10.12., 9:00 - 11:00: Due to updates GitLab may be unavailable for some minutes between 09:00 and 11:00.

syntaxtree.py 19.5 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.
"""

Eckhart Arnold's avatar
Eckhart Arnold committed
19
import collections.abc
20
import copy
21
import os
22 23
from functools import partial

24
from DHParser.toolkit import is_logging, log_dir, identity, re, typing
25
from DHParser.error import Error, linebreaks, line_col
26 27 28 29
from DHParser.stringview import StringView

from typing import AbstractSet, Any, ByteString, Callable, cast, Container, Dict, \
        Iterator, Iterable, List, NamedTuple, Sequence, Union, Text, Tuple, Hashable
30

31 32 33 34 35 36 37
__all__ = ('ParserBase',
           'WHITESPACE_PTYPE',
           'TOKEN_PTYPE',
           'MockParser',
           'ZombieParser',
           'ZOMBIE_PARSER',
           'Node',
38 39
           'mock_syntax_tree',
           'TransformationFunc')
Eckhart Arnold's avatar
Eckhart Arnold committed
40 41


42 43 44 45 46 47 48 49 50 51 52 53 54 55
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


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()):
56
        self._name = name  # type: str
57 58 59 60 61 62 63 64
        self._ptype = ':' + self.__class__.__name__  # type: str

    def __repr__(self):
        return self.name + self.ptype

    def __str__(self):
        return self.name + (' = ' if self.name else '') + repr(self)

65 66 67 68
    @property
    def name(self):
        return self._name

69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
    @property
    def ptype(self) -> str:
        return self._ptype

    @property
    def repr(self) -> str:
        return self.name if self.name else repr(self)


WHITESPACE_PTYPE = ':Whitespace'
TOKEN_PTYPE = ':Token'


class MockParser(ParserBase):
    """
    MockParser objects can be used to reconstruct syntax trees from a
    serialized form like S-expressions or XML. Mock objects can mimic
    different parser types by assigning them a ptype on initialization.

    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.
    """
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
94
        super().__init__(name)
95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136
        self._ptype = ptype or ':' + self.__class__.__name__


class ZombieParser(MockParser):
    """
    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):
        super(ZombieParser, self).__init__("__ZOMBIE__")
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
        self.__class__.alive = True

    def __copy__(self):
        return self

    def __deepcopy__(self, memo):
        return self

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


ZOMBIE_PARSER = ZombieParser()


#######################################################################
#
# syntaxtree nodes
#
#######################################################################


137
ChildrenType = Tuple['Node', ...]
138 139
StrictResultType = Union[ChildrenType, StringView, str]
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
140 141


Eckhart Arnold's avatar
Eckhart Arnold committed
142 143
def flatten_sxpr(sxpr: str) -> str:
    """Returns S-expression `sxpr` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
144 145 146
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
147
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
148 149 150 151 152
    '(a (b c))'
    """
    return re.sub('\s(?=\))', '', re.sub('\s+', ' ', sxpr)).strip()


Eckhart Arnold's avatar
Eckhart Arnold committed
153
class Node(collections.abc.Sized):
154 155 156 157 158 159 160 161 162 163 164
    """
    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!
165 166 167 168 169 170
        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, ...)``.
171 172
        errors (list):  A list of parser- or compiler-errors:
            tuple(position, string) attached to this node
173 174 175
        error_flag (int):  0 if no error occurred in either the node
            itself or any of its descendants. Otherwise contains the
            highest warning or error level or all errors that occurred.
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195
        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.
    """
196

197 198
    __slots__ = ['_result', 'children', '_errors', '_len', '_pos', 'parser', 'error_flag']

199

Eckhart Arnold's avatar
Eckhart Arnold committed
200
    def __init__(self, parser, result: ResultType) -> None:
201 202 203
        """Initializes the ``Node``-object with the ``Parser``-Instance
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
204
        # self._result = ''  # type: StrictResultType
205
        # self.children = ()  # type: ChildrenType
206
        self.error_flag = 0   # type: bool
207
        self._errors = []  # type: List[Error]
208
        self.result = result
Eckhart Arnold's avatar
Eckhart Arnold committed
209
        self._len = len(self._result) if not self.children else \
210
            sum(child._len for child in self.children)  # type: int
Eckhart Arnold's avatar
Eckhart Arnold committed
211
        # self.pos: int  = 0  # continuous updating of pos values wastes a lot of time
212
        self._pos = -1  # type: int
213 214
        self.parser = parser or ZOMBIE_PARSER

Eckhart Arnold's avatar
Eckhart Arnold committed
215

216 217
    def __str__(self):
        if self.children:
218
            return "".join(str(child) for child in self.children)
219 220 221
        elif isinstance(self.result, StringView):
            self.result = str(self.result)
        return self.result
222

Eckhart Arnold's avatar
Eckhart Arnold committed
223

224 225 226 227 228 229
    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)
230

Eckhart Arnold's avatar
Eckhart Arnold committed
231

232 233 234 235 236 237 238 239 240
    def __len__(self):
        return self._len


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


241
    def __eq__(self, other):
242 243
        # return str(self.parser) == str(other.parser) and self.result == other.result
        return self.tag_name == other.tag_name and self.result == other.result
244

Eckhart Arnold's avatar
Eckhart Arnold committed
245

246
    def __hash__(self):
247
        return hash(self.tag_name)
248

Eckhart Arnold's avatar
Eckhart Arnold committed
249

250 251 252 253 254 255
    def __deepcopy__(self, memodict={}):
        result = copy.deepcopy(self.result)
        other = Node(self.parser, result)
        other._pos = self._pos
        return other

Eckhart Arnold's avatar
Eckhart Arnold committed
256

257
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
258
    def tag_name(self) -> str:
259
        return self.parser.name or self.parser.ptype
260

Eckhart Arnold's avatar
Eckhart Arnold committed
261

262
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
263
    def result(self) -> StrictResultType:
264 265 266
        return self._result

    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
267
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
268
        # # made obsolete by static type checking with mypy
269 270 271
        # 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
272 273
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
274 275 276
        self._result = (result,) if isinstance(result, Node) else result or ''  # type: StrictResultType
        # self._result = (result,) if isinstance(result, Node) else str(result) \
        #     if isinstance(result, StringView) else result or ''  # type: StrictResultType
277
        self.children = cast(ChildrenType, self._result) \
Eckhart Arnold's avatar
Eckhart Arnold committed
278
            if isinstance(self._result, tuple) else cast(ChildrenType, ())  # type: ChildrenType
279 280 281
        if self.children:
            self.error_flag = max(self.error_flag,
                                  max(child.error_flag for child in self.children))  # type: bool
282 283

    @property
284
    def pos(self) -> int:
285 286 287 288
        assert self._pos >= 0, "position value not initialized!"
        return self._pos

    @pos.setter
289
    def pos(self, pos: int):
290 291
        self._pos = pos
        offset = 0
292
        # recursively adjust pos-values of all children
293 294
        for child in self.children:
            child.pos = pos + offset
295
            offset += len(child)
296 297 298
        # add pos-values to Error-objects
        for err in self._errors:
            err.pos = pos
299

Eckhart Arnold's avatar
Eckhart Arnold committed
300

301
    @property
302
    def errors(self) -> List[Error]:
303 304
        return self._errors.copy()

305
    def add_error(self, message: str, level: int = Error.ERROR, code: Hashable = 0) -> 'Node':
Eckhart Arnold's avatar
Eckhart Arnold committed
306
        self._errors.append(Error(message, level, code))
307
        self.error_flag = max(self.error_flag, self._errors[-1].level)
Eckhart Arnold's avatar
Eckhart Arnold committed
308 309
        return self

310 311
    def collect_errors(self, document: Union[StringView, str] = '', clear_errors=False) -> List[
        Error]:
Eckhart Arnold's avatar
Eckhart Arnold committed
312
        """
313
        Recursively adds line- and column-numbers to all error objects.
Eckhart Arnold's avatar
Eckhart Arnold committed
314 315 316 317
        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.
        """
318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339
        if self.error_flag:
            lbreaks = linebreaks(document) if document else []
            return self._collect_errors(lbreaks, clear_errors)
        else:
            return []

    def _collect_errors(self, lbreaks: List[int] = [], clear_errors=False) -> List[Error]:
        if self.error_flag:
            errors = self.errors
            if lbreaks:
                for err in errors:
                    err.pos = self.pos
                    err.line, err.column = line_col(lbreaks, err.pos)
            if clear_errors:
                self._errors = []
                self.error_flag = 0
            if self.children:
                for child in self.children:
                    errors.extend(child._collect_errors(lbreaks, clear_errors))
            return errors
        else:
            return []
340

Eckhart Arnold's avatar
Eckhart Arnold committed
341

342
    def _tree_repr(self, tab, openF, closeF, dataF=identity, density=0) -> str:
343 344 345 346 347 348 349 350
        """
        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.

351
        Args:
352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367
            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:
368
            return head.rstrip() + tail.lstrip()
369

370
        D = None if density & 2 else ''
371 372 373

        if self.children:
            content = []
374
            for child in self.children:
375
                subtree = child._tree_repr(tab, openF, closeF, dataF, density).split('\n')
376
                content.append('\n'.join((tab + s) for s in subtree))
377
            return head + '\n'.join(content) + tail.lstrip(D)
378

379
        res = cast(str, self.result)  # safe, because if there are no children, result is a string
380 381 382 383
        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()
384
        else:
385
            return head + '\n'.join([tab + dataF(s) for s in res.split('\n')]) + tail.lstrip(D)
386

Eckhart Arnold's avatar
Eckhart Arnold committed
387

388
    def as_sxpr(self, src: str=None, compact: bool=False) -> str:
389 390 391
        """
        Returns content as S-expression, i.e. in lisp-like form.

392
        Args:
393 394 395
            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.
Eckhart Arnold's avatar
Eckhart Arnold committed
396 397 398
            compact:  If True a compact representation is returned where
                brackets are ommited and only the indentation indicates the
                tree structure.
399 400
        """

401 402
        lB, rB, D = ('', '', 1) if compact else ('(', '\n)', 0)

403
        def opening(node) -> str:
404
            s = lB + node.tag_name
405 406
            # s += " '(pos %i)" % node.pos
            if src:
407
                s += " '(pos %i " % node.pos  # + " %i %i)" % line_col(src, node.pos)
408 409 410
            if node.errors:
                s += " '(err '(%s))" % ' '.join(str(err).replace('"', r'\"')
                                                for err in node.errors)
411
            return s + '\n'
412 413 414 415 416 417

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

418
        return self._tree_repr('    ', opening, lambda node: rB, pretty, density=D)
419

Eckhart Arnold's avatar
Eckhart Arnold committed
420

421
    def as_xml(self, src: str=None) -> str:
422 423 424
        """
        Returns content as XML-tree.

425
        Args:
426 427 428 429 430
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
        """

431
        def opening(node) -> str:
432 433 434 435 436 437
            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)
438
            return s + ">\n"
439 440

        def closing(node):
441
            return '\n</' + node.tag_name + '>'
442

443
        return self._tree_repr('    ', opening, closing, density=1)
444

Eckhart Arnold's avatar
Eckhart Arnold committed
445

Eckhart Arnold's avatar
Eckhart Arnold committed
446 447 448
    def structure(self) -> str:
        """Return structure (and content) as S-expression on a single line
        without any line breaks."""
Eckhart Arnold's avatar
Eckhart Arnold committed
449 450
        return flatten_sxpr(self.as_sxpr())

451

Eckhart Arnold's avatar
Eckhart Arnold committed
452
    def content(self) -> str:
453
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
454 455
        Returns content as string, inserting error messages where
        errors occurred.
456
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
457 458
        s = "".join(child.content() for child in self.children) if self.children \
            else str(self.result)
459 460
        return (' <<< Error on "%s" | %s >>> '
                % (s, '; '.join(e.message for e in self._errors))) if self._errors else s
461

Eckhart Arnold's avatar
Eckhart Arnold committed
462

463
    def find(self, match_function: Callable) -> Iterator['Node']:
464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483
        """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

Eckhart Arnold's avatar
Eckhart Arnold committed
484

485 486 487 488 489
    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


Eckhart Arnold's avatar
Eckhart Arnold committed
490
    def log(self, log_file_name):
491
        if is_logging():
Eckhart Arnold's avatar
Eckhart Arnold committed
492 493 494 495
            path = os.path.join(log_dir(), log_file_name)
            if os.path.exists(path):
                print('WARNING: Log-file "%s" already exists and will be overwritten!' % path)
            with open(path, "w", encoding="utf-8") as f:
496
                f.write(self.as_sxpr())
Eckhart Arnold's avatar
Eckhart Arnold committed
497

498

499
def mock_syntax_tree(sxpr):
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523
    """
    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()

524 525 526 527 528 529 530 531 532
    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))
533 534
    else:
        lines = []
535
        while sxpr and sxpr[0] != ')':
536
            for qm in ['"""', "'''", '"', "'"]:
537
                m = re.match(qm + r'.*?' + qm, sxpr, re.DOTALL)
538 539
                if m:
                    i = len(qm)
540 541
                    lines.append(sxpr[i:m.end() - i])
                    sxpr = sxpr[m.end():].strip()
542 543
                    break
            else:
544 545 546
                m = re.match(r'(?:(?!\)).)*', sxpr, re.DOTALL)
                lines.append(sxpr[:m.end()])
                sxpr = sxpr[m.end():]
547 548 549 550
        result = "\n".join(lines)
    return Node(MockParser(name, ':' + class_name), result)


551
TransformationFunc = Union[Callable[[Node], Any], partial]
552

553 554 555 556 557

# 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())