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

syntaxtree.py 36 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
__all__ = ('ParserBase',
           'WHITESPACE_PTYPE',
di68kap's avatar
di68kap committed
39
           'PLAINTEXT_PTYPE',
40 41 42 43
           'TOKEN_PTYPE',
           'MockParser',
           'ZombieParser',
           'ZOMBIE_PARSER',
eckhart's avatar
eckhart committed
44
           'ZOMBIE_NODE',
45
           'Node',
di68kap's avatar
di68kap committed
46
           'RootNode',
47
           'parse_sxpr',
48 49 50
           'parse_xml',
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
51 52


53 54 55 56 57 58 59 60 61 62 63 64 65
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


class ParserBase:
    """
    ParserBase is the base class for all real and mock parser classes.
    It is defined here, because Node objects require a parser object
    for instantiation.
    """
Eckhart Arnold's avatar
Eckhart Arnold committed
66 67
    __slots__ = '_name', '_ptype'

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

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

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

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

81 82
    @property
    def name(self):
83 84
        """Returns the name of the parser or the empty string '' for unnamed
        parsers."""
85 86
        return self._name

87 88
    @property
    def ptype(self) -> str:
89 90
        """Returns the type of the parser. By default this is the parser's
        class name preceded by a colon, e.g. ':ZeroOrMore'."""
91 92 93 94
        return self._ptype

    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
95
        """Returns the parser's name if it has a name and repr()"""
96 97
        return self.name if self.name else repr(self)

98
    def reset(self):
99
        """Resets any parser variables. (Should be overridden.)"""
100 101
        pass

eckhart's avatar
eckhart committed
102
    def grammar(self) -> Optional[object]:
103 104
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
105 106 107
        return None

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
108
        """Applies the function `func` to the parser. Returns False, if
109
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
110
        otherwise."""
111 112
        return False

113 114

WHITESPACE_PTYPE = ':Whitespace'
di68kap's avatar
di68kap committed
115
PLAINTEXT_PTYPE = ':PlainText'
116 117 118 119 120 121 122 123 124 125 126 127 128
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.
    """
Eckhart Arnold's avatar
Eckhart Arnold committed
129 130
    __slots__ = ()

131 132
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
133
        super().__init__(name)
134 135 136 137 138 139 140 141 142 143 144 145 146 147
        self._ptype = ptype or ':' + self.__class__.__name__


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

    ``ZombieParser`` is the class of the singelton object
    ``ZOMBIE_PARSER``. The  ``ZOMBIE_PARSER`` has a name and can be
    called, but it never matches. It serves as a substitute where only
    these (or one of these properties) is needed, but no real Parser-
    object is instantiated.
    """
    alive = False
Eckhart Arnold's avatar
Eckhart Arnold committed
148
    __slots__ = ()
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176

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


177
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
178
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
179
StrictResultType = Union[ChildrenType, StringView, str]
180
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
181 182


Eckhart Arnold's avatar
Eckhart Arnold committed
183
def flatten_sxpr(sxpr: str) -> str:
184
    """Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
185 186 187
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
188
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
189 190
    '(a (b c))'
    """
191
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
192 193


194 195 196 197 198 199 200
def flatten_xml(xml: str) -> str:
    """Returns an XML-tree as a one linter without unnecessary whitespace,
    i.e. only whitespace within leaf-nodes is preserved.
    """
    return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?<=</\w+>)\s+', '', xml))


Eckhart Arnold's avatar
Eckhart Arnold committed
201
class Node(collections.abc.Sized):
202 203 204
    """
    Represents a node in the concrete or abstract syntax tree.

205
    Attributes:
206 207
        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
208

209 210 211
        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.
212

213 214
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
215

eckhart's avatar
eckhart committed
216 217 218 219
        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
220
        parser (Parser):  The parser which generated this node.
221 222
            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
223 224
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
225
            example by calling ``isinstance(node.parer, ...)``.
226

227 228 229
        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
230
            the length before AST-transformation and will never change
231
            through AST-transformation. READ ONLY!
232

233 234
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
235
            The value of ``pos`` is -1 meaning invalid by default.
236
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
237
            nodes relative to this value.
238 239

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
240
            pos value of the root node should be set to 0 right
241 242
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
243
            Other than that, this value should be considered READ ONLY.
244 245
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
246

247 248
        errors (list):  A list of all errors that occured on this node.

249
        attributes (dict): An optional dictionary of XML-attributes. This
250 251 252
            dictionary is created lazily upon first usage. The attributes
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
253
    """
254

Eckhart Arnold's avatar
Eckhart Arnold committed
255
    __slots__ = '_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content'
256

257 258 259
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
260 261
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
262
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
263
        self._pos = -1                  # type: int
di68kap's avatar
di68kap committed
264 265
        # 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
266
        if leafhint:
eckhart's avatar
eckhart committed
267
            self._result = result       # type: StrictResultType
eckhart's avatar
eckhart committed
268
            self._content = None        # type: Optional[str]
269
            self.children = NoChildren  # type: ChildrenType
270
            self._len = -1              # type: int  # lazy evaluation
271 272
        else:
            self.result = result
273 274
        self.parser = parser or ZOMBIE_PARSER

Eckhart Arnold's avatar
Eckhart Arnold committed
275

276
    def __str__(self):
Eckhart Arnold's avatar
Eckhart Arnold committed
277 278 279
        # s = self._content if self._content is not None else \
        #     "".join(str(child) for child in self.children) if self.children else self.content
        s = self.content
eckhart's avatar
eckhart committed
280
        if self.errors:
281
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
282
                   (s, '; '.join(e.message for e in self.errors))
283
        return s
284

Eckhart Arnold's avatar
Eckhart Arnold committed
285

286 287 288 289 290 291
    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)
292

Eckhart Arnold's avatar
Eckhart Arnold committed
293

294
    def __len__(self):
295
        if self._len < 0:
di68kap's avatar
di68kap committed
296
            self._len = sum(len(child) for child in self.children) \
297
                if self.children else len(self._result)
298 299 300 301 302 303 304 305
        return self._len


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


306
    def __eq__(self, other):
307 308 309 310
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
        """
311
        return self.tag_name == other.tag_name and self.result == other.result
312

Eckhart Arnold's avatar
Eckhart Arnold committed
313

314
    def __hash__(self):
315
        return hash(self.tag_name)
316

Eckhart Arnold's avatar
Eckhart Arnold committed
317

318 319 320
    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
321
        an integer or the first child node with the given tag name. Examples::
322

323
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
324 325
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
326 327
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
328 329 330 331

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
332
        Returns:
333 334
            Node: All nodes which have a given tag name.
        """
335 336 337
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
338
            else:
339 340 341 342 343
                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!')
344 345 346 347


    def __contains__(self, tag_name: str) -> bool:
        """
348
        Returns true if a child with the given tag name exists.
349
        Args:
350 351
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
352 353 354 355
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
356 357 358 359 360
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
361
            return False
362
        raise ValueError('Leave node cannot contain other nodes')
363
        # generator = self.select_by_tag(tag_name, False)
364 365 366 367 368
        # try:
        #     generator.__next__()
        #     return True
        # except StopIteration:
        #     return False
369 370


371
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
372
    def tag_name(self) -> str:
373 374 375 376 377 378
        """
        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`.
        """
379
        return self.parser.name or self.parser.ptype
380

Eckhart Arnold's avatar
Eckhart Arnold committed
381

382
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
383
    def result(self) -> StrictResultType:
384 385 386 387 388
        """
        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.
        """
389 390
        return self._result

391

392
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
393
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
394
        # # made obsolete by static type checking with mypy
395 396 397
        # 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
398 399
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
400 401
        self._len = -1        # lazy evaluation
        self._content = None
402 403 404 405 406 407 408 409 410
        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
411
                self._result = result
412

413 414

    @property
eckhart's avatar
eckhart committed
415
    def content(self) -> str:
416
        """
417
        Returns content as string, omitting error messages.
418
        """
419 420 421 422
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
423
                # self._content = self._result
424 425 426
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
427 428 429 430 431 432 433 434 435 436 437


    @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))


438
    @property
439
    def pos(self) -> int:
eckhart's avatar
eckhart committed
440 441
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
442
            raise AssertionError("Position value not initialized!")
443 444
        return self._pos

eckhart's avatar
eckhart committed
445

eckhart's avatar
eckhart committed
446
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
447 448 449 450 451 452 453 454 455
        """
        (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
456 457
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
458
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
459
        offset = self.pos
460
        for child in self.children:
eckhart's avatar
eckhart committed
461
            child.init_pos(offset)
eckhart's avatar
eckhart committed
462 463 464
            offset = child.pos + len(child)
        return self

465

466 467
    @property
    def attributes(self):
468 469 470
        """
        Returns a dictionary of XML-Attributes attached to the Node.
        """
471 472 473 474
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
475

476
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
477
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
478 479 480 481 482 483 484 485
        """
        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.

486
        Args:
487
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
488
            open_fn:   (Node->str) A function that returns an opening
489
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
490
            close_fn:  (Node->str) A function that returns a closeF
491
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
492
            data_fn:   (str->str) A function that filters the data string
493 494 495 496 497 498
                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
499 500
        head = open_fn(self)
        tail = close_fn(self)
501 502

        if not self.result:
503
            return head.rstrip() + tail.lstrip()
504

eckhart's avatar
eckhart committed
505
        tail = tail.lstrip(None if density & 2 else '')
506

507
        outer_tab = '' if inline else tab
508
        inline = inline or inline_fn(self)
509
        sep, inner_tab = ('', '') if inline else ('\n', tab)
510

511 512
        if self.children:
            content = []
513
            for child in self.children:
514
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
515
                                           density, inline, inline_fn).split('\n')
516
                content.append((sep + inner_tab).join(s for s in subtree))
517
            return head + outer_tab + (sep + inner_tab).join(content) + tail
518

519
        res = cast(str, self.result)  # safe, because if there are no children, result is a string
520 521
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
eckhart's avatar
eckhart committed
522
            gap = ' ' if head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
523
            return head.rstrip() + gap + data_fn(self.result) + tail.lstrip()
524
        else:
eckhart's avatar
eckhart committed
525
            return head + '\n'.join([tab + data_fn(s) for s in res.split('\n')]) + tail
526

Eckhart Arnold's avatar
Eckhart Arnold committed
527

528 529
    def as_sxpr(self, src: str = None, showerrors: bool = True, indentation: int = 2,
                compact: bool = False) -> str:
530 531 532
        """
        Returns content as S-expression, i.e. in lisp-like form.

533
        Args:
534 535 536
            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
537
            compact:  If True a compact representation is returned where
538
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
539
                tree structure.
540 541
        """

eckhart's avatar
eckhart committed
542
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
543
        lbreaks = linebreaks(src) if src else []  # type: List[int]
544

545
        def opening(node) -> str:
eckhart's avatar
eckhart committed
546
            """Returns the opening string for the representation of `node`."""
547
            txt = [left_bracket,  node.tag_name]
548
            # s += " '(pos %i)" % node.add_pos
549
            if hasattr(node, '_xml_attr'):
550
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attributes.items())
551
            if src:
eckhart's avatar
eckhart committed
552
                txt.append(" `(pos %i %i %i)" % (node.pos, *line_col(lbreaks, node.pos)))
553
            if showerrors and node.errors:
554
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
555
            return "".join(txt) + '\n'
556

eckhart's avatar
eckhart committed
557 558 559
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
560

eckhart's avatar
eckhart committed
561 562 563 564 565
        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'\"')
566

567
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
568

eckhart's avatar
eckhart committed
569

570
    def as_xml(self, src: str = None, showerrors: bool = True, indentation: int = 2,
Eckhart Arnold's avatar
Eckhart Arnold committed
571
               inline_tags: Set[str]=set(), omit_tags: Set[str]=set()) -> str:
572 573 574
        """
        Returns content as XML-tree.

575
        Args:
576 577 578
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
579 580 581 582 583 584 585
            inline_tags:  A set of tag names, the content of which will always be written
                on a single line, unless it contains explicit line feeds ('\n').
            omit_tags:  A set of tags from which only the content will be printed, but
                neither the opening tag nor its attributes nor the closing tag. This
                allows producing a mix of plain text and child tags in the output,
                which otherwise is not supported by the Node object, because it
                requires its content to be either a tuple of children or string content.
586 587
        """

588
        def opening(node) -> str:
589 590 591
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
592
            txt = ['<', node.tag_name]
593 594
            has_reserved_attrs = hasattr(node, '_xml_attr') \
                and any (r in node.attributes for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
595
            if hasattr(node, '_xml_attr'):
596
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attributes.items())
597
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
598
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
599
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
600 601
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
602
            return "".join(txt + [">\n"])
603 604

        def closing(node):
605 606 607
            """Returns the closing string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
608
            return ('\n</') + node.tag_name + '>'
609

610 611 612 613 614 615 616
        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.
            """
            return node.tag_name in inline_tags

617
        line_breaks = linebreaks(src) if src else []
618 619
        return self._tree_repr(' ' * indentation, opening, closing,
                               density=1, inline_fn=inlining)
620

Eckhart Arnold's avatar
Eckhart Arnold committed
621

622
    def select(self, match_function: Callable, include_root: bool=False) -> Iterator['Node']:
eckhart's avatar
eckhart committed
623
        """
624
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
625

626
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
627
        given `match_function` evaluates to True. The tree is
628 629 630
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
631

632 633 634
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
635 636
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
637
        Yields:
638
            Node: All nodes of the tree for which
639 640
            ``match_function(node)`` returns True
        """
641
        if include_root and match_function(self):
642
            yield self
eckhart's avatar
eckhart committed
643 644 645
        for child in self.children:
            for node in child.select(match_function, True):
                yield node
646

Eckhart Arnold's avatar
Eckhart Arnold committed
647

eckhart's avatar
eckhart committed
648
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
649
                      include_root: bool=False) -> Iterator['Node']:
650
        """
651 652 653 654
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
655

656
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
657
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
658
            ['(X (c "d"))', '(X "F")']
659
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
660
            ['(b "X")', '(X (c "d"))', '(X "F")']
661
            >>> any(tree.select_by_tag('a', False))
662
            False
663
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
664
            ['(a (b "X") (X (c "d")) (e (X "F")))']
665 666
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
667 668

        Args:
669 670 671 672
            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.
673 674 675
        Yields:
            Node: All nodes which have a given tag name.
        """
676
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
677
            tag_names = frozenset({tag_names})
678
        return self.select(lambda node: node.tag_name in tag_names, include_root)
679 680


681 682 683 684 685 686 687 688 689 690 691 692 693 694 695
    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


696
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
697 698 699
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
700 701 702
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
703
class RootNode(Node):
704
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
705

Eckhart Arnold's avatar
Eckhart Arnold committed
706
        all_errors (list):  A list of all errors that have occured so far during
707 708
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
709

710 711
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
712
    """
eckhart's avatar
eckhart committed
713
    def __init__(self, node: Optional[Node] = None) -> 'RootNode':
di68kap's avatar
di68kap committed
714
        super().__init__(ZOMBIE_PARSER, '')
715
        self.all_errors = []
di68kap's avatar
di68kap committed
716
        self.error_flag = 0
eckhart's avatar
eckhart committed
717 718
        if node is not None:
            self.swallow(node)
di68kap's avatar
di68kap committed
719

720
    def swallow(self, node: Node) -> 'RootNode':
721 722 723 724 725 726 727 728 729 730
        """Put `self` in the place of `node` by copying all its data.
        Returns self.

        This is done by the parse.Grammar object after
        parsing has finished, so that the Grammar object always
        returns a syntax tree rooted in a RootNode object.

        It is possible to add errors to a RootNode object, before it
        has actually swallowed the root of the syntax tree.
        """
di68kap's avatar
di68kap committed
731 732 733 734 735
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
736 737
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
738
        self._content = node._content
739
        return self
di68kap's avatar
di68kap committed
740

eckhart's avatar
eckhart committed
741 742
    def add_error(self, node: Node, error: Error) -> 'RootNode':
        """Adds an Error object to the tree, locating it at a specific node."""
di68kap's avatar
di68kap committed
743 744 745 746 747
        self.all_errors.append(error)
        self.error_flag = max(self.error_flag, error.code)
        node.errors.append(error)
        return self

eckhart's avatar
eckhart committed
748
    def new_error(self,
749
                  node: Node,
di68kap's avatar
di68kap committed
750
                  message: str,
di68kap's avatar
di68kap committed
751
                  code: int = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
752
        """
eckhart's avatar
eckhart committed
753
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
754 755 756 757 758
        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
759
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
760 761
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
762

eckhart's avatar
eckhart committed
763
    def collect_errors(self) -> List[Error]:
di68kap's avatar
di68kap committed
764 765
        """Returns the list of errors, ordered bv their position.
        """
766
        self.all_errors.sort(key=lambda e: e.pos)
Eckhart Arnold's avatar
Eckhart Arnold committed
767
        return self.all_errors
di68kap's avatar
di68kap committed
768 769


eckhart's avatar
eckhart committed
770 771 772
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


773
def parse_sxpr(sxpr: str) -> Node:
774
    """
775 776 777 778 779
    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.
780 781

    Example:
782
    >>> parse_sxpr("(a (b c))").as_sxpr()
783 784
    '(a\\n    (b\\n        "c"\\n    )\\n)'
    """
785 786
    sxpr = StringView(sxpr).strip()
    mock_parsers = dict()
787

788
    def next_block(s: StringView):
eckhart's avatar
eckhart committed
789 790 791
        """Generator that yields all characters until the next closing bracket
        that does not match an opening bracket matched earlier within the same
        package."""
792
        s = s.strip()
793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811
        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)
812

813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859
    def inner_parser(sxpr: StringView) -> Node:
        if sxpr[0] != '(':
            raise ValueError('"(" expected, not ' + sxpr[:10])
        # assert sxpr[0] == '(', sxpr
        sxpr = sxpr[1:].strip()
        match = sxpr.match(re.compile(r'[\w:]+'))
        if match is None:
            raise AssertionError('Malformed S-expression Node-tagname or identifier expected, '
                                 'not "%s"' % sxpr[:40].replace('\n', ''))
        end = match.end() - sxpr.begin
        tagname = sxpr[:end]
        name, class_name = (tagname.split(':') + [''])[:2]
        sxpr = sxpr[end:].strip()
        attributes = OrderedDict()
        if sxpr[0] == '(':
            result = tuple(inner_parser(block) for block in next_block(sxpr))
        else:
            lines = []
            while sxpr and sxpr[0:1] != ')':
                # parse attributes
                while sxpr[:2] == "`(":
                    i = sxpr.find('"')
                    k = sxpr.find(')')
                    # read very special attribute pos
                    if sxpr[2:5] == "pos" and 0 < i < k:
                        pos = int(sxpr[5:k].strip().split(' ')[0])
                    # ignore very special attribute err
                    elif sxpr[2:5] == "err" and 0 <= sxpr.find('`', 5) < k:
                        m = sxpr.find('(', 5)
                        while m >= 0 and m < k:
                            m = sxpr.find('(', k)
                            k = max(k, sxpr.find(')', max(m, 0)))
                    # read attributes
                    else:
                        attr = sxpr[2:i].strip()
                        value = sxpr[i:k].strip()[1:-1]
                        attributes[attr] = value
                    sxpr = sxpr[k+1:].strip()
                # parse content
                for qtmark in ['"""', "'''", '"', "'"]:
                    match = sxpr.match(re.compile(qtmark + r'.*?' + qtmark, re.DOTALL))
                    if match:
                        end = match.end() - sxpr.begin
                        i = len(qtmark)
                        lines.append(str(sxpr[i:end - i]))
                        sxpr = sxpr[end:].strip()
                        break
860
                else:
861 862 863 864 865 866 867 868 869 870 871 872 873
                    match = sxpr.match(re.compile(r'(?:(?!\)).)*', re.DOTALL))
                    end = match.end() - sxpr.begin
                    lines.append(str(sxpr[:end]))
                    sxpr = sxpr[end:]
            result = "\n".join(lines)
        node = Node(mock_parsers.setdefault(tagname, MockParser(name, ':' + class_name)), result)
        if attributes:
            node.attributes.update(attributes)
        return node

    return inner_parser(sxpr)


di68kap's avatar
di68kap committed
874 875 876
RX_WHITESPACE_TAIL = re.compile(r'\s*$')


877 878 879 880 881
def parse_xml(xml: str) -> Node:
    """
    Generates a tree of nodes from a (Pseudo-)XML-source.
    """
    xml = StringView(xml)
di68kap's avatar
di68kap committed
882 883
    PlainText = MockParser('', PLAINTEXT_PTYPE)
    mock_parsers = {PLAINTEXT_PTYPE: PlainText}
884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903

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

    def parse_opening_tag(s: StringView) -> Tuple[StringView, str, OrderedDict, bool]:
        """Parses an opening tag. Returns the string segment following the
        the opening tag, the tag name, a dictionary of attributes and
        a flag indicating whether the tag is actually a solitary tag as
        indicated by a slash at the end, i.e. <br/>."""
        match = s.match(re.compile(r'<\s*(?P<tagname>[\w:]+)\s*'))
        assert match
        tagname = match.groupdict()['tagname']
di68kap's avatar
di68kap committed
904 905
        section = s[match.end() - s.begin:]
        s, attributes = parse_attributes(section)
906 907
        i = s.find('>')
        assert i >= 0
908
        return s[i+1:], tagname, attributes, s[i-1] == "/"
909 910

    def parse_closing_tag(s: StringView) -> Tuple[StringView, str]: