syntaxtree.py 47.3 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
# syntaxtree.py - syntax tree classes for DHParser
#
# Copyright 2016  by Eckhart Arnold (arnold@badw.de)
#                 Bavarian Academy of Sciences an Humanities (badw.de)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied.  See the License for the specific language governing
# permissions and limitations under the License.
17 18 19


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

26
from collections import OrderedDict
eckhart's avatar
eckhart committed
27
import copy
28
from typing import Callable, cast, Iterator, List, AbstractSet, Set, Union, Tuple, Optional, Dict
29

eckhart's avatar
eckhart committed
30
from DHParser.error import Error, ErrorCode, linebreaks, line_col
31
from DHParser.stringview import StringView
32
from DHParser.toolkit import get_config_value, re
di68kap's avatar
di68kap committed
33

34

di68kap's avatar
di68kap committed
35
__all__ = ('WHITESPACE_PTYPE',
36
           'TOKEN_PTYPE',
37 38
           'ZOMBIE_TAG',
           'PLACEHOLDER',
eckhart's avatar
eckhart committed
39 40 41
           'ResultType',
           'StrictResultType',
           'ChildrenType',
42
           'Node',
43
           'serialize',
44
           'FrozenNode',
45
           'tree_sanity_check',
di68kap's avatar
di68kap committed
46
           'RootNode',
47
           'parse_sxpr',
48
           'parse_xml',
eckhart's avatar
eckhart committed
49
           'parse_tree',
50 51
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
52 53


54 55 56 57 58 59 60 61
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


WHITESPACE_PTYPE = ':Whitespace'
62
TOKEN_PTYPE = ':Token'
63

64
ZOMBIE_TAG = "__ZOMBIE__"
65

66 67 68 69 70 71 72
#######################################################################
#
# syntaxtree nodes
#
#######################################################################


73
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
74
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
75
StrictResultType = Union[ChildrenType, StringView, str]
76
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
77 78


Eckhart Arnold's avatar
Eckhart Arnold committed
79
def flatten_sxpr(sxpr: str) -> str:
80 81
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
82 83 84
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
85
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
86 87
    '(a (b c))'
    """
88

89
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
90 91


92
def flatten_xml(xml: str) -> str:
93 94
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
95
    i.e. only whitespace within leaf-nodes is preserved.
96 97
    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`.
98
    """
99

100 101 102 103
    # 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']
104
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
105 106


eckhart's avatar
eckhart committed
107
RX_AMP = re.compile(r'&(?!\w+;)')
108 109


Eckhart Arnold's avatar
Eckhart Arnold committed
110
class Node:  # (collections.abc.Sized): Base class omitted for cython-compatibility
111 112 113
    """
    Represents a node in the concrete or abstract syntax tree.

114 115
    TODO: Add some documentation and doc-tests here...

116
    Attributes:
117 118
        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
119

120 121 122
        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.
123

124 125
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
126

eckhart's avatar
eckhart committed
127 128 129 130
        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
131
        parser (Parser):  The parser which generated this node.
132 133
            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
134 135
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
136
            example by calling ``isinstance(node.parer, ...)``.
137

138 139 140
        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
141
            the length before AST-transformation and will never change
142
            through AST-transformation. READ ONLY!
143

144 145
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
146
            The value of ``pos`` is -1 meaning invalid by default.
147
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
148
            nodes relative to this value.
149 150

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
151
            pos value of the root node should be set to 0 right
152 153
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
154
            Other than that, this value should be considered READ ONLY.
155 156
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
157

158 159
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
160 161
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
162
    """
163

164
    __slots__ = '_result', 'children', '_len', '_pos', 'tag_name', '_xml_attr'
165

166
    def __init__(self, tag_name: str, result: ResultType, leafhint: bool = False) -> None:
167 168
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
169 170
        that generated the node and the parser's result.
        """
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)
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
        self.tag_name = tag_name        # type: str
182

eckhart's avatar
eckhart committed
183 184
    def __deepcopy__(self, memo):
        if self.children:
185
            duplicate = self.__class__(self.tag_name, copy.deepcopy(self.children), False)
eckhart's avatar
eckhart committed
186
        else:
187
            duplicate = self.__class__(self.tag_name, self.result, True)
eckhart's avatar
eckhart committed
188 189
        duplicate._pos = self._pos
        duplicate._len = self._len
eckhart's avatar
eckhart committed
190
        if self.attr_active():
di68kap's avatar
di68kap committed
191 192
            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
193
        return duplicate
Eckhart Arnold's avatar
Eckhart Arnold committed
194

195
    def __str__(self):
196 197
        if isinstance(self, RootNode):
            root = cast(RootNode, self)
198
            errors = root.errors_sorted
199 200 201 202 203 204
            if errors:
                e_pos = errors[0].pos
                return self.content[:e_pos] + \
                   ' <<< Error on "%s" | %s >>> ' % \
                   (self.content[e_pos - self.pos:], '; '.join(e.message for e in errors))
        return self.content
Eckhart Arnold's avatar
Eckhart Arnold committed
205

206
    def __repr__(self):
207
        # mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
208 209
        # name, ptype = (self._tag_name.split(':') + [''])[:2]
        # parg = "MockParser({name}, {ptype})".format(name=name, ptype=ptype)
210
        rarg = str(self) if not self.children else \
211
            "(" + ", ".join(child.__repr__() for child in self.children) + ")"
212
        return "Node(%s, %s)" % (self.tag_name, rarg)
213

Eckhart Arnold's avatar
Eckhart Arnold committed
214

215
    def __len__(self):
216
        if self._len < 0:
di68kap's avatar
di68kap committed
217
            self._len = sum(len(child) for child in self.children) \
218
                if self.children else len(self._result)
219 220 221 222 223 224 225 226
        return self._len


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


227 228 229 230 231 232 233 234 235
    # can lead to obscure mistakes, where default object comparison behaviour is expected
    # def __eq__(self, other):
    #     """
    #     Equality of nodes: Two nodes are considered as equal, if their tag
    #     name is the same, if their results are equal and if their attributes
    #     and attribute values are the same.
    #     """
    #     return self.tag_name == other.tag_name and self.result == other.result \
    #         and self.compare_attr(other)
236

Eckhart Arnold's avatar
Eckhart Arnold committed
237

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

Eckhart Arnold's avatar
Eckhart Arnold committed
241

242 243 244
    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
245
        an integer or the first child node with the given tag name. Examples::
246

247
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
248 249
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
250 251
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
252 253 254 255

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
256
        Returns:
257 258
            Node: All nodes which have a given tag name.
        """
259 260 261
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
262
            else:
263 264 265 266 267
                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!')
268 269 270 271


    def __contains__(self, tag_name: str) -> bool:
        """
272
        Returns true if a child with the given tag name exists.
273
        Args:
274 275
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
276 277 278 279
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
280 281 282 283 284
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
285
            return False
286
        raise ValueError('Leave node cannot contain other nodes')
287 288


289
    def equals(self, other: 'Node') -> bool:
290
        """
291 292 293 294 295 296
        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`.
297 298 299 300 301 302 303 304 305 306
        """
        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


307 308 309 310 311 312 313 314 315 316 317 318 319 320
    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


321
    def is_anonymous(self) -> bool:
Eckhart Arnold's avatar
Eckhart Arnold committed
322
        return not self.tag_name or self.tag_name[0] == ':'
323

Eckhart Arnold's avatar
Eckhart Arnold committed
324

325
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
326
    def result(self) -> StrictResultType:
327 328 329 330 331
        """
        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.
        """
332 333
        return self._result

334

335
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
336
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
337
        # # made obsolete by static type checking with mypy
Eckhart Arnold's avatar
Eckhart Arnold committed
338 339 340 341
        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
342 343
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
344
        self._len = -1        # lazy evaluation
345
        # self._content = None
346 347 348 349 350 351 352 353 354
        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
355
                self._result = result  # cast(StrictResultType, result)
356

357 358 359 360 361 362 363 364 365 366 367 368
    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]
369 370

    @property
eckhart's avatar
eckhart committed
371
    def content(self) -> str:
372
        """
373 374 375
        Returns content as string. If the node has child-nodes, the
        string content of the child-nodes is recursively read and then
        concatenated.
376
        """
377 378 379 380 381
        return ''.join(self._content())
        ## unoptimized
        # return "".join(child.content for child in self.children) if self.children else str(
        # self._result)
        ## obsolete
382 383 384 385 386 387 388 389
        # if self._content is None:
        #     if self.children:
        #         self._content = "".join(child.content for child in self.children)
        #     else:
        #         # self._content = self._result
        #         self._content = str(self._result)
        #         self._result = self._content  # self._result might be more efficient as a string!?
        # return self._content
390 391 392 393 394
    #
    #
    # @content.setter
    # def content(self, content: str):
    #     self.result = content
di68kap's avatar
di68kap committed
395 396


397 398 399 400 401 402
    @property
    def structure(self) -> str:
        """
        Return structure (and content) as S-expression on a single line
        without any line breaks.
        """
403
        return flatten_sxpr(self.as_sxpr())
404 405


406
    @property
407
    def pos(self) -> int:
eckhart's avatar
eckhart committed
408 409
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
Eckhart Arnold's avatar
Eckhart Arnold committed
410
            raise AssertionError("Position value not initialized! Use Node.with_pos()")
411 412
        return self._pos

eckhart's avatar
eckhart committed
413

Eckhart Arnold's avatar
Eckhart Arnold committed
414
    def with_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
415
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
416
        Initialize position value. Usually, the parser guard
eckhart's avatar
eckhart committed
417 418
        (`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
419
        when Nodes are created outside the reach of the parser
eckhart's avatar
eckhart committed
420
        guard, their document-position must be assigned manually.
Eckhart Arnold's avatar
Eckhart Arnold committed
421 422
        Position values of the child nodes are assigned recursively, too.
        Returns the node itself for convenience.
eckhart's avatar
eckhart committed
423
        """
424 425 426 427
        # 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
428 429 430 431 432 433 434 435 436 437
        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
438 439
        return self

440

441 442 443 444 445 446 447 448 449 450 451 452 453
    @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
454 455 456 457 458 459 460 461 462 463 464 465 466
    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


467
    def compare_attr(self, other: 'Node') -> bool:
468
        """
469 470
        Returns True, if `self` and `other` have the same attributes with the
        same attribute values.
471
        """
472 473 474 475 476 477 478 479 480
        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
481

Eckhart Arnold's avatar
Eckhart Arnold committed
482

483
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
484
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
485 486 487 488 489 490 491 492
        """
        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.

493
        Args:
494
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
495
            open_fn:   (Node->str) A function that returns an opening
496
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
497
            close_fn:  (Node->str) A function that returns a closeF
498
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
499
            data_fn:   (str->str) A function that filters the data string
500 501 502 503 504 505
                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
506 507
        head = open_fn(self)
        tail = close_fn(self)
508 509

        if not self.result:
510
            return head.rstrip() + tail.lstrip()
511

eckhart's avatar
eckhart committed
512
        tail = tail.lstrip(None if density & 2 else '')
513

514
        inline = inline or inline_fn(self)
515 516 517 518 519
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
520 521
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
522

523 524
        if self.children:
            content = []
525
            for child in self.children:
526
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
527
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
528
                if subtree:
eckhart's avatar
eckhart committed
529 530
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
531
            return head + usetab + (sep + usetab).join(content) + tail
532

eckhart's avatar
eckhart committed
533
        res = self.content
eckhart's avatar
eckhart committed
534
        if not inline and not head:
eckhart's avatar
eckhart committed
535
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
536
            res = res.strip()
537 538
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
539
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
540
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
541
        else:
542
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
543

Eckhart Arnold's avatar
Eckhart Arnold committed
544

545 546
    def as_sxpr(self, src: str = None,
                indentation: int = 2,
547
                compact: bool = False) -> str:
548
        """
549 550
        Returns content as S-expression, i.e. in lisp-like form. If this
        method is callad on a RootNode-object,
551

552
        Args:
553 554 555
            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.
556 557
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
558
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
559
                tree structure.
560 561
        """

eckhart's avatar
eckhart committed
562
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
563
        lbreaks = linebreaks(src) if src else []  # type: List[int]
564
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[RootNode]
565

566
        def opening(node: Node) -> str:
eckhart's avatar
eckhart committed
567
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
568
            txt = [left_bracket, node.tag_name]
569
            # s += " '(pos %i)" % node.add_pos
570
            # txt.append(str(id(node)))  # for debugging
eckhart's avatar
eckhart committed
571
            if node.attr_active():
572
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
573
            if src:
574 575
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
576 577
            if root and id(node) in root.error_nodes:
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in root.get_errors(node)))
578
            return "".join(txt) + '\n'
579

580
        def closing(node: Node) -> str:
eckhart's avatar
eckhart committed
581 582
            """Returns the closing string for the representation of `node`."""
            return right_bracket
583

584
        def pretty(strg: str) -> str:
eckhart's avatar
eckhart committed
585 586 587 588
            """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'\"')
589

590
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
591

eckhart's avatar
eckhart committed
592

593 594
    def as_xml(self, src: str = None,
               indentation: int = 2,
eckhart's avatar
eckhart committed
595 596 597
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
598 599 600
        """
        Returns content as XML-tree.

601
        Args:
602 603 604
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
605
            indentation: The number of whitespaces for indentation
606 607 608
            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
609
                neither the opening tag nor its attr nor the closing tag. This
610 611 612
                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.
613 614
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
615
        """
616
        root = cast(RootNode, self) if isinstance(self, RootNode) else None  # type: Optional[RootNode]
617

618
        def opening(node: Node) -> str:
619 620 621
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
622
            txt = ['<', node.tag_name]
eckhart's avatar
eckhart committed
623
            has_reserved_attrs = node.attr_active() \
eckhart's avatar
eckhart committed
624
                and any(r in node.attr for r in {'err', 'line', 'col'})
eckhart's avatar
eckhart committed
625
            if node.attr_active():
626
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
627
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
628
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
629
            if root and id(node) in root.error_nodes and not has_reserved_attrs:
eckhart's avatar
eckhart committed
630
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
eckhart's avatar
eckhart committed
631
                                                 for err in root.get_errors(node)))
632 633 634
            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
635
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
636 637 638
            else:
                ending = ">\n"
            return "".join(txt + [ending])
639

640
        def closing(node: Node):
641
            """Returns the closing string for the representation of `node`."""
642
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
643
                return ''
644
            return ('\n</') + node.tag_name + '>'
645

646 647 648 649 650 651
        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

652
        def inlining(node: Node):
653 654 655 656
            """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
657
            return node.tag_name in inline_tags \
eckhart's avatar
eckhart committed
658
                or (node.attr_active()
eckhart's avatar
eckhart committed
659
                    and node.attr.get('xml:space', 'default') == 'preserve')
660

661
        line_breaks = linebreaks(src) if src else []
662
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
663
                               density=1, inline_fn=inlining)
664

Eckhart Arnold's avatar
Eckhart Arnold committed
665

eckhart's avatar
eckhart committed
666
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
667
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
668
        """
669
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
670

671
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
672
        given `match_function` evaluates to True. The tree is
673 674 675
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
676

677 678 679
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
680 681
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
682 683
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
684
        Yields:
685
            Node: All nodes of the tree for which
686 687
            ``match_function(node)`` returns True
        """
688
        if include_root and match_function(self):
689
            yield self
690 691 692
        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
693
                yield node
694

Eckhart Arnold's avatar
Eckhart Arnold committed
695

eckhart's avatar
eckhart committed
696
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
eckhart's avatar
eckhart committed
697
                      include_root: bool = False) -> Iterator['Node']:
698
        """
699 700 701 702
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
703

704
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
705
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
706
            ['(X (c "d"))', '(X "F")']
707
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
708
            ['(b "X")', '(X (c "d"))', '(X "F")']
709
            >>> any(tree.select_by_tag('a', False))
710
            False
711
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
712
            ['(a (b "X") (X (c "d")) (e (X "F")))']
713 714
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
715 716

        Args:
717 718 719 720
            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.
721 722 723
        Yields:
            Node: All nodes which have a given tag name.
        """
724
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
725
            tag_names = frozenset({tag_names})
726
        return self.select(lambda node: node.tag_name in tag_names, include_root)
727 728


729 730 731 732
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

733
        This function is mostly just syntactic sugar for
734 735 736 737 738 739 740 741 742 743
        ``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


744
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
745 746 747
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
748 749 750
        return sum(child.tree_size() for child in self.children) + 1


751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777
def serialize(node: Node, how: str='default') -> str:
    """
    Serializes the tree starting with `node` either as S-expression, XML
    or in compact form. Possible values for `how` are 'S-expression',
    'XML', '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 == 's-expression':
        return node.as_sxpr()
    elif switch == 'xml':
        return node.as_xml()
    elif switch == 'compact':
        return node.as_sxpr(compact=True)
    else:
        raise ValueError('Unknown serialization %s, %s' % (how, switch))


778
class FrozenNode(Node):
779 780 781 782 783 784 785 786 787 788
    """
    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()`.
    """
789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805

    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
806 807 808 809
    @property
    def attr(self):
        raise AssertionError("Attributes cannot be accessed on a frozen node")

Eckhart Arnold's avatar
Eckhart Arnold committed
810
    def with_pos(self, pos: int) -> 'Node':
811 812 813
        pass


814
PLACEHOLDER = FrozenNode('__PLACEHOLDER__', '')
di68kap's avatar
di68kap committed
815 816


817
def tree_sanity_check(tree: Node) -> bool:
818 819 820 821 822 823 824 825 826
    """
    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.
    """
827
    node_set = set()  # type: Set[Node]
828
    for node in tree.select(lambda nd: True, include_root=True):
829
        if node in node_set or isinstance(Node, FrozenNode):
830 831 832 833 834
            return False
        node_set.add(node)
    return True


di68kap's avatar
di68kap committed
835
class RootNode(Node):
836 837 838 839 840 841 842 843 844 845 846 847
    """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
848 849
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
850

851 852 853 854 855 856
        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

857 858
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
859 860 861 862 863 864

        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
865
    """
866

eckhart's avatar
eckhart committed
867
    def __init__(self, node: Optional[Node] = None):
868
        super().__init__('__not_yet_ready__', '')
869
        self.errors = []               # type: List[Error]
870 871
        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)
di68kap's avatar
di68kap committed
872
        self.error_flag = 0
eckhart's avatar
eckhart committed
873 874
        if node is not None:
            self.swallow(node)
875
        # customization for XML-Representation
eckhart's avatar
eckhart committed
876
        self.inline_tags = set()  # type: Set[str]
877 878
        self.omit_tags = set()    # type: Set[str]
        self.empty_tags = set()   # type: Set[str]