syntaxtree.py 41.3 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
# syntaxtree.py - syntax tree classes for DHParser
#
# Copyright 2016  by Eckhart Arnold (arnold@badw.de)
#                 Bavarian Academy of Sciences an Humanities (badw.de)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied.  See the License for the specific language governing
# permissions and limitations under the License.
17 18 19


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

26

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

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

36

37 38 39 40 41 42
__all__ = ('ParserBase',
           'WHITESPACE_PTYPE',
           'TOKEN_PTYPE',
           'MockParser',
           'ZombieParser',
           'ZOMBIE_PARSER',
eckhart's avatar
eckhart committed
43
           'ZOMBIE_NODE',
eckhart's avatar
eckhart committed
44 45 46
           'ResultType',
           'StrictResultType',
           'ChildrenType',
47
           'Node',
di68kap's avatar
di68kap committed
48
           'RootNode',
eckhart's avatar
eckhart committed
49
           'ZOMBIE_ROOTNODE',
50
           'parse_sxpr',
51 52 53
           'parse_xml',
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
54 55


56 57 58 59 60 61 62 63 64 65 66 67 68
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


class ParserBase:
    """
    ParserBase is the base class for all real and mock parser classes.
    It is defined here, because Node objects require a parser object
    for instantiation.
    """
69

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

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

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

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

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

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

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

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

eckhart's avatar
eckhart committed
100 101 102 103
    def apply(self, func: Callable):
        """Applies the function `func` recursively to the parser and all
        descendant parsers, if any exist."""
        pass
104

105 106

WHITESPACE_PTYPE = ':Whitespace'
107
TOKEN_PTYPE = ':Token'
108 109 110 111 112 113


class MockParser(ParserBase):
    """
    MockParser objects can be used to reconstruct syntax trees from a
    serialized form like S-expressions or XML. Mock objects can mimic
eckhart's avatar
eckhart committed
114
    different parser types by assigning them a `ptype` on initialization.
115 116 117 118 119

    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.
    """
120

Eckhart Arnold's avatar
Eckhart Arnold committed
121 122
    __slots__ = ()

123 124
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
125 126
        super().__init__()
        self.name = name
di68kap's avatar
di68kap committed
127 128
        if ptype:
            self.ptype = ptype  # or ':' + self.__class__.__name__
129 130 131 132 133 134 135 136 137 138 139 140


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.
    """
141

142
    alive = False
Eckhart Arnold's avatar
Eckhart Arnold committed
143
    __slots__ = ()
144 145

    def __init__(self):
146
        super(ZombieParser, self).__init__()
147 148
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
149
        self.name = "__ZOMBIE__"
150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
        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
#
#######################################################################


173
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
174
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
175
StrictResultType = Union[ChildrenType, StringView, str]
176
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
177 178


Eckhart Arnold's avatar
Eckhart Arnold committed
179
def flatten_sxpr(sxpr: str) -> str:
180 181
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
182 183 184
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
185
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
186 187
    '(a (b c))'
    """
188

189
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
190 191


192
def flatten_xml(xml: str) -> str:
193 194
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
195
    i.e. only whitespace within leaf-nodes is preserved.
196 197
    A more precise alternative to `flatten_xml` is to use Node.as_xml()
    ans passing a set containing the top level tag to parameter `inline_tags`.
198
    """
199

200 201 202 203
    # works only with regex
    # return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?<=</\w+>)\s+', '', xml))
    def tag_only(m):
        return m.groupdict()['closing_tag']
204
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
205 206


eckhart's avatar
eckhart committed
207
RX_AMP = re.compile(r'&(?!\w+;)')
208 209


Eckhart Arnold's avatar
Eckhart Arnold committed
210
class Node:  # (collections.abc.Sized): Base class omitted for cython-compatibility
211 212 213
    """
    Represents a node in the concrete or abstract syntax tree.

214 215
    TODO: Add some documentation and doc-tests here...

216
    Attributes:
217 218
        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
219

220 221 222
        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.
223

224 225
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
226

eckhart's avatar
eckhart committed
227 228 229 230
        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
231
        parser (Parser):  The parser which generated this node.
232 233
            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
234 235
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
236
            example by calling ``isinstance(node.parer, ...)``.
237

238 239 240
        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
241
            the length before AST-transformation and will never change
242
            through AST-transformation. READ ONLY!
243

244 245
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
246
            The value of ``pos`` is -1 meaning invalid by default.
247
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
248
            nodes relative to this value.
249 250

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
251
            pos value of the root node should be set to 0 right
252 253
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
254
            Other than that, this value should be considered READ ONLY.
255 256
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
257

258 259
        errors (list):  A list of all errors that occured on this node.

260 261
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
262 263
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
264
    """
265

Eckhart Arnold's avatar
Eckhart Arnold committed
266
    __slots__ = '_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content'
267

268 269 270
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
271 272
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
273
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
274
        self._pos = -1                  # type: int
275
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
276
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
277
        if leafhint:
eckhart's avatar
eckhart committed
278
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
eckhart's avatar
eckhart committed
279
            self._content = None        # type: Optional[str]
280
            self.children = NoChildren  # type: ChildrenType
281
            self._len = -1              # type: int  # lazy evaluation
282 283
        else:
            self.result = result
284 285
        self.parser = parser or ZOMBIE_PARSER

eckhart's avatar
eckhart committed
286 287 288 289 290 291 292 293 294 295 296
    def __deepcopy__(self, memo):
        if self.children:
            duplicate = self.__class__(self.parser, copy.deepcopy(self.children), False)
        else:
            duplicate = self.__class__(self.parser, self.result, True)
        duplicate.errors = copy.deepcopy(self.errors) if self.errors else []
        duplicate._pos = self._pos
        duplicate._len = self._len
        if hasattr(self, '_xml_attr'):
            duplicate._xml_attr = copy.deepcopy(self._xml_attr)
        return duplicate
Eckhart Arnold's avatar
Eckhart Arnold committed
297

298
    def __str__(self):
299
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
300
        if self.errors:
301
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
302
                   (s, '; '.join(e.message for e in self.errors))
303
        return s
304

Eckhart Arnold's avatar
Eckhart Arnold committed
305

306 307 308 309
    def __repr__(self):
        mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
        parg = "MockParser({name}, {ptype})".format(**mpargs)
        rarg = str(self) if not self.children else \
eckhart's avatar
eckhart committed
310
            "(" + ", ".join(repr(child) for child in self.children) + ")"
311
        return "Node(%s, %s)" % (parg, rarg)
312

Eckhart Arnold's avatar
Eckhart Arnold committed
313

314
    def __len__(self):
315
        if self._len < 0:
di68kap's avatar
di68kap committed
316
            self._len = sum(len(child) for child in self.children) \
317
                if self.children else len(self._result)
318 319 320 321 322 323 324 325
        return self._len


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


326
    def __eq__(self, other):
327 328 329
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
330 331 332 333

        Note: It is not required that two nodes have the same errors attached.
        In case you need to check for error equality as well, compare a
        serialization that includes error messages, as_sxpr() will do!
334
        """
335
        return self.tag_name == other.tag_name and self.result == other.result
336

Eckhart Arnold's avatar
Eckhart Arnold committed
337

338
    def __hash__(self):
339
        return hash(self.tag_name)
340

Eckhart Arnold's avatar
Eckhart Arnold committed
341

342 343 344
    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
345
        an integer or the first child node with the given tag name. Examples::
346

347
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
348 349
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
350 351
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
352 353 354 355

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
356
        Returns:
357 358
            Node: All nodes which have a given tag name.
        """
359 360 361
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
362
            else:
363 364 365 366 367
                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!')
368 369 370 371


    def __contains__(self, tag_name: str) -> bool:
        """
372
        Returns true if a child with the given tag name exists.
373
        Args:
374 375
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
376 377 378 379
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
380 381 382 383 384
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
385
            return False
386
        raise ValueError('Leave node cannot contain other nodes')
387 388


389 390 391 392 393 394 395 396 397 398 399 400 401 402
    def get(self, index_or_tagname: Union[int, str],
            surrogate: Union['Node', Iterator['Node']]) -> Union['Node', Iterator['Node']]:
        """Returns the child node with the given index if ``index_or_tagname``
        is an integer or the first child node with the given tag name. If no
        child with the given index or tag_name exists, the ``surrogate`` is
        returned instead. This mimics the behaviour of Python's dictionary's
        get-method.
        """
        try:
            return self[index_or_tagname]
        except KeyError:
            return surrogate


403
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
404
    def tag_name(self) -> str:
405 406 407 408 409 410
        """
        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`.
        """
411
        return self.parser.name or self.parser.ptype
412

Eckhart Arnold's avatar
Eckhart Arnold committed
413

414
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
415
    def result(self) -> StrictResultType:
416 417 418 419 420
        """
        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.
        """
421 422
        return self._result

423

424
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
425
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
426
        # # made obsolete by static type checking with mypy
427 428 429
        # 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
430 431
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
432 433
        self._len = -1        # lazy evaluation
        self._content = None
434 435 436 437 438 439 440 441 442
        if isinstance(result, Node):
            self.children = (result,)
            self._result = self.children
        else:
            if isinstance(result, tuple):
                self.children = result
                self._result = result or ''
            else:
                self.children = NoChildren
eckhart's avatar
eckhart committed
443
                self._result = result  # cast(StrictResultType, result)
444

445 446

    @property
eckhart's avatar
eckhart committed
447
    def content(self) -> str:
448
        """
di68kap's avatar
di68kap committed
449 450 451
        Returns content as string, omitting error messages. If the node has
        child-nodes, the string content of the child-nodes is recursively read
        and then concatenated.
452
        """
453 454 455 456
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
457
                # self._content = self._result
458 459 460
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
461 462


di68kap's avatar
di68kap committed
463 464 465 466 467
    @content.setter
    def content(self, content: str):
        self.result = content


468 469 470 471 472 473 474 475 476
    @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))


477
    @property
478
    def pos(self) -> int:
eckhart's avatar
eckhart committed
479 480
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
481
            raise AssertionError("Position value not initialized!")
482 483
        return self._pos

eckhart's avatar
eckhart committed
484

eckhart's avatar
eckhart committed
485
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
486 487 488 489 490 491 492 493 494
        """
        (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
495 496
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
497
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
498
        offset = self.pos
499
        for child in self.children:
eckhart's avatar
eckhart committed
500
            child.init_pos(offset)
eckhart's avatar
eckhart committed
501 502 503
            offset = child.pos + len(child)
        return self

504

505
    @property
506
    def attr(self):
507
        """
508
        Returns a dictionary of XML-attr attached to the node.
509
        """
510 511 512 513
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
514

515
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
516
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
517 518 519 520 521 522 523 524
        """
        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.

525
        Args:
526
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
527
            open_fn:   (Node->str) A function that returns an opening
528
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
529
            close_fn:  (Node->str) A function that returns a closeF
530
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
531
            data_fn:   (str->str) A function that filters the data string
532 533 534 535 536 537
                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
538 539
        head = open_fn(self)
        tail = close_fn(self)
540 541

        if not self.result:
542
            return head.rstrip() + tail.lstrip()
543

eckhart's avatar
eckhart committed
544
        tail = tail.lstrip(None if density & 2 else '')
545

546
        inline = inline or inline_fn(self)
547 548 549 550 551
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
552 553
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
554

555 556
        if self.children:
            content = []
557
            for child in self.children:
558
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
559
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
560
                if subtree:
eckhart's avatar
eckhart committed
561 562
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
563
            return head + usetab + (sep + usetab).join(content) + tail
564

eckhart's avatar
eckhart committed
565
        res = self.content
eckhart's avatar
eckhart committed
566
        if not inline and not head:
eckhart's avatar
eckhart committed
567
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
568
            res = res.strip()
569 570
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
571
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
572
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
573
        else:
574
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
575

Eckhart Arnold's avatar
Eckhart Arnold committed
576

577 578 579
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
580
                compact: bool = False) -> str:
581 582 583
        """
        Returns content as S-expression, i.e. in lisp-like form.

584
        Args:
585 586 587
            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.
588 589 590
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
591
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
592
                tree structure.
593 594
        """

eckhart's avatar
eckhart committed
595
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
596
        lbreaks = linebreaks(src) if src else []  # type: List[int]
597

598
        def opening(node) -> str:
eckhart's avatar
eckhart committed
599
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
600
            txt = [left_bracket, node.tag_name]
601
            # s += " '(pos %i)" % node.add_pos
602
            if hasattr(node, '_xml_attr'):
603
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
604
            if src:
605 606
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
607
            if showerrors and node.errors:
608
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
609
            return "".join(txt) + '\n'
610

eckhart's avatar
eckhart committed
611 612 613
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
614

eckhart's avatar
eckhart committed
615 616 617 618 619
        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'\"')
620

621
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
622

eckhart's avatar
eckhart committed
623

624 625 626
    def as_xml(self, src: str = None,
               showerrors: bool = True,
               indentation: int = 2,
eckhart's avatar
eckhart committed
627 628 629
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
630 631 632
        """
        Returns content as XML-tree.

633
        Args:
634 635 636
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
637 638
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
639 640 641
            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
642
                neither the opening tag nor its attr nor the closing tag. This
643 644 645
                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.
646 647
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
648 649
        """

650
        def opening(node) -> str:
651 652 653
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
654
            txt = ['<', node.tag_name]
655
            has_reserved_attrs = hasattr(node, '_xml_attr') \
eckhart's avatar
eckhart committed
656
                and any(r in node.attr for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
657
            if hasattr(node, '_xml_attr'):
658
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
659
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
660
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
661
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
662 663
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
664 665 666
            if node.tag_name in empty_tags:
                assert not node.result, ("Node %s with content %s is not an empty element!" %
                                         (node.tag_name, str(node)))
eckhart's avatar
eckhart committed
667
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
668 669 670
            else:
                ending = ">\n"
            return "".join(txt + [ending])
671 672

        def closing(node):
673
            """Returns the closing string for the representation of `node`."""
674
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
675
                return ''
676
            return ('\n</') + node.tag_name + '>'
677

678 679 680 681 682 683 684
        def sanitizer(content: str) -> str:
            """Substitute "&", "<", ">" in XML-content by the respective entities."""
            content = RX_AMP.sub('&amp;', content)
            content = content.replace('<', '&lt;').replace('>', '&gt;')
            return content


685 686 687 688 689
        def inlining(node):
            """Returns True, if `node`'s tag name is contained in `inline_tags`,
            thereby signalling that the children of this node shall not be
            printed on several lines to avoid unwanted gaps in the output.
            """
eckhart's avatar
eckhart committed
690 691 692
            return node.tag_name in inline_tags \
                or (hasattr(node, '_xml_attr')
                    and node.attr.get('xml:space', 'default') == 'preserve')
693

694
        line_breaks = linebreaks(src) if src else []
695
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
696
                               density=1, inline_fn=inlining)
697

Eckhart Arnold's avatar
Eckhart Arnold committed
698

eckhart's avatar
eckhart committed
699
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
700
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
701
        """
702
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
703

704
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
705
        given `match_function` evaluates to True. The tree is
706 707 708
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
709

710 711 712
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
713 714
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
715 716
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
717
        Yields:
718
            Node: All nodes of the tree for which
719 720
            ``match_function(node)`` returns True
        """
721
        if include_root and match_function(self):
722
            yield self
723 724 725
        child_iterator = reversed(self.children) if reverse else self.children
        for child in child_iterator:
            for node in child.select(match_function, True, reverse):
eckhart's avatar
eckhart committed
726
                yield node
727

Eckhart Arnold's avatar
Eckhart Arnold committed
728

eckhart's avatar
eckhart committed
729
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
eckhart's avatar
eckhart committed
730
                      include_root: bool = False) -> Iterator['Node']:
731
        """
732 733 734 735
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
736

737
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
738
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
739
            ['(X (c "d"))', '(X "F")']
740
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
741
            ['(b "X")', '(X (c "d"))', '(X "F")']
742
            >>> any(tree.select_by_tag('a', False))
743
            False
744
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
745
            ['(a (b "X") (X (c "d")) (e (X "F")))']
746 747
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
748 749

        Args:
750 751 752 753
            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.
754 755 756
        Yields:
            Node: All nodes which have a given tag name.
        """
757
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
758
            tag_names = frozenset({tag_names})
759
        return self.select(lambda node: node.tag_name in tag_names, include_root)
760 761


762 763 764 765
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

766
        This function is mostly just syntactic sugar for
767 768 769 770 771 772 773 774 775 776
        ``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


777
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
778 779 780
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
781 782 783
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
784 785 786
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


di68kap's avatar
di68kap committed
787
class RootNode(Node):
788
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
789

Eckhart Arnold's avatar
Eckhart Arnold committed
790
        all_errors (list):  A list of all errors that have occured so far during
791 792
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
793

794 795
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
796
    """
797

eckhart's avatar
eckhart committed
798
    def __init__(self, node: Optional[Node] = None):
di68kap's avatar
di68kap committed
799
        super().__init__(ZOMBIE_PARSER, '')
eckhart's avatar
eckhart committed
800
        self.all_errors = []  # type: List[Error]
di68kap's avatar
di68kap committed
801
        self.error_flag = 0
eckhart's avatar
eckhart committed
802 803
        if node is not None:
            self.swallow(node)
804
        # customization for XML-Representation
eckhart's avatar
eckhart committed
805 806 807
        self.inline_tags = set()  # type: Set[str]
        self.omit_tags = set()  # type: Set[str]
        self.empty_tags = set()  # type: Set[str]
di68kap's avatar
di68kap committed
808

809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830
    def __deepcopy__(self, memodict={}):
        duplicate = self.__class__(None)
        if self.children:
            duplicate.children = copy.deepcopy(self.children)
            duplicate._result = duplicate.children
        else:
            duplicate.children = NoChildren
            duplicate._result = self._result
        duplicate.errors = copy.deepcopy(self.errors) if self.errors else []
        duplicate._pos = self._pos
        duplicate._len = self._len
        if hasattr(self, '_xml_attr'):
            duplicate._xml_attr = copy.deepcopy(self._xml_attr)
        duplicate.all_errors = copy.deepcopy(self.all_errors)
        duplicate.error_flag = self.error_flag
        duplicate.inline_tags = self.inline_tags
        duplicate.omit_tags = self.omit_tags
        duplicate.empty_tags = self.empty_tags
        duplicate.parser = self.parser
        return duplicate


831
    def swallow(self, node: Node) -> 'RootNode':
832 833
        """
        Put `self` in the place of `node` by copying all its data.
834 835 836 837 838 839 840 841 842
        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
843 844 845 846 847
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
848 849
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
850
        self._content = node._content
851
        return self
di68kap's avatar
di68kap committed
852

eckhart's avatar
eckhart committed
853
    def add_error(self, node: Node, error: Error) -> 'RootNode':
854 855 856
        """
        Adds an Error object to the tree, locating it at a specific node.
        """
di68kap's avatar
di68kap committed
857 858 859 860 861
        self.all_errors.append(error)
        self.error_flag = max(self.error_flag, error.code)
        node.errors.append(error)
        return self