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

syntaxtree.py 19.1 KB
Newer Older
1
"""syntaxtree.py - syntax tree classes for DHParser
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

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.
"""

19
import copy
20
import os
21 22
from functools import partial

23 24 25 26
try:
    import regex as re
except ImportError:
    import re
27 28 29 30 31 32
try:
    from typing import AbstractSet, Any, ByteString, Callable, cast, Container, Dict, \
        Iterator, List, NamedTuple, Sequence, Union, Text, Tuple
except ImportError:
    from .typing34 import AbstractSet, Any, ByteString, Callable, cast, Container, Dict, \
        Iterator, List, NamedTuple, Sequence, Union, Text, Tuple
33

34
from DHParser.toolkit import log_dir, line_col, identity
35

36
__all__ = ('WHITESPACE_PTYPE',
37
           'TOKEN_PTYPE',
Eckhart Arnold's avatar
Eckhart Arnold committed
38
           'ZOMBIE_PARSER',
Eckhart Arnold's avatar
Eckhart Arnold committed
39
           'ParserBase',
Eckhart Arnold's avatar
Eckhart Arnold committed
40 41
           'Error',
           'Node',
42 43
           'mock_syntax_tree',
           'TransformationFunc')
Eckhart Arnold's avatar
Eckhart Arnold committed
44 45


Eckhart Arnold's avatar
Eckhart Arnold committed
46 47 48 49 50 51 52 53 54 55
class ParserBase:
    """
    ParserBase is the base class for all real and mock parser classes.
    It is defined here, because Node objects require a parser object
    for instantiation.
    """
    def __init__(self, name=''):  # , pbases=frozenset()):
        self.name = name  # type: str
        self._ptype = ':' + self.__class__.__name__  # type: str

56 57 58
    def __repr__(self):
        return self.name + self.ptype

Eckhart Arnold's avatar
Eckhart Arnold committed
59
    def __str__(self):
60
        return self.name + (' = ' if self.name else '') + repr(self)
Eckhart Arnold's avatar
Eckhart Arnold committed
61 62 63 64 65

    @property
    def ptype(self) -> str:
        return self._ptype

66 67 68
    @property
    def repr(self) -> str:
        return self.name if self.name else repr(self)
Eckhart Arnold's avatar
Eckhart Arnold committed
69

70

71 72 73 74
WHITESPACE_PTYPE = ':Whitespace'
TOKEN_PTYPE = ':Token'


Eckhart Arnold's avatar
Eckhart Arnold committed
75
class MockParser(ParserBase):
76 77
    """
    MockParser objects can be used to reconstruct syntax trees from a
Eckhart Arnold's avatar
Eckhart Arnold committed
78 79
    serialized form like S-expressions or XML. Mock objects can mimic
    different parser types by assigning them a ptype on initialization.
80 81 82 83 84
    
    Mock objects should not be used for anything other than 
    syntax tree (re-)construction. In all other cases where a parser
    object substitute is needed, chose the singleton ZOMBIE_PARSER.
    """
Eckhart Arnold's avatar
Eckhart Arnold committed
85
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
86
        assert not ptype or ptype[0] == ':'
Eckhart Arnold's avatar
Eckhart Arnold committed
87
        super(MockParser, self).__init__(name)
88
        self.name = name
Eckhart Arnold's avatar
Eckhart Arnold committed
89
        self._ptype = ptype or ':' + self.__class__.__name__
90 91 92


class ZombieParser(MockParser):
93 94 95 96 97 98 99 100 101 102 103 104
    """
    Serves as a substitute for a Parser instance.

    ``ZombieParser`` is the class of the singelton object
    ``ZOMBIE_PARSER``. The  ``ZOMBIE_PARSER`` has a name and can be
    called, but it never matches. It serves as a substitute where only
    these (or one of these properties) is needed, but no real Parser-
    object is instantiated.
    """
    alive = False

    def __init__(self):
105
        super(ZombieParser, self).__init__("__ZOMBIE__")
106 107 108 109
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
        self.__class__.alive = True

Eckhart Arnold's avatar
Eckhart Arnold committed
110 111 112 113 114 115
    def __copy__(self):
        return self

    def __deepcopy__(self, memo):
        return self

116 117 118 119
    def __call__(self, text):
        """Better call Saul ;-)"""
        return None, text

120 121 122 123

ZOMBIE_PARSER = ZombieParser()


124 125 126 127 128
# # Python 3.6:
# class Error(NamedTuple):
#     pos: int
#     msg: str
Error = NamedTuple('Error', [('pos', int), ('msg', str)])
129

130 131 132
ChildrenType = Tuple['Node', ...]
StrictResultType = Union[ChildrenType, str]
ResultType = Union[ChildrenType, 'Node', str, None]
133 134


Eckhart Arnold's avatar
Eckhart Arnold committed
135 136 137 138 139 140 141 142 143 144 145
def oneliner_sxpr(sxpr: str) -> str:
    """Returns S-expression `sxpr` as a one liner without unnecessary
    whitespace.

    Example:
    >>> oneliner_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
    '(a (b c))'
    """
    return re.sub('\s(?=\))', '', re.sub('\s+', ' ', sxpr)).strip()


146 147 148 149 150 151 152 153 154 155 156 157
class Node:
    """
    Represents a node in the concrete or abstract syntax tree.

    Attributes:
        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
        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.
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
158 159 160 161 162 163
        parser (Parser):  The parser which generated this node. 
            WARNING: In case you use mock syntax trees for testing or
            parser replacement during the AST-transformation: DO NOT
            rely on this being a real parser object in any phase after 
            parsing (i.e. AST-transformation and compiling), for 
            example by calling ``isinstance(node.parer, ...)``.
164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185
        errors (list):  A list of parser- or compiler-errors:
            tuple(position, string) attached to this node
        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
            the length before AST-transformation ans will never change
            through AST-transformation. READ ONLY!
        pos (int):  the position of the node within the parsed text.

            The value of ``pos`` is -1 meaning invalid by default. 
            Setting this value will set the positions of all child
            nodes relative to this value.  

            To set the pos values of all nodes in a syntax tree, the
            pos value of the root node should be set to 0 right 
            after parsing.

            Other than that, this value should be considered READ ONLY. 
            At any rate, it should only be reassigned only during
            parsing stage and never during or after the
            AST-transformation.
    """
186

Eckhart Arnold's avatar
Eckhart Arnold committed
187
    def __init__(self, parser, result: ResultType) -> None:
188 189 190
        """Initializes the ``Node``-object with the ``Parser``-Instance
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
191 192
        # self._result = ''  # type: StrictResultType
        # self._children = ()  # type: ChildrenType
193
        self._errors = []  # type: List[str]
194
        self.result = result
Eckhart Arnold's avatar
Eckhart Arnold committed
195 196 197
        self._len = len(result) if not self._children else \
            sum(child._len for child in self._children)  # type: int
        # self.pos: int  = 0  # continuous updating of pos values wastes a lot of time
198
        self._pos = -1  # type: int
199
        self.parser = parser or ZOMBIE_PARSER
Eckhart Arnold's avatar
Eckhart Arnold committed
200 201
        self.error_flag = any(r.error_flag for r in self._children) \
            if self._children else False  # type: bool
202 203 204

    def __str__(self):
        if self.children:
205
            return "".join(str(child) for child in self.children)
206 207 208 209 210 211 212 213
        return str(self.result)

    def __repr__(self):
        mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
        parg = "MockParser({name}, {ptype})".format(**mpargs)
        rarg = str(self) if not self.children else \
               "(" + ", ".join(repr(child) for child in self.children) + ")"
        return "Node(%s, %s)" % (parg, rarg)
214

215
    def __eq__(self, other):
216 217
        # return str(self.parser) == str(other.parser) and self.result == other.result
        return self.tag_name == other.tag_name and self.result == other.result
218

219
    def __hash__(self):
220
        return hash(self.tag_name)
221 222 223 224 225 226 227

    def __deepcopy__(self, memodict={}):
        result = copy.deepcopy(self.result)
        other = Node(self.parser, result)
        other._pos = self._pos
        return other

228
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
229
    def tag_name(self) -> str:
230
        return self.parser.name or self.parser.ptype
231 232

    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
233
    def result(self) -> StrictResultType:
234 235 236
        return self._result

    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
237
    def result(self, result: ResultType):
238 239 240 241
        # # made obsolete by static type checking with mypy is done
        # assert ((isinstance(result, tuple) and all(isinstance(child, Node) for child in result))
        #         or isinstance(result, Node)
        #         or isinstance(result, str)), str(result)
Eckhart Arnold's avatar
Eckhart Arnold committed
242
        self._result = (result,) if isinstance(result, Node) else result or ''  # type: StrictResultType
243
        self._children = cast(ChildrenType, self._result) \
Eckhart Arnold's avatar
Eckhart Arnold committed
244
            if isinstance(self._result, tuple) else cast(ChildrenType, ())  # type: ChildrenType
245 246

    @property
247
    def children(self) -> ChildrenType:
248 249 250
        return self._children

    @property
251
    def len(self) -> int:
252
        # DEBUGGING:  print(self.tag_name, str(self.pos), str(self._len), str(self)[:10].replace('\n','.'))
253 254 255
        return self._len

    @property
256
    def pos(self) -> int:
257 258 259 260
        assert self._pos >= 0, "position value not initialized!"
        return self._pos

    @pos.setter
261 262
    def pos(self, pos: int):
        # assert isinstance(pos, int)
263 264 265 266 267 268 269
        self._pos = pos
        offset = 0
        for child in self.children:
            child.pos = pos + offset
            offset += child.len

    @property
270
    def errors(self) -> List[Error]:
271 272
        return [Error(self.pos, err) for err in self._errors]

Eckhart Arnold's avatar
Eckhart Arnold committed
273 274 275 276 277 278 279 280 281
    def add_error(self, error_str) -> 'Node':
        self._errors.append(error_str)
        self.error_flag = True
        return self

    def propagate_error_flags(self) -> None:
        """Recursively propagates error flags set on child nodes to its
        parents. This can be used if errors are added to descendant
        nodes after syntaxtree construction, i.e. in the compile phase.
282
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300
        for child in self.children:
            child.propagate_error_flags()
            self.error_flag = self.error_flag or child.error_flag

    def collect_errors(self, clear_errors=False) -> List[Error]:
        """
        Returns all errors of this node or any child node in the form
        of a set of tuples (position, error_message), where position
        is always relative to this node.
        """
        errors = self.errors
        if clear_errors:
            self._errors = []
            self.error_flag = False
        if self.children:
            for child in self.children:
                errors.extend(child.collect_errors(clear_errors))
        return errors
301

302
    def _tree_repr(self, tab, openF, closeF, dataF=identity, density=0) -> str:
303 304 305 306 307 308 309 310
        """
        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.

311
        Args:
312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327
            tab (str):  The indentation string, e.g. '\t' or '    '
            openF:  (Node->str) A function that returns an opening
                string (e.g. an XML-tag_name) for a given node
            closeF:  (Node->str) A function that returns a closeF
                string (e.g. an XML-tag_name) for a given node.
            dataF:  (str->str) A function that filters the data string
                before printing, e.g. to add quotation marks

        Returns (str):
            A string that contains a (serialized) tree representation
            of the node and its children.
        """
        head = openF(self)
        tail = closeF(self)

        if not self.result:
328
            return head.rstrip() + tail.lstrip()
329

330
        D = None if density & 2 else ''
331 332 333

        if self.children:
            content = []
334
            for child in self.children:
335
                subtree = child._tree_repr(tab, openF, closeF, dataF, density).split('\n')
336
                content.append('\n'.join((tab + s) for s in subtree))
337
            return head + '\n'.join(content) + tail.lstrip(D)
338

339
        res = cast(str, self.result)  # safe, because if there are no children, result is a string
340 341 342 343
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
            gap = ' ' if head.rstrip()[-1] != '>' else ''
            return head.rstrip() + gap + dataF(self.result) + tail.lstrip()
344
        else:
345
            return head + '\n'.join([tab + dataF(s) for s in res.split('\n')]) + tail.lstrip(D)
346

347
    def as_sxpr(self, src: str=None) -> str:
348 349 350
        """
        Returns content as S-expression, i.e. in lisp-like form.

351
        Args:
352 353 354 355 356
            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.
        """

357
        def opening(node) -> str:
358 359 360
            s = '(' + node.tag_name
            # s += " '(pos %i)" % node.pos
            if src:
Eckhart Arnold's avatar
Eckhart Arnold committed
361
                s += " '(pos %i " % node.pos + " %i %i)" % line_col(src, node.pos)
362 363 364
            if node.errors:
                s += " '(err '(%s))" % ' '.join(str(err).replace('"', r'\"')
                                                for err in node.errors)
365
            return s + '\n'
366 367 368 369 370 371

        def pretty(s):
            return '"%s"' % s if s.find('"') < 0 \
                else "'%s'" % s if s.find("'") < 0 \
                else '"%s"' % s.replace('"', r'\"')

372
        return self._tree_repr('    ', opening, lambda node: '\n)', pretty, density=0)
373

374
    def as_xml(self, src: str=None) -> str:
375 376 377
        """
        Returns content as XML-tree.

378
        Args:
379 380 381 382 383
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
        """

384
        def opening(node) -> str:
385 386 387 388 389 390
            s = '<' + node.tag_name
            # s += ' pos="%i"' % node.pos
            if src:
                s += ' line="%i" col="%i"' % line_col(src, node.pos)
            if node.errors:
                s += ' err="%s"' % ''.join(str(err).replace('"', r'\"') for err in node.errors)
391
            return s + ">\n"
392 393

        def closing(node):
394
            return '\n</' + node.tag_name + '>'
395

396
        return self._tree_repr('    ', opening, closing, density=1)
397

Eckhart Arnold's avatar
Eckhart Arnold committed
398 399 400 401
    def structure(self) -> str:
        """Return structure (and content) as S-expression on a single line
        without any line breaks."""
        return oneliner_sxpr(self.as_sxpr())
402

Eckhart Arnold's avatar
Eckhart Arnold committed
403
    def content(self) -> str:
404
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
405 406
        Returns content as string, inserting error messages where
        errors occurred.
407
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
408 409 410 411
        s = "".join(child.content() for child in self.children) if self.children \
            else str(self.result)
        return (
        ' <<< Error on "%s" | %s >>> ' % (s, '; '.join(self._errors))) if self._errors else s
412

413
    def find(self, match_function: Callable) -> Iterator['Node']:
414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433
        """Finds nodes in the tree that match a specific criterion.
        
        ``find`` is a generator that yields all nodes for which the
        given ``match_function`` evaluates to True. The tree is 
        traversed pre-order.
        
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
        Yields:
            Node: all nodes of the tree for which 
            ``match_function(node)`` returns True
        """
        if match_function(self):
            yield self
        else:
            for child in self.children:
                for nd in child.find(match_function):
                    yield nd

434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450
    # def range(self, match_first, match_last):
    #     """Iterates over the range of nodes, starting from the first
    #     node for which ``match_first`` becomes True until the first node
    #     after this one for which ``match_last`` becomes true or until
    #     the end if it never does.
    #
    #     Args:
    #         match_first (function): A function  that takes as Node
    #             object as argument and returns True or False
    #         match_last (function): A function  that takes as Node
    #             object as argument and returns True or False
    #     Yields:
    #         Node: all nodes of the tree for which
    #         ``match_function(node)`` returns True
    #     """


451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472
    # def navigate(self, path):
    #     """Yields the results of all descendant elements matched by
    #     ``path``, e.g.
    #     'd/s' yields 'l' from (d (s l)(e (r x1) (r x2))
    #     'e/r' yields 'x1', then 'x2'
    #     'e'   yields (r x1)(r x2)
    #
    #     Args:
    #         path (str):  The path of the object, e.g. 'a/b/c'. The
    #             components of ``path`` can be regular expressions
    #
    #     Returns:
    #         The object at the path, either a string or a Node or
    #         ``None``, if the path did not match.
    #     """
    #     def nav(node, pl):
    #         if pl:
    #             return itertools.chain(nav(child, pl[1:]) for child in node.children
    #                                    if re.match(pl[0], child.tag_name))
    #         else:
    #             return self.result,
    #     return nav(path.split('/'))
473

Eckhart Arnold's avatar
Eckhart Arnold committed
474 475 476 477 478
    def log(self, log_file_name):
        st_file_name = log_file_name
        with open(os.path.join(log_dir(), st_file_name), "w", encoding="utf-8") as f:
            f.write(self.as_sxpr())

479

480
def mock_syntax_tree(sxpr):
481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504
    """
    Generates a tree of nodes from an S-expression.

    Example:
    >>> mock_syntax_tree("(a (b c))").as_sxpr()
    '(a\\n    (b\\n        "c"\\n    )\\n)'
    """

    def next_block(s):
        s = s.strip()
        while s[0] != ')':
            if s[0] != '(': raise ValueError('"(" expected, not ' + s[:10])
            # assert s[0] == '(', s
            level = 1
            i = 1
            while level > 0:
                if s[i] == '(':
                    level += 1
                elif s[i] == ')':
                    level -= 1
                i += 1
            yield s[:i]
            s = s[i:].strip()

505 506 507 508 509 510 511 512 513
    sxpr = sxpr.strip()
    if sxpr[0] != '(': raise ValueError('"(" expected, not ' + sxpr[:10])
    # assert sxpr[0] == '(', sxpr
    sxpr = sxpr[1:].strip()
    m = re.match('[\w:]+', sxpr)
    name, class_name = (sxpr[:m.end()].split(':') + [''])[:2]
    sxpr = sxpr[m.end():].strip()
    if sxpr[0] == '(':
        result = tuple(mock_syntax_tree(block) for block in next_block(sxpr))
514 515
    else:
        lines = []
516
        while sxpr and sxpr[0] != ')':
517
            for qm in ['"""', "'''", '"', "'"]:
518
                m = re.match(qm + r'.*?' + qm, sxpr, re.DOTALL)
519 520
                if m:
                    i = len(qm)
521 522
                    lines.append(sxpr[i:m.end() - i])
                    sxpr = sxpr[m.end():].strip()
523 524
                    break
            else:
525 526 527
                m = re.match(r'(?:(?!\)).)*', sxpr, re.DOTALL)
                lines.append(sxpr[:m.end()])
                sxpr = sxpr[m.end():]
528 529 530 531
        result = "\n".join(lines)
    return Node(MockParser(name, ':' + class_name), result)


532
TransformationFunc = Union[Callable[[Node], Any], partial]
533

534 535 536 537 538

# if __name__ == "__main__":
#     st = mock_syntax_tree("(alpha (beta (gamma i\nj\nk) (delta y)) (epsilon z))")
#     print(st.as_sxpr())
#     print(st.as_xml())