05.11., 9:00 - 11:00: Due to updates GitLab may be unavailable for some minutes between 09:00 and 11:00.

syntaxtree.py 33.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
# syntaxtree.py - syntax tree classes for DHParser
#
# Copyright 2016  by Eckhart Arnold (arnold@badw.de)
#                 Bavarian Academy of Sciences an Humanities (badw.de)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied.  See the License for the specific language governing
# permissions and limitations under the License.
17 18 19


"""
20 21
Module ``syntaxtree`` defines the ``Node``-class for syntax trees as well
as an abstract base class for parser-objects. The latter is defined
eckhart's avatar
eckhart committed
22
here, because node-objects refer to parser-objects. All concrete
23 24 25
parser classes are defined in the ``parse`` module.
"""

26

Eckhart Arnold's avatar
Eckhart Arnold committed
27
import collections.abc
28
from collections import OrderedDict
29
import copy
30

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

36

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


50 51 52 53 54 55 56 57 58 59 60 61 62 63
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


class ParserBase:
    """
    ParserBase is the base class for all real and mock parser classes.
    It is defined here, because Node objects require a parser object
    for instantiation.
    """
    def __init__(self, name=''):  # , pbases=frozenset()):
64
        self._name = name  # type: str
65 66 67
        self._ptype = ':' + self.__class__.__name__  # type: str

    def __repr__(self):
di68kap's avatar
di68kap committed
68
         return self.name + self.ptype
69 70 71 72

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

73 74 75
    def __call__(self, text: StringView) -> Tuple[Optional['Node'], StringView]:
        return None, text

76 77
    @property
    def name(self):
78 79
        """Returns the name of the parser or the empty string '' for unnamed
        parsers."""
80 81
        return self._name

82 83
    @property
    def ptype(self) -> str:
84 85
        """Returns the type of the parser. By default this is the parser's
        class name preceded by a colon, e.g. ':ZeroOrMore'."""
86 87 88 89
        return self._ptype

    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
90
        """Returns the parser's name if it has a name and repr()"""
91 92
        return self.name if self.name else repr(self)

93
    def reset(self):
94
        """Resets any parser variables. (Should be overridden.)"""
95 96
        pass

eckhart's avatar
eckhart committed
97
    def grammar(self) -> Optional[object]:
98 99
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
100 101 102
        return None

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
103
        """Applies the function `func` to the parser. Returns False, if
104
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
105
        otherwise."""
106 107
        return False

108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124

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


class MockParser(ParserBase):
    """
    MockParser objects can be used to reconstruct syntax trees from a
    serialized form like S-expressions or XML. Mock objects can mimic
    different parser types by assigning them a ptype on initialization.

    Mock objects should not be used for anything other than
    syntax tree (re-)construction. In all other cases where a parser
    object substitute is needed, chose the singleton ZOMBIE_PARSER.
    """
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
125
        super().__init__(name)
126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167
        self._ptype = ptype or ':' + self.__class__.__name__


class ZombieParser(MockParser):
    """
    Serves as a substitute for a Parser instance.

    ``ZombieParser`` is the class of the singelton object
    ``ZOMBIE_PARSER``. The  ``ZOMBIE_PARSER`` has a name and can be
    called, but it never matches. It serves as a substitute where only
    these (or one of these properties) is needed, but no real Parser-
    object is instantiated.
    """
    alive = False

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

    def __copy__(self):
        return self

    def __deepcopy__(self, memo):
        return self

    def __call__(self, text):
        """Better call Saul ;-)"""
        return None, text


ZOMBIE_PARSER = ZombieParser()


#######################################################################
#
# syntaxtree nodes
#
#######################################################################


168
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
169
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
170
StrictResultType = Union[ChildrenType, StringView, str]
171
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
172 173


Eckhart Arnold's avatar
Eckhart Arnold committed
174
def flatten_sxpr(sxpr: str) -> str:
175
    """Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
176 177 178
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
179
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
180 181
    '(a (b c))'
    """
182
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
183 184


Eckhart Arnold's avatar
Eckhart Arnold committed
185
class Node(collections.abc.Sized):
186 187 188
    """
    Represents a node in the concrete or abstract syntax tree.

189
    Attributes:
190 191
        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
192

193 194 195
        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.
196

197 198
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
199

eckhart's avatar
eckhart committed
200 201 202 203
        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
204
        parser (Parser):  The parser which generated this node.
205 206
            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
207 208
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
209
            example by calling ``isinstance(node.parer, ...)``.
210

211 212 213
        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
214
            the length before AST-transformation and will never change
215
            through AST-transformation. READ ONLY!
216

217 218
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
219
            The value of ``pos`` is -1 meaning invalid by default.
220
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
221
            nodes relative to this value.
222 223

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
224
            pos value of the root node should be set to 0 right
225 226
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
227
            Other than that, this value should be considered READ ONLY.
228 229
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
230

231
        attributes (dict): An optional dictionary of XML-attributes. This
232 233 234
            dictionary is created lazily upon first usage. The attributes
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
eckhart's avatar
eckhart committed
235 236 237


        _parent (Node): SLOT RESERVED FOR FUTURE USE!
238
    """
239

di68kap's avatar
di68kap committed
240
    __slots__ = ['_result', 'children', '_len', '_pos', 'parser', '_xml_attr', '_content']
241

242 243 244
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
245 246
        that generated the node and the parser's result.
        """
eckhart's avatar
eckhart committed
247
        self._pos = -1                  # type: int
di68kap's avatar
di68kap committed
248 249
        # Assignment to self.result initializes the attributes _result, children and _len
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
250
        if leafhint:
eckhart's avatar
eckhart committed
251
            self._result = result       # type: StrictResultType
eckhart's avatar
eckhart committed
252
            self._content = None        # type: Optional[str]
253
            self.children = NoChildren  # type: ChildrenType
254
            self._len = -1              # type: int  # lazy evaluation
255 256
        else:
            self.result = result
257 258
        self.parser = parser or ZOMBIE_PARSER

Eckhart Arnold's avatar
Eckhart Arnold committed
259

260
    def __str__(self):
261
        s = "".join(str(child) for child in self.children) if self.children else self.content
262 263 264 265
        if self._errors:
            return ' <<< Error on "%s" | %s >>> ' % \
                   (s, '; '.join(e.message for e in self._errors))
        return s
266

Eckhart Arnold's avatar
Eckhart Arnold committed
267

268 269 270 271 272 273
    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)
274

Eckhart Arnold's avatar
Eckhart Arnold committed
275

276
    def __len__(self):
277
        if self._len < 0:
di68kap's avatar
di68kap committed
278
            self._len = sum(len(child) for child in self.children) \
279
                if self.children else len(self._result)
280 281 282 283 284 285 286 287
        return self._len


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


288
    def __eq__(self, other):
289 290 291 292
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
        """
293
        return self.tag_name == other.tag_name and self.result == other.result
294

Eckhart Arnold's avatar
Eckhart Arnold committed
295

296
    def __hash__(self):
297
        return hash(self.tag_name)
298

Eckhart Arnold's avatar
Eckhart Arnold committed
299

300 301 302 303 304 305
    def __deepcopy__(self, memodict={}):
        result = copy.deepcopy(self.result)
        other = Node(self.parser, result)
        other._pos = self._pos
        return other

Eckhart Arnold's avatar
Eckhart Arnold committed
306

307 308 309
    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
310
        an integer or the first child node with the given tag name. Examples::
311

312
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
313 314
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
315 316
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
317 318 319 320

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
321
        Returns:
322 323
            Node: All nodes which have a given tag name.
        """
324 325 326
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
327
            else:
328 329 330 331 332
                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!')
333 334 335 336


    def __contains__(self, tag_name: str) -> bool:
        """
337
        Returns true if a child with the given tag name exists.
338
        Args:
339 340
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
341 342 343 344
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
345 346 347 348 349
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
350
            return False
351
        raise ValueError('Leave node cannot contain other nodes')
352
        # generator = self.select_by_tag(tag_name, False)
353 354 355 356 357
        # try:
        #     generator.__next__()
        #     return True
        # except StopIteration:
        #     return False
358 359


360
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
361
    def tag_name(self) -> str:
362 363 364 365 366 367
        """
        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`.
        """
368
        return self.parser.name or self.parser.ptype
369

Eckhart Arnold's avatar
Eckhart Arnold committed
370

371
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
372
    def result(self) -> StrictResultType:
373 374 375 376 377
        """
        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.
        """
378 379
        return self._result

380

381
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
382
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
383
        # # made obsolete by static type checking with mypy
384 385 386
        # 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
387 388
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
389 390
        self._len = -1        # lazy evaluation
        self._content = None
391 392 393 394 395 396 397 398
        if isinstance(result, Node):
            self.children = (result,)
            self._result = self.children
            self.error_flag = result.error_flag
        else:
            if isinstance(result, tuple):
                self.children = result
                self._result = result or ''
399 400 401
                if result:
                    if self.error_flag == 0:
                        self.error_flag = max(child.error_flag for child in self.children)
402 403
            else:
                self.children = NoChildren
404
                self._result = result
405

406 407

    @property
eckhart's avatar
eckhart committed
408
    def content(self) -> str:
409
        """
410
        Returns content as string, omitting error messages.
411
        """
412 413 414 415
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
416
                # self._content = self._result
417 418 419
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
420 421 422 423 424 425 426 427 428 429 430


    @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))


431
    @property
432
    def pos(self) -> int:
eckhart's avatar
eckhart committed
433 434
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
435
            raise AssertionError("Position value not initialized!")
436 437
        return self._pos

eckhart's avatar
eckhart committed
438

eckhart's avatar
eckhart committed
439
    def init_pos(self, pos: int, overwrite: bool = False) -> 'Node':
eckhart's avatar
eckhart committed
440 441 442 443 444 445 446 447 448
        """
        (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
449 450 451 452 453 454
        if overwrite or self._pos < 0:
            self._pos = pos
            for err in self._errors:
                err.pos = pos
        else:
            assert self._pos == pos, str("%i != %i" % (self._pos, pos))
455
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
456
        offset = self.pos
457
        for child in self.children:
eckhart's avatar
eckhart committed
458
            child.init_pos(offset, overwrite)
eckhart's avatar
eckhart committed
459 460 461
            offset = child.pos + len(child)
        return self

462

di68kap's avatar
di68kap committed
463 464 465 466 467 468 469
    # @property
    # def errors(self) -> List[Error]:
    #     """
    #     Returns the errors that occurred at this Node,
    #     not including any errors from child nodes.
    #     """
    #     return self._errors.copy()
470

471 472 473 474 475 476 477
    @property
    def attributes(self):
        """Returns a dictionary of XML-Attributes attached to the Node."""
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
478

479
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i, density=0) -> str:
480 481 482 483 484 485 486 487
        """
        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.

488
        Args:
489
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
490
            open_fn:   (Node->str) A function that returns an opening
491
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
492
            close_fn:  (Node->str) A function that returns a closeF
493
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
494
            data_fn:   (str->str) A function that filters the data string
495 496 497 498 499 500
                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
501 502
        head = open_fn(self)
        tail = close_fn(self)
503 504

        if not self.result:
505
            return head.rstrip() + tail.lstrip()
506

eckhart's avatar
eckhart committed
507
        tail = tail.lstrip(None if density & 2 else '')
508 509 510

        if self.children:
            content = []
511
            for child in self.children:
eckhart's avatar
eckhart committed
512
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn, density).split('\n')
513
                content.append('\n'.join((tab + s) for s in subtree))
eckhart's avatar
eckhart committed
514
            return head + '\n'.join(content) + tail
515

516
        res = cast(str, self.result)  # safe, because if there are no children, result is a string
517 518 519
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
            gap = ' ' if head.rstrip()[-1] != '>' else ''
eckhart's avatar
eckhart committed
520
            return head.rstrip() + gap + data_fn(self.result) + tail.lstrip()
521
        else:
eckhart's avatar
eckhart committed
522
            return head + '\n'.join([tab + data_fn(s) for s in res.split('\n')]) + tail
523

Eckhart Arnold's avatar
Eckhart Arnold committed
524

525
    def as_sxpr(self, src: str = None, compact: bool = False, showerrors: bool = True) -> str:
526 527 528
        """
        Returns content as S-expression, i.e. in lisp-like form.

529
        Args:
530 531 532
            src:  The source text or `None`. In case the source text is
                given the position of the element in the text will be
                reported as line and column.
Eckhart Arnold's avatar
Eckhart Arnold committed
533
            compact:  If True a compact representation is returned where
534
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
535
                tree structure.
536 537
        """

eckhart's avatar
eckhart committed
538
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
539

540
        def opening(node) -> str:
eckhart's avatar
eckhart committed
541
            """Returns the opening string for the representation of `node`."""
542
            txt = [left_bracket,  node.tag_name]
543
            # s += " '(pos %i)" % node.pos
544
            if hasattr(node, '_xml_attr'):
545
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attributes.items())
546
            if src:
547
                txt.append(" `(pos %i %i %i)" % (node.pos, *line_col(src, node.pos)))
di68kap's avatar
di68kap committed
548 549
            # if node.error_flag:   # just for debugging error collecting
            #     txt += " HAS ERRORS"
550
            if showerrors and node.errors:
551
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
552
            return "".join(txt) + '\n'
553

eckhart's avatar
eckhart committed
554 555 556
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
557

eckhart's avatar
eckhart committed
558 559 560 561 562
        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'\"')
563

eckhart's avatar
eckhart committed
564
        return self._tree_repr('    ', opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
565

eckhart's avatar
eckhart committed
566

567
    def as_xml(self, src: str = None, showerrors: bool = True) -> str:
568 569 570
        """
        Returns content as XML-tree.

571
        Args:
572 573 574 575 576
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
        """

577
        def opening(node) -> str:
eckhart's avatar
eckhart committed
578
            """Returns the opening string for the representation of `node`."""            
eckhart's avatar
eckhart committed
579
            txt = ['<', node.tag_name]
580 581
            has_reserved_attrs = hasattr(node, '_xml_attr') \
                and any (r in node.attributes for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
582
            if hasattr(node, '_xml_attr'):
583
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attributes.items())
584
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
585
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
586
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
587 588 589
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
            return "".join(txt + [">\n"])
590 591

        def closing(node):
eckhart's avatar
eckhart committed
592
            """Returns the closing string for the representation of `node`."""            
593
            return '\n</' + node.tag_name + '>'
594

595
        line_breaks = linebreaks(src) if src else []
596
        return self._tree_repr('    ', opening, closing, density=1)
597

Eckhart Arnold's avatar
Eckhart Arnold committed
598

599
    def select(self, match_function: Callable, include_root: bool=True) -> Iterator['Node']:
eckhart's avatar
eckhart committed
600
        """
601
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
602

603
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
604
        given `match_function` evaluates to True. The tree is
605 606 607
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
608

609 610 611
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
612 613
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
614
        Yields:
615
            Node: All nodes of the tree for which
616 617
            ``match_function(node)`` returns True
        """
618
        if include_root and match_function(self):
619 620 621
            yield self
        else:
            for child in self.children:
622
                for node in child.select(match_function, True):
eckhart's avatar
eckhart committed
623
                    yield node
624

Eckhart Arnold's avatar
Eckhart Arnold committed
625

eckhart's avatar
eckhart committed
626
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
627
                      include_root: bool=True) -> Iterator['Node']:
628
        """
629 630 631 632
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
633

634
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
635
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
636
            ['(X (c "d"))', '(X "F")']
637
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
638
            ['(b "X")', '(X (c "d"))', '(X "F")']
639
            >>> any(tree.select_by_tag('a', False))
640
            False
641
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
642
            ['(a (b "X") (X (c "d")) (e (X "F")))']
643 644
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
645 646

        Args:
647 648 649 650
            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.
651 652 653
        Yields:
            Node: All nodes which have a given tag name.
        """
654
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
655
            tag_names = frozenset({tag_names})
656
        return self.select(lambda node: node.tag_name in tag_names, include_root)
657 658


659 660 661 662 663 664 665 666 667 668 669 670 671 672 673
    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


674
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
675 676 677
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
678 679 680
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730
class RootNode(Node):
    """

        errors (list):  A list of parser- or compiler-errors:
            tuple(position, string) attached to this node

        error_flag (int):  0 if no error occurred in either the node
            itself or any of its descendants. Otherwise contains the
            highest warning or error level or all errors that occurred.
    """
    def __init__(self):
        super().__init__(ZOMBIE_PARSER, '')
        self.errors = []
        self.error_flag = 0

    def swallow(self, node):
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
        self._xml_attr = node._xml_attr
        self._content = node._content

    def add_error(self,
                  pos: int,
                  message: str,
                  code: int = Error.ERROR) -> 'Node':
        """
        Adds an error to this tree.
        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
        """
        self.errors.append(Error(message, code, pos))
        self.error_flag = max(self.error_flag, code)
        return self

    def collect_errors(self, clear_errors=False) -> List[Error]:
        """Returns the list of errors, ordered bv their position.
        """
        self.errors.sort(key=lambda e: e.pos)
        errors = self.errors
        if clear_errors:
            self.errors = []
            self.error_flag = 0
        return errors


eckhart's avatar
eckhart committed
731 732 733
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


734
def parse_sxpr(sxpr: str) -> Node:
735
    """
736 737 738 739 740
    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.
741 742

    Example:
743
    >>> parse_sxpr("(a (b c))").as_sxpr()
744 745
    '(a\\n    (b\\n        "c"\\n    )\\n)'
    """
746 747
    sxpr = StringView(sxpr).strip()
    mock_parsers = dict()
748

749
    def next_block(s: StringView):
eckhart's avatar
eckhart committed
750 751 752
        """Generator that yields all characters until the next closing bracket
        that does not match an opening bracket matched earlier within the same
        package."""
753
        s = s.strip()
754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772
        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)
773

774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824
    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()
        pos = 0
        attributes = OrderedDict()
        if sxpr[0] == '(':
            result = tuple(inner_parser(block) for block in next_block(sxpr))
            for node in result:
                node._pos = pos
                pos += len(node)
        else:
            lines = []
            while sxpr and sxpr[0:1] != ')':
                # parse attributes
                while sxpr[:2] == "`(":
                    i = sxpr.find('"')
                    k = sxpr.find(')')
                    # read very special attribute pos
                    if sxpr[2:5] == "pos" and 0 < i < k:
                        pos = int(sxpr[5:k].strip().split(' ')[0])
                    # ignore very special attribute err
                    elif sxpr[2:5] == "err" and 0 <= sxpr.find('`', 5) < k:
                        m = sxpr.find('(', 5)
                        while m >= 0 and m < k:
                            m = sxpr.find('(', k)
                            k = max(k, sxpr.find(')', max(m, 0)))
                    # read attributes
                    else:
                        attr = sxpr[2:i].strip()
                        value = sxpr[i:k].strip()[1:-1]
                        attributes[attr] = value
                    sxpr = sxpr[k+1:].strip()
                # parse content
                for qtmark in ['"""', "'''", '"', "'"]:
                    match = sxpr.match(re.compile(qtmark + r'.*?' + qtmark, re.DOTALL))
                    if match:
                        end = match.end() - sxpr.begin
                        i = len(qtmark)
                        lines.append(str(sxpr[i:end - i]))
                        sxpr = sxpr[end:].strip()
                        break
825
                else:
826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911
                    match = sxpr.match(re.compile(r'(?:(?!\)).)*', re.DOTALL))
                    end = match.end() - sxpr.begin
                    lines.append(str(sxpr[:end]))
                    sxpr = sxpr[end:]
            result = "\n".join(lines)
        node = Node(mock_parsers.setdefault(tagname, MockParser(name, ':' + class_name)), result)
        if attributes:
            node.attributes.update(attributes)
        node._pos = pos
        return node

    return inner_parser(sxpr)


def parse_xml(xml: str) -> Node:
    """
    Generates a tree of nodes from a (Pseudo-)XML-source.
    """
    xml = StringView(xml)
    PlainText = MockParser('', 'PlainText')
    mock_parsers = {':PlainText': PlainText}

    def parse_attributes(s: StringView) -> Tuple[StringView, OrderedDict]:
        """Parses a sqeuence of XML-Attributes. Returns the string-slice
        beginning after the end of the attributes."""
        attributes = OrderedDict()
        restart = 0
        for match in s.finditer(re.compile(r'\s*(?P<attr>\w+)\s*=\s*"(?P<value>.*)"\s*')):
            d = match.groupdict()
            attributes[d['attr']] = d['value']
            restart = match.end() - s.begin
        return (s[restart:], attributes)

    def parse_opening_tag(s: StringView) -> Tuple[StringView, str, OrderedDict, bool]:
        """Parses an opening tag. Returns the string segment following the
        the opening tag, the tag name, a dictionary of attributes and
        a flag indicating whether the tag is actually a solitary tag as
        indicated by a slash at the end, i.e. <br/>."""
        match = s.match(re.compile(r'<\s*(?P<tagname>[\w:]+)\s*'))
        assert match
        tagname = match.groupdict()['tagname']
        s, attributes = parse_attributes(s[match.end() - s.begin:])
        i = s.find('>')
        assert i >= 0
        return s[i+1,], tagname, attributes, s[i-1] == "/"

    def parse_closing_tag(s: StringView) -> Tuple[StringView, str]:
        """Parses a closing tag returns the string segment, just after
        the closing tag."""
        match = s.match(re.compile(r'</\s*(?P<tagname>[\w:]+)>'))
        assert match
        tagname = match.groupdict()['tagname']
        return s[match.end() - s.begin:], tagname

    def parse_leaf_content(s: StringView) -> Tuple[StringView, str]:
        """Parses a piece of the content of a tag, just until the next opening,
        closing or solitary tag is reached."""
        i = 0
        while s[i] != "<" or s[max(0, i-1)] == "\\":
            i = s.find("<", i)
        return s[i:], s[:i]

    def parse_full_content(s: StringView) -> Tuple[StringView, Node]:
        """Parses the full content of a tag, starting right at the beginning
        of the opening tag and ending right after the closing tag.
        """
        result = []
        s, tagname, attributes, solitary = parse_opening_tag(s)
        name, class_name = (tagname.split(":") + [''])[:2]
        if not solitary:
            while s and not s[:2] == "</":
                s, leaf = parse_leaf_content(s)
                if not s.match(re.compile("\s*$")):
                    result.append(Node(PlainText, leaf))
                if s[:1] == "<" and s[:2] != "</":
                    s, child = parse_full_content(s)
                    result.append(child)
            s, closing_tagname = parse_closing_tag(s)
            assert tagname == closing_tagname
        if len(result) == 1 and isinstance(result[0].parser == PlainText):
            result = result[0].result
        else:
            result = tuple(result)
        return Node(mock_parsers.setdefault(tagname, MockParser(name, ":" + class_name)), result)

    return parse_full_content(xml[xml.search(re.compile(r'<(?!\?)')):])
912

913
# if __name__ == "__main__":
914
#     st = parse_sxpr("(alpha (beta (gamma i\nj\nk) (delta y)) (epsilon z))")
915 916
#     print(st.as_sxpr())
#     print(st.as_xml())