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
__all__ = ('ParserBase',
           'WHITESPACE_PTYPE',
di68kap's avatar
di68kap committed
38
           'PLAINTEXT_PTYPE',
39
40
41
42
           '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
#######################################################################
#
# 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.
    """
Eckhart Arnold's avatar
Eckhart Arnold committed
65
66
    __slots__ = '_name', '_ptype'

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

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

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

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

80
81
    @property
    def name(self):
82
83
        """Returns the name of the parser or the empty string '' for unnamed
        parsers."""
84
85
        return self._name

86
87
    @property
    def ptype(self) -> str:
88
89
        """Returns the type of the parser. By default this is the parser's
        class name preceded by a colon, e.g. ':ZeroOrMore'."""
90
91
92
93
        return self._ptype

    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
94
        """Returns the parser's name if it has a name and repr()"""
95
96
        return self.name if self.name else repr(self)

97
    def reset(self):
98
        """Resets any parser variables. (Should be overridden.)"""
99
100
        pass

eckhart's avatar
eckhart committed
101
    def grammar(self) -> Optional[object]:
102
103
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
104
105
106
        return None

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
107
        """Applies the function `func` to the parser. Returns False, if
108
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
109
        otherwise."""
110
111
        return False

112
113

WHITESPACE_PTYPE = ':Whitespace'
di68kap's avatar
di68kap committed
114
PLAINTEXT_PTYPE = ':PlainText'
115
116
117
118
119
120
121
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
eckhart's avatar
eckhart committed
122
    different parser types by assigning them a `ptype` on initialization.
123
124
125
126
127

    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
128
129
    __slots__ = ()

130
131
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
132
        super().__init__(name)
133
134
135
136
137
138
139
140
141
142
143
144
145
146
        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
Eckhart Arnold's avatar
Eckhart Arnold committed
147
    __slots__ = ()
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175

    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
#
#######################################################################


176
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
177
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
178
StrictResultType = Union[ChildrenType, StringView, str]
179
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
180
181


Eckhart Arnold's avatar
Eckhart Arnold committed
182
def flatten_sxpr(sxpr: str) -> str:
183
    """Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
184
185
186
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
187
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
188
189
    '(a (b c))'
    """
190
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
191
192


193
def flatten_xml(xml: str) -> str:
194
    """Returns an XML-tree as a one liner without unnecessary whitespace,
195
196
    i.e. only whitespace within leaf-nodes is preserved.
    """
197
198
199
200
201
    # works only with regex
    # return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?<=</\w+>)\s+', '', xml))
    def tag_only(m):
        return m.groupdict()['closing_tag']
    return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?P<closing_tag></\w+>)\s+', tag_only, xml))
202
203


Eckhart Arnold's avatar
Eckhart Arnold committed
204
class Node(collections.abc.Sized):
205
206
207
    """
    Represents a node in the concrete or abstract syntax tree.

208
    Attributes:
209
210
        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
211

212
213
214
        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.
215

216
217
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
218

eckhart's avatar
eckhart committed
219
220
221
222
        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
223
        parser (Parser):  The parser which generated this node.
224
225
            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
226
227
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
228
            example by calling ``isinstance(node.parer, ...)``.
229

230
231
232
        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
233
            the length before AST-transformation and will never change
234
            through AST-transformation. READ ONLY!
235

236
237
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
238
            The value of ``pos`` is -1 meaning invalid by default.
239
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
240
            nodes relative to this value.
241
242

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
243
            pos value of the root node should be set to 0 right
244
245
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
246
            Other than that, this value should be considered READ ONLY.
247
248
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
249

250
251
        errors (list):  A list of all errors that occured on this node.

252
        attributes (dict): An optional dictionary of XML-attributes. This
253
254
255
            dictionary is created lazily upon first usage. The attributes
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
256
    """
257

Eckhart Arnold's avatar
Eckhart Arnold committed
258
    __slots__ = '_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content'
259

260
261
262
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
263
264
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
265
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
266
        self._pos = -1                  # type: int
di68kap's avatar
di68kap committed
267
268
        # 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
269
        if leafhint:
eckhart's avatar
eckhart committed
270
            self._result = result       # type: StrictResultType
eckhart's avatar
eckhart committed
271
            self._content = None        # type: Optional[str]
272
            self.children = NoChildren  # type: ChildrenType
273
            self._len = -1              # type: int  # lazy evaluation
274
275
        else:
            self.result = result
276
277
        self.parser = parser or ZOMBIE_PARSER

Eckhart Arnold's avatar
Eckhart Arnold committed
278

279
    def __str__(self):
280
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
281
        if self.errors:
282
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
283
                   (s, '; '.join(e.message for e in self.errors))
284
        return s
285

Eckhart Arnold's avatar
Eckhart Arnold committed
286

287
288
289
290
291
292
    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)
293

Eckhart Arnold's avatar
Eckhart Arnold committed
294

295
    def __len__(self):
296
        if self._len < 0:
di68kap's avatar
di68kap committed
297
            self._len = sum(len(child) for child in self.children) \
298
                if self.children else len(self._result)
299
300
301
302
303
304
305
306
        return self._len


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


307
    def __eq__(self, other):
308
309
310
311
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
        """
312
        return self.tag_name == other.tag_name and self.result == other.result
313

Eckhart Arnold's avatar
Eckhart Arnold committed
314

315
    def __hash__(self):
316
        return hash(self.tag_name)
317

Eckhart Arnold's avatar
Eckhart Arnold committed
318

319
320
321
    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
322
        an integer or the first child node with the given tag name. Examples::
323

324
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
325
326
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
327
328
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
329
330
331
332

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
333
        Returns:
334
335
            Node: All nodes which have a given tag name.
        """
336
337
338
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
339
            else:
340
341
342
343
344
                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!')
345
346
347
348


    def __contains__(self, tag_name: str) -> bool:
        """
349
        Returns true if a child with the given tag name exists.
350
        Args:
351
352
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
353
354
355
356
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
357
358
359
360
361
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
362
            return False
363
        raise ValueError('Leave node cannot contain other nodes')
364
        # generator = self.select_by_tag(tag_name, False)
365
366
367
368
369
        # try:
        #     generator.__next__()
        #     return True
        # except StopIteration:
        #     return False
370
371


372
373
374
375
376
377
378
379
380
381
382
383
384
385
    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


386
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
387
    def tag_name(self) -> str:
388
389
390
391
392
393
        """
        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`.
        """
394
        return self.parser.name or self.parser.ptype
395

Eckhart Arnold's avatar
Eckhart Arnold committed
396

397
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
398
    def result(self) -> StrictResultType:
399
400
401
402
403
        """
        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.
        """
404
405
        return self._result

406

407
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
408
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
409
        # # made obsolete by static type checking with mypy
410
411
412
        # 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
413
414
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
415
416
        self._len = -1        # lazy evaluation
        self._content = None
417
418
419
420
421
422
423
424
425
        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
426
                self._result = result
427

428
429

    @property
eckhart's avatar
eckhart committed
430
    def content(self) -> str:
431
        """
432
        Returns content as string, omitting error messages.
433
        """
434
435
436
437
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
438
                # self._content = self._result
439
440
441
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
442
443
444
445
446
447
448
449
450
451
452


    @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))


453
    @property
454
    def pos(self) -> int:
eckhart's avatar
eckhart committed
455
456
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
457
            raise AssertionError("Position value not initialized!")
458
459
        return self._pos

eckhart's avatar
eckhart committed
460

eckhart's avatar
eckhart committed
461
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
462
463
464
465
466
467
468
469
470
        """
        (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
471
472
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
473
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
474
        offset = self.pos
475
        for child in self.children:
eckhart's avatar
eckhart committed
476
            child.init_pos(offset)
eckhart's avatar
eckhart committed
477
478
479
            offset = child.pos + len(child)
        return self

480

481
482
    @property
    def attributes(self):
483
484
485
        """
        Returns a dictionary of XML-Attributes attached to the Node.
        """
486
487
488
489
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
490

491
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
492
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
493
494
495
496
497
498
499
500
        """
        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.

501
        Args:
502
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
503
            open_fn:   (Node->str) A function that returns an opening
504
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
505
            close_fn:  (Node->str) A function that returns a closeF
506
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
507
            data_fn:   (str->str) A function that filters the data string
508
509
510
511
512
513
                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
514
515
        head = open_fn(self)
        tail = close_fn(self)
516
517

        if not self.result:
518
            return head.rstrip() + tail.lstrip()
519

eckhart's avatar
eckhart committed
520
        tail = tail.lstrip(None if density & 2 else '')
521

522
        inline = inline or inline_fn(self)
523
524
525
526
527
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
528
529
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
530

531
532
        if self.children:
            content = []
533
            for child in self.children:
534
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
535
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
536
537
538
                if subtree:
                    subtree = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in subtree))
539
            return head + usetab + (sep + usetab).join(content) + tail
540

541
        res = cast(str, self.result)  # safe, because if there are no children, result is a string
eckhart's avatar
eckhart committed
542
        if not inline and not head:
eckhart's avatar
eckhart committed
543
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
544
            res = res.strip()
545
546
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
547
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
548
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
549
        else:
550
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
551

Eckhart Arnold's avatar
Eckhart Arnold committed
552

553
554
555
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
556
                compact: bool = False) -> str:
557
558
559
        """
        Returns content as S-expression, i.e. in lisp-like form.

560
        Args:
561
562
563
            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.
564
565
566
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
567
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
568
                tree structure.
569
570
        """

eckhart's avatar
eckhart committed
571
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
572
        lbreaks = linebreaks(src) if src else []  # type: List[int]
573

574
        def opening(node) -> str:
eckhart's avatar
eckhart committed
575
            """Returns the opening string for the representation of `node`."""
576
            txt = [left_bracket,  node.tag_name]
577
            # s += " '(pos %i)" % node.add_pos
578
            if hasattr(node, '_xml_attr'):
579
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attributes.items())
580
            if src:
eckhart's avatar
eckhart committed
581
                txt.append(" `(pos %i %i %i)" % (node.pos, *line_col(lbreaks, node.pos)))
582
            if showerrors and node.errors:
583
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
584
            return "".join(txt) + '\n'
585

eckhart's avatar
eckhart committed
586
587
588
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
589

eckhart's avatar
eckhart committed
590
591
592
593
594
        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'\"')
595

596
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
597

eckhart's avatar
eckhart committed
598

599
600
601
602
603
604
    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:
605
606
607
        """
        Returns content as XML-tree.

608
        Args:
609
610
611
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
612
613
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
614
615
616
617
618
619
620
            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.
621
622
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
623
624
        """

625
        def opening(node) -> str:
626
627
628
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
629
            txt = ['<', node.tag_name]
630
631
            has_reserved_attrs = hasattr(node, '_xml_attr') \
                and any (r in node.attributes for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
632
            if hasattr(node, '_xml_attr'):
633
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attributes.items())
634
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
635
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
636
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
637
638
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
639
640
641
            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
642
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
643
644
645
            else:
                ending = ">\n"
            return "".join(txt + [ending])
646
647

        def closing(node):
648
            """Returns the closing string for the representation of `node`."""
649
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
650
                return ''
651
            return ('\n</') + node.tag_name + '>'
652

653
654
655
656
657
        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.
            """
658
659
            return node.tag_name in inline_tags or (hasattr(node, '_xml_attr') \
                and node.attributes.get('xml:space', 'default') == 'preserve')
660

661
        line_breaks = linebreaks(src) if src else []
662
663
        return self._tree_repr(' ' * indentation, opening, closing,
                               density=1, inline_fn=inlining)
664

Eckhart Arnold's avatar
Eckhart Arnold committed
665

666
    def select(self, match_function: Callable, include_root: bool=False) -> Iterator['Node']:
eckhart's avatar
eckhart committed
667
        """
668
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
669

670
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
671
        given `match_function` evaluates to True. The tree is
672
673
674
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
675

676
677
678
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
679
680
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
681
        Yields:
682
            Node: All nodes of the tree for which
683
684
            ``match_function(node)`` returns True
        """
685
        if include_root and match_function(self):
686
            yield self
eckhart's avatar
eckhart committed
687
688
689
        for child in self.children:
            for node in child.select(match_function, True):
                yield node
690

Eckhart Arnold's avatar
Eckhart Arnold committed
691

eckhart's avatar
eckhart committed
692
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
693
                      include_root: bool=False) -> Iterator['Node']:
694
        """
695
696
697
698
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
699

700
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
701
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
702
            ['(X (c "d"))', '(X "F")']
703
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
704
            ['(b "X")', '(X (c "d"))', '(X "F")']
705
            >>> any(tree.select_by_tag('a', False))
706
            False
707
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
708
            ['(a (b "X") (X (c "d")) (e (X "F")))']
709
710
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
711
712

        Args:
713
714
715
716
            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.
717
718
719
        Yields:
            Node: All nodes which have a given tag name.
        """
720
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
721
            tag_names = frozenset({tag_names})
722
        return self.select(lambda node: node.tag_name in tag_names, include_root)
723
724


725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
    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


740
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
741
742
743
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
744
745
746
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
747
class RootNode(Node):
748
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
749

Eckhart Arnold's avatar
Eckhart Arnold committed
750
        all_errors (list):  A list of all errors that have occured so far during
751
752
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
753

754
755
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
756
    """
eckhart's avatar
eckhart committed
757
    def __init__(self, node: Optional[Node] = None) -> 'RootNode':
di68kap's avatar
di68kap committed
758
        super().__init__(ZOMBIE_PARSER, '')
759
        self.all_errors = []
di68kap's avatar
di68kap committed
760
        self.error_flag = 0
eckhart's avatar
eckhart committed
761
762
        if node is not None:
            self.swallow(node)
763
764
765
766
        # customization for XML-Representation
        self.inline_tags = set()
        self.omit_tags = set()
        self.empty_tags = set()
di68kap's avatar
di68kap committed
767

768
    def swallow(self, node: Node) -> 'RootNode':
769
770
771
772
773
774
775
776
777
778
        """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
779
780
781
782
783
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
784
785
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
786
        self._content = node._content
787
        return self
di68kap's avatar
di68kap committed
788

eckhart's avatar
eckhart committed
789
790
    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
791
792
793
794
795
        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
796
    def new_error(self,
797
                  node: Node,
di68kap's avatar
di68kap committed
798
                  message: str,
di68kap's avatar
di68kap committed
799
                  code: int = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
800
        """
eckhart's avatar
eckhart committed
801
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
802
803
804
805
806
        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
807
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
808
809
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
810

eckhart's avatar
eckhart committed
811
    def collect_errors(self) -> List[Error]:
di68kap's avatar
di68kap committed
812
813
        """Returns the list of errors, ordered bv their position.
        """
814
        self.all_errors.sort(key=lambda e: e.pos)
Eckhart Arnold's avatar
Eckhart Arnold committed
815
        return self.all_errors
di68kap's avatar
di68kap committed
816

817
818
819
820
821
822
823
824
    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
825

eckhart's avatar
eckhart committed
826
827
828
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


829
def parse_sxpr(sxpr: str) -> Node:
830
    """
831
832
833
834
835
    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.
836
837

    Example:
838
    >>> parse_sxpr("(a (b c))").as_sxpr()
839
    '(a\\n  (b\\n    "c"\\n  )\\n)'
840
    """
841
842
    sxpr = StringView(sxpr).strip()
    mock_parsers = dict()
843

844
    def next_block(s: StringView):
eckhart's avatar
eckhart committed
845
846
847
        """Generator that yields all characters until the next closing bracket
        that does not match an opening bracket matched earlier within the same
        package."""
848
        s = s.strip()
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
        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)
868

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
    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
916
                else:
917
918
919
920
921
922
923
924
925
926
927
928
929
                    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
930
931
932
RX_WHITESPACE_TAIL = re.compile(r'\s*$')


933
934
935
936
937
def parse_xml(xml: str) -> Node:
    """
    Generates a tree of nodes from a (Pseudo-)XML-source.
    """
    xml = StringView(xml)
di68kap's avatar
di68kap committed
938
939
    PlainText = MockParser('', PLAINTEXT_PTYPE)
    mock_parsers = {PLAINTEXT_PTYPE: PlainText}
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959

    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
960
961
        section = s[match.end() - s.begin:]
        s, attributes = parse_attributes(section)
962
963
        i = s.find('>')
        assert i >= 0
964
        return s[i+1:], tagname, attributes, s[i-1] == "/"
965
966

    def parse_closing_tag(s: StringView) -> Tuple[StringView, str]:
967
        """Parses a closing tag and returns the string segment, just after
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
        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
992
                if not leaf.match(RX_WHITESPACE_TAIL):
993
994
995
996
997
998
                    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
di68kap's avatar
di68kap committed
999
        if len(result) == 1 and result[0].parser.ptype == PLAINTEXT_PTYPE:
Eckhart Arnold's avatar