syntaxtree.py 39.8 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
eckhart's avatar
eckhart committed
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
#######################################################################
#
# 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.
    """
65

66
    __slots__ = 'name', 'ptype'
Eckhart Arnold's avatar
Eckhart Arnold committed
67

68
69
70
    def __init__(self,):  # , pbases=frozenset()):
        self.name = ''  # type: str
        self.ptype = ':' + self.__class__.__name__  # type: str
71
72

    def __repr__(self):
di68kap's avatar
di68kap committed
73
         return self.name + self.ptype
74
75
76
77

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

78
79
80
    def __call__(self, text: StringView) -> Tuple[Optional['Node'], StringView]:
        return None, text

81
82
    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
83
        """Returns the parser's name if it has a name and repr()"""
84
85
        return self.name if self.name else repr(self)

86
    def reset(self):
87
        """Resets any parser variables. (Should be overridden.)"""
88
89
        pass

eckhart's avatar
eckhart committed
90
    def grammar(self) -> Optional[object]:
91
92
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
93
94
95
        return None

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
96
        """Applies the function `func` to the parser. Returns False, if
97
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
98
        otherwise."""
99
100
        return False

101
102

WHITESPACE_PTYPE = ':Whitespace'
103
TOKEN_PTYPE = ':Token'
104
105
106
107
108
109


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
110
    different parser types by assigning them a `ptype` on initialization.
111
112
113
114
115

    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.
    """
116

Eckhart Arnold's avatar
Eckhart Arnold committed
117
118
    __slots__ = ()

119
120
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
121
122
        super().__init__()
        self.name = name
di68kap's avatar
di68kap committed
123
124
        if ptype:
            self.ptype = ptype  # or ':' + self.__class__.__name__
125
126
127
128
129
130
131
132
133
134
135
136


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.
    """
137

138
    alive = False
Eckhart Arnold's avatar
Eckhart Arnold committed
139
    __slots__ = ()
140
141

    def __init__(self):
142
        super(ZombieParser, self).__init__()
143
144
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
145
        self.name = "__ZOMBIE__"
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
        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
#
#######################################################################


169
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
170
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
171
StrictResultType = Union[ChildrenType, StringView, str]
172
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
173
174


Eckhart Arnold's avatar
Eckhart Arnold committed
175
def flatten_sxpr(sxpr: str) -> str:
176
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

185
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
186
187


188
def flatten_xml(xml: str) -> str:
189
190
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
191
    i.e. only whitespace within leaf-nodes is preserved.
192
193
    A more precise alternative to `flatten_xml` is to use Node.as_xml()
    ans passing a set containing the top level tag to parameter `inline_tags`.
194
    """
195

196
197
198
199
    # 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']
200
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
201
202


203
204
205
RX_AMP = re.compile('&(?!\w+;)')


Eckhart Arnold's avatar
Eckhart Arnold committed
206
class Node(collections.abc.Sized):
207
208
209
    """
    Represents a node in the concrete or abstract syntax tree.

210
211
    TODO: Add some documentation and doc-tests here...

212
    Attributes:
213
214
        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
215

216
217
218
        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.
219

220
221
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
222

eckhart's avatar
eckhart committed
223
224
225
226
        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
227
        parser (Parser):  The parser which generated this node.
228
229
            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
230
231
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
232
            example by calling ``isinstance(node.parer, ...)``.
233

234
235
236
        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
237
            the length before AST-transformation and will never change
238
            through AST-transformation. READ ONLY!
239

240
241
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
242
            The value of ``pos`` is -1 meaning invalid by default.
243
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
244
            nodes relative to this value.
245
246

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
247
            pos value of the root node should be set to 0 right
248
249
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
250
            Other than that, this value should be considered READ ONLY.
251
252
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
253

254
255
        errors (list):  A list of all errors that occured on this node.

256
257
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
258
259
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
260
    """
261

Eckhart Arnold's avatar
Eckhart Arnold committed
262
    __slots__ = '_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content'
263

264
265
266
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
267
268
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
269
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
270
        self._pos = -1                  # type: int
271
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
272
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
273
        if leafhint:
eckhart's avatar
eckhart committed
274
            self._result = result       # type: StrictResultType
eckhart's avatar
eckhart committed
275
            self._content = None        # type: Optional[str]
276
            self.children = NoChildren  # type: ChildrenType
277
            self._len = -1              # type: int  # lazy evaluation
278
279
        else:
            self.result = result
280
281
        self.parser = parser or ZOMBIE_PARSER

eckhart's avatar
eckhart committed
282
283
284
285
286
287
288
289
290
291
292
    def __deepcopy__(self, memo):
        if self.children:
            duplicate = self.__class__(self.parser, copy.deepcopy(self.children), False)
        else:
            duplicate = self.__class__(self.parser, self.result, True)
        duplicate.errors = copy.deepcopy(self.errors) if self.errors else []
        duplicate._pos = self._pos
        duplicate._len = self._len
        if hasattr(self, '_xml_attr'):
            duplicate._xml_attr = copy.deepcopy(self._xml_attr)
        return duplicate
Eckhart Arnold's avatar
Eckhart Arnold committed
293

294
    def __str__(self):
295
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
296
        if self.errors:
297
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
298
                   (s, '; '.join(e.message for e in self.errors))
299
        return s
300

Eckhart Arnold's avatar
Eckhart Arnold committed
301

302
303
304
305
306
307
    def __repr__(self):
        mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
        parg = "MockParser({name}, {ptype})".format(**mpargs)
        rarg = str(self) if not self.children else \
               "(" + ", ".join(repr(child) for child in self.children) + ")"
        return "Node(%s, %s)" % (parg, rarg)
308

Eckhart Arnold's avatar
Eckhart Arnold committed
309

310
    def __len__(self):
311
        if self._len < 0:
di68kap's avatar
di68kap committed
312
            self._len = sum(len(child) for child in self.children) \
313
                if self.children else len(self._result)
314
315
316
317
318
319
320
321
        return self._len


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


322
    def __eq__(self, other):
323
324
325
326
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
        """
327
        return self.tag_name == other.tag_name and self.result == other.result
328

Eckhart Arnold's avatar
Eckhart Arnold committed
329

330
    def __hash__(self):
331
        return hash(self.tag_name)
332

Eckhart Arnold's avatar
Eckhart Arnold committed
333

334
335
336
    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
337
        an integer or the first child node with the given tag name. Examples::
338

339
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
340
341
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
342
343
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
344
345
346
347

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
348
        Returns:
349
350
            Node: All nodes which have a given tag name.
        """
351
352
353
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
354
            else:
355
356
357
358
359
                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!')
360
361
362
363


    def __contains__(self, tag_name: str) -> bool:
        """
364
        Returns true if a child with the given tag name exists.
365
        Args:
366
367
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
368
369
370
371
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
372
373
374
375
376
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
377
            return False
378
        raise ValueError('Leave node cannot contain other nodes')
379
380


381
382
383
384
385
386
387
388
389
390
391
392
393
394
    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


395
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
396
    def tag_name(self) -> str:
397
398
399
400
401
402
        """
        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`.
        """
403
        return self.parser.name or self.parser.ptype
404

Eckhart Arnold's avatar
Eckhart Arnold committed
405

406
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
407
    def result(self) -> StrictResultType:
408
409
410
411
412
        """
        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.
        """
413
414
        return self._result

415

416
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
417
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
418
        # # made obsolete by static type checking with mypy
419
420
421
        # 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
422
423
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
424
425
        self._len = -1        # lazy evaluation
        self._content = None
426
427
428
429
430
431
432
433
434
        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
435
                self._result = result
436

437
438

    @property
eckhart's avatar
eckhart committed
439
    def content(self) -> str:
440
        """
di68kap's avatar
di68kap committed
441
442
443
        Returns content as string, omitting error messages. If the node has
        child-nodes, the string content of the child-nodes is recursively read
        and then concatenated.
444
        """
445
446
447
448
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
449
                # self._content = self._result
450
451
452
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
453
454


di68kap's avatar
di68kap committed
455
456
457
458
459
    @content.setter
    def content(self, content: str):
        self.result = content


460
461
462
463
464
465
466
467
468
    @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))


469
    @property
470
    def pos(self) -> int:
eckhart's avatar
eckhart committed
471
472
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
473
            raise AssertionError("Position value not initialized!")
474
475
        return self._pos

eckhart's avatar
eckhart committed
476

eckhart's avatar
eckhart committed
477
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
478
479
480
481
482
483
484
485
486
        """
        (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
487
488
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
489
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
490
        offset = self.pos
491
        for child in self.children:
eckhart's avatar
eckhart committed
492
            child.init_pos(offset)
eckhart's avatar
eckhart committed
493
494
495
            offset = child.pos + len(child)
        return self

496

497
    @property
498
    def attr(self):
499
        """
500
        Returns a dictionary of XML-attr attached to the node.
501
        """
502
503
504
505
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
506

507
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
508
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
509
510
511
512
513
514
515
516
        """
        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.

517
        Args:
518
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
519
            open_fn:   (Node->str) A function that returns an opening
520
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
521
            close_fn:  (Node->str) A function that returns a closeF
522
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
523
            data_fn:   (str->str) A function that filters the data string
524
525
526
527
528
529
                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
530
531
        head = open_fn(self)
        tail = close_fn(self)
532
533

        if not self.result:
534
            return head.rstrip() + tail.lstrip()
535

eckhart's avatar
eckhart committed
536
        tail = tail.lstrip(None if density & 2 else '')
537

538
        inline = inline or inline_fn(self)
539
540
541
542
543
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
544
545
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
546

547
548
        if self.children:
            content = []
549
            for child in self.children:
550
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
551
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
552
553
554
                if subtree:
                    subtree = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in subtree))
555
            return head + usetab + (sep + usetab).join(content) + tail
556

557
        res = self.content  # cast(str, self.result)  # safe, because if there are no children, result is a string
eckhart's avatar
eckhart committed
558
        if not inline and not head:
eckhart's avatar
eckhart committed
559
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
560
            res = res.strip()
561
562
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
563
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
564
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
565
        else:
566
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
567

Eckhart Arnold's avatar
Eckhart Arnold committed
568

569
570
571
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
572
                compact: bool = False) -> str:
573
574
575
        """
        Returns content as S-expression, i.e. in lisp-like form.

576
        Args:
577
578
579
            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.
580
581
582
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
583
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
584
                tree structure.
585
586
        """

eckhart's avatar
eckhart committed
587
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
588
        lbreaks = linebreaks(src) if src else []  # type: List[int]
589

590
        def opening(node) -> str:
eckhart's avatar
eckhart committed
591
            """Returns the opening string for the representation of `node`."""
592
            txt = [left_bracket,  node.tag_name]
593
            # s += " '(pos %i)" % node.add_pos
594
            if hasattr(node, '_xml_attr'):
595
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
596
            if src:
597
598
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
599
            if showerrors and node.errors:
600
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
601
            return "".join(txt) + '\n'
602

eckhart's avatar
eckhart committed
603
604
605
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
606

eckhart's avatar
eckhart committed
607
608
609
610
611
        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'\"')
612

613
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
614

eckhart's avatar
eckhart committed
615

616
617
618
619
620
621
    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:
622
623
624
        """
        Returns content as XML-tree.

625
        Args:
626
627
628
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
629
630
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
631
632
633
            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
634
                neither the opening tag nor its attr nor the closing tag. This
635
636
637
                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.
638
639
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
640
641
        """

642
        def opening(node) -> str:
643
644
645
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
646
            txt = ['<', node.tag_name]
647
            has_reserved_attrs = hasattr(node, '_xml_attr') \
648
                and any (r in node.attr for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
649
            if hasattr(node, '_xml_attr'):
650
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
651
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
652
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
653
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
654
655
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
656
657
658
            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
659
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
660
661
662
            else:
                ending = ">\n"
            return "".join(txt + [ending])
663
664

        def closing(node):
665
            """Returns the closing string for the representation of `node`."""
666
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
667
                return ''
668
            return ('\n</') + node.tag_name + '>'
669

670
671
672
673
674
675
676
        def sanitizer(content: str) -> str:
            """Substitute "&", "<", ">" in XML-content by the respective entities."""
            content = RX_AMP.sub('&amp;', content)
            content = content.replace('<', '&lt;').replace('>', '&gt;')
            return content


677
678
679
680
681
        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.
            """
682
            return node.tag_name in inline_tags or (hasattr(node, '_xml_attr') \
683
                                                    and node.attr.get('xml:space', 'default') == 'preserve')
684

685
        line_breaks = linebreaks(src) if src else []
686
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
687
                               density=1, inline_fn=inlining)
688

Eckhart Arnold's avatar
Eckhart Arnold committed
689

690
691
    def select(self, match_function: Callable, include_root: bool=False, reverse: bool=False) \
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
692
        """
693
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
694

695
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
696
        given `match_function` evaluates to True. The tree is
697
698
699
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
700

701
702
703
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
704
705
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
706
707
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
708
        Yields:
709
            Node: All nodes of the tree for which
710
711
            ``match_function(node)`` returns True
        """
712
        if include_root and match_function(self):
713
            yield self
714
715
716
        child_iterator = reversed(self.children) if reverse else self.children
        for child in child_iterator:
            for node in child.select(match_function, True, reverse):
eckhart's avatar
eckhart committed
717
                yield node
718

Eckhart Arnold's avatar
Eckhart Arnold committed
719

eckhart's avatar
eckhart committed
720
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
721
                      include_root: bool=False) -> Iterator['Node']:
722
        """
723
724
725
726
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
727

728
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
729
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
730
            ['(X (c "d"))', '(X "F")']
731
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
732
            ['(b "X")', '(X (c "d"))', '(X "F")']
733
            >>> any(tree.select_by_tag('a', False))
734
            False
735
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
736
            ['(a (b "X") (X (c "d")) (e (X "F")))']
737
738
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
739
740

        Args:
741
742
743
744
            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.
745
746
747
        Yields:
            Node: All nodes which have a given tag name.
        """
748
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
749
            tag_names = frozenset({tag_names})
750
        return self.select(lambda node: node.tag_name in tag_names, include_root)
751
752


753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
    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


768
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
769
770
771
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
772
773
774
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
775
776
777
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


di68kap's avatar
di68kap committed
778
class RootNode(Node):
779
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
780

Eckhart Arnold's avatar
Eckhart Arnold committed
781
        all_errors (list):  A list of all errors that have occured so far during
782
783
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
784

785
786
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
787
    """
788

eckhart's avatar
eckhart committed
789
    def __init__(self, node: Optional[Node] = None) -> 'RootNode':
di68kap's avatar
di68kap committed
790
        super().__init__(ZOMBIE_PARSER, '')
791
        self.all_errors = []
di68kap's avatar
di68kap committed
792
        self.error_flag = 0
eckhart's avatar
eckhart committed
793
794
        if node is not None:
            self.swallow(node)
795
796
797
798
        # customization for XML-Representation
        self.inline_tags = set()
        self.omit_tags = set()
        self.empty_tags = set()
di68kap's avatar
di68kap committed
799

800
    def swallow(self, node: Node) -> 'RootNode':
801
802
        """
        Put `self` in the place of `node` by copying all its data.
803
804
805
806
807
808
809
810
811
        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
812
813
814
815
816
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
817
818
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
819
        self._content = node._content
820
        return self
di68kap's avatar
di68kap committed
821

eckhart's avatar
eckhart committed
822
    def add_error(self, node: Node, error: Error) -> 'RootNode':
823
824
825
        """
        Adds an Error object to the tree, locating it at a specific node.
        """
di68kap's avatar
di68kap committed
826
827
828
829
830
        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
831
    def new_error(self,
832
                  node: Node,
di68kap's avatar
di68kap committed
833
                  message: str,
di68kap's avatar
di68kap committed
834
                  code: int = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
835
        """
eckhart's avatar
eckhart committed
836
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
837
838
839
840
841
        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
842
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
843
844
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
845

eckhart's avatar
eckhart committed
846
    def collect_errors(self) -> List[Error]:
847
848
        """
        Returns the list of errors, ordered bv their position.
di68kap's avatar
di68kap committed
849
        """
850
        self.all_errors.sort(key=lambda e: e.pos)
Eckhart Arnold's avatar
Eckhart Arnold committed
851
        return self.all_errors
di68kap's avatar
di68kap committed
852

853
    def customized_XML(self):
854
855
        """
        Returns a customized XML representation of the tree.
856
        See the docstring of `Node.as_xml()` for an explanation of the
857
858
        customizations.
        """
859
860
861
862
        return self.as_xml(inline_tags = self.inline_tags,
                           omit_tags=self.omit_tags,
                           empty_tags=self.empty_tags)

di68kap's avatar
di68kap committed
863

di68kap's avatar
di68kap committed
864
865
866
867
868
#######################################################################
#
# S-expression- and XML-parsers
#
#######################################################################
eckhart's avatar
eckhart committed
869
870


871
def parse_sxpr(sxpr: str) -> Node:
872
    """
873
874
875
876
877
    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.
878
879

    Example:
880
    >>> parse_sxpr("(a (b c))").as_sxpr()
881
    '(a\\n  (b\\n    "c"\\n  )\\n)'
882
    """
883

884
885
    sxpr = StringView(sxpr).strip()
    mock_parsers = dict()
886

887
    def next_block(s: StringView):
eckhart's avatar
eckhart committed
888
889
        """Generator that yields all characters until the next closing bracket
        that does not match an opening bracket matched earlier within the same
890
891
        package.
        """
892
        s = s.strip()
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
        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)
912

913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
    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] != ')':
932
                # parse attr
933
934
935
936
937
938