In January 2021 we will introduce a 10 GB quota for project repositories. Higher limits for individual projects will be available on request. Please see https://doku.lrz.de/display/PUBLIC/GitLab for more information.

syntaxtree.py 38.2 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
from collections import OrderedDict
eckhart's avatar
eckhart committed
27
import copy
28

eckhart's avatar
eckhart committed
29
from DHParser.error import Error, ErrorCode, linebreaks, line_col
30
from DHParser.stringview import StringView
di68kap's avatar
di68kap committed
31 32
from DHParser.toolkit import re
from typing import Callable, cast, Iterator, List, AbstractSet, Set, Union, Tuple, Optional
di68kap's avatar
di68kap committed
33

34

di68kap's avatar
di68kap committed
35
__all__ = ('WHITESPACE_PTYPE',
36
           'TOKEN_PTYPE',
37
           'ZOMBIE',
eckhart's avatar
eckhart committed
38
           'ZOMBIE_NODE',
eckhart's avatar
eckhart committed
39 40 41
           'ResultType',
           'StrictResultType',
           'ChildrenType',
42
           'Node',
di68kap's avatar
di68kap committed
43
           'RootNode',
eckhart's avatar
eckhart committed
44
           'ZOMBIE_ROOTNODE',
45
           'parse_sxpr',
46 47 48
           'parse_xml',
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
49 50


51 52 53 54 55 56 57 58
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


WHITESPACE_PTYPE = ':Whitespace'
59
TOKEN_PTYPE = ':Token'
60

61 62
ZOMBIE = "__ZOMBIE__"

63 64 65 66 67 68 69
#######################################################################
#
# syntaxtree nodes
#
#######################################################################


70
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
71
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
72
StrictResultType = Union[ChildrenType, StringView, str]
73
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
74 75


Eckhart Arnold's avatar
Eckhart Arnold committed
76
def flatten_sxpr(sxpr: str) -> str:
77 78
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
79 80 81
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
82
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
83 84
    '(a (b c))'
    """
85

86
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
87 88


89
def flatten_xml(xml: str) -> str:
90 91
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
92
    i.e. only whitespace within leaf-nodes is preserved.
93 94
    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`.
95
    """
96

97 98 99 100
    # 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']
101
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
102 103


eckhart's avatar
eckhart committed
104
RX_AMP = re.compile(r'&(?!\w+;)')
105 106


Eckhart Arnold's avatar
Eckhart Arnold committed
107
class Node:  # (collections.abc.Sized): Base class omitted for cython-compatibility
108 109 110
    """
    Represents a node in the concrete or abstract syntax tree.

111 112
    TODO: Add some documentation and doc-tests here...

113
    Attributes:
114 115
        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
116

117 118 119
        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.
120

121 122
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
123

eckhart's avatar
eckhart committed
124 125 126 127
        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
128
        parser (Parser):  The parser which generated this node.
129 130
            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
131 132
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
133
            example by calling ``isinstance(node.parer, ...)``.
134

135 136 137
        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
138
            the length before AST-transformation and will never change
139
            through AST-transformation. READ ONLY!
140

141 142
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
143
            The value of ``pos`` is -1 meaning invalid by default.
144
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
145
            nodes relative to this value.
146 147

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
148
            pos value of the root node should be set to 0 right
149 150
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
151
            Other than that, this value should be considered READ ONLY.
152 153
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
154

155 156
        errors (list):  A list of all errors that occured on this node.

157 158
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
159 160
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
161
    """
162

163
    __slots__ = '_result', 'children', '_len', '_pos', 'tag_name', 'errors', '_xml_attr', '_content'
164

165
    def __init__(self, tag_name: Optional[str], result: ResultType, leafhint: bool = False) -> None:
166 167
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
168 169
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
170
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
171
        self._pos = -1                  # type: int
172
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
173
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
174
        if leafhint:
eckhart's avatar
eckhart committed
175
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
eckhart's avatar
eckhart committed
176
            self._content = None        # type: Optional[str]
177
            self.children = NoChildren  # type: ChildrenType
178
            self._len = -1              # type: int  # lazy evaluation
179 180
        else:
            self.result = result
181
        assert tag_name is None or isinstance(tag_name, str)   # TODO: Delete this line
182
        self.tag_name = tag_name if tag_name else ZOMBIE
183 184 185 186
        # if parser is None:
        #     self._tag_name = ZOMBIE
        # else:
        #     self._tag_name = parser.name or parser.ptype
187

eckhart's avatar
eckhart committed
188 189
    def __deepcopy__(self, memo):
        if self.children:
190
            duplicate = self.__class__(self.tag_name, copy.deepcopy(self.children), False)
eckhart's avatar
eckhart committed
191
        else:
192
            duplicate = self.__class__(self.tag_name, self.result, True)
eckhart's avatar
eckhart committed
193 194 195 196 197 198
        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
199

200
    def __str__(self):
201
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
202
        if self.errors:
203
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
204
                   (s, '; '.join(e.message for e in self.errors))
205
        return s
206

Eckhart Arnold's avatar
Eckhart Arnold committed
207

208
    def __repr__(self):
209
        # mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
210 211
        # name, ptype = (self._tag_name.split(':') + [''])[:2]
        # parg = "MockParser({name}, {ptype})".format(name=name, ptype=ptype)
212
        rarg = str(self) if not self.children else \
eckhart's avatar
eckhart committed
213
            "(" + ", ".join(repr(child) for child in self.children) + ")"
214
        return "Node(%s, %s)" % (self.tag_name, rarg)
215

Eckhart Arnold's avatar
Eckhart Arnold committed
216

217
    def __len__(self):
218
        if self._len < 0:
di68kap's avatar
di68kap committed
219
            self._len = sum(len(child) for child in self.children) \
220
                if self.children else len(self._result)
221 222 223 224 225 226 227 228
        return self._len


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


229
    def __eq__(self, other):
230 231 232
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
233 234 235 236

        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!
237
        """
238
        return self.tag_name == other.tag_name and self.result == other.result
239

Eckhart Arnold's avatar
Eckhart Arnold committed
240

241
    def __hash__(self):
242
        return hash(self.tag_name)
243

Eckhart Arnold's avatar
Eckhart Arnold committed
244

245 246 247
    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
248
        an integer or the first child node with the given tag name. Examples::
249

250
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
251 252
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
253 254
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
255 256 257 258

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
259
        Returns:
260 261
            Node: All nodes which have a given tag name.
        """
262 263 264
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
265
            else:
266 267 268 269 270
                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!')
271 272 273 274


    def __contains__(self, tag_name: str) -> bool:
        """
275
        Returns true if a child with the given tag name exists.
276
        Args:
277 278
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
279 280 281 282
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
283 284 285 286 287
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
288
            return False
289
        raise ValueError('Leave node cannot contain other nodes')
290 291


292 293 294 295 296 297 298 299 300 301 302 303 304 305
    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


306
    def is_anonymous(self):
307
        return self.tag_name[0] == ':'
308

Eckhart Arnold's avatar
Eckhart Arnold committed
309

310
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
311
    def result(self) -> StrictResultType:
312 313 314 315 316
        """
        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.
        """
317 318
        return self._result

319

320
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
321
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
322
        # # made obsolete by static type checking with mypy
323 324 325
        # 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
326 327
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
328 329
        self._len = -1        # lazy evaluation
        self._content = None
330 331 332 333 334 335 336 337 338
        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
339
                self._result = result  # cast(StrictResultType, result)
340

341 342

    @property
eckhart's avatar
eckhart committed
343
    def content(self) -> str:
344
        """
di68kap's avatar
di68kap committed
345 346 347
        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.
348
        """
349 350 351 352
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
353
                # self._content = self._result
354 355 356
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
357 358


di68kap's avatar
di68kap committed
359 360 361 362 363
    @content.setter
    def content(self, content: str):
        self.result = content


364 365 366 367 368 369 370 371 372
    @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))


373
    @property
374
    def pos(self) -> int:
eckhart's avatar
eckhart committed
375 376
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
377
            raise AssertionError("Position value not initialized!")
378 379
        return self._pos

eckhart's avatar
eckhart committed
380

eckhart's avatar
eckhart committed
381
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
382 383 384 385 386 387 388 389 390
        """
        (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
391 392
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
393
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
394
        offset = self.pos
395
        for child in self.children:
eckhart's avatar
eckhart committed
396
            child.init_pos(offset)
eckhart's avatar
eckhart committed
397 398 399
            offset = child.pos + len(child)
        return self

400

401
    @property
402
    def attr(self):
403
        """
404
        Returns a dictionary of XML-attr attached to the node.
405
        """
406 407 408 409
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
410

411
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
412
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
413 414 415 416 417 418 419 420
        """
        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.

421
        Args:
422
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
423
            open_fn:   (Node->str) A function that returns an opening
424
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
425
            close_fn:  (Node->str) A function that returns a closeF
426
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
427
            data_fn:   (str->str) A function that filters the data string
428 429 430 431 432 433
                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
434 435
        head = open_fn(self)
        tail = close_fn(self)
436 437

        if not self.result:
438
            return head.rstrip() + tail.lstrip()
439

eckhart's avatar
eckhart committed
440
        tail = tail.lstrip(None if density & 2 else '')
441

442
        inline = inline or inline_fn(self)
443 444 445 446 447
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
448 449
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
450

451 452
        if self.children:
            content = []
453
            for child in self.children:
454
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
455
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
456
                if subtree:
eckhart's avatar
eckhart committed
457 458
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
459
            return head + usetab + (sep + usetab).join(content) + tail
460

eckhart's avatar
eckhart committed
461
        res = self.content
eckhart's avatar
eckhart committed
462
        if not inline and not head:
eckhart's avatar
eckhart committed
463
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
464
            res = res.strip()
465 466
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
467
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
468
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
469
        else:
470
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
471

Eckhart Arnold's avatar
Eckhart Arnold committed
472

473 474 475
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
476
                compact: bool = False) -> str:
477 478 479
        """
        Returns content as S-expression, i.e. in lisp-like form.

480
        Args:
481 482 483
            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.
484 485 486
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
487
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
488
                tree structure.
489 490
        """

eckhart's avatar
eckhart committed
491
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
492
        lbreaks = linebreaks(src) if src else []  # type: List[int]
493

494
        def opening(node) -> str:
eckhart's avatar
eckhart committed
495
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
496
            txt = [left_bracket, node.tag_name]
497
            # s += " '(pos %i)" % node.add_pos
498
            if hasattr(node, '_xml_attr'):
499
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
500
            if src:
501 502
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
503
            if showerrors and node.errors:
504
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
505
            return "".join(txt) + '\n'
506

eckhart's avatar
eckhart committed
507 508 509
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
510

eckhart's avatar
eckhart committed
511 512 513 514 515
        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'\"')
516

517
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
518

eckhart's avatar
eckhart committed
519

520 521 522
    def as_xml(self, src: str = None,
               showerrors: bool = True,
               indentation: int = 2,
eckhart's avatar
eckhart committed
523 524 525
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
526 527 528
        """
        Returns content as XML-tree.

529
        Args:
530 531 532
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
533 534
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
535 536 537
            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
538
                neither the opening tag nor its attr nor the closing tag. This
539 540 541
                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.
542 543
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
544 545
        """

546
        def opening(node) -> str:
547 548 549
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
550
            txt = ['<', node.tag_name]
551
            has_reserved_attrs = hasattr(node, '_xml_attr') \
eckhart's avatar
eckhart committed
552
                and any(r in node.attr for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
553
            if hasattr(node, '_xml_attr'):
554
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
555
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
556
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
557
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
558 559
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
560 561 562
            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
563
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
564 565 566
            else:
                ending = ">\n"
            return "".join(txt + [ending])
567 568

        def closing(node):
569
            """Returns the closing string for the representation of `node`."""
570
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
571
                return ''
572
            return ('\n</') + node.tag_name + '>'
573

574 575 576 577 578 579 580
        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


581 582 583 584 585
        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
586 587 588
            return node.tag_name in inline_tags \
                or (hasattr(node, '_xml_attr')
                    and node.attr.get('xml:space', 'default') == 'preserve')
589

590
        line_breaks = linebreaks(src) if src else []
591
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
592
                               density=1, inline_fn=inlining)
593

Eckhart Arnold's avatar
Eckhart Arnold committed
594

eckhart's avatar
eckhart committed
595
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
596
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
597
        """
598
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
599

600
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
601
        given `match_function` evaluates to True. The tree is
602 603 604
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
605

606 607 608
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
609 610
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
611 612
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
613
        Yields:
614
            Node: All nodes of the tree for which
615 616
            ``match_function(node)`` returns True
        """
617
        if include_root and match_function(self):
618
            yield self
619 620 621
        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
622
                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]],
eckhart's avatar
eckhart committed
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
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

662
        This function is mostly just syntactic sugar for
663 664 665 666 667 668 669 670 671 672
        ``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


680
ZOMBIE_NODE = Node(ZOMBIE, '')
di68kap's avatar
di68kap committed
681 682


di68kap's avatar
di68kap committed
683
class RootNode(Node):
684
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
685

Eckhart Arnold's avatar
Eckhart Arnold committed
686
        all_errors (list):  A list of all errors that have occured so far during
687 688
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
689

690 691
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
692
    """
693

eckhart's avatar
eckhart committed
694
    def __init__(self, node: Optional[Node] = None):
695
        super().__init__(ZOMBIE, '')
eckhart's avatar
eckhart committed
696
        self.all_errors = []  # type: List[Error]
di68kap's avatar
di68kap committed
697
        self.error_flag = 0
eckhart's avatar
eckhart committed
698 699
        if node is not None:
            self.swallow(node)
700
        # customization for XML-Representation
eckhart's avatar
eckhart committed
701 702 703
        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
704

705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722
    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
723
        duplicate.tag_name = self.tag_name
724 725 726
        return duplicate


727
    def swallow(self, node: Node) -> 'RootNode':
728 729
        """
        Put `self` in the place of `node` by copying all its data.
730 731 732 733 734 735 736 737 738
        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
739 740 741 742
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
743
        self.tag_name = node.tag_name
744 745
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
746
        self._content = node._content
747
        return self
di68kap's avatar
di68kap committed
748

eckhart's avatar
eckhart committed
749
    def add_error(self, node: Node, error: Error) -> 'RootNode':
750 751 752
        """
        Adds an Error object to the tree, locating it at a specific node.
        """
di68kap's avatar
di68kap committed
753 754 755 756 757
        self.all_errors.append(error)
        self.error_flag = max(self.error_flag, error.code)
        node.errors.append(error)
        return self

eckhart's avatar
eckhart committed
758
    def new_error(self,
759
                  node: Node,
di68kap's avatar
di68kap committed
760
                  message: str,
eckhart's avatar
eckhart committed
761
                  code: ErrorCode = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
762
        """
eckhart's avatar
eckhart committed
763
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
764 765 766 767 768
        Parameters:
            pos(int):     The position of the error in the source text
            message(str): A string with the error message.abs
            code(int):    An error code to identify the kind of error
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
769
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
770 771
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
772

eckhart's avatar
eckhart committed
773
    def collect_errors(self) -> List[Error]:
774 775
        """
        Returns the list of errors, ordered bv their position.
di68kap's avatar
di68kap committed
776
        """
777
        self.all_errors.sort(key=lambda e: e.pos)
Eckhart Arnold's avatar
Eckhart Arnold committed
778
        return self.all_errors
di68kap's avatar
di68kap committed
779

780
    def customized_XML(self):
781 782
        """
        Returns a customized XML representation of the tree.
783
        See the docstring of `Node.as_xml()` for an explanation of the
784 785
        customizations.
        """
eckhart's avatar
eckhart committed
786
        return self.as_xml(inline_tags=self.inline_tags,
787 788 789
                           omit_tags=self.omit_tags,
                           empty_tags=self.empty_tags)

di68kap's avatar
di68kap committed
790

eckhart's avatar
eckhart committed
791 792
ZOMBIE_ROOTNODE = RootNode()

di68kap's avatar
di68kap committed
793 794 795 796 797
#######################################################################
#
# S-expression- and XML-parsers
#
#######################################################################
eckhart's avatar
eckhart committed
798 799


eckhart's avatar
eckhart committed
800
def parse_sxpr(sxpr: Union[str, StringView]) -> Node:
801
    """
802 803 804 805 806
    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.
807 808

    Example:
809
    >>> parse_sxpr("(a (b c))").as_sxpr()
810
    '(a\\n  (b\\n    "c"\\n  )\\n)'
811
    """
812

eckhart's avatar
eckhart committed
813
    sxpr = StringView(sxpr).strip() if isinstance(sxpr, str) else sxpr.strip()
814
    # mock_parsers = dict()  # type: Dict[StringView, MockParser]