syntaxtree.py 56 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
import json
29
from typing import Callable, cast, Iterator, List, AbstractSet, Set, Union, Tuple, Optional, Dict
30

31 32
from DHParser.configuration import SERIALIZATIONS, XML_SERIALIZATION, SXPRESSION_SERIALIZATION, \
    COMPACT_SERIALIZATION, JSON_SERIALIZATION
eckhart's avatar
eckhart committed
33
from DHParser.error import Error, ErrorCode, linebreaks, line_col
34
from DHParser.stringview import StringView
35
from DHParser.toolkit import JSONSerializable, get_config_value, gen_id, re
di68kap's avatar
di68kap committed
36

37

di68kap's avatar
di68kap committed
38
__all__ = ('WHITESPACE_PTYPE',
39
           'TOKEN_PTYPE',
40 41
           'ZOMBIE_TAG',
           'PLACEHOLDER',
eckhart's avatar
eckhart committed
42 43 44
           'ResultType',
           'StrictResultType',
           'ChildrenType',
45
           'Node',
46
           'FrozenNode',
47
           'tree_sanity_check',
di68kap's avatar
di68kap committed
48
           'RootNode',
49 50
           'DHParser_JSONEncoder',
           'dhparser_obj_hook',
51
           'parse_sxpr',
52
           'parse_xml',
53
           'parse_json_syntaxtree',
eckhart's avatar
eckhart committed
54
           'parse_tree',
55 56
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
57 58


59 60 61 62 63 64 65 66
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


WHITESPACE_PTYPE = ':Whitespace'
67
TOKEN_PTYPE = ':Token'
68

69
ZOMBIE_TAG = "__ZOMBIE__"
70

71 72 73 74 75 76 77
#######################################################################
#
# syntaxtree nodes
#
#######################################################################


eckhart's avatar
eckhart committed
78 79
RX_IS_SXPR = re.compile(r'\s*\(')
RX_IS_XML = re.compile(r'\s*<')
80
RX_ATTR_NAME = re.compile(r'[\w.:-]')
eckhart's avatar
eckhart committed
81 82


83
def flatten_sxpr(sxpr: str, threshold: int = -1) -> str:
84 85
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
86 87
    whitespace.

88 89 90 91 92 93 94
    The ``threshold`` value is a maximum number of
    characters allowed in the flattened expression. If this number
    is exceeded the the unflattened S-expression is returned. A
    negative number means that the S-expression will always be
    flattened. Zero or (any postive integer <= 3) essentially means
    that the expression will not be flattened.

Eckhart Arnold's avatar
Eckhart Arnold committed
95
    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
96
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
97 98
    '(a (b c))'
    """
eckhart's avatar
eckhart committed
99 100 101
    assert RX_IS_SXPR.match(sxpr)
    if threshold == 0:
        return sxpr
102
    flat = re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
eckhart's avatar
eckhart committed
103
    if len(flat) > threshold >= 0:
104 105
        return sxpr.strip()
    return flat
Eckhart Arnold's avatar
Eckhart Arnold committed
106 107


108
def flatten_xml(xml: str) -> str:
109 110
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
111
    i.e. only whitespace within leaf-nodes is preserved.
112 113
    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`.
114
    """
115 116
    # works only with regex
    # return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?<=</\w+>)\s+', '', xml))
eckhart's avatar
eckhart committed
117
    assert RX_IS_XML.match(xml)
eckhart's avatar
eckhart committed
118

119
    def tag_only(m):
eckhart's avatar
eckhart committed
120
        """Return only the tag, drop the whitespace."""
121
        return m.groupdict()['closing_tag']
122
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
123 124


125 126 127 128 129
ChildrenType = Tuple['Node', ...]
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
StrictResultType = Union[ChildrenType, StringView, str]
ResultType = Union[ChildrenType, 'Node', StringView, str, None]

eckhart's avatar
eckhart committed
130
RX_AMP = re.compile(r'&(?!\w+;)')
131 132


133
class Node(JSONSerializable):  # (collections.abc.Sized): Base class omitted for cython-compatibility
134 135 136
    """
    Represents a node in the concrete or abstract syntax tree.

137 138
    TODO: Add some documentation and doc-tests here...

eckhart's avatar
eckhart committed
139
    Attributes and Properties:
140 141
        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
142

143 144 145
        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.
146

147 148
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
149

eckhart's avatar
eckhart committed
150 151 152 153
        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.

154 155 156
        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
157
            the length before AST-transformation and will never change
158
            through AST-transformation. READ ONLY!
159

160 161
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
162
            The value of ``pos`` is -1 meaning invalid by default.
163
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
164
            nodes relative to this value.
165 166

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
167
            pos value of the root node should be set to 0 right
168 169
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
170
            Other than that, this value should be considered READ ONLY.
171 172
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
173

174 175
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
176 177
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
178
    """
179

180
    __slots__ = '_result', 'children', '_pos', 'tag_name', '_xml_attr', '_id'
181

182
    def __init__(self, tag_name: str, result: ResultType, leafhint: bool = False) -> None:
183 184
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
185 186
        that generated the node and the parser's result.
        """
eckhart's avatar
eckhart committed
187
        self._pos = -1                  # type: int
188
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
189
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
190
        if leafhint:
eckhart's avatar
eckhart committed
191
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
192 193 194
            self.children = NoChildren  # type: ChildrenType
        else:
            self.result = result
195
        self.tag_name = tag_name        # type: str
196

eckhart's avatar
eckhart committed
197 198
    def __deepcopy__(self, memo):
        if self.children:
199
            duplicate = self.__class__(self.tag_name, copy.deepcopy(self.children), False)
eckhart's avatar
eckhart committed
200
        else:
201
            duplicate = self.__class__(self.tag_name, self.result, True)
eckhart's avatar
eckhart committed
202
        duplicate._pos = self._pos
203
        if self.has_attr():
di68kap's avatar
di68kap committed
204 205
            duplicate.attr.update(copy.deepcopy(self._xml_attr))
            # duplicate._xml_attr = copy.deepcopy(self._xml_attr)  # this is not cython compatible
eckhart's avatar
eckhart committed
206
        return duplicate
Eckhart Arnold's avatar
Eckhart Arnold committed
207

208
    def __str__(self):
209 210
        if isinstance(self, RootNode):
            root = cast(RootNode, self)
211
            errors = root.errors_sorted
212 213
            if errors:
                e_pos = errors[0].pos
Eckhart Arnold's avatar
Eckhart Arnold committed
214
                content = self.content
eckhart's avatar
eckhart committed
215 216
                return content[:e_pos] + ' <<< Error on "%s" | %s >>> ' % \
                    (content[e_pos - self.pos:], '; '.join(e.message for e in errors))
217
        return self.content
Eckhart Arnold's avatar
Eckhart Arnold committed
218

219
    def __repr__(self):
220
        # mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
221 222
        # name, ptype = (self._tag_name.split(':') + [''])[:2]
        # parg = "MockParser({name}, {ptype})".format(name=name, ptype=ptype)
223
        rarg = str(self) if not self.children else \
224
            "(" + ", ".join(child.__repr__() for child in self.children) + ")"
225
        return "Node(%s, %s)" % (self.tag_name, rarg)
226

227
    def __len__(self):
228 229
        return (sum(len(child) for child in self.children)
                if self.children else len(self._result))
230 231

    def __bool__(self):
232 233 234 235
        """Returns the bool value of a node, which is always True. The reason
        for this is that a boolean test on a variable that can contain a node
        or None will only yield `False` in case of None.
        """
236 237
        return True

238
    def __hash__(self):
239
        return hash(self.tag_name)
240

241 242 243
    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
244
        an integer or the first child node with the given tag name. Examples::
245

246
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
247 248
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
249 250
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
251 252 253 254

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
255
        Returns:
256 257
            Node: All nodes which have a given tag name.
        """
258 259 260
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
261
            else:
262 263 264 265 266
                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!')
267 268 269

    def __contains__(self, tag_name: str) -> bool:
        """
270
        Returns true if a child with the given tag name exists.
271
        Args:
272 273
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
274 275 276 277
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
278 279 280 281 282
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
283
            return False
284
        raise ValueError('Leave node cannot contain other nodes')
285

286 287 288 289 290 291 292 293 294 295
    @property
    def id(self) -> int:
        """Returns the unique id of the Node."""
        try:
            return self._id
        except AttributeError:
            self._id = gen_id()
            return self._id


296
    def equals(self, other: 'Node') -> bool:
297
        """
298 299 300 301 302 303
        Equality of value: Two nodes are considered as having the same value,
        if their tag name is the same, if their results are equal and
        if their attributes and attribute values are the same.

        Returns True, if the tree originating in node `self` is equal by
        value to the tree originating in node `other`.
304 305 306 307 308 309 310 311 312
        """
        if self.tag_name == other.tag_name and self.compare_attr(other):
            if self.children:
                return (len(self.children) == len(other.children)
                        and all(a.equals(b) for a, b in zip(self.children, other.children)))
            else:
                return self.result == other.result
        return False

313 314 315 316 317 318 319 320 321 322 323 324 325
    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

326
    def is_anonymous(self) -> bool:
eckhart's avatar
eckhart committed
327 328 329 330 331 332 333 334
        """Returns True, if the Node is an "anonymous" Node, i.e. a node that
        has not been created by a named parser.

        The tag name of anonymous node is a colon followed by the class name
        of the parser that created the node, i.e. ":Series". It is recommended
        practice to remove (or name) all anonymous nodes during the
        AST-transformation.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
335
        return not self.tag_name or self.tag_name[0] == ':'
336

337
    ## node content
Eckhart Arnold's avatar
Eckhart Arnold committed
338

339
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
340
    def result(self) -> StrictResultType:
341 342 343 344 345
        """
        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.
        """
346 347 348
        return self._result

    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
349
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
350
        # # made obsolete by static type checking with mypy
eckhart's avatar
eckhart committed
351 352 353 354
        # assert ((isinstance(result, tuple) and all(isinstance(child, Node) for child in result))
        #         or isinstance(result, Node)
        #         or isinstance(result, str)
        #         or isinstance(result, StringView)), "%s (%s)" % (str(result), str(type(result)))
Eckhart Arnold's avatar
Eckhart Arnold committed
355 356
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
357
        # self._content = None
358 359 360 361 362 363 364 365 366
        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
367
                self._result = result  # cast(StrictResultType, result)
368

369 370 371 372 373 374 375 376 377 378 379 380
    def _content(self) -> List[str]:
        """
        Returns string content as list of string fragments
        that are gathered from all child nodes in order.
        """
        if self.children:
            fragments = []
            for child in self.children:
                fragments.extend(child._content())
            return fragments
        self._result = str(self._result)
        return [self._result]
381 382

    @property
eckhart's avatar
eckhart committed
383
    def content(self) -> str:
384
        """
385 386 387
        Returns content as string. If the node has child-nodes, the
        string content of the child-nodes is recursively read and then
        concatenated.
388
        """
389 390 391 392 393 394 395 396
        if self.children:
            fragments = []
            for child in self.children:
                fragments.extend(child._content())
            return ''.join(fragments)
        self._result = str(self._result)
        return self._result
        # unoptimized
Eckhart Arnold's avatar
Eckhart Arnold committed
397 398
        # return "".join(child.content for child in self.children) if self.children \
        #     else str(self._result)
di68kap's avatar
di68kap committed
399

400
    ## node position
di68kap's avatar
di68kap committed
401

402
    @property
403
    def pos(self) -> int:
eckhart's avatar
eckhart committed
404 405
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
Eckhart Arnold's avatar
Eckhart Arnold committed
406
            raise AssertionError("Position value not initialized! Use Node.with_pos()")
407 408
        return self._pos

Eckhart Arnold's avatar
Eckhart Arnold committed
409
    def with_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
410
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
411
        Initialize position value. Usually, the parser guard
eckhart's avatar
eckhart committed
412 413
        (`parsers.add_parser_guard()`) takes care of assigning the
        position in the document to newly created nodes. However,
Eckhart Arnold's avatar
Eckhart Arnold committed
414
        when Nodes are created outside the reach of the parser
eckhart's avatar
eckhart committed
415
        guard, their document-position must be assigned manually.
Eckhart Arnold's avatar
Eckhart Arnold committed
416 417
        Position values of the child nodes are assigned recursively, too.
        Returns the node itself for convenience.
eckhart's avatar
eckhart committed
418
        """
419 420 421 422
        # condition self.pos == pos cannot be assumed when tokens or whitespace
        # are dropped early!
        # assert self._pos < 0 or self.pos == pos, ("pos mismatch %i != %i at Node: %s"
        #                                           % (self._pos, pos, repr(self)))
Eckhart Arnold's avatar
Eckhart Arnold committed
423 424 425 426 427 428 429 430 431 432
        if pos != self._pos >= 0:
            raise AssertionError("Position value cannot be reassigned to a different value!")
        if self._pos < 0:
            self._pos = pos
            # recursively adjust pos-values of all children
            offset = self.pos
            for child in self.children:
                if child._pos < 0:
                    child.with_pos(offset)
                offset = child.pos + len(child)
eckhart's avatar
eckhart committed
433 434
        return self

435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451
    ## (XML-)attributes

    def has_attr(self) -> bool:
        """
        Returns `True`, if the node has any attributes, `False` otherwise.

        This function does not create an attribute dictionary, therefore
        it should be preferred to querying node.attr when testing for the
        existence of any attributes.
        """
        try:
            # if self._xml_attr is not None:
            #     return True
            return bool(self._xml_attr)
        except AttributeError:
            pass
        return False
452

453 454 455 456
    @property
    def attr(self):
        """
        Returns a dictionary of XML-attr attached to the node.
457 458 459 460 461 462 463 464 465 466

        Examples:
            >>> node = Node(None, '')
            >>> print('Any attributes present?', node.has_attr())
            Any attributes present? False
            >>> node.attr['id'] = 'identificator'
            >>> node.attr
            OrderedDict([('id', 'identificator')])
            >>> node.attr['id']
            'identificator'
467
            >>> del node.attr['id']
468 469 470 471 472 473 474
            >>> node.attr
            OrderedDict()

        NOTE: Use `node.attr_active()` rather than bool(node.attr) to check the
        presence of any attributes. Attribute dictionaries are created lazily
        and node.attr would create a dictionary, even though it may never be
        needed any more.
475 476 477 478 479 480 481 482 483
        """
        try:
            if self._xml_attr is None:          # cython compatibility
                self._xml_attr = OrderedDict()
        except AttributeError:
            self._xml_attr = OrderedDict()
        return self._xml_attr

    def compare_attr(self, other: 'Node') -> bool:
484
        """
485 486
        Returns True, if `self` and `other` have the same attributes with the
        same attribute values.
487
        """
488 489
        if self.has_attr():
            if other.has_attr():
490 491 492
                return self.attr == other.attr
            return len(self.attr) == 0
            # self has empty dictionary and other has no attributes
493
        elif other.has_attr():
494 495 496
            return len(other.attr) == 0
            # other has empty attribute dictionary and self as no attributes
        return True  # neither self nor other have any attributes
497

498
    ## tree traversal and node selection
Eckhart Arnold's avatar
Eckhart Arnold committed
499

500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
            -> Iterator['Node']:
        """
        Finds nodes in the tree that fulfill a given criterion.

        `select` is a generator that yields all nodes for which the
        given `match_function` evaluates to True. The tree is
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.

        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
        Yields:
            Node: All nodes of the tree for which
            ``match_function(node)`` returns True
        """
        if include_root and match_function(self):
            yield self
        child_iterator = reversed(self.children) if reverse else self.children
        for child in child_iterator:
            if match_function(child):
                yield child
            yield from child.select(match_function, False, reverse)
        # The above variant is slightly faster
        # for child in child_iterator:
        #     yield from child.select(match_function, True, reverse)

    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
                      include_root: bool = False) -> Iterator['Node']:
        """
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::

            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
            ['(X (c "d"))', '(X "F")']
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
            ['(b "X")', '(X (c "d"))', '(X "F")']
            >>> any(tree.select_by_tag('a', False))
            False
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
            ['(a (b "X") (X (c "d")) (e (X "F")))']
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'

        Args:
eckhart's avatar
eckhart committed
554
            tag_names(set): A tag name or set of tag names that is being
555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578
                searched for
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
        Yields:
            Node: All nodes which have a given tag name.
        """
        if isinstance(tag_names, str):
            tag_names = frozenset({tag_names})
        return self.select(lambda node: node.tag_name in tag_names, include_root)

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

        This function is mostly just syntactic sugar for
        ``next(node.select_by_tag(tag_names, False))``. However, rather than
        raising a StopIterationError if no descendant with the given tag-name
        exists, it returns None.
        """
        try:
            return next(self.select_by_tag(tag_names, False))
        except StopIteration:
            return None

579
    ## serialization methods
580

581
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
582
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
583 584 585 586 587 588 589 590
        """
        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.

591
        Args:
592
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
593
            open_fn:   (Node->str) A function that returns an opening
594
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
595
            close_fn:  (Node->str) A function that returns a closeF
596
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
597
            data_fn:   (str->str) A function that filters the data string
598 599 600 601 602 603
                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
604 605
        head = open_fn(self)
        tail = close_fn(self)
606 607

        if not self.result:
608
            return head.rstrip() + tail.lstrip()
609

eckhart's avatar
eckhart committed
610
        tail = tail.lstrip(None if density & 2 else '')
611

612
        inline = inline or inline_fn(self)
613 614 615 616 617
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
618 619
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
620

621 622
        if self.children:
            content = []
623
            for child in self.children:
624
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
625
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
626
                if subtree:
eckhart's avatar
eckhart committed
627 628
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
629
            return head + usetab + (sep + usetab).join(content) + tail
630

eckhart's avatar
eckhart committed
631
        res = self.content
eckhart's avatar
eckhart committed
632
        if not inline and not head:
eckhart's avatar
eckhart committed
633
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
634
            res = res.strip()
635 636
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
637
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
638
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
639
        else:
640
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
641

642
    def as_sxpr(self, src: Optional[str] = None,
643
                indentation: int = 2,
644
                compact: bool = False,
eckhart's avatar
eckhart committed
645
                flatten_threshold: int = 92) -> str:
646
        """
647
        Returns content as S-expression, i.e. in lisp-like form. If this
eckhart's avatar
eckhart committed
648
        method is called on a RootNode-object,
649

650
        Args:
651 652
            src:  The source text or `None`. In case the source text is
                given the position of the element in the text will be
653 654 655
                reported as position, line, column. In case the empty string is
                given rather than None, only the position value will be
                reported in case it has been initialized, i.e. pos >= 0.
656 657
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
658
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
659
                tree structure.
660 661 662
            flatten_threshold:  Return the S-expression in flattened form if
                the flattened expression does not exceed the threshold length.
                A negative number means that it will always be flattened.
663 664
        """

eckhart's avatar
eckhart committed
665
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
666
        lbreaks = linebreaks(src) if src else []  # type: List[int]
667
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[RootNode]
668

669
        def opening(node: Node) -> str:
eckhart's avatar
eckhart committed
670
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
671
            txt = [left_bracket, node.tag_name]
672
            # s += " '(pos %i)" % node.add_pos
673
            # txt.append(str(node.id))  # for debugging
674
            if node.has_attr():
675
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
676
            if src:
677
                line, col = line_col(lbreaks, node.pos)
678 679 680
                txt.append(' `(pos %i %i %i)' % (node.pos, line, col))
            elif src is not None and node._pos >= 0:
                txt.append(' `(pos %i)' % node.pos)
681
            if root and node.id in root.error_nodes:
682
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in root.get_errors(node)))
683
            return "".join(txt) + '\n'
684

685
        def closing(node: Node) -> str:
eckhart's avatar
eckhart committed
686 687
            """Returns the closing string for the representation of `node`."""
            return right_bracket
688

689
        def pretty(strg: str) -> str:
eckhart's avatar
eckhart committed
690 691 692 693
            """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'\"')
694

695
        sxpr = self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
eckhart's avatar
eckhart committed
696
        return sxpr if compact else flatten_sxpr(sxpr, flatten_threshold)
Eckhart Arnold's avatar
Eckhart Arnold committed
697

698 699
    def as_xml(self, src: str = None,
               indentation: int = 2,
eckhart's avatar
eckhart committed
700 701 702
               inline_tags: Set[str] = frozenset(),
               omit_tags: Set[str] = frozenset(),
               empty_tags: Set[str] = frozenset()) -> str:
703 704 705
        """
        Returns content as XML-tree.

706
        Args:
707 708
            src:  The source text or `None`. In case the source text is given,
                the position will also be reported as line and column.
709
            indentation: The number of whitespaces for indentation
710 711 712
            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
713
                neither the opening tag nor its attr nor the closing tag. This
714 715 716
                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.
717 718
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
719
        """
720
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[RootNode]
721

722
        def opening(node: Node) -> str:
723 724 725
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
726
            txt = ['<', node.tag_name]
727
            has_reserved_attrs = node.has_attr() \
eckhart's avatar
eckhart committed
728
                and any(r in node.attr for r in {'err', 'line', 'col'})
729
            if node.has_attr():
730
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
731
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
732
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
733
            if src == '' and not (node.has_attr() and '_pos' in node.attr) and node.pos >= 0:
734
                txt.append(' _pos="%i"' % node.pos)
735
            if root and node.id in root.error_nodes and not has_reserved_attrs:
eckhart's avatar
eckhart committed
736
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
eckhart's avatar
eckhart committed
737
                                                 for err in root.get_errors(node)))
738 739 740
            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
741
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
742 743 744
            else:
                ending = ">\n"
            return "".join(txt + [ending])
745

746
        def closing(node: Node):
747
            """Returns the closing string for the representation of `node`."""
748
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
749
                return ''
eckhart's avatar
eckhart committed
750
            return '\n</' + node.tag_name + '>'
751

752 753 754 755 756 757
        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

758
        def inlining(node: Node):
759 760 761 762
            """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
763
            return node.tag_name in inline_tags \
764
                or (node.has_attr()
eckhart's avatar
eckhart committed
765
                    and node.attr.get('xml:space', 'default') == 'preserve')
766

767
        line_breaks = linebreaks(src) if src else []
768
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
769
                               density=1, inline_fn=inlining)
770

771
    ## JSON reading and writing
Eckhart Arnold's avatar
Eckhart Arnold committed
772

773
    def to_json_obj(self) -> Dict:
eckhart's avatar
eckhart committed
774
        """Serialize a node or tree as json-object"""
775
        json_obj = {'__class__': 'DHParser.Node', 'tag_name': self.tag_name }
776
        if self.children:
777
            json_obj['result'] = [child.to_json_obj() for child in self.children]
778
        else:
779
            json_obj['result'] = str(self._result)
780
        if self.has_attr():
781
            json_obj['attr'] = dict(self._xml_attr)
782
        if self._pos >= 0:
783 784 785
            json_obj['pos'] = self._pos
        json_obj['id'] = self.id
        return json_obj
786

787 788
    @staticmethod
    def from_json_obj(json_obj: Dict) -> 'Node':
789 790 791
        """Convert a json object representing a node (or tree) back into a
        Node object. Raises a ValueError, if `json_obj` does not represent
        a node."""
792
        assert isinstance(json_obj, Dict)
793
        if json_obj.get('__class__', '') not in ('DHParser.Node', 'DHParser.RootNode'):
794 795
            raise ValueError('JSON object: ' + str(json_obj) +
                             ' does not represent a Node object.')
796 797
        tag_name = json_obj['tag_name']
        result = json_obj['result']
798
        if isinstance(result, str):
799 800 801 802 803
            leafhint = True
        else:
            leafhint = False
            result = tuple(Node.from_json_obj(child) for child in result)
        node = Node(tag_name, result, leafhint)
804 805
        node._pos = json_obj.get('pos', -1)
        attr = json_obj.get('attr', {})
806
        if attr:
807
            node.attr.update(attr)
808
        node._id = json_obj['id']
809
        return node
810

811 812 813 814
    def as_json(self, indent: Optional[int] = 2, ensure_ascii=False) -> str:
        return json.dumps(self.to_json_obj(), indent=indent, ensure_ascii=ensure_ascii,
                          separators=(', ', ': ') if indent is not None else (',', ':'))

815
    ## generalized serialization methoed
816

817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844
    def serialize_as(self: 'Node', how: str = 'default') -> str:
        """
        Serializes the tree starting with `node` either as S-expression, XML, JSON,
        or in compact form. Possible values for `how` are 'S-expression',
        'XML', 'JSON', 'compact' accordingly, or 'AST', 'CST', 'default' in which case
        the value of respective configuration variable determines the
        serialization format. (See module `configuration.py`.)
        """
        switch = how.lower()

        if switch == 'ast':
            switch = get_config_value('ast_serialization').lower()
        elif switch == 'cst':
            switch = get_config_value('cst_serialization').lower()
        elif switch == 'default':
            switch = get_config_value('default_serialization').lower()

        if switch == SXPRESSION_SERIALIZATION.lower():
            return self.as_sxpr(flatten_threshold=get_config_value('flatten_sxpr_threshold'))
        elif switch == XML_SERIALIZATION.lower():
            return self.as_xml()
        elif switch == JSON_SERIALIZATION.lower():
            return self.as_json()
        elif switch == COMPACT_SERIALIZATION.lower():
            return self.as_sxpr(compact=True)
        else:
            raise ValueError('Unknown serialization %s. Allowed values are either: %s or : %s'
                             % (how, "'ast', 'cst', 'default'", ", ".join(list(SERIALIZATIONS))))
845 846


847
class FrozenNode(Node):
848 849 850 851
    """
    FrozenNode is an immutable kind of Node, i.e. it must not be changed
    after initialization. The purpose is mainly to allow certain kinds of
    optimization, like not having to instantiate empty nodes (because they
852 853 854
    are always the same and will be dropped while parsing, anyway) or,
    rather, throw errors if the program tries to treat a node that is
    supposed to be a temporary (frozen) node as if it was a regular node.
855

856
    Frozen nodes must only be used temporarily during parsing or
857 858 859
    tree-transformation and should not occur in the product of the
    transformation any more. This can be verified with `tree_sanity_check()`.
    """
860 861 862 863 864

    def __init__(self, tag_name: str, result: ResultType) -> None:
        if isinstance(result, str) or isinstance(result, StringView):
            result = str(result)
        else:
865
            raise TypeError('FrozenNode only accepts string as result. '
866 867 868 869 870 871 872 873 874 875 876
                            '(Only leaf-nodes can be frozen nodes.)')
        super(FrozenNode, self).__init__(tag_name, result, True)

    @property
    def result(self) -> StrictResultType:
        return self._result

    @result.setter
    def result(self, result: ResultType):
        raise TypeError('FrozenNode does not allow re-assignment of results.')

eckhart's avatar
eckhart committed
877 878 879 880
    @property
    def attr(self):
        raise AssertionError("Attributes cannot be accessed on a frozen node")

Eckhart Arnold's avatar
Eckhart Arnold committed
881
    def with_pos(self, pos: int) -> 'Node':
882 883
        pass

884 885 886 887 888 889 890
    def to_json_obj(self) -> Dict:
        raise NotImplementedError("Frozen nodes cannot and should not be serialized!")

    @staticmethod
    def from_json_obj(json_obj: Dict) -> 'JSONSerializable':
        raise NotImplementedError("Frozen nodes cannot and should not be deserialized!")

891

892
PLACEHOLDER = FrozenNode('__PLACEHOLDER__', '')
di68kap's avatar
di68kap committed
893 894


895
def tree_sanity_check(tree: Node) -> bool:
896 897 898 899 900 901 902 903 904
    """
    Sanity check for syntax trees: One and the same node must never appear
    twice in the syntax tree. Frozen Nodes (EMTPY_NODE, PLACEHOLDER)
    should only exist temporarily and must have been dropped or eliminated
    before any kind of tree generation (i.e. parsing) or transformation
    is finished.
    :param tree: the root of the tree to be checked
    :return: True, if the tree is `sane`, False otherwise.
    """
905
    node_set = set()  # type: Set[Node]
906