syntaxtree.py 51.1 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

eckhart's avatar
eckhart committed
31
from DHParser.error import Error, ErrorCode, linebreaks, line_col
32
from DHParser.stringview import StringView
33
from DHParser.toolkit import get_config_value, re
di68kap's avatar
di68kap committed
34

35

di68kap's avatar
di68kap committed
36
__all__ = ('WHITESPACE_PTYPE',
37
           'TOKEN_PTYPE',
38 39
           'ZOMBIE_TAG',
           'PLACEHOLDER',
eckhart's avatar
eckhart committed
40 41 42
           'ResultType',
           'StrictResultType',
           'ChildrenType',
43
           'Node',
44
           'serialize',
45
           'FrozenNode',
46
           'tree_sanity_check',
di68kap's avatar
di68kap committed
47
           'RootNode',
48
           'parse_sxpr',
49
           'parse_xml',
50
           'parse_json_syntaxtree',
eckhart's avatar
eckhart committed
51
           'parse_tree',
52 53
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
54 55


56 57 58 59 60 61 62 63
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


WHITESPACE_PTYPE = ':Whitespace'
64
TOKEN_PTYPE = ':Token'
65

66
ZOMBIE_TAG = "__ZOMBIE__"
67

68 69 70 71 72 73 74
#######################################################################
#
# syntaxtree nodes
#
#######################################################################


eckhart's avatar
eckhart committed
75 76
RX_IS_SXPR = re.compile(r'\s*\(')
RX_IS_XML = re.compile(r'\s*<')
77
RX_ATTR_NAME = re.compile(r'[\w.:-]')
eckhart's avatar
eckhart committed
78 79


80
def flatten_sxpr(sxpr: str, threshold: int = -1) -> str:
81 82
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
83 84
    whitespace.

85 86 87 88 89 90 91
    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
92
    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
93
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
94 95
    '(a (b c))'
    """
eckhart's avatar
eckhart committed
96 97 98
    assert RX_IS_SXPR.match(sxpr)
    if threshold == 0:
        return sxpr
99
    flat = re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
eckhart's avatar
eckhart committed
100
    if len(flat) > threshold >= 0:
101 102
        return sxpr.strip()
    return flat
Eckhart Arnold's avatar
Eckhart Arnold committed
103 104


105
def flatten_xml(xml: str) -> str:
106 107
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
108
    i.e. only whitespace within leaf-nodes is preserved.
109 110
    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`.
111
    """
112 113
    # works only with regex
    # return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?<=</\w+>)\s+', '', xml))
eckhart's avatar
eckhart committed
114
    assert RX_IS_XML.match(xml)
eckhart's avatar
eckhart committed
115

116
    def tag_only(m):
eckhart's avatar
eckhart committed
117
        """Return only the tag, drop the whitespace."""
118
        return m.groupdict()['closing_tag']
119
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
120 121


122 123 124 125 126
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
127
RX_AMP = re.compile(r'&(?!\w+;)')
128 129


Eckhart Arnold's avatar
Eckhart Arnold committed
130
class Node:  # (collections.abc.Sized): Base class omitted for cython-compatibility
131 132 133
    """
    Represents a node in the concrete or abstract syntax tree.

134 135
    TODO: Add some documentation and doc-tests here...

eckhart's avatar
eckhart committed
136
    Attributes and Properties:
137 138
        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
139

140 141 142
        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.
143

144 145
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
146

eckhart's avatar
eckhart committed
147 148 149 150
        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.

151 152 153
        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
154
            the length before AST-transformation and will never change
155
            through AST-transformation. READ ONLY!
156

157 158
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
159
            The value of ``pos`` is -1 meaning invalid by default.
160
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
161
            nodes relative to this value.
162 163

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
164
            pos value of the root node should be set to 0 right
165 166
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
167
            Other than that, this value should be considered READ ONLY.
168 169
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
170

171 172
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
173 174
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
175
    """
176

177
    __slots__ = '_result', 'children', '_pos', 'tag_name', '_xml_attr'
178

179
    def __init__(self, tag_name: str, result: ResultType, leafhint: bool = False) -> None:
180 181
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
182 183
        that generated the node and the parser's result.
        """
eckhart's avatar
eckhart committed
184
        self._pos = -1                  # type: int
185
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
186
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
187
        if leafhint:
eckhart's avatar
eckhart committed
188
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
189 190 191
            self.children = NoChildren  # type: ChildrenType
        else:
            self.result = result
192
        self.tag_name = tag_name        # type: str
193

eckhart's avatar
eckhart committed
194

eckhart's avatar
eckhart committed
195 196
    def __deepcopy__(self, memo):
        if self.children:
197
            duplicate = self.__class__(self.tag_name, copy.deepcopy(self.children), False)
eckhart's avatar
eckhart committed
198
        else:
199
            duplicate = self.__class__(self.tag_name, self.result, True)
eckhart's avatar
eckhart committed
200
        duplicate._pos = self._pos
eckhart's avatar
eckhart committed
201
        if self.attr_active():
di68kap's avatar
di68kap committed
202 203
            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
204
        return duplicate
Eckhart Arnold's avatar
Eckhart Arnold committed
205

eckhart's avatar
eckhart committed
206

207
    def __str__(self):
208 209
        if isinstance(self, RootNode):
            root = cast(RootNode, self)
210
            errors = root.errors_sorted
211 212
            if errors:
                e_pos = errors[0].pos
Eckhart Arnold's avatar
Eckhart Arnold committed
213
                content = self.content
eckhart's avatar
eckhart committed
214 215
                return content[:e_pos] + ' <<< Error on "%s" | %s >>> ' % \
                    (content[e_pos - self.pos:], '; '.join(e.message for e in errors))
216
        return self.content
Eckhart Arnold's avatar
Eckhart Arnold committed
217

eckhart's avatar
eckhart 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

Eckhart Arnold's avatar
Eckhart Arnold committed
227

228
    def __len__(self):
229 230
        return (sum(len(child) for child in self.children)
                if self.children else len(self._result))
231 232 233


    def __bool__(self):
234 235 236 237
        """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.
        """
238 239 240
        return True


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
    def equals(self, other: 'Node') -> bool:
293
        """
294 295 296 297 298 299
        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`.
300 301 302 303 304 305 306 307 308 309
        """
        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


310 311 312 313 314 315 316 317 318 319 320 321 322 323
    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


324
    def is_anonymous(self) -> bool:
eckhart's avatar
eckhart committed
325 326 327 328 329 330 331 332
        """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
333
        return not self.tag_name or self.tag_name[0] == ':'
334

Eckhart Arnold's avatar
Eckhart Arnold committed
335

336
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
337
    def result(self) -> StrictResultType:
338 339 340 341 342
        """
        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.
        """
343 344
        return self._result

345

346
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
347
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
348
        # # made obsolete by static type checking with mypy
eckhart's avatar
eckhart committed
349 350 351 352
        # 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
353 354
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
355
        # self._content = None
356 357 358 359 360 361 362 363 364
        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
365
                self._result = result  # cast(StrictResultType, result)
366

367

368 369 370 371 372 373 374 375 376 377 378 379
    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]
380

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


401
    @property
402
    def pos(self) -> int:
eckhart's avatar
eckhart committed
403 404
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
Eckhart Arnold's avatar
Eckhart Arnold committed
405
            raise AssertionError("Position value not initialized! Use Node.with_pos()")
406 407
        return self._pos

eckhart's avatar
eckhart committed
408

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
    @property
    def attr(self):
        """
        Returns a dictionary of XML-attr attached to the node.
        """
        try:
            if self._xml_attr is None:          # cython compatibility
                self._xml_attr = OrderedDict()
        except AttributeError:
            self._xml_attr = OrderedDict()
        return self._xml_attr


eckhart's avatar
eckhart committed
449 450 451 452 453 454 455 456 457 458 459 460 461
    def attr_active(self) -> bool:
        """
        Returns True, if XML-Attributes of this node have ever been set
        or queried, even if unsuccessfully.
        """
        try:
            if self._xml_attr is not None:
                return True
        except AttributeError:
            pass
        return False


462
    def compare_attr(self, other: 'Node') -> bool:
463
        """
464 465
        Returns True, if `self` and `other` have the same attributes with the
        same attribute values.
466
        """
467 468 469 470 471 472 473 474 475
        if self.attr_active():
            if other.attr_active():
                return self.attr == other.attr
            return len(self.attr) == 0
            # self has empty dictionary and other has no attributes
        elif other.attr_active():
            return len(other.attr) == 0
            # other has empty attribute dictionary and self as no attributes
        return True  # neither self nor other have any attributes
476

Eckhart Arnold's avatar
Eckhart Arnold committed
477

478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532
    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
533
            tag_names(set): A tag name or set of tag names that is being
534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569
                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


    def tree_size(self) -> int:
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
        return sum(child.tree_size() for child in self.children) + 1

    #
    # serialization methods
    #

570
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
571
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
572 573 574 575 576 577 578 579
        """
        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.

580
        Args:
581
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
582
            open_fn:   (Node->str) A function that returns an opening
583
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
584
            close_fn:  (Node->str) A function that returns a closeF
585
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
586
            data_fn:   (str->str) A function that filters the data string
587 588 589 590 591 592
                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
593 594
        head = open_fn(self)
        tail = close_fn(self)
595 596

        if not self.result:
597
            return head.rstrip() + tail.lstrip()
598

eckhart's avatar
eckhart committed
599
        tail = tail.lstrip(None if density & 2 else '')
600

601
        inline = inline or inline_fn(self)
602 603 604 605 606
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
607 608
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
609

610 611
        if self.children:
            content = []
612
            for child in self.children:
613
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
614
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
615
                if subtree:
eckhart's avatar
eckhart committed
616 617
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
618
            return head + usetab + (sep + usetab).join(content) + tail
619

eckhart's avatar
eckhart committed
620
        res = self.content
eckhart's avatar
eckhart committed
621
        if not inline and not head:
eckhart's avatar
eckhart committed
622
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
623
            res = res.strip()
624 625
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
626
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
627
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
628
        else:
629
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
630

Eckhart Arnold's avatar
Eckhart Arnold committed
631

632
    def as_sxpr(self, src: Optional[str] = None,
633
                indentation: int = 2,
634
                compact: bool = False,
eckhart's avatar
eckhart committed
635
                flatten_threshold: int = 92) -> str:
636
        """
637
        Returns content as S-expression, i.e. in lisp-like form. If this
eckhart's avatar
eckhart committed
638
        method is called on a RootNode-object,
639

640
        Args:
641 642
            src:  The source text or `None`. In case the source text is
                given the position of the element in the text will be
643 644 645
                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.
646 647
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
648
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
649
                tree structure.
650 651 652
            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.
653 654
        """

eckhart's avatar
eckhart committed
655
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
656
        lbreaks = linebreaks(src) if src else []  # type: List[int]
657
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[RootNode]
658

659
        def opening(node: Node) -> str:
eckhart's avatar
eckhart committed
660
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
661
            txt = [left_bracket, node.tag_name]
662
            # s += " '(pos %i)" % node.add_pos
663
            # txt.append(str(id(node)))  # for debugging
eckhart's avatar
eckhart committed
664
            if node.attr_active():
665
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
666
            if src:
667
                line, col = line_col(lbreaks, node.pos)
668 669 670
                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)
671 672
            if root and id(node) in root.error_nodes:
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in root.get_errors(node)))
673
            return "".join(txt) + '\n'
674

675
        def closing(node: Node) -> str:
eckhart's avatar
eckhart committed
676 677
            """Returns the closing string for the representation of `node`."""
            return right_bracket
678

679
        def pretty(strg: str) -> str:
eckhart's avatar
eckhart committed
680 681 682 683
            """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'\"')
684

685
        sxpr = self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
eckhart's avatar
eckhart committed
686
        return sxpr if compact else flatten_sxpr(sxpr, flatten_threshold)
Eckhart Arnold's avatar
Eckhart Arnold committed
687

eckhart's avatar
eckhart committed
688

689 690
    def as_xml(self, src: str = None,
               indentation: int = 2,
eckhart's avatar
eckhart committed
691 692 693
               inline_tags: Set[str] = frozenset(),
               omit_tags: Set[str] = frozenset(),
               empty_tags: Set[str] = frozenset()) -> str:
694 695 696
        """
        Returns content as XML-tree.

697
        Args:
698 699
            src:  The source text or `None`. In case the source text is given,
                the position will also be reported as line and column.
700
            indentation: The number of whitespaces for indentation
701 702 703
            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
704
                neither the opening tag nor its attr nor the closing tag. This
705 706 707
                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.
708 709
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
710
        """
711
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[RootNode]
712

713
        def opening(node: Node) -> str:
714 715 716
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
717
            txt = ['<', node.tag_name]
eckhart's avatar
eckhart committed
718
            has_reserved_attrs = node.attr_active() \
eckhart's avatar
eckhart committed
719
                and any(r in node.attr for r in {'err', 'line', 'col'})
eckhart's avatar
eckhart committed
720
            if node.attr_active():
721
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
722
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
723
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
724 725
            if src == '' and not (node.attr_active() and '_pos' in node.attr) and node.pos >= 0:
                txt.append(' _pos="%i"' % node.pos)
726
            if root and id(node) in root.error_nodes and not has_reserved_attrs:
eckhart's avatar
eckhart committed
727
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
eckhart's avatar
eckhart committed
728
                                                 for err in root.get_errors(node)))
729 730 731
            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
732
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
733 734 735
            else:
                ending = ">\n"
            return "".join(txt + [ending])
736

737
        def closing(node: Node):
738
            """Returns the closing string for the representation of `node`."""
739
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
740
                return ''
eckhart's avatar
eckhart committed
741
            return '\n</' + node.tag_name + '>'
742

743 744 745 746 747 748
        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

749
        def inlining(node: Node):
750 751 752 753
            """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
754
            return node.tag_name in inline_tags \
eckhart's avatar
eckhart committed
755
                or (node.attr_active()
eckhart's avatar
eckhart committed
756
                    and node.attr.get('xml:space', 'default') == 'preserve')
757

758
        line_breaks = linebreaks(src) if src else []
759
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
760
                               density=1, inline_fn=inlining)
761

Eckhart Arnold's avatar
Eckhart Arnold committed
762

763
    def to_json_obj(self) -> Dict:
eckhart's avatar
eckhart committed
764
        """Serialize a node or tree as json-object"""
Eckhart Arnold's avatar
Eckhart Arnold committed
765 766 767
        data = [self.tag_name,
                [child.to_json_obj() for child in self.children]
                if self.children else str(self._result)]
768 769 770 771 772
        has_attr = self.attr_active()
        if self._pos >= 0 or has_attr:
            data.append(self._pos)
        if has_attr:
            data.append(dict(self._xml_attr))
eckhart's avatar
eckhart committed
773
        return {'__class__': 'DHParser.Node', 'data': data}
774 775


776 777
    @staticmethod
    def from_json_obj(json_obj: Dict) -> 'Node':
778 779 780 781 782 783
        """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."""
        if json_obj.get('__class__', '') != 'DHParser.Node':
            raise ValueError('JSON object: ' + str(json_obj) +
                             ' does not represent a Node object.')
784
        tag_name, result, pos, attr = (json_obj['data'] + [-1, None])[:4]
785
        if isinstance(result, str):
786 787 788 789 790 791 792
            leafhint = True
        else:
            leafhint = False
            result = tuple(Node.from_json_obj(child) for child in result)
        node = Node(tag_name, result, leafhint)
        node._pos = pos
        if attr:
793
            node.attr.update(attr)
794
        return node
795

796 797 798 799
    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 (',', ':'))

800

eckhart's avatar
eckhart committed
801
def serialize(node: Node, how: str = 'default') -> str:
802
    """
803
    Serializes the tree starting with `node` either as S-expression, XML, JSON,
804
    or in compact form. Possible values for `how` are 'S-expression',
805
    'XML', 'JSON', 'compact' accordingly, or 'AST', 'CST', 'default' in which case
806 807 808 809 810 811 812 813 814 815 816 817 818
    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 == 's-expression':
819
        return node.as_sxpr(flatten_threshold=get_config_value('flatten_sxpr_threshold'))
820 821
    elif switch == 'xml':
        return node.as_xml()
822 823
    elif switch == 'json':
        return node.as_json()
824 825 826 827 828 829
    elif switch == 'compact':
        return node.as_sxpr(compact=True)
    else:
        raise ValueError('Unknown serialization %s, %s' % (how, switch))


830
class FrozenNode(Node):
831 832 833 834 835 836 837 838 839 840
    """
    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
    are always the same and will be dropped while parsing, anyway).

    Frozen nodes must be used only temporarily during parsing or
    tree-transformation and should not occur in the product of the
    transformation any more. This can be verified with `tree_sanity_check()`.
    """
841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857

    def __init__(self, tag_name: str, result: ResultType) -> None:
        if isinstance(result, str) or isinstance(result, StringView):
            result = str(result)
        else:
            raise TypeError('FrozenNode only accepts string as results. '
                            '(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
858 859 860 861
    @property
    def attr(self):
        raise AssertionError("Attributes cannot be accessed on a frozen node")

Eckhart Arnold's avatar
Eckhart Arnold committed
862
    def with_pos(self, pos: int) -> 'Node':
863 864 865
        pass


866
PLACEHOLDER = FrozenNode('__PLACEHOLDER__', '')
di68kap's avatar
di68kap committed
867 868


869
def tree_sanity_check(tree: Node) -> bool:
870 871 872 873 874 875 876 877 878
    """
    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.
    """
879
    node_set = set()  # type: Set[Node]
880
    for node in tree.select(lambda nd: True, include_root=True):
881
        if node in node_set or isinstance(Node, FrozenNode):
882 883 884 885 886
            return False
        node_set.add(node)
    return True


di68kap's avatar
di68kap committed
887
class RootNode(Node):
888 889 890 891 892 893 894 895 896 897 898 899
    """The root node for the syntax tree is a special kind of node that keeps
    and manages global properties of the tree as a whole. These are first and
    foremost the list off errors that occurred during tree generation
    (i.e. parsing) or any transformation of the tree. Other properties concern
    the customization of the XML-serialization.

    The root node can be instantiated before the tree is fully parsed. This is
    necessary, because the root node is needed for managing error messages
    during the parsing process, already. In order to connect the root node to
    the tree, when parsing is finished, the swallow()-method must be called.

        errors (list):  A list of all errors that have occurred so far during
900 901
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
902

903 904 905 906 907 908
        error_nodes (dict): A mapping of node-ids to a list of errors that
                occurred on the node with the respective id.

        error_positions (dict): A mapping of locations to a set of ids of
                nodes that contain an error at that particular location

909 910
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
911 912 913 914 915 916

        inline_tags (set of strings): see `Node.as_xml()` for an explanation.

        omit_tags (set of strings): see `Node.as_xml()` for an explanation.

        empty_tags (set oif strings): see `Node.as_xml()` for an explanation.
di68kap's avatar
di68kap committed
917
    """
918

eckhart's avatar
eckhart committed
919
    def __init__(self, node: Optional[Node] = None):
920
        super().__init__('__not_yet_ready__', '')
921
        self.errors = []               # type: List[Error]
922 923
        self.error_nodes = dict()      # type: Dict[int, List[Error]]  # id(node) -> error list
        self.error_positions = dict()  # type: Dict[int, Set[int]]  # pos -> set of id(node)