
304

305
    def __hash__(self):
306
        return hash(self.tag_name)
307

Eckhart Arnold's avatar
Eckhart Arnold committed
308

309
310
311
312
313
314
    def __deepcopy__(self, memodict={}):
        result = copy.deepcopy(self.result)
        other = Node(self.parser, result)
        other._pos = self._pos
        return other

Eckhart Arnold's avatar
Eckhart Arnold committed
315

316
317
318
    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
319
        an integer or the first child node with the given tag name. Examples::
320

321
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
322
323
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
324
325
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
326
327
328
329

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
330
        Returns:
331
332
            Node: All nodes which have a given tag name.
        """
333
334
335
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
336
            else:
337
338
339
340
341
                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!')
342
343
344
345


    def __contains__(self, tag_name: str) -> bool:
        """
346
        Returns true if a child with the given tag name exists.
347
        Args:
348
349
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
350
351
352
353
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
354
355
356
357
358
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
359
            return False
360
        raise ValueError('Leave node cannot contain other nodes')
361
        # generator = self.select_by_tag(tag_name, False)
362
363
364
365
366
        # try:
        #     generator.__next__()
        #     return True
        # except StopIteration:
        #     return False
367
368


369
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
370
    def tag_name(self) -> str:
371
372
373
374
375
376
        """
        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`.
        """
377
        return self.parser.name or self.parser.ptype
378

Eckhart Arnold's avatar
Eckhart Arnold committed
379

380
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
381
    def result(self) -> StrictResultType:
382
383
384
385
386
        """
        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.
        """
387
388
        return self._result

389

390
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
391
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
392
        # # made obsolete by static type checking with mypy
393
394
395
        # 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
396
397
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
398
399
        self._len = -1        # lazy evaluation
        self._content = None
400
401
402
403
404
405
406
407
408
        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
409
                self._result = result
410

411
412

    @property
eckhart's avatar
eckhart committed
413
    def content(self) -> str:
414
        """
415
        Returns content as string, omitting error messages.
416
        """
417
418
419
420
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
421
                # self._content = self._result
422
423
424
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
425
426
427
428
429
430
431
432
433
434
435


    @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))


436
    @property
437
    def pos(self) -> int:
eckhart's avatar
eckhart committed
438
439
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
440
            raise AssertionError("Position value not initialized!")
441
442
        return self._pos

eckhart's avatar
eckhart committed
443

eckhart's avatar
eckhart committed
444
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
445
446
447
448
449
450
451
452
453
        """
        (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
454
455
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
456
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
457
        offset = self.pos
458
        for child in self.children:
eckhart's avatar
eckhart committed
459
            child.init_pos(offset)
eckhart's avatar
eckhart committed
460
461
462
            offset = child.pos + len(child)
        return self

463

464
465
    @property
    def attributes(self):
466
467
468
        """
        Returns a dictionary of XML-Attributes attached to the Node.
        """
469
470
471
472
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
473

474
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i, density=0) -> str:
475
476
477
478
479
480
481
482
        """
        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.

483
        Args:
484
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
485
            open_fn:   (Node->str) A function that returns an opening
486
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
487
            close_fn:  (Node->str) A function that returns a closeF
488
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
489
            data_fn:   (str->str) A function that filters the data string
490
491
492
493
494
495
                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
496
497
        head = open_fn(self)
        tail = close_fn(self)
498
499

        if not self.result:
500
            return head.rstrip() + tail.lstrip()
501

eckhart's avatar
eckhart committed
502
        tail = tail.lstrip(None if density & 2 else '')
503
504
505

        if self.children:
            content = []
506
            for child in self.children:
eckhart's avatar
eckhart committed
507
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn, density).split('\n')
508
                content.append('\n'.join((tab + s) for s in subtree))
eckhart's avatar
eckhart committed
509
            return head + '\n'.join(content) + tail
510

511
        res = cast(str, self.result)  # safe, because if there are no children, result is a string
512
513
514
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
            gap = ' ' if head.rstrip()[-1] != '>' else ''
eckhart's avatar
eckhart committed
515
            return head.rstrip() + gap + data_fn(self.result) + tail.lstrip()
516
        else:
eckhart's avatar
eckhart committed
517
            return head + '\n'.join([tab + data_fn(s) for s in res.split('\n')]) + tail
518

Eckhart Arnold's avatar
Eckhart Arnold committed
519

520
521
    def as_sxpr(self, src: str = None, compact: bool = False, showerrors: bool = True,
                indentation: int = 2) -> str:
522
523
524
        """
        Returns content as S-expression, i.e. in lisp-like form.

525
        Args:
526
527
528
            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.
Eckhart Arnold's avatar
Eckhart Arnold committed
529
            compact:  If True a compact representation is returned where
530
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
531
                tree structure.
532
533
        """

eckhart's avatar
eckhart committed
534
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
535
        lbreaks = linebreaks(src) if src else []  # type: List[int]
536

537
        def opening(node) -> str:
eckhart's avatar
eckhart committed
538
            """Returns the opening string for the representation of `node`."""
539
            txt = [left_bracket,  node.tag_name]
540
            # s += " '(pos %i)" % node.add_pos
541
            if hasattr(node, '_xml_attr'):
542
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attributes.items())
543
            if src:
eckhart's avatar
eckhart committed
544
                txt.append(" `(pos %i %i %i)" % (node.pos, *line_col(lbreaks, node.pos)))
di68kap's avatar
di68kap committed
545
546
            # if node.error_flag:   # just for debugging error collecting
            #     txt += " HAS ERRORS"
547
            if showerrors and node.errors:
548
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
549
            return "".join(txt) + '\n'
550

eckhart's avatar
eckhart committed
551
552
553
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
554

eckhart's avatar
eckhart committed
555
556
557
558
559
        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'\"')
560

561
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
562

eckhart's avatar
eckhart committed
563

564
    def as_xml(self, src: str = None, showerrors: bool = True, indentation: int = 2) -> str:
565
566
567
        """
        Returns content as XML-tree.

568
        Args:
569
570
571
572
573
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
        """

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

        def closing(node):
eckhart's avatar
eckhart committed
589
            """Returns the closing string for the representation of `node`."""            
590
            return '\n</' + node.tag_name + '>'
591

592
        line_breaks = linebreaks(src) if src else []
593
        return self._tree_repr(' ' * indentation, opening, closing, density=1)
594

Eckhart Arnold's avatar
Eckhart Arnold committed
595

596
    def select(self, match_function: Callable, include_root: bool=True) -> Iterator['Node']:
eckhart's avatar
eckhart committed
597
        """
598
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
599

600
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
601
        given `match_function` evaluates to True. The tree is
602
603
604
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
605

606
607
608
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
609
610
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
611
        Yields:
612
            Node: All nodes of the tree for which
613
614
            ``match_function(node)`` returns True
        """
615
        if include_root and match_function(self):
616
617
618
            yield self
        else:
            for child in self.children:
619
                for node in child.select(match_function, True):
eckhart's avatar
eckhart committed
620
                    yield node
621

Eckhart Arnold's avatar
Eckhart Arnold committed
622

eckhart's avatar
eckhart committed
623
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
624
                      include_root: bool=True) -> Iterator['Node']:
625
        """
626
627
628
629
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
630

631
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
632
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
633
            ['(X (c "d"))', '(X "F")']
634
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
635
            ['(b "X")', '(X (c "d"))', '(X "F")']
636
            >>> any(tree.select_by_tag('a', False))
637
            False
638
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
639
            ['(a (b "X") (X (c "d")) (e (X "F")))']
640
641
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
642
643

        Args:
644
645
646
647
            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.
648
649
650
        Yields:
            Node: All nodes which have a given tag name.
        """
651
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
652
            tag_names = frozenset({tag_names})
653
        return self.select(lambda node: node.tag_name in tag_names, include_root)
654
655


656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
    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


671
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
672
673
674
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
675
676
677
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
678
class RootNode(Node):
679
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
680

681
682
683
        errors (list):  A list of all errors that have occured so far during
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
684

685
686
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
687
    """
eckhart's avatar
eckhart committed
688
    def __init__(self, node: Optional[Node] = None) -> 'RootNode':
di68kap's avatar
di68kap committed
689
        super().__init__(ZOMBIE_PARSER, '')
690
        self.all_errors = []
eckhart's avatar
eckhart committed
691
        self.err_nodes_keep = []
di68kap's avatar
di68kap committed
692
        self.error_flag = 0
eckhart's avatar
eckhart committed
693
694
        if node is not None:
            self.swallow(node)
di68kap's avatar
di68kap committed
695

696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
    # def _propagate_errors(self):
    #     if not self.all_errors or not self.error_propagation:
    #         return
    #     self.all_errors.sort(key=lambda e: e.pos)
    #     i = 0
    #     for leaf in self.select(lambda nd: not nd.children, False):
    #         leaf.errors = []
    #         while i < len(self.all_errors) \
    #                 and leaf.pos <= self.all_errors[i].add_pos < leaf.add_pos + leaf.len:
    #             leaf._errors.append(self.all_errors[i])
    #             i += 1
    #         if i >= len(self.all_errors):
    #             break
    #
    # def _propagate_new_error(self, error):
    #     if self.error_propagation:
    #         for leaf in self.select(lambda nd: not nd.children, True):
    #             if leaf.pos <= error.add_pos < leaf.add_pos + leaf.len:
    #                 leaf._errors.append(error)
    #                 break
    #         else:
    #             assert False, "Error %s at pos %i out of bounds" % (str(error), error.add_pos)

    def swallow(self, node: Node) -> 'RootNode':
di68kap's avatar
di68kap committed
720
721
722
723
724
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
725
726
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
727
        self._content = node._content
728
        return self
di68kap's avatar
di68kap committed
729

eckhart's avatar
eckhart committed
730
731
    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
732
733
734
735
736
737
        self.all_errors.append(error)
        self.error_flag = max(self.error_flag, error.code)
        node.errors.append(error)
        self.err_nodes_keep.append(node)
        return self

eckhart's avatar
eckhart committed
738
    def new_error(self,
739
                  node: Node,
di68kap's avatar
di68kap committed
740
                  message: str,
di68kap's avatar
di68kap committed
741
                  code: int = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
742
        """
eckhart's avatar
eckhart committed
743
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
744
745
746
747
748
        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's avatar
eckhart committed
749
        error = Error(message, code, node=node)
eckhart's avatar
eckhart committed
750
751
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
752

eckhart's avatar
eckhart committed
753
    def collect_errors(self) -> List[Error]:
di68kap's avatar
di68kap committed
754
755
        """Returns the list of errors, ordered bv their position.
        """
eckhart's avatar
eckhart committed
756
757
758
        # for node in self.err_nodes:  # lazy evaluation of positions
        #     for err in node.errors:  # moved to error.Error.pos
        #         err.pos = node.pos
759
        self.all_errors.sort(key=lambda e: e.pos)
eckhart's avatar
eckhart committed
760
761
        for node in self.err_nodes_keep:  # redundant: consider removing Error.Error._node_keep
            for error in node.errors:
eckhart's avatar
eckhart committed
762
                assert error._pos < 0 or node.pos <= error._pos <= node.pos + len(node)
763
764
                if error._pos < 0:
                    error._pos = node.pos
eckhart's avatar
eckhart committed
765
        self.err_nodes_keep = []
766
        errors = self.all_errors
eckhart's avatar
eckhart committed
767
768
        # for error in self.all_errors:
        #     _ = error.pos
di68kap's avatar
di68kap committed
769
770
771
        return errors


eckhart's avatar
eckhart committed
772
773
774
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


775
def parse_sxpr(sxpr: str) -> Node:
776
    """
777
778
779
780
781
    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.
782
783

    Example:
784
    >>> parse_sxpr("(a (b c))").as_sxpr()
785
786
    '(a\\n    (b\\n        "c"\\n    )\\n)'
    """
787
788
    sxpr = StringView(sxpr).strip()
    mock_parsers = dict()
789

790
    def next_block(s: StringView):
eckhart's avatar
eckhart committed
791
792
793
        """Generator that yields all characters until the next closing bracket
        that does not match an opening bracket matched earlier within the same
        package."""
794
        s = s.strip()
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
        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)
814

815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
    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
862
                else:
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
                    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)


def parse_xml(xml: str) -> Node:
    """
    Generates a tree of nodes from a (Pseudo-)XML-source.
    """
    xml = StringView(xml)
881
    PlainText = MockParser('', ':PlainText')
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
    mock_parsers = {':PlainText': PlainText}

    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']
        s, attributes = parse_attributes(s[match.end() - s.begin:])
        i = s.find('>')
        assert i >= 0
906
        return s[i+1:], tagname, attributes, s[i-1] == "/"
907
908

    def parse_closing_tag(s: StringView) -> Tuple[StringView, str]:
909
        """Parses a closing tag and returns the string segment, just after
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
        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)
                if not s.match(re.compile("\s*$")):
                    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
        if len(result) == 1 and isinstance(result[0].parser == PlainText):
            result = result[0].result
        else:
            result = tuple(result)
        return Node(mock_parsers.setdefault(tagname, MockParser(name, ":" + class_name)), result)

947
948
949
    match_header = xml.search(re.compile(r'<(?!\?)'))
    start = match_header.start() if match_header else 0
    return parse_full_content(xml[start:])
950

951
# if __name__ == "__main__":
952
#     st = parse_sxpr("(alpha (beta (gamma i\nj\nk) (delta y)) (epsilon z))")
953
954
#     print(st.as_sxpr())
#     print(st.as_xml())