syntaxtree.py 36.2 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 66
#######################################################################
#
# 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()):
67
        self._name = name  # type: str
68 69 70
        self._ptype = ':' + self.__class__.__name__  # type: str

    def __repr__(self):
di68kap's avatar
di68kap committed
71
         return self.name + self.ptype
72 73 74 75

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

76 77 78
    def __call__(self, text: StringView) -> Tuple[Optional['Node'], StringView]:
        return None, text

79 80
    @property
    def name(self):
81 82
        """Returns the name of the parser or the empty string '' for unnamed
        parsers."""
83 84
        return self._name

85 86
    @property
    def ptype(self) -> str:
87 88
        """Returns the type of the parser. By default this is the parser's
        class name preceded by a colon, e.g. ':ZeroOrMore'."""
89 90 91 92
        return self._ptype

    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
93
        """Returns the parser's name if it has a name and repr()"""
94 95
        return self.name if self.name else repr(self)

96
    def reset(self):
97
        """Resets any parser variables. (Should be overridden.)"""
98 99
        pass

eckhart's avatar
eckhart committed
100
    def grammar(self) -> Optional[object]:
101 102
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
103 104 105
        return None

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
106
        """Applies the function `func` to the parser. Returns False, if
107
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
108
        otherwise."""
109 110
        return False

111 112

WHITESPACE_PTYPE = ':Whitespace'
di68kap's avatar
di68kap committed
113
PLAINTEXT_PTYPE = ':PlainText'
114 115 116 117 118 119 120 121 122 123 124 125 126 127 128
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.
    """
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
129
        super().__init__(name)
130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171
        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
#
#######################################################################


172
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
173
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
174
StrictResultType = Union[ChildrenType, StringView, str]
175
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
176 177


Eckhart Arnold's avatar
Eckhart Arnold committed
178
def flatten_sxpr(sxpr: str) -> str:
179
    """Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
180 181 182
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
183
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
184 185
    '(a (b c))'
    """
186
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
187 188


189 190 191 192 193 194 195
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
196
class Node(collections.abc.Sized):
197 198 199
    """
    Represents a node in the concrete or abstract syntax tree.

200
    Attributes:
201 202
        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
203

204 205 206
        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.
207

208 209
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
210

eckhart's avatar
eckhart committed
211 212 213 214
        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
215
        parser (Parser):  The parser which generated this node.
216 217
            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
218 219
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
220
            example by calling ``isinstance(node.parer, ...)``.
221

222 223 224
        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
225
            the length before AST-transformation and will never change
226
            through AST-transformation. READ ONLY!
227

228 229
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
230
            The value of ``pos`` is -1 meaning invalid by default.
231
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
232
            nodes relative to this value.
233 234

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
235
            pos value of the root node should be set to 0 right
236 237
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
238
            Other than that, this value should be considered READ ONLY.
239 240
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
241

242 243
        errors (list):  A list of all errors that occured on this node.

244
        attributes (dict): An optional dictionary of XML-attributes. This
245 246 247
            dictionary is created lazily upon first usage. The attributes
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
248
    """
249

eckhart's avatar
eckhart committed
250
    __slots__ = ['_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content']
251

252 253 254
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
255 256
        that generated the node and the parser's result.
        """
eckhart's avatar
eckhart committed
257
        self.errors = []               # type: List[Error]
eckhart's avatar
eckhart committed
258
        self._pos = -1                  # type: int
di68kap's avatar
di68kap committed
259 260
        # 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
261
        if leafhint:
eckhart's avatar
eckhart committed
262
            self._result = result       # type: StrictResultType
eckhart's avatar
eckhart committed
263
            self._content = None        # type: Optional[str]
264
            self.children = NoChildren  # type: ChildrenType
265
            self._len = -1              # type: int  # lazy evaluation
266 267
        else:
            self.result = result
268 269
        self.parser = parser or ZOMBIE_PARSER

Eckhart Arnold's avatar
Eckhart Arnold committed
270

271
    def __str__(self):
272
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
273
        if self.errors:
274
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
275
                   (s, '; '.join(e.message for e in self.errors))
276
        return s
277

Eckhart Arnold's avatar
Eckhart Arnold committed
278

279 280 281 282 283 284
    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)
285

Eckhart Arnold's avatar
Eckhart Arnold committed
286

287
    def __len__(self):
288
        if self._len < 0:
di68kap's avatar
di68kap committed
289
            self._len = sum(len(child) for child in self.children) \
290
                if self.children else len(self._result)
291 292 293 294 295 296 297 298
        return self._len


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


299
    def __eq__(self, other):
300 301 302 303
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
        """
304
        return self.tag_name == other.tag_name and self.result == other.result
305

Eckhart Arnold's avatar
Eckhart Arnold committed
306

307
    def __hash__(self):
308
        return hash(self.tag_name)
309

Eckhart Arnold's avatar
Eckhart Arnold committed
310

311 312 313
    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
314
        an integer or the first child node with the given tag name. Examples::
315

316
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
317 318
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
319 320
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
321 322 323 324

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
325
        Returns:
326 327
            Node: All nodes which have a given tag name.
        """
328 329 330
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
331
            else:
332 333 334 335 336
                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!')
337 338 339 340


    def __contains__(self, tag_name: str) -> bool:
        """
341
        Returns true if a child with the given tag name exists.
342
        Args:
343 344
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
345 346 347 348
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
349 350 351 352 353
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
354
            return False
355
        raise ValueError('Leave node cannot contain other nodes')
356
        # generator = self.select_by_tag(tag_name, False)
357 358 359 360 361
        # try:
        #     generator.__next__()
        #     return True
        # except StopIteration:
        #     return False
362 363


364
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
365
    def tag_name(self) -> str:
366 367 368 369 370 371
        """
        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`.
        """
372
        return self.parser.name or self.parser.ptype
373

Eckhart Arnold's avatar
Eckhart Arnold committed
374

375
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
376
    def result(self) -> StrictResultType:
377 378 379 380 381
        """
        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.
        """
382 383
        return self._result

384

385
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
386
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
387
        # # made obsolete by static type checking with mypy
388 389 390
        # 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
391 392
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
393 394
        self._len = -1        # lazy evaluation
        self._content = None
395 396 397 398 399 400 401 402 403
        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
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) -> '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
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
451
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
452
        offset = self.pos
453
        for child in self.children:
eckhart's avatar
eckhart committed
454
            child.init_pos(offset)
eckhart's avatar
eckhart committed
455 456 457
            offset = child.pos + len(child)
        return self

458

459 460
    @property
    def attributes(self):
461 462 463
        """
        Returns a dictionary of XML-Attributes attached to the Node.
        """
464 465 466 467
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
468

469 470
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
                   density=0, inline=False) -> str:
471 472 473 474 475 476 477 478
        """
        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.

479
        Args:
480
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
481
            open_fn:   (Node->str) A function that returns an opening
482
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
483
            close_fn:  (Node->str) A function that returns a closeF
484
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
485
            data_fn:   (str->str) A function that filters the data string
486 487 488 489 490 491
                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
492 493
        head = open_fn(self)
        tail = close_fn(self)
494 495

        if not self.result:
496
            return head.rstrip() + tail.lstrip()
497

eckhart's avatar
eckhart committed
498
        tail = tail.lstrip(None if density & 2 else '')
499

500 501 502
        outer_tab = '' if inline else tab
        inline = inline or hasattr(self, '_xml_attr') and '_inline' in self.attributes
        sep, inner_tab = ('', '') if inline else ('\n', tab)
503

504 505
        if self.children:
            content = []
506
            for child in self.children:
507 508
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
                                           density, inline).split('\n')
509
                content.append((sep + inner_tab).join(s for s in subtree))
510
            return head + outer_tab + (sep + inner_tab).join(content) + tail
511

512
        res = cast(str, self.result)  # safe, because if there are no children, result is a string
513 514 515
        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
516
            return head.rstrip() + gap + data_fn(self.result) + tail.lstrip()
517
        else:
eckhart's avatar
eckhart committed
518
            return head + '\n'.join([tab + data_fn(s) for s in res.split('\n')]) + tail
519

Eckhart Arnold's avatar
Eckhart Arnold committed
520

521 522
    def as_sxpr(self, src: str = None, compact: bool = False, showerrors: bool = True,
                indentation: int = 2) -> str:
523 524 525
        """
        Returns content as S-expression, i.e. in lisp-like form.

526
        Args:
527 528 529
            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
530
            compact:  If True a compact representation is returned where
531
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
532
                tree structure.
533 534
        """

eckhart's avatar
eckhart committed
535
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
536
        lbreaks = linebreaks(src) if src else []  # type: List[int]
537

538
        def opening(node) -> str:
eckhart's avatar
eckhart committed
539
            """Returns the opening string for the representation of `node`."""
540
            txt = [left_bracket,  node.tag_name]
541
            # s += " '(pos %i)" % node.add_pos
542
            if hasattr(node, '_xml_attr'):
543 544
                txt.extend(' `(%s "%s")' % (k, v)
                           for k, v in node.attributes.items() if k != '_inline')
545
            if src:
eckhart's avatar
eckhart committed
546
                txt.append(" `(pos %i %i %i)" % (node.pos, *line_col(lbreaks, node.pos)))
di68kap's avatar
di68kap committed
547 548
            # if node.error_flag:   # just for debugging error collecting
            #     txt += " HAS ERRORS"
549
            if showerrors and node.errors:
550
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
551
            return "".join(txt) + '\n'
552

eckhart's avatar
eckhart committed
553 554 555
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
556

eckhart's avatar
eckhart committed
557 558 559 560 561
        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'\"')
562

563
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
564

eckhart's avatar
eckhart committed
565

566
    def as_xml(self, src: str = None, showerrors: bool = True, indentation: int = 2) -> str:
567 568 569
        """
        Returns content as XML-tree.

570
        Args:
571 572 573 574 575
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
        """

576
        def opening(node) -> str:
eckhart's avatar
eckhart committed
577
            """Returns the opening string for the representation of `node`."""            
eckhart's avatar
eckhart committed
578
            txt = ['<', node.tag_name]
579 580
            has_reserved_attrs = hasattr(node, '_xml_attr') \
                and any (r in node.attributes for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
581
            if hasattr(node, '_xml_attr'):
582 583
                txt.extend(' %s="%s"' % (k, v)
                           for k, v in node.attributes.items() if k != '_inline')
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
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
589
            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(' ' * indentation, opening, closing, density=1)
597

Eckhart Arnold's avatar
Eckhart Arnold committed
598

599
    def select(self, match_function: Callable, include_root: bool=False) -> 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
            yield self
eckhart's avatar
eckhart committed
620 621 622
        for child in self.children:
            for node in child.select(match_function, True):
                yield node
623

Eckhart Arnold's avatar
Eckhart Arnold committed
624

eckhart's avatar
eckhart committed
625
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
626
                      include_root: bool=False) -> Iterator['Node']:
627
        """
628 629 630 631
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
632

633
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
634
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
635
            ['(X (c "d"))', '(X "F")']
636
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
637
            ['(b "X")', '(X (c "d"))', '(X "F")']
638
            >>> any(tree.select_by_tag('a', False))
639
            False
640
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
641
            ['(a (b "X") (X (c "d")) (e (X "F")))']
642 643
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
644 645

        Args:
646 647 648 649
            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.
650 651 652
        Yields:
            Node: All nodes which have a given tag name.
        """
653
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
654
            tag_names = frozenset({tag_names})
655
        return self.select(lambda node: node.tag_name in tag_names, include_root)
656 657


658 659 660 661 662 663 664 665 666 667 668 669 670 671 672
    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


673
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
674 675 676
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
677 678 679
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
680
class RootNode(Node):
681
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
682

683 684 685
        errors (list):  A list of all errors that have occured so far during
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
686

687 688
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
689
    """
eckhart's avatar
eckhart committed
690
    def __init__(self, node: Optional[Node] = None) -> 'RootNode':
di68kap's avatar
di68kap committed
691
        super().__init__(ZOMBIE_PARSER, '')
692
        self.all_errors = []
eckhart's avatar
eckhart committed
693
        self.err_nodes_keep = []
di68kap's avatar
di68kap committed
694
        self.error_flag = 0
eckhart's avatar
eckhart committed
695 696
        if node is not None:
            self.swallow(node)
di68kap's avatar
di68kap committed
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
    # def _propagate_errors(self):
    #     if not self.all_errors or not self.error_propagation:
    #         return
    #     self.all_errors.sort(key=lambda e: e.pos)
    #     i = 0
    #     for leaf in self.select(lambda nd: not nd.children, False):
    #         leaf.errors = []
    #         while i < len(self.all_errors) \
    #                 and leaf.pos <= self.all_errors[i].add_pos < leaf.add_pos + leaf.len:
    #             leaf._errors.append(self.all_errors[i])
    #             i += 1
    #         if i >= len(self.all_errors):
    #             break
    #
    # def _propagate_new_error(self, error):
    #     if self.error_propagation:
    #         for leaf in self.select(lambda nd: not nd.children, True):
    #             if leaf.pos <= error.add_pos < leaf.add_pos + leaf.len:
    #                 leaf._errors.append(error)
    #                 break
    #         else:
    #             assert False, "Error %s at pos %i out of bounds" % (str(error), error.add_pos)

    def swallow(self, node: Node) -> 'RootNode':
di68kap's avatar
di68kap committed
722 723 724 725 726
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
727 728
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
729
        self._content = node._content
730
        return self
di68kap's avatar
di68kap committed
731

eckhart's avatar
eckhart committed
732 733
    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
734 735 736 737 738 739
        self.all_errors.append(error)
        self.error_flag = max(self.error_flag, error.code)
        node.errors.append(error)
        self.err_nodes_keep.append(node)
        return self

eckhart's avatar
eckhart committed
740
    def new_error(self,
741
                  node: Node,
di68kap's avatar
di68kap committed
742
                  message: str,
di68kap's avatar
di68kap committed
743
                  code: int = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
744
        """
eckhart's avatar
eckhart committed
745
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
746 747 748 749 750
        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's avatar
eckhart committed
751
        error = Error(message, code, node=node)
eckhart's avatar
eckhart committed
752 753
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
754

eckhart's avatar
eckhart committed
755
    def collect_errors(self) -> List[Error]:
di68kap's avatar
di68kap committed
756 757
        """Returns the list of errors, ordered bv their position.
        """
eckhart's avatar
eckhart committed
758 759 760
        # for node in self.err_nodes:  # lazy evaluation of positions
        #     for err in node.errors:  # moved to error.Error.pos
        #         err.pos = node.pos
761
        self.all_errors.sort(key=lambda e: e.pos)
eckhart's avatar
eckhart committed
762 763
        for node in self.err_nodes_keep:  # redundant: consider removing Error.Error._node_keep
            for error in node.errors:
eckhart's avatar
eckhart committed
764
                assert error._pos < 0 or node.pos <= error._pos <= node.pos + len(node)
765 766
                if error._pos < 0:
                    error._pos = node.pos
eckhart's avatar
eckhart committed
767
        self.err_nodes_keep = []
768
        errors = self.all_errors
eckhart's avatar
eckhart committed
769 770
        # for error in self.all_errors:
        #     _ = error.pos
di68kap's avatar
di68kap committed
771 772 773
        return errors


eckhart's avatar
eckhart committed
774 775 776
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


777
def parse_sxpr(sxpr: str) -> Node:
778
    """
779 780 781 782 783
    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.
784 785

    Example:
786
    >>> parse_sxpr("(a (b c))").as_sxpr()
787 788
    '(a\\n    (b\\n        "c"\\n    )\\n)'
    """
789 790
    sxpr = StringView(sxpr).strip()
    mock_parsers = dict()
791

792
    def next_block(s: StringView):
eckhart's avatar
eckhart committed
793 794 795
        """Generator that yields all characters until the next closing bracket
        that does not match an opening bracket matched earlier within the same
        package."""
796
        s = s.strip()
797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815
        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)
816

817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863
    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
864
                else:
865 866 867 868 869 870 871 872 873 874 875 876 877
                    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
878 879 880
RX_WHITESPACE_TAIL = re.compile(r'\s*$')


881 882 883 884 885
def parse_xml(xml: str) -> Node:
    """
    Generates a tree of nodes from a (Pseudo-)XML-source.
    """
    xml = StringView(xml)
di68kap's avatar
di68kap committed
886 887
    PlainText = MockParser('', PLAINTEXT_PTYPE)
    mock_parsers = {PLAINTEXT_PTYPE: PlainText}
888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907

    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
908 909
        section = s[match.end() - s.begin:]
        s, attributes = parse_attributes(section)
910 911
        i = s.find('>')
        assert i >= 0
912
        return s[i+1:], tagname, attributes, s[i-1] == "/"
913 914

    def parse_closing_tag(s: StringView) -> Tuple[StringView, str]:
915
        """Parses a closing tag and returns the string segment, just after
916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939
        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)