syntaxtree.py 41.9 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
from collections import OrderedDict
eckhart's avatar
eckhart committed
27
import copy
28

eckhart's avatar
eckhart committed
29
from DHParser.error import Error, ErrorCode, linebreaks, line_col
30
from DHParser.stringview import StringView
di68kap's avatar
di68kap committed
31
32
from DHParser.toolkit import re
from typing import Callable, cast, Iterator, List, AbstractSet, Set, Union, Tuple, Optional
di68kap's avatar
di68kap committed
33

34

di68kap's avatar
di68kap committed
35
__all__ = ('WHITESPACE_PTYPE',
36
           'TOKEN_PTYPE',
37
38
           'ZOMBIE_TAG',
           'PLACEHOLDER',
eckhart's avatar
eckhart committed
39
40
41
           'ResultType',
           'StrictResultType',
           'ChildrenType',
42
           'Node',
43
           'FrozenNode',
di68kap's avatar
di68kap committed
44
           'RootNode',
45
           'parse_sxpr',
46
47
48
           'parse_xml',
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
49
50


51
52
53
54
55
56
57
58
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


WHITESPACE_PTYPE = ':Whitespace'
59
TOKEN_PTYPE = ':Token'
60

61
ZOMBIE_TAG = "__ZOMBIE__"
62

63
64
65
66
67
68
69
#######################################################################
#
# syntaxtree nodes
#
#######################################################################


70
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
71
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
72
StrictResultType = Union[ChildrenType, StringView, str]
73
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
74
75


Eckhart Arnold's avatar
Eckhart Arnold committed
76
def flatten_sxpr(sxpr: str) -> str:
77
78
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
79
80
81
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
82
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
83
84
    '(a (b c))'
    """
85

86
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
87
88


89
def flatten_xml(xml: str) -> str:
90
91
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
92
    i.e. only whitespace within leaf-nodes is preserved.
93
94
    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`.
95
    """
96

97
98
99
100
    # 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']
101
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
102
103


eckhart's avatar
eckhart committed
104
RX_AMP = re.compile(r'&(?!\w+;)')
105
106


Eckhart Arnold's avatar
Eckhart Arnold committed
107
class Node:  # (collections.abc.Sized): Base class omitted for cython-compatibility
108
109
110
    """
    Represents a node in the concrete or abstract syntax tree.

111
112
    TODO: Add some documentation and doc-tests here...

113
    Attributes:
114
115
        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
116

117
118
119
        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.
120

121
122
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
123

eckhart's avatar
eckhart committed
124
125
126
127
        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
128
        parser (Parser):  The parser which generated this node.
129
130
            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
131
132
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
133
            example by calling ``isinstance(node.parer, ...)``.
134

135
136
137
        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
138
            the length before AST-transformation and will never change
139
            through AST-transformation. READ ONLY!
140

141
142
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
143
            The value of ``pos`` is -1 meaning invalid by default.
144
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
145
            nodes relative to this value.
146
147

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
148
            pos value of the root node should be set to 0 right
149
150
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
151
            Other than that, this value should be considered READ ONLY.
152
153
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
154

155
156
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
157
158
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
159
    """
160

161
    __slots__ = '_result', 'children', '_len', '_pos', 'tag_name', '_xml_attr', '_content'
162

163
    def __init__(self, tag_name: str, result: ResultType, leafhint: bool = False) -> None:
164
165
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
166
167
        that generated the node and the parser's result.
        """
eckhart's avatar
eckhart committed
168
        self._pos = -1                  # type: int
169
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
170
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
171
        if leafhint:
eckhart's avatar
eckhart committed
172
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
eckhart's avatar
eckhart committed
173
            self._content = None        # type: Optional[str]
174
            self.children = NoChildren  # type: ChildrenType
175
            self._len = -1              # type: int  # lazy evaluation
176
177
        else:
            self.result = result
178
        self.tag_name = tag_name        # type: str
179

eckhart's avatar
eckhart committed
180
181
    def __deepcopy__(self, memo):
        if self.children:
182
            duplicate = self.__class__(self.tag_name, copy.deepcopy(self.children), False)
eckhart's avatar
eckhart committed
183
        else:
184
            duplicate = self.__class__(self.tag_name, self.result, True)
eckhart's avatar
eckhart committed
185
186
        duplicate._pos = self._pos
        duplicate._len = self._len
eckhart's avatar
eckhart committed
187
        if self.attr_active():
di68kap's avatar
di68kap committed
188
189
            duplicate.attr.update(copy.deepcopy(self._xml_attr))
            # duplicate._xml_attr = copy.deepcopy(self._xml_attr)  # this is not cython compatible
eckhart's avatar
eckhart committed
190
        return duplicate
Eckhart Arnold's avatar
Eckhart Arnold committed
191

192
    def __str__(self):
193
194
        if isinstance(self, RootNode):
            root = cast(RootNode, self)
195
            errors = root.errors_sorted
196
197
198
199
200
201
            if errors:
                e_pos = errors[0].pos
                return self.content[:e_pos] + \
                   ' <<< Error on "%s" | %s >>> ' % \
                   (self.content[e_pos - self.pos:], '; '.join(e.message for e in errors))
        return self.content
Eckhart Arnold's avatar
Eckhart Arnold committed
202

203
    def __repr__(self):
204
        # mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
205
206
        # name, ptype = (self._tag_name.split(':') + [''])[:2]
        # parg = "MockParser({name}, {ptype})".format(name=name, ptype=ptype)
207
        rarg = str(self) if not self.children else \
208
            "(" + ", ".join(child.__repr__() for child in self.children) + ")"
209
        return "Node(%s, %s)" % (self.tag_name, rarg)
210

Eckhart Arnold's avatar
Eckhart Arnold committed
211

212
    def __len__(self):
213
        if self._len < 0:
di68kap's avatar
di68kap committed
214
            self._len = sum(len(child) for child in self.children) \
215
                if self.children else len(self._result)
216
217
218
219
220
221
222
223
        return self._len


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


224
    def __eq__(self, other):
225
226
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
227
228
        name is the same, if their results are equal and if their attributes
        and attribute values are the same.
229
        """
230
231
        return self.tag_name == other.tag_name and self.result == other.result \
            and self.compare_attr(other)
232

Eckhart Arnold's avatar
Eckhart Arnold committed
233

234
    def __hash__(self):
235
        return hash(self.tag_name)
236

Eckhart Arnold's avatar
Eckhart Arnold committed
237

238
239
240
    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
241
        an integer or the first child node with the given tag name. Examples::
242

243
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
244
245
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
246
247
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
248
249
250
251

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
252
        Returns:
253
254
            Node: All nodes which have a given tag name.
        """
255
256
257
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
258
            else:
259
260
261
262
263
                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!')
264
265
266
267


    def __contains__(self, tag_name: str) -> bool:
        """
268
        Returns true if a child with the given tag name exists.
269
        Args:
270
271
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
272
273
274
275
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
276
277
278
279
280
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
281
            return False
282
        raise ValueError('Leave node cannot contain other nodes')
283
284


285
286
287
288
289
290
291
292
293
294
295
296
297
298
    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


299
    def is_anonymous(self):
Eckhart Arnold's avatar
Eckhart Arnold committed
300
        return not self.tag_name or self.tag_name[0] == ':'
301

Eckhart Arnold's avatar
Eckhart Arnold committed
302

303
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
304
    def result(self) -> StrictResultType:
305
306
307
308
309
        """
        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.
        """
310
311
        return self._result

312

313
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
314
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
315
        # # made obsolete by static type checking with mypy
316
317
318
        # 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
319
320
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
321
322
        self._len = -1        # lazy evaluation
        self._content = None
323
324
325
326
327
328
329
330
331
        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
eckhart's avatar
eckhart committed
332
                self._result = result  # cast(StrictResultType, result)
333

334
335

    @property
eckhart's avatar
eckhart committed
336
    def content(self) -> str:
337
        """
338
339
340
        Returns content as string. If the node has child-nodes, the
        string content of the child-nodes is recursively read and then
        concatenated.
341
        """
342
343
344
345
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
346
                # self._content = self._result
347
348
349
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
350
351
352
353
354
    #
    #
    # @content.setter
    # def content(self, content: str):
    #     self.result = content
di68kap's avatar
di68kap committed
355
356


357
358
359
360
361
362
    @property
    def structure(self) -> str:
        """
        Return structure (and content) as S-expression on a single line
        without any line breaks.
        """
363
        return flatten_sxpr(self.as_sxpr())
364
365


366
    @property
367
    def pos(self) -> int:
eckhart's avatar
eckhart committed
368
369
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
Eckhart Arnold's avatar
Eckhart Arnold committed
370
            raise AssertionError("Position value not initialized! Use Node.with_pos()")
371
372
        return self._pos

eckhart's avatar
eckhart committed
373

Eckhart Arnold's avatar
Eckhart Arnold committed
374
    def with_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
375
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
376
        Initialize position value. Usually, the parser guard
eckhart's avatar
eckhart committed
377
378
        (`parsers.add_parser_guard()`) takes care of assigning the
        position in the document to newly created nodes. However,
Eckhart Arnold's avatar
Eckhart Arnold committed
379
        when Nodes are created outside the reach of the parser
eckhart's avatar
eckhart committed
380
        guard, their document-position must be assigned manually.
Eckhart Arnold's avatar
Eckhart Arnold committed
381
382
        Position values of the child nodes are assigned recursively, too.
        Returns the node itself for convenience.
eckhart's avatar
eckhart committed
383
        """
384
385
386
387
        # condition self.pos == pos cannot be assumed when tokens or whitespace
        # are dropped early!
        # assert self._pos < 0 or self.pos == pos, ("pos mismatch %i != %i at Node: %s"
        #                                           % (self._pos, pos, repr(self)))
Eckhart Arnold's avatar
Eckhart Arnold committed
388
389
390
391
392
393
394
395
396
397
        if pos != self._pos >= 0:
            raise AssertionError("Position value cannot be reassigned to a different value!")
        if self._pos < 0:
            self._pos = pos
            # recursively adjust pos-values of all children
            offset = self.pos
            for child in self.children:
                if child._pos < 0:
                    child.with_pos(offset)
                offset = child.pos + len(child)
eckhart's avatar
eckhart committed
398
399
        return self

400

401
402
403
404
405
406
407
408
409
410
411
412
413
    @property
    def attr(self):
        """
        Returns a dictionary of XML-attr attached to the node.
        """
        try:
            if self._xml_attr is None:          # cython compatibility
                self._xml_attr = OrderedDict()
        except AttributeError:
            self._xml_attr = OrderedDict()
        return self._xml_attr


eckhart's avatar
eckhart committed
414
415
416
417
418
419
420
421
422
423
424
425
426
    def attr_active(self) -> bool:
        """
        Returns True, if XML-Attributes of this node have ever been set
        or queried, even if unsuccessfully.
        """
        try:
            if self._xml_attr is not None:
                return True
        except AttributeError:
            pass
        return False


427
    def compare_attr(self, other: 'Node') -> bool:
428
        """
429
430
        Returns True, if `self` and `other` have the same attributes with the
        same attribute values.
431
        """
432
433
434
435
436
437
438
439
440
        if self.attr_active():
            if other.attr_active():
                return self.attr == other.attr
            return len(self.attr) == 0
            # self has empty dictionary and other has no attributes
        elif other.attr_active():
            return len(other.attr) == 0
            # other has empty attribute dictionary and self as no attributes
        return True  # neither self nor other have any attributes
441

Eckhart Arnold's avatar
Eckhart Arnold committed
442

443
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
444
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
445
446
447
448
449
450
451
452
        """
        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.

453
        Args:
454
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
455
            open_fn:   (Node->str) A function that returns an opening
456
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
457
            close_fn:  (Node->str) A function that returns a closeF
458
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
459
            data_fn:   (str->str) A function that filters the data string
460
461
462
463
464
465
                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
466
467
        head = open_fn(self)
        tail = close_fn(self)
468
469

        if not self.result:
470
            return head.rstrip() + tail.lstrip()
471

eckhart's avatar
eckhart committed
472
        tail = tail.lstrip(None if density & 2 else '')
473

474
        inline = inline or inline_fn(self)
475
476
477
478
479
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
480
481
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
482

483
484
        if self.children:
            content = []
485
            for child in self.children:
486
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
487
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
488
                if subtree:
eckhart's avatar
eckhart committed
489
490
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
491
            return head + usetab + (sep + usetab).join(content) + tail
492

eckhart's avatar
eckhart committed
493
        res = self.content
eckhart's avatar
eckhart committed
494
        if not inline and not head:
eckhart's avatar
eckhart committed
495
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
496
            res = res.strip()
497
498
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
499
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
500
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
501
        else:
502
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
503

Eckhart Arnold's avatar
Eckhart Arnold committed
504

505
506
    def as_sxpr(self, src: str = None,
                indentation: int = 2,
507
                compact: bool = False) -> str:
508
        """
509
510
        Returns content as S-expression, i.e. in lisp-like form. If this
        method is callad on a RootNode-object,
511

512
        Args:
513
514
515
            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.
516
517
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
518
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
519
                tree structure.
520
521
        """

eckhart's avatar
eckhart committed
522
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
523
        lbreaks = linebreaks(src) if src else []  # type: List[int]
524
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[Node]
525

526
        def opening(node) -> str:
eckhart's avatar
eckhart committed
527
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
528
            txt = [left_bracket, node.tag_name]
529
            # s += " '(pos %i)" % node.add_pos
eckhart's avatar
eckhart committed
530
            if node.attr_active():
531
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
532
            if src:
533
534
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
535
536
            if root and id(node) in root.error_nodes:
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in root.get_errors(node)))
537
            return "".join(txt) + '\n'
538

eckhart's avatar
eckhart committed
539
540
541
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
542

eckhart's avatar
eckhart committed
543
544
545
546
547
        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'\"')
548

549
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
550

eckhart's avatar
eckhart committed
551

552
553
    def as_xml(self, src: str = None,
               indentation: int = 2,
eckhart's avatar
eckhart committed
554
555
556
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
557
558
559
        """
        Returns content as XML-tree.

560
        Args:
561
562
563
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
564
            indentation: The number of whitespaces for indentation
565
566
567
            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
568
                neither the opening tag nor its attr nor the closing tag. This
569
570
571
                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.
572
573
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
574
        """
575
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[Node]
576

577
        def opening(node) -> str:
578
579
580
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
581
            txt = ['<', node.tag_name]
eckhart's avatar
eckhart committed
582
            has_reserved_attrs = node.attr_active() \
eckhart's avatar
eckhart committed
583
                and any(r in node.attr for r in {'err', 'line', 'col'})
eckhart's avatar
eckhart committed
584
            if node.attr_active():
585
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
586
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
587
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
588
            if root and id(node) in root.error_nodes and not has_reserved_attrs:
eckhart's avatar
eckhart committed
589
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
590
                                                 for err in root.get_error(node)))
591
592
593
            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
594
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
595
596
597
            else:
                ending = ">\n"
            return "".join(txt + [ending])
598
599

        def closing(node):
600
            """Returns the closing string for the representation of `node`."""
601
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
602
                return ''
603
            return ('\n</') + node.tag_name + '>'
604

605
606
607
608
609
610
611
        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


612
613
614
615
616
        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.
            """
eckhart's avatar
eckhart committed
617
            return node.tag_name in inline_tags \
eckhart's avatar
eckhart committed
618
                or (node.attr_active()
eckhart's avatar
eckhart committed
619
                    and node.attr.get('xml:space', 'default') == 'preserve')
620

621
        line_breaks = linebreaks(src) if src else []
622
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
623
                               density=1, inline_fn=inlining)
624

Eckhart Arnold's avatar
Eckhart Arnold committed
625

eckhart's avatar
eckhart committed
626
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
627
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
628
        """
629
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
630

631
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
632
        given `match_function` evaluates to True. The tree is
633
634
635
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
636

637
638
639
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
640
641
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
642
643
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
644
        Yields:
645
            Node: All nodes of the tree for which
646
647
            ``match_function(node)`` returns True
        """
648
        if include_root and match_function(self):
649
            yield self
650
651
652
        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
653
                yield node
654

Eckhart Arnold's avatar
Eckhart Arnold committed
655

eckhart's avatar
eckhart committed
656
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
eckhart's avatar
eckhart committed
657
                      include_root: bool = False) -> Iterator['Node']:
658
        """
659
660
661
662
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
663

664
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
665
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
666
            ['(X (c "d"))', '(X "F")']
667
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
668
            ['(b "X")', '(X (c "d"))', '(X "F")']
669
            >>> any(tree.select_by_tag('a', False))
670
            False
671
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
672
            ['(a (b "X") (X (c "d")) (e (X "F")))']
673
674
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
675
676

        Args:
677
678
679
680
            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.
681
682
683
        Yields:
            Node: All nodes which have a given tag name.
        """
684
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
685
            tag_names = frozenset({tag_names})
686
        return self.select(lambda node: node.tag_name in tag_names, include_root)
687
688


689
690
691
692
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

693
        This function is mostly just syntactic sugar for
694
695
696
697
698
699
700
701
702
703
        ``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


704
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
705
706
707
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
708
709
710
        return sum(child.tree_size() for child in self.children) + 1


711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
class FrozenNode(Node):

    def __init__(self, tag_name: str, result: ResultType) -> None:
        if isinstance(result, str) or isinstance(result, StringView):
            result = str(result)
        else:
            raise TypeError('FrozenNode only accepts string as results. '
                            '(Only leaf-nodes can be frozen nodes.)')
        super(FrozenNode, self).__init__(tag_name, result, True)

    @property
    def result(self) -> StrictResultType:
        return self._result

    @result.setter
    def result(self, result: ResultType):
        raise TypeError('FrozenNode does not allow re-assignment of results.')

eckhart's avatar
eckhart committed
729
730
731
732
    @property
    def attr(self):
        raise AssertionError("Attributes cannot be accessed on a frozen node")

733
734
735
736
737
738
739
740
    # @property
    # def errors(self) -> List[Error]:
    #     return ()
    #
    # @errors.setter
    # def errors(self, errors: List[Error]):
    #     if errors:
    #         raise AssertionError('Cannot assign error list to frozen node')
eckhart's avatar
eckhart committed
741

Eckhart Arnold's avatar
Eckhart Arnold committed
742
    def with_pos(self, pos: int) -> 'Node':
743
744
745
746
        pass


PLACEHOLDER = Node('__PLACEHOLDER__', '')
di68kap's avatar
di68kap committed
747
748


di68kap's avatar
di68kap committed
749
class RootNode(Node):
750
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
751

752
        errors (list):  A list of all errors that have occured so far during
753
754
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
755

756
757
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
758
    """
759

eckhart's avatar
eckhart committed
760
    def __init__(self, node: Optional[Node] = None):
761
        super().__init__('__not_yet_ready__', '')
762
        self.errors = []           # type: List[Error]
763
764
        self.error_nodes = dict()      # type: Dict[int, List[Error]]  # id(node) -> error list
        self.error_positions = dict()  # type: Dict[int, Set[int]]  # pos -> set of id(node)
di68kap's avatar
di68kap committed
765
        self.error_flag = 0
eckhart's avatar
eckhart committed
766
767
        if node is not None:
            self.swallow(node)
768
        # customization for XML-Representation
eckhart's avatar
eckhart committed
769
770
771
        self.inline_tags = set()  # type: Set[str]
        self.omit_tags = set()  # type: Set[str]
        self.empty_tags = set()  # type: Set[str]
di68kap's avatar
di68kap committed
772

773
774
775
776
777
778
779
780
781
782
    def __deepcopy__(self, memodict={}):
        duplicate = self.__class__(None)
        if self.children:
            duplicate.children = copy.deepcopy(self.children)
            duplicate._result = duplicate.children
        else:
            duplicate.children = NoChildren
            duplicate._result = self._result
        duplicate._pos = self._pos
        duplicate._len = self._len
eckhart's avatar
eckhart committed
783
        if self.attr_active():
di68kap's avatar
di68kap committed
784
            duplicate.attr.update(copy.deepcopy(self._xml_attr))
785
            # duplicate._xml_attr = copy.deepcopy(self._xml_attr)  # this is blocked by cython
786
        duplicate.errors = copy.copy(self.errors)
787
788
        duplicate.error_nodes = copy.copy(self.error_nodes)
        duplicate.error_positions = copy.deepcopy(self.error_positions)
789
790
791
792
        duplicate.error_flag = self.error_flag
        duplicate.inline_tags = self.inline_tags
        duplicate.omit_tags = self.omit_tags
        duplicate.empty_tags = self.empty_tags
793
        duplicate.tag_name = self.tag_name
794
795
796
        return duplicate


797
    def swallow(self, node: Node) -> 'RootNode':
798
799
        """
        Put `self` in the place of `node` by copying all its data.
800
801
802
803
804
805
806
807
808
        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
809
810
811
812
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
813
        self.tag_name = node.tag_name
eckhart's avatar
eckhart committed
814
        if node.attr_active():
815
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
816
        self._content = node._content
817
818
        if id(node) in self.error_nodes:
            self.error_nodes[id(self)] = self.error_nodes[id(node)]
819
        return self
di68kap's avatar
di68kap committed
820

eckhart's avatar
eckhart committed
821
    def add_error(self, node: Node, error: Error) -> 'RootNode':
822
823
824
        """
        Adds an Error object to the tree, locating it at a specific node.
        """
eckhart's avatar
eckhart committed
825
        assert not isinstance(node, FrozenNode)
826
        assert node.pos == error.pos
827
        self.errors.append(error)
di68kap's avatar
di68kap committed
828
        self.error_flag = max(self.error_flag, error.code)
829
830
        self.error_nodes.setdefault(id(node), []).append(error)
        self.error_positions.setdefault(node.pos, set()).add(id(node))
di68kap's avatar
di68kap committed
831
832
        return self

eckhart's avatar
eckhart committed
833
    def new_error(self,
834
                  node: Node,
di68kap's avatar
di68kap committed
835
                  message: str,
eckhart's avatar
eckhart committed
836
                  code: ErrorCode = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
837
        """
eckhart's avatar
eckhart committed
838
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
839
840
841
842
843
        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
844
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
845
846
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
847

848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
    def get_errors(self, node: Node) -> List[Error]:
        """
        Returns the List of errors that occured on the node or any child node
        at the same position that has already been removed from the tree,
        for example, because it was an anonymous empty child node.
        """
        node_id = id(node)           # type: int
        errors = []                  # type: List[Error]
        for nid in self.error_positions[node.pos]:
            if nid == node_id:
                errors.extend(self.error_nodes[nid])
            else:
                for nd in node.select(lambda n: id(n) == nid):
                    break
                else:
                    # node is not connected to tree any more => display its errors on its parent
                    errors.extend(self.error_nodes[nid])
        return errors

867
    @property
868
    def errors_sorted(self) -> List[Error]:
869
870
        """
        Returns the list of errors, ordered bv their position.
di68kap's avatar
di68kap committed
871
        """
872
873
        self.errors.sort(key=lambda e: e.pos)
        return self.errors
di68kap's avatar
di68kap committed
874

875
    def customized_XML(self):
876
877
        """
        Returns a customized XML representation of the tree.
878
        See the docstring of `Node.as_xml()` for an explanation of the
879
880
        customizations.
        """
eckhart's avatar
eckhart committed
881
        return self.as_xml(inline_tags=self.inline_tags,
882
883
884
                           omit_tags=self.omit_tags,
                           empty_tags=self.empty_tags)

di68kap's avatar
di68kap committed
885

di68kap's avatar
di68kap committed
886
887
888
889
890
#######################################################################
#
# S-expression- and XML-parsers
#
#######################################################################
eckhart's avatar
eckhart committed
891
892


eckhart's avatar
eckhart committed
893
def parse_sxpr(sxpr: Union[str, StringView]) -> Node:
894
    """
895
896
897
898
899
    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.
900
901

    Example:
902
    >>> parse_sxpr("(a (b c))").as_sxpr()