12.8.2021, 9:00 - 11:00: Due to updates GitLab may be unavailable for some minutes between 09:00 and 11:00.

syntaxtree.py 41.3 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

eckhart's avatar
eckhart committed
31
from DHParser.error import Error, ErrorCode, 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, Dict, 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',
eckhart's avatar
eckhart committed
44 45 46
           'ResultType',
           'StrictResultType',
           'ChildrenType',
47
           'Node',
di68kap's avatar
di68kap committed
48
           'RootNode',
eckhart's avatar
eckhart committed
49
           'ZOMBIE_ROOTNODE',
50
           'parse_sxpr',
51 52 53
           'parse_xml',
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
54 55


56 57 58 59 60 61 62 63 64 65 66 67 68
#######################################################################
#
# 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.
    """
69

70
    __slots__ = 'name', 'ptype'
Eckhart Arnold's avatar
Eckhart Arnold committed
71

72 73 74
    def __init__(self,):  # , pbases=frozenset()):
        self.name = ''  # type: str
        self.ptype = ':' + self.__class__.__name__  # type: str
75 76

    def __repr__(self):
eckhart's avatar
eckhart committed
77
        return self.name + self.ptype
78 79 80 81

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

82 83 84
    def __call__(self, text: StringView) -> Tuple[Optional['Node'], StringView]:
        return None, text

85 86
    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
87
        """Returns the parser's name if it has a name and repr()"""
88 89
        return self.name if self.name else repr(self)

90
    def reset(self):
91
        """Resets any parser variables. (Should be overridden.)"""
92 93
        pass

Eckhart Arnold's avatar
Eckhart Arnold committed
94
    @property
eckhart's avatar
eckhart committed
95
    def grammar(self) -> 'Grammar':
96 97
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
eckhart's avatar
eckhart committed
98
        raise NotImplementedError
99 100

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
101
        """Applies the function `func` to the parser. Returns False, if
102
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
103
        otherwise."""
104 105
        return False

106 107

WHITESPACE_PTYPE = ':Whitespace'
108
TOKEN_PTYPE = ':Token'
109 110 111 112 113 114


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
115
    different parser types by assigning them a `ptype` on initialization.
116 117 118 119 120

    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.
    """
121

Eckhart Arnold's avatar
Eckhart Arnold committed
122 123
    __slots__ = ()

124 125
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
126 127
        super().__init__()
        self.name = name
di68kap's avatar
di68kap committed
128 129
        if ptype:
            self.ptype = ptype  # or ':' + self.__class__.__name__
130 131 132 133 134 135 136 137 138 139 140 141


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.
    """
142

143
    alive = False
Eckhart Arnold's avatar
Eckhart Arnold committed
144
    __slots__ = ()
145 146

    def __init__(self):
147
        super(ZombieParser, self).__init__()
148 149
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
150
        self.name = "__ZOMBIE__"
151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173
        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
#
#######################################################################


174
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
175
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
176
StrictResultType = Union[ChildrenType, StringView, str]
177
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
178 179


Eckhart Arnold's avatar
Eckhart Arnold committed
180
def flatten_sxpr(sxpr: str) -> str:
181 182
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
183 184 185
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
186
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
187 188
    '(a (b c))'
    """
189

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:
194 195
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
196
    i.e. only whitespace within leaf-nodes is preserved.
197 198
    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`.
199
    """
200

201 202 203 204
    # 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']
205
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
206 207


eckhart's avatar
eckhart committed
208
RX_AMP = re.compile(r'&(?!\w+;)')
209 210


Eckhart Arnold's avatar
Eckhart Arnold committed
211
class Node(collections.abc.Sized):
212 213 214
    """
    Represents a node in the concrete or abstract syntax tree.

215 216
    TODO: Add some documentation and doc-tests here...

217
    Attributes:
218 219
        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
220

221 222 223
        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.
224

225 226
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
227

eckhart's avatar
eckhart committed
228 229 230 231
        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
232
        parser (Parser):  The parser which generated this node.
233 234
            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
235 236
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
237
            example by calling ``isinstance(node.parer, ...)``.
238

239 240 241
        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
242
            the length before AST-transformation and will never change
243
            through AST-transformation. READ ONLY!
244

245 246
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
247
            The value of ``pos`` is -1 meaning invalid by default.
248
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
249
            nodes relative to this value.
250 251

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
252
            pos value of the root node should be set to 0 right
253 254
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
255
            Other than that, this value should be considered READ ONLY.
256 257
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
258

259 260
        errors (list):  A list of all errors that occured on this node.

261 262
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
263 264
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
265
    """
266

Eckhart Arnold's avatar
Eckhart Arnold committed
267
    __slots__ = '_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content'
268

269 270 271
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
272 273
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
274
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
275
        self._pos = -1                  # type: int
276
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
277
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
278
        if leafhint:
eckhart's avatar
eckhart committed
279
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
eckhart's avatar
eckhart committed
280
            self._content = None        # type: Optional[str]
281
            self.children = NoChildren  # type: ChildrenType
282
            self._len = -1              # type: int  # lazy evaluation
283 284
        else:
            self.result = result
285 286
        self.parser = parser or ZOMBIE_PARSER

eckhart's avatar
eckhart committed
287 288 289 290 291 292 293 294 295 296 297
    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
298

299
    def __str__(self):
300
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
301
        if self.errors:
302
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
303
                   (s, '; '.join(e.message for e in self.errors))
304
        return s
305

Eckhart Arnold's avatar
Eckhart Arnold committed
306

307 308 309 310
    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 \
eckhart's avatar
eckhart committed
311
            "(" + ", ".join(repr(child) for child in self.children) + ")"
312
        return "Node(%s, %s)" % (parg, rarg)
313

Eckhart Arnold's avatar
Eckhart Arnold committed
314

315
    def __len__(self):
316
        if self._len < 0:
di68kap's avatar
di68kap committed
317
            self._len = sum(len(child) for child in self.children) \
318
                if self.children else len(self._result)
319 320 321 322 323 324 325 326
        return self._len


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


327
    def __eq__(self, other):
328 329 330
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
331 332 333 334

        Note: It is not required that two nodes have the same errors attached.
        In case you need to check for error equality as well, compare a
        serialization that includes error messages, as_sxpr() will do!
335
        """
336
        return self.tag_name == other.tag_name and self.result == other.result
337

Eckhart Arnold's avatar
Eckhart Arnold committed
338

339
    def __hash__(self):
340
        return hash(self.tag_name)
341

Eckhart Arnold's avatar
Eckhart Arnold committed
342

343 344 345
    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
346
        an integer or the first child node with the given tag name. Examples::
347

348
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
349 350
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
351 352
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
353 354 355 356

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
357
        Returns:
358 359
            Node: All nodes which have a given tag name.
        """
360 361 362
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
363
            else:
364 365 366 367 368
                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!')
369 370 371 372


    def __contains__(self, tag_name: str) -> bool:
        """
373
        Returns true if a child with the given tag name exists.
374
        Args:
375 376
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
377 378 379 380
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
381 382 383 384 385
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
386
            return False
387
        raise ValueError('Leave node cannot contain other nodes')
388 389


390 391 392 393 394 395 396 397 398 399 400 401 402 403
    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


404
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
405
    def tag_name(self) -> str:
406 407 408 409 410 411
        """
        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`.
        """
412
        return self.parser.name or self.parser.ptype
413

Eckhart Arnold's avatar
Eckhart Arnold committed
414

415
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
416
    def result(self) -> StrictResultType:
417 418 419 420 421
        """
        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.
        """
422 423
        return self._result

424

425
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
426
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
427
        # # made obsolete by static type checking with mypy
428 429 430
        # 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
431 432
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
433 434
        self._len = -1        # lazy evaluation
        self._content = None
435 436 437 438 439 440 441 442 443
        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
444
                self._result = result  # cast(StrictResultType, result)
445

446 447

    @property
eckhart's avatar
eckhart committed
448
    def content(self) -> str:
449
        """
di68kap's avatar
di68kap committed
450 451 452
        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.
453
        """
454 455 456 457
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
458
                # self._content = self._result
459 460 461
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
462 463


di68kap's avatar
di68kap committed
464 465 466 467 468
    @content.setter
    def content(self, content: str):
        self.result = content


469 470 471 472 473 474 475 476 477
    @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))


478
    @property
479
    def pos(self) -> int:
eckhart's avatar
eckhart committed
480 481
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
482
            raise AssertionError("Position value not initialized!")
483 484
        return self._pos

eckhart's avatar
eckhart committed
485

eckhart's avatar
eckhart committed
486
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
487 488 489 490 491 492 493 494 495
        """
        (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
496 497
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
498
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
499
        offset = self.pos
500
        for child in self.children:
eckhart's avatar
eckhart committed
501
            child.init_pos(offset)
eckhart's avatar
eckhart committed
502 503 504
            offset = child.pos + len(child)
        return self

505

506
    @property
507
    def attr(self):
508
        """
509
        Returns a dictionary of XML-attr attached to the node.
510
        """
511 512 513 514
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
515

516
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
517
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
518 519 520 521 522 523 524 525
        """
        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.

526
        Args:
527
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
528
            open_fn:   (Node->str) A function that returns an opening
529
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
530
            close_fn:  (Node->str) A function that returns a closeF
531
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
532
            data_fn:   (str->str) A function that filters the data string
533 534 535 536 537 538
                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
539 540
        head = open_fn(self)
        tail = close_fn(self)
541 542

        if not self.result:
543
            return head.rstrip() + tail.lstrip()
544

eckhart's avatar
eckhart committed
545
        tail = tail.lstrip(None if density & 2 else '')
546

547
        inline = inline or inline_fn(self)
548 549 550 551 552
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
553 554
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
555

556 557
        if self.children:
            content = []
558
            for child in self.children:
559
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
560
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
561
                if subtree:
eckhart's avatar
eckhart committed
562 563
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
564
            return head + usetab + (sep + usetab).join(content) + tail
565

eckhart's avatar
eckhart committed
566
        res = self.content
eckhart's avatar
eckhart committed
567
        if not inline and not head:
eckhart's avatar
eckhart committed
568
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
569
            res = res.strip()
570 571
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
572
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
573
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
574
        else:
575
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
576

Eckhart Arnold's avatar
Eckhart Arnold committed
577

578 579 580
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
581
                compact: bool = False) -> str:
582 583 584
        """
        Returns content as S-expression, i.e. in lisp-like form.

585
        Args:
586 587 588
            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.
589 590 591
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
592
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
593
                tree structure.
594 595
        """

eckhart's avatar
eckhart committed
596
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
597
        lbreaks = linebreaks(src) if src else []  # type: List[int]
598

599
        def opening(node) -> str:
eckhart's avatar
eckhart committed
600
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
601
            txt = [left_bracket, node.tag_name]
602
            # s += " '(pos %i)" % node.add_pos
603
            if hasattr(node, '_xml_attr'):
604
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
605
            if src:
606 607
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
608
            if showerrors and node.errors:
609
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
610
            return "".join(txt) + '\n'
611

eckhart's avatar
eckhart committed
612 613 614
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
615

eckhart's avatar
eckhart committed
616 617 618 619 620
        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'\"')
621

622
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
623

eckhart's avatar
eckhart committed
624

625 626 627
    def as_xml(self, src: str = None,
               showerrors: bool = True,
               indentation: int = 2,
eckhart's avatar
eckhart committed
628 629 630
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
631 632 633
        """
        Returns content as XML-tree.

634
        Args:
635 636 637
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
638 639
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
640 641 642
            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
643
                neither the opening tag nor its attr nor the closing tag. This
644 645 646
                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.
647 648
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
649 650
        """

651
        def opening(node) -> str:
652 653 654
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
655
            txt = ['<', node.tag_name]
656
            has_reserved_attrs = hasattr(node, '_xml_attr') \
eckhart's avatar
eckhart committed
657
                and any(r in node.attr for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
658
            if hasattr(node, '_xml_attr'):
659
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
660
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
661
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
662
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
663 664
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
665 666 667
            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
668
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
669 670 671
            else:
                ending = ">\n"
            return "".join(txt + [ending])
672 673

        def closing(node):
674
            """Returns the closing string for the representation of `node`."""
675
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
676
                return ''
677
            return ('\n</') + node.tag_name + '>'
678

679 680 681 682 683 684 685
        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


686 687 688 689 690
        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
691 692 693
            return node.tag_name in inline_tags \
                or (hasattr(node, '_xml_attr')
                    and node.attr.get('xml:space', 'default') == 'preserve')
694

695
        line_breaks = linebreaks(src) if src else []
696
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
697
                               density=1, inline_fn=inlining)
698

Eckhart Arnold's avatar
Eckhart Arnold committed
699

eckhart's avatar
eckhart committed
700
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
701
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
702
        """
703
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
704

705
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
706
        given `match_function` evaluates to True. The tree is
707 708 709
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
710

711 712 713
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
714 715
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
716 717
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
718
        Yields:
719
            Node: All nodes of the tree for which
720 721
            ``match_function(node)`` returns True
        """
722
        if include_root and match_function(self):
723
            yield self
724 725 726
        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
727
                yield node
728

Eckhart Arnold's avatar
Eckhart Arnold committed
729

eckhart's avatar
eckhart committed
730
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
eckhart's avatar
eckhart committed
731
                      include_root: bool = False) -> Iterator['Node']:
732
        """
733 734 735 736
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
737

738
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
739
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
740
            ['(X (c "d"))', '(X "F")']
741
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
742
            ['(b "X")', '(X (c "d"))', '(X "F")']
743
            >>> any(tree.select_by_tag('a', False))
744
            False
745
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
746
            ['(a (b "X") (X (c "d")) (e (X "F")))']
747 748
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
749 750

        Args:
751 752 753 754
            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.
755 756 757
        Yields:
            Node: All nodes which have a given tag name.
        """
758
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
759
            tag_names = frozenset({tag_names})
760
        return self.select(lambda node: node.tag_name in tag_names, include_root)
761 762


763 764 765 766
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

767
        This function is mostly just syntactic sugar for
768 769 770 771 772 773 774 775 776 777
        ``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


778
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
779 780 781
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
782 783 784
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
785 786 787
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


di68kap's avatar
di68kap committed
788
class RootNode(Node):
789
    """TODO: Add Documentation!!!
di68kap's avatar