syntaxtree.py 41.9 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
__all__ = ('ParserBase',
           'WHITESPACE_PTYPE',
           'TOKEN_PTYPE',
           'MockParser',
           'ZombieParser',
42
           'ZOMBIE',
43
           'ZOMBIE_PARSER',
eckhart's avatar
eckhart committed
44
           'ZOMBIE_NODE',
eckhart's avatar
eckhart committed
45 46 47
           'ResultType',
           'StrictResultType',
           'ChildrenType',
48
           'Node',
di68kap's avatar
di68kap committed
49
           'RootNode',
eckhart's avatar
eckhart committed
50
           'ZOMBIE_ROOTNODE',
51
           'parse_sxpr',
52 53 54
           'parse_xml',
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
55 56


57 58 59 60 61 62 63 64 65 66 67 68 69
#######################################################################
#
# 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.
    """
70

71
    __slots__ = 'name', 'ptype'
Eckhart Arnold's avatar
Eckhart Arnold committed
72

73 74 75
    def __init__(self,):  # , pbases=frozenset()):
        self.name = ''  # type: str
        self.ptype = ':' + self.__class__.__name__  # type: str
76 77

    def __repr__(self):
eckhart's avatar
eckhart committed
78
        return self.name + self.ptype
79 80 81 82

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

83 84 85
    def __call__(self, text: StringView) -> Tuple[Optional['Node'], StringView]:
        return None, text

86 87
    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
88
        """Returns the parser's name if it has a name and repr()"""
Eckhart Arnold's avatar
Eckhart Arnold committed
89
        return self.name if self.name else self.__repr__()
90

91
    def reset(self):
92
        """Resets any parser variables. (Should be overridden.)"""
93 94
        pass

Eckhart Arnold's avatar
Eckhart Arnold committed
95
    @property
eckhart's avatar
eckhart committed
96
    def grammar(self) -> 'Grammar':
97 98
        """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
99
        raise NotImplementedError
100

eckhart's avatar
eckhart committed
101 102 103 104
    def apply(self, func: Callable):
        """Applies the function `func` recursively to the parser and all
        descendant parsers, if any exist."""
        pass
105

106 107

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


class MockParser(ParserBase):
    """
    MockParser objects can be used to reconstruct syntax trees from a
    serialized form like S-expressions or XML. Mock objects can mimic
eckhart's avatar
eckhart committed
115
    different parser types by assigning them a `ptype` on initialization.
116 117 118 119 120

    Mock objects should not be used for anything other than
    syntax tree (re-)construction. In all other cases where a parser
    object substitute is needed, chose the singleton ZOMBIE_PARSER.
    """
121

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

124 125
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
126 127
        super().__init__()
        self.name = name
di68kap's avatar
di68kap committed
128 129
        if ptype:
            self.ptype = ptype  # or ':' + self.__class__.__name__
130 131


132 133 134
ZOMBIE = "__ZOMBIE__"


135 136 137 138 139 140 141 142 143 144
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.
    """
145

146
    alive = False
Eckhart Arnold's avatar
Eckhart Arnold committed
147
    __slots__ = ()
148 149

    def __init__(self):
150
        super(ZombieParser, self).__init__()
151 152
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
153
        self.name = ZOMBIE
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176
        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:
184 185
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
186 187 188
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
189
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
190 191
    '(a (b c))'
    """
192

193
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
194 195


196
def flatten_xml(xml: str) -> str:
197 198
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
199
    i.e. only whitespace within leaf-nodes is preserved.
200 201
    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`.
202
    """
203

204 205 206 207
    # 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']
208
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
209 210


eckhart's avatar
eckhart committed
211
RX_AMP = re.compile(r'&(?!\w+;)')
212 213


Eckhart Arnold's avatar
Eckhart Arnold committed
214
class Node:  # (collections.abc.Sized): Base class omitted for cython-compatibility
215 216 217
    """
    Represents a node in the concrete or abstract syntax tree.

218 219
    TODO: Add some documentation and doc-tests here...

220
    Attributes:
221 222
        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
223

224 225 226
        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.
227

228 229
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
230

eckhart's avatar
eckhart committed
231 232 233 234
        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
235
        parser (Parser):  The parser which generated this node.
236 237
            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
238 239
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
240
            example by calling ``isinstance(node.parer, ...)``.
241

242 243 244
        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
245
            the length before AST-transformation and will never change
246
            through AST-transformation. READ ONLY!
247

248 249
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
250
            The value of ``pos`` is -1 meaning invalid by default.
251
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
252
            nodes relative to this value.
253 254

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
255
            pos value of the root node should be set to 0 right
256 257
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
258
            Other than that, this value should be considered READ ONLY.
259 260
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
261

262 263
        errors (list):  A list of all errors that occured on this node.

264 265
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
266 267
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
268
    """
269

270
    __slots__ = '_result', 'children', '_len', '_pos', '_tag_name', 'errors', '_xml_attr', '_content'
271

272 273 274
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
275 276
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
277
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
278
        self._pos = -1                  # type: int
279
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
280
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
281
        if leafhint:
eckhart's avatar
eckhart committed
282
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
eckhart's avatar
eckhart committed
283
            self._content = None        # type: Optional[str]
284
            self.children = NoChildren  # type: ChildrenType
285
            self._len = -1              # type: int  # lazy evaluation
286 287
        else:
            self.result = result
288 289 290 291
        if parser is None:
            self._tag_name = ZOMBIE
        else:
            self._tag_name = parser.name or parser.ptype
292

eckhart's avatar
eckhart committed
293 294 295 296 297 298 299 300 301 302 303
    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
304

305
    def __str__(self):
306
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
307
        if self.errors:
308
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
309
                   (s, '; '.join(e.message for e in self.errors))
310
        return s
311

Eckhart Arnold's avatar
Eckhart Arnold committed
312

313
    def __repr__(self):
314 315 316
        # mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
        name, ptype = (self._tag_name.split(':') + [''])[:2]
        parg = "MockParser({name}, {ptype})".format(name=name, ptype=ptype)
317
        rarg = str(self) if not self.children else \
eckhart's avatar
eckhart committed
318
            "(" + ", ".join(repr(child) for child in self.children) + ")"
319
        return "Node(%s, %s)" % (parg, rarg)
320

Eckhart Arnold's avatar
Eckhart Arnold committed
321

322
    def __len__(self):
323
        if self._len < 0:
di68kap's avatar
di68kap committed
324
            self._len = sum(len(child) for child in self.children) \
325
                if self.children else len(self._result)
326 327 328 329 330 331 332 333
        return self._len


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


334
    def __eq__(self, other):
335 336 337
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
338 339 340 341

        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!
342
        """
343
        return self.tag_name == other.tag_name and self.result == other.result
344

Eckhart Arnold's avatar
Eckhart Arnold committed
345

346
    def __hash__(self):
347
        return hash(self.tag_name)
348

Eckhart Arnold's avatar
Eckhart Arnold committed
349

350 351 352
    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
353
        an integer or the first child node with the given tag name. Examples::
354

355
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
356 357
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
358 359
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
360 361 362 363

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
364
        Returns:
365 366
            Node: All nodes which have a given tag name.
        """
367 368 369
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
370
            else:
371 372 373 374 375
                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!')
376 377 378 379


    def __contains__(self, tag_name: str) -> bool:
        """
380
        Returns true if a child with the given tag name exists.
381
        Args:
382 383
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
384 385 386 387
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
388 389 390 391 392
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
393
            return False
394
        raise ValueError('Leave node cannot contain other nodes')
395 396


397 398 399 400 401 402 403 404 405 406 407 408 409 410
    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


411
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
412
    def tag_name(self) -> str:
413 414 415 416 417 418
        """
        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`.
        """
419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434
        return self._tag_name

    @tag_name.setter
    def tag_name(self, tag_name: str):
        assert tag_name
        self._tag_name = tag_name


    def is_anonymous(self):
        return self._tag_name[0] == ':'


    @property
    def parser(self) -> MockParser:
        name, ptype = (self.tag_name.split(':') + [''])[:2]
        return MockParser(name, ':' + ptype)
435

Eckhart Arnold's avatar
Eckhart Arnold committed
436

437
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
438
    def result(self) -> StrictResultType:
439 440 441 442 443
        """
        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.
        """
444 445
        return self._result

446

447
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
448
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
449
        # # made obsolete by static type checking with mypy
450 451 452
        # 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
453 454
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
455 456
        self._len = -1        # lazy evaluation
        self._content = None
457 458 459 460 461 462 463 464 465
        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
466
                self._result = result  # cast(StrictResultType, result)
467

468 469

    @property
eckhart's avatar
eckhart committed
470
    def content(self) -> str:
471
        """
di68kap's avatar
di68kap committed
472 473 474
        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.
475
        """
476 477 478 479
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
480
                # self._content = self._result
481 482 483
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
484 485


di68kap's avatar
di68kap committed
486 487 488 489 490
    @content.setter
    def content(self, content: str):
        self.result = content


491 492 493 494 495 496 497 498 499
    @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))


500
    @property
501
    def pos(self) -> int:
eckhart's avatar
eckhart committed
502 503
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
504
            raise AssertionError("Position value not initialized!")
505 506
        return self._pos

eckhart's avatar
eckhart committed
507

eckhart's avatar
eckhart committed
508
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
509 510 511 512 513 514 515 516 517
        """
        (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
518 519
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
520
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
521
        offset = self.pos
522
        for child in self.children:
eckhart's avatar
eckhart committed
523
            child.init_pos(offset)
eckhart's avatar
eckhart committed
524 525 526
            offset = child.pos + len(child)
        return self

527

528
    @property
529
    def attr(self):
530
        """
531
        Returns a dictionary of XML-attr attached to the node.
532
        """
533 534 535 536
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
537

538
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
539
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
540 541 542 543 544 545 546 547
        """
        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.

548
        Args:
549
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
550
            open_fn:   (Node->str) A function that returns an opening
551
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
552
            close_fn:  (Node->str) A function that returns a closeF
553
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
554
            data_fn:   (str->str) A function that filters the data string
555 556 557 558 559 560
                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
561 562
        head = open_fn(self)
        tail = close_fn(self)
563 564

        if not self.result:
565
            return head.rstrip() + tail.lstrip()
566

eckhart's avatar
eckhart committed
567
        tail = tail.lstrip(None if density & 2 else '')
568

569
        inline = inline or inline_fn(self)
570 571 572 573 574
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
575 576
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
577

578 579
        if self.children:
            content = []
580
            for child in self.children:
581
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
582
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
583
                if subtree:
eckhart's avatar
eckhart committed
584 585
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
586
            return head + usetab + (sep + usetab).join(content) + tail
587

eckhart's avatar
eckhart committed
588
        res = self.content
eckhart's avatar
eckhart committed
589
        if not inline and not head:
eckhart's avatar
eckhart committed
590
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
591
            res = res.strip()
592 593
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
594
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
595
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
596
        else:
597
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
598

Eckhart Arnold's avatar
Eckhart Arnold committed
599

600 601 602
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
603
                compact: bool = False) -> str:
604 605 606
        """
        Returns content as S-expression, i.e. in lisp-like form.

607
        Args:
608 609 610
            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.
611 612 613
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
614
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
615
                tree structure.
616 617
        """

eckhart's avatar
eckhart committed
618
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
619
        lbreaks = linebreaks(src) if src else []  # type: List[int]
620

621
        def opening(node) -> str:
eckhart's avatar
eckhart committed
622
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
623
            txt = [left_bracket, node.tag_name]
624
            # s += " '(pos %i)" % node.add_pos
625
            if hasattr(node, '_xml_attr'):
626
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
627
            if src:
628 629
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
630
            if showerrors and node.errors:
631
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
632
            return "".join(txt) + '\n'
633

eckhart's avatar
eckhart committed
634 635 636
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
637

eckhart's avatar
eckhart committed
638 639 640 641 642
        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'\"')
643

644
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
645

eckhart's avatar
eckhart committed
646

647 648 649
    def as_xml(self, src: str = None,
               showerrors: bool = True,
               indentation: int = 2,
eckhart's avatar
eckhart committed
650 651 652
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
653 654 655
        """
        Returns content as XML-tree.

656
        Args:
657 658 659
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
660 661
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
662 663 664
            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
665
                neither the opening tag nor its attr nor the closing tag. This
666 667 668
                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.
669 670
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
671 672
        """

673
        def opening(node) -> str:
674 675 676
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
677
            txt = ['<', node.tag_name]
678
            has_reserved_attrs = hasattr(node, '_xml_attr') \
eckhart's avatar
eckhart committed
679
                and any(r in node.attr for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
680
            if hasattr(node, '_xml_attr'):
681
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
682
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
683
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
684
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
685 686
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
687 688 689
            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
690
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
691 692 693
            else:
                ending = ">\n"
            return "".join(txt + [ending])
694 695

        def closing(node):
696
            """Returns the closing string for the representation of `node`."""
697
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
698
                return ''
699
            return ('\n</') + node.tag_name + '>'
700

701 702 703 704 705 706 707
        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


708 709 710 711 712
        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
713 714 715
            return node.tag_name in inline_tags \
                or (hasattr(node, '_xml_attr')
                    and node.attr.get('xml:space', 'default') == 'preserve')
716

717
        line_breaks = linebreaks(src) if src else []
718
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
719
                               density=1, inline_fn=inlining)
720

Eckhart Arnold's avatar
Eckhart Arnold committed
721

eckhart's avatar
eckhart committed
722
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
723
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
724
        """
725
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
726

727
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
728
        given `match_function` evaluates to True. The tree is
729 730 731
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
732

733 734 735
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
736 737
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
738 739
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
740
        Yields:
741
            Node: All nodes of the tree for which
742 743
            ``match_function(node)`` returns True
        """
744
        if include_root and match_function(self):
745
            yield self
746 747 748
        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
749
                yield node
750

Eckhart Arnold's avatar
Eckhart Arnold committed
751

eckhart's avatar
eckhart committed
752
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
eckhart's avatar
eckhart committed
753
                      include_root: bool = False) -> Iterator['Node']:
754
        """
755 756 757 758
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
759

760
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
761
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
762
            ['(X (c "d"))', '(X "F")']
763
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
764
            ['(b "X")', '(X (c "d"))', '(X "F")']
765
            >>> any(tree.select_by_tag('a', False))
766
            False
767
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
768
            ['(a (b "X") (X (c "d")) (e (X "F")))']
769 770
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
771 772

        Args:
773 774 775 776
            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.
777 778 779
        Yields:
            Node: All nodes which have a given tag name.
        """
780
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
781
            tag_names = frozenset({tag_names})
782
        return self.select(lambda node: node.tag_name in tag_names, include_root)
783 784


785 786 787 788
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

789
        This function is mostly just syntactic sugar for
790 791 792 793 794 795 796 797 798 799
        ``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


800
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
801 802 803
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
804 805 806
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
807 808 809
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


di68kap's avatar
di68kap committed
810
class RootNode(Node):
811
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
812

Eckhart Arnold's avatar
Eckhart Arnold committed
813
        all_errors (list):  A list of all errors that have occured so far during
814 815
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
816

817 818
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
819
    """
820

eckhart's avatar
eckhart committed
821
    def __init__(self, node: Optional[Node] = None):
di68kap's avatar
di68kap committed
822
        super().__init__(ZOMBIE_PARSER, '')
eckhart's avatar
eckhart committed
823
        self.all_errors = []  # type: List[Error]
di68kap's avatar
di68kap committed
824
        self.error_flag = 0
eckhart's avatar
eckhart committed
825 826
        if node is not None:
            self.swallow(node)
827
        # customization for XML-Representation
eckhart's avatar
eckhart committed
828 829 830
        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
831

832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849
    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
850
        duplicate.tag_name = self.tag_name
851 852 853
        return duplicate


854
    def swallow(self, node: Node) -> 'RootNode':
855 856
        """
        Put `self` in the place of `node` by copying all its data.
857 858 859 860 861