ebnf.py 24.2 KB
Newer Older
1
2
#!/usr/bin/python3

3
"""ebnf.py - EBNF -> Python-Parser compilation for DHParser
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

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
21
# import collections
22
23
import keyword
from functools import partial
24

25
26
27
28
29
try:
    import regex as re
except ImportError:
    import re

30
31
from .__init__ import __version__
from .toolkit import load_if_file, escape_re, md5, sane_parser_name
32
from .parsers import GrammarBase, mixin_comment, Forward, RE, NegativeLookahead, \
33
    Alternative, Sequence, Optional, Required, OneOrMore, ZeroOrMore, Token, CompilerBase
34
from .syntaxtree import Node, remove_enclosing_delimiters, reduce_single_child, \
di68kap's avatar
di68kap committed
35
    replace_by_single_child, TOKEN_KEYWORD, remove_expendables, remove_tokens, flatten, \
36
    forbid, assert_content, WHITESPACE_KEYWORD
37
38


Eckhart Arnold's avatar
Eckhart Arnold committed
39
40
41
42
43
__all__ = ['EBNFGrammar',
           'EBNFTransTable',
           'EBNFCompilerError',
           # 'Scanner',
           'EBNFCompiler']
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64


class EBNFGrammar(GrammarBase):
    r"""Parser for an EBNF source file, with this grammar:

    # EBNF-Grammar in EBNF

    @ comment    =  /#.*(?:\n|$)/                    # comments start with '#' and eat all chars up to and including '\n'
    @ whitespace =  /\s*/                            # whitespace includes linefeed
    @ literalws  =  right                            # trailing whitespace of literals will be ignored tacitly

    syntax     =  [~//] { definition | directive } §EOF
    definition =  symbol §"=" expression
    directive  =  "@" §symbol §"=" ( regexp | literal | list_ )

    expression =  term { "|" term }
    term       =  { factor }+
    factor     =  [flowmarker] [retrieveop] symbol !"="   # negative lookahead to be sure it's not a definition
                | [flowmarker] literal
                | [flowmarker] regexp
                | [flowmarker] group
65
                | [flowmarker] regexchain
66
67
68
69
70
71
72
73
74
                | [flowmarker] oneormore
                | repetition
                | option

    flowmarker =  "!"  | "&"  | "§" |                # '!' negative lookahead, '&' positive lookahead, '§' required
                  "-!" | "-&"                        # '-' negative lookbehind, '-&' positive lookbehind
    retrieveop =  "::" | ":"                         # '::' pop, ':' retrieve

    group      =  "(" expression §")"
75
    regexchain =  "<" expression §">"                # compiles "expression" into a singular regular expression
76
77
    oneormore  =  "{" expression "}+"
    repetition =  "{" expression §"}"
78
79
80
    option     =  "[" expression §"]"

    link       = regexp | symbol | literal           # semantic restriction: symbol must evaluate to a regexp or chain
81
82
83
84
85
86
87

    symbol     =  /(?!\d)\w+/~                       # e.g. expression, factor, parameter_list
    literal    =  /"(?:[^"]|\\")*?"/~                # e.g. "(", '+', 'while'
                | /'(?:[^']|\\')*?'/~                # whitespace following literals will be ignored tacitly.
    regexp     =  /~?\/(?:[^\/]|(?<=\\)\/)*\/~?/~    # e.g. /\w+/, ~/#.*(?:\n|$)/~
                                                     # '~' is a whitespace-marker, if present leading or trailing
                                                     # whitespace of a regular expression will be ignored tacitly.
88
    list_      =  /\w+/~ { "," /\w+/~ }              # comma separated list of symbols, e.g. BEGIN_LIST, END_LIST,
89
90
91
92
                                                     # BEGIN_QUOTE, END_QUOTE ; see CommonMark/markdown.py for an exmaple
    EOF =  !/./
    """
    expression = Forward()
93
    source_hash__ = "a410e1727fb7575e98ff8451dbf8f3bd"
94
    parser_initialization__ = "upon instatiation"
95
96
    COMMENT__ = r'#.*(?:\n|$)'
    WSP__ = mixin_comment(whitespace=r'\s*', comment=r'#.*(?:\n|$)')
97
    wspL__ = ''
98
    wspR__ = WSP__
99
    EOF = NegativeLookahead(RE('.', wR=''))
100
    list_ = Sequence(RE('\\w+'), ZeroOrMore(Sequence(Token(","), RE('\\w+'))))
101
102
103
    regexp = RE('~?/(?:[^/]|(?<=\\\\)/)*/~?')
    literal = Alternative(RE('"(?:[^"]|\\\\")*?"'), RE("'(?:[^']|\\\\')*?'"))
    symbol = RE('(?!\\d)\\w+')
104
105
    link = Alternative(regexp, symbol, literal)
    option = Sequence(Token("["), expression, Required(Token("]")))
106
107
    repetition = Sequence(Token("{"), expression, Required(Token("}")))
    oneormore = Sequence(Token("{"), expression, Token("}+"))
108
    regexchain = Sequence(Token("<"), expression, Required(Token(">")))
109
110
111
112
113
    group = Sequence(Token("("), expression, Required(Token(")")))
    retrieveop = Alternative(Token("::"), Token(":"))
    flowmarker = Alternative(Token("!"), Token("&"), Token("§"), Token("-!"), Token("-&"))
    factor = Alternative(Sequence(Optional(flowmarker), Optional(retrieveop), symbol, NegativeLookahead(Token("="))),
                         Sequence(Optional(flowmarker), literal), Sequence(Optional(flowmarker), regexp),
114
115
                         Sequence(Optional(flowmarker), group), Sequence(Optional(flowmarker), regexchain),
                         Sequence(Optional(flowmarker), oneormore), repetition, option)
116
117
118
119
    term = OneOrMore(factor)
    expression.set(Sequence(term, ZeroOrMore(Sequence(Token("|"), term))))
    directive = Sequence(Token("@"), Required(symbol), Required(Token("=")), Alternative(regexp, literal, list_))
    definition = Sequence(symbol, Required(Token("=")), expression)
120
    syntax = Sequence(Optional(RE('', wR='', wL=WSP__)), ZeroOrMore(Alternative(definition, directive)), Required(EOF))
121
122
123
    root__ = syntax


di68kap's avatar
di68kap committed
124
EBNF_ASTTransform = {
125
126
127
128
129
130
131
132
133
134
135
136
137
    # AST Transformations for EBNF-grammar
    "syntax":
        remove_expendables,
    "directive, definition":
        partial(remove_tokens, tokens={'@', '='}),
    "expression, chain":
        [replace_by_single_child, flatten,
         partial(remove_tokens, tokens={'|', '--'})],
    "term":
        [replace_by_single_child, flatten],  # supports both idioms:  "{ factor }+" and "factor { factor }"
    "factor, flowmarker, retrieveop":
        replace_by_single_child,
    "group":
138
139
140
        [remove_enclosing_delimiters, replace_by_single_child],
    "oneormore, repetition, option, regexchain":
        [reduce_single_child, remove_enclosing_delimiters],
141
    "symbol, literal, regexp":
142
143
144
        [remove_expendables, reduce_single_child],
    (TOKEN_KEYWORD, WHITESPACE_KEYWORD):
        [remove_expendables, reduce_single_child],
145
146
    "list_":
        [partial(remove_tokens, tokens={','})],
147
148
149
150
    "":
        [remove_expendables, replace_by_single_child]
}

151
152
153
154
155
156
EBNF_semantic_validation = {
    # Semantic validation on the AST
    "repetition, option, oneormore":
        [partial(forbid, child_tags=['repetition', 'option', 'oneormore']),
         partial(assert_content, regex=r'(?!§)')],
}
157

158
EBNF_ASTPipeline = [EBNF_ASTTransform, EBNF_semantic_validation]
di68kap's avatar
di68kap committed
159
160


161
162
163
164
165
166
167
168
169
170
171
172
173
174
class EBNFCompilerError(Exception):
    """Error raised by `EBNFCompiler` class. (Not compilation errors
    in the strict sense, see `CompilationError` below)"""
    pass


class EBNFCompiler(CompilerBase):
    """Generates a Parser from an abstract syntax tree of a grammar specified
    in EBNF-Notation.
    """
    COMMENT_KEYWORD = "COMMENT__"
    RESERVED_SYMBOLS = {TOKEN_KEYWORD, WHITESPACE_KEYWORD, COMMENT_KEYWORD}
    AST_ERROR = "Badly structured syntax tree. " \
                "Potentially due to erroneuos AST transformation."
175
176
177
178
    PREFIX_TABLE = {'§': 'Required',
                    '&': 'Lookahead', '!': 'NegativeLookahead',
                    '-&': 'Lookbehind', '-!': 'NegativeLookbehind',
                    '::': 'Pop', ':': 'Retrieve'}
179
180
181
    WHITESPACE = {'horizontal': r'[\t ]*',  # default: horizontal
                  'linefeed': r'[ \t]*\n?(?!\s*\n)[ \t]*',
                  'vertical': r'\s*'}
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196

    def __init__(self, grammar_name="", source_text=""):
        super(EBNFCompiler, self).__init__()
        assert grammar_name == "" or re.match('\w+\Z', grammar_name)
        self.grammar_name = grammar_name
        self.source_text = load_if_file(source_text)
        self._reset()

    def _reset(self):
        self.rules = set()
        self.symbols = set()
        self.variables = set()
        self.definition_names = []
        self.recursive = set()
        self.root = ""
197
        self.directives = {'whitespace': self.WHITESPACE['horizontal'],
198
                           'comment': '',
199
                           'literalws': ['right'],
200
201
                           'tokens': set(),     # alt. 'scanner_tokens'
                           'complement': set()}     # alt. 'retrieve_complement'
202
203
204
205
206
207
208
209
210

    def gen_scanner_skeleton(self):
        name = self.grammar_name + "Scanner"
        return "def %s(text):\n    return text\n" % name

    def gen_AST_skeleton(self):
        if not self.definition_names:
            raise EBNFCompilerError('Compiler has not been run before calling '
                                    '"gen_AST_Skeleton()"!')
di68kap's avatar
di68kap committed
211
212
213
        tt_name = self.grammar_name + '_ASTTransform'
        pl_name = self.grammar_name + '_ASTPipeline'
        transtable = [tt_name + ' = {',
214
215
216
                      '    # AST Transformations for the ' +
                      self.grammar_name + '-grammar']
        for name in self.definition_names:
di68kap's avatar
di68kap committed
217
            transtable.append('    "' + name + '": no_operation,')
218
        transtable += ['    "": no_operation', '}', '',  pl_name + ' = [%s]' % tt_name, '']
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
        return '\n'.join(transtable)

    def gen_compiler_skeleton(self):
        if not self.definition_names:
            raise EBNFCompilerError('Compiler has not been run before calling '
                                    '"gen_Compiler_Skeleton()"!')
        compiler = ['class ' + self.grammar_name + 'Compiler(CompilerBase):',
                    '    """Compiler for the abstract-syntax-tree of a ' +
                    self.grammar_name + ' source file.',
                    '    """', '',
                    '    def __init__(self, grammar_name="' +
                    self.grammar_name + '"):',
                    '        super(' + self.grammar_name +
                    'Compiler, self).__init__()',
                    "        assert re.match('\w+\Z', grammar_name)", '']
        for name in self.definition_names:
            if name == self.root:
                compiler += ['    def ' + name + '(self, node):',
                             '        return node', '']
            else:
                compiler += ['    def ' + name + '(self, node):',
                             '        pass', '']
241
        return '\n'.join(compiler)
242
243
244
245
246
247

    def gen_parser(self, definitions):
        # fix capture of variables that have been defined before usage [sic!]
        if self.variables:
            for i in range(len(definitions)):
                if definitions[i][0] in self.variables:
248
                    definitions[i] = (definitions[i][0], 'Capture(%s)' % definitions[1])
249
250

        self.definition_names = [defn[0] for defn in definitions]
Eckhart Arnold's avatar
Eckhart Arnold committed
251
252
253
254
        definitions.append(('wspR__', WHITESPACE_KEYWORD
                            if 'right' in self.directives['literalws'] else "''"))
        definitions.append(('wspL__', WHITESPACE_KEYWORD
                            if 'left' in self.directives['literalws'] else "''"))
255
256
257
258
259
260
261
262
        definitions.append((WHITESPACE_KEYWORD,
                            ("mixin_comment(whitespace="
                             "r'{whitespace}', comment=r'{comment}')").
                            format(**self.directives)))
        definitions.append((self.COMMENT_KEYWORD, "r'{comment}'".format(**self.directives)))

        # prepare parser class header and docstring and
        # add EBNF grammar to the doc string of the parser class
263
        article = 'an ' if self.grammar_name[0:1] in "AaEeIiOoUu" else 'a '  # what about 'hour', 'universe' etc.?
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
        declarations = ['class ' + self.grammar_name +
                        'Grammar(GrammarBase):',
                        'r"""Parser for ' + article + self.grammar_name +
                        ' source file' +
                        (', with this grammar:' if self.source_text else '.')]
        definitions.append(('parser_initialization__', '"upon instatiation"'))
        if self.source_text:
            definitions.append(('source_hash__',
                                '"%s"' % md5(self.source_text, __version__)))
            declarations.append('')
            declarations += [line for line in self.source_text.split('\n')]
            while declarations[-1].strip() == '':
                declarations = declarations[:-1]
        declarations.append('"""')

279
280
281
282
283
284
285
        # add default functions for complement filters of pop or retrieve operators
        for symbol in self.directives['complement']:
            declarations.append('@staticmethod\n'
                                'def complement_%s(value): \n' % symbol +
                                '    return value.replace("(", ")").replace("[", "]")'
                                '.replace("{", "}").replace(">", "<")\n')

286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
        # turn definitions into declarations in reverse order
        self.root = definitions[0][0] if definitions else ""
        definitions.reverse()
        declarations += [symbol + ' = Forward()'
                         for symbol in sorted(list(self.recursive))]
        for symbol, statement in definitions:
            if symbol in self.recursive:
                declarations += [symbol + '.set(' + statement + ')']
            else:
                declarations += [symbol + ' = ' + statement]
        for nd in self.symbols:
            if nd.result not in self.rules:
                nd.add_error("Missing production for symbol '%s'" % nd.result)
        if self.root and 'root__' not in self.symbols:
            declarations.append('root__ = ' + self.root)
        declarations.append('')
        return '\n    '.join(declarations)

    def syntax(self, node):
        self._reset()
        definitions = []

        # drop the wrapping sequence node
        if isinstance(node.parser, Sequence) and \
                isinstance(node.result[0].parser, ZeroOrMore):
            node = node.result[0]

        # compile definitions and directives and collect definitions
        for nd in node.result:
            if nd.parser.name == "definition":
                definitions.append(self.compile__(nd))
            else:
                assert nd.parser.name == "directive", nd.as_sexpr()
                self.compile__(nd)

        return self.gen_parser(definitions)

    def definition(self, node):
        rule = node.result[0].result
325
326
327
        if rule in self.rules:
            node.add_error('A rule with name "%s" has already been defined.' % rule)
        elif rule in EBNFCompiler.RESERVED_SYMBOLS:
328
329
330
331
            node.add_error('Symbol "%s" is a reserved symbol.' % rule)
        elif not sane_parser_name(rule):
            node.add_error('Illegal symbol "%s". Symbols must not start or '
                           ' end with a doube underscore "__".' % rule)
332
        elif rule in self.directives['tokens']:
333
334
335
336
337
338
339
340
341
            node.add_error('Symbol "%s" has already been defined as '
                           'a scanner token.' % rule)
        elif keyword.iskeyword(rule):
            node.add_error('Python keyword "%s" may not be used as a symbol. '
                           % rule + '(This may change in the furute.)')
        try:
            self.rules.add(rule)
            defn = self.compile__(node.result[1])
            if rule in self.variables:
342
                defn = 'Capture(%s)' % defn
343
344
345
346
347
                self.variables.remove(rule)
        except TypeError as error:
            errmsg = EBNFCompiler.AST_ERROR + " (" + str(error) + ")\n" + node.as_sexpr()
            node.add_error(errmsg)
            rule, defn = rule + ':error', '"' + errmsg + '"'
Eckhart Arnold's avatar
Eckhart Arnold committed
348
        return rule, defn
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365

    @staticmethod
    def _check_rx(node, rx):
        """Checks whether the string `rx` represents a valid regular
        expression. Makes sure that multiline regular expressions are
        prepended by the multiline-flag. Returns the regular expression string.
        """
        rx = rx if rx.find('\n') < 0 or rx[0:4] == '(?x)' else '(?x)' + rx
        try:
            re.compile(rx)
        except Exception as re_error:
            node.add_error("malformed regular expression %s: %s" %
                           (repr(rx), str(re_error)))
        return rx

    def directive(self, node):
        key = node.result[0].result.lower()
366
        assert key not in self.directives['tokens']
367
        if key in {'comment', 'whitespace'}:
368
369
            if node.result[1].parser.name == "list_":
                if len(node.result[1].result) != 1:
Eckhart Arnold's avatar
Eckhart Arnold committed
370
                    node.add_error('Directive "%s" must have one, but not %i values.' %
371
372
                                   (key, len(node.result[1])))
                value = self.compile__(node.result[1]).pop()
373
374
                if key == 'whitespace' and value in EBNFCompiler.WHITESPACE:
                    value = EBNFCompiler.WHITESPACE[value]  # replace whitespace-name by regex
375
                else:
376
                    node.add_error('Value "%s" not allowed for directive "%s".' % (value, key))
377
            else:
378
379
380
381
382
383
384
385
                value = node.result[1].result.strip("~")
                if value != node.result[1].result:
                    node.add_error("Whitespace marker '~' not allowed in definition of "
                                   "%s regular expression." % key)
                if value[0] + value[-1] in {'""', "''"}:
                    value = escape_re(value[1:-1])
                elif value[0] + value[-1] == '//':
                    value = self._check_rx(node, value[1:-1])
386
387
388
                if key == 'whitespace' and not re.match(value, ''):
                    node.add_error("Implicit whitespace should always match the empty string, "
                                   "/%s/ does not." % value)
389
            self.directives[key] = value
390

391
392
393
        elif key == 'literalws':
            value = {item.lower() for item in self.compile__(node.result[1])}
            if (len(value - {'left', 'right', 'both', 'none'}) > 0
Eckhart Arnold's avatar
Eckhart Arnold committed
394
                    or ('none' in value and len(value) > 1)):
395
396
397
398
399
400
401
                node.add_error('Directive "literalws" allows the values '
                               '`left`, `right`, `both` or `none`, '
                               'but not `%s`' % ", ".join(value))
            ws = {'left', 'right'} if 'both' in value \
                else {} if 'none' in value else value
            self.directives[key] = list(ws)

402
        elif key in {'tokens', 'scanner_tokens'}:
403
            self.directives['tokens'] |= self.compile__(node.result[1])
404
405
406
407

        elif key in {'complement', 'retrieve_complement'}:
            self.directives['complement'] |= self.compile__(node.result[1])

408
409
410
        else:
            node.add_error('Unknown directive %s ! (Known ones are %s .)' %
                           (key,
411
                            ', '.join(list(self.directives.keys()))))
412
413
        return ""

414
    def non_terminal(self, node, parser_class, custom_args=[]):
415
416
417
        """Compiles any non-terminal, where `parser_class` indicates the Parser class
        name for the particular non-terminal.
        """
418
        arguments = [self.compile__(r) for r in node.result] + custom_args
419
420
421
422
423
424
425
426
427
428
429
430
431
        return parser_class + '(' + ', '.join(arguments) + ')'

    def expression(self, node):
        return self.non_terminal(node, 'Alternative')

    def term(self, node):
        return self.non_terminal(node, 'Sequence')

    def factor(self, node):
        assert isinstance(node.parser, Sequence), node.as_sexpr()  # these assert statements can be removed
        assert node.children
        assert len(node.result) >= 2, node.as_sexpr()
        prefix = node.result[0].result
432
        custom_args = []
433
434
435

        if prefix in {'::', ':'}:
            assert len(node.result) == 2
436
437
            arg = node.result[-1]
            argstr = str(arg)
438
            if arg.parser.name != 'symbol':
Eckhart Arnold's avatar
Eckhart Arnold committed
439
                node.add_error(('Retrieve Operator "%s" requires a symbol, '
440
441
                                'and not a %s.') % (prefix, str(arg.parser)))
                return str(arg.result)
442
443
            if str(arg) in self.directives['complement']:
                custom_args = ['complement=%s_complement' % str(arg)]
444
445
            self.variables.add(arg.result)

446
        elif len(node.result) > 2:
447
448
            # shift = (Node(node.parser, node.result[1].result),)
            # node.result[1].result = shift + node.result[2:]
449
            node.result[1].result = (Node(node.result[1].parser, node.result[1].result),) \
450
451
452
453
454
                                    + node.result[2:]
            node.result[1].parser = node.parser
            node.result = (node.result[0], node.result[1])

        node.result = node.result[1:]
455
456
457
458
459
        try:
            parser_class = self.PREFIX_TABLE[prefix]
            return self.non_terminal(node, parser_class, custom_args)
        except KeyError:
            node.add_error('Unknown prefix "%s".' % prefix)
460
461
462
463
464
465
466
467
468
469

    def option(self, node):
        return self.non_terminal(node, 'Optional')

    def repetition(self, node):
        return self.non_terminal(node, 'ZeroOrMore')

    def oneormore(self, node):
        return self.non_terminal(node, 'OneOrMore')

470
471
472
    def regexchain(self, node):
        raise EBNFCompilerError("Not yet implemented!")

473
474
475
476
477
    def group(self, node):
        raise EBNFCompilerError("Group nodes should have been eliminated by "
                                "AST transformation!")

    def symbol(self, node):
478
        if node.result in self.directives['tokens']:
479
480
481
482
483
484
485
486
            return 'ScannerToken("' + node.result + '")'
        else:
            self.symbols.add(node)
            if node.result in self.rules:
                self.recursive.add(node.result)
            return node.result

    def literal(self, node):
487
        return 'Token(' + node.result.replace('\\', r'\\') + ')'  # return 'Token(' + ', '.join([node.result]) + ')' ?
488
489
490
491
492
493
494
495
496
497
498

    def regexp(self, node):
        rx = node.result
        name = []
        if rx[:2] == '~/':
            if not 'left' in self.directives['literalws']:
                name = ['wL=' + WHITESPACE_KEYWORD] + name
            rx = rx[1:]
        elif 'left' in self.directives['literalws']:
            name = ["wL=''"] + name
        if rx[-2:] == '/~':
Eckhart Arnold's avatar
Eckhart Arnold committed
499
            if 'right' not in self.directives['literalws']:
500
501
502
503
504
505
506
507
508
509
510
511
512
513
                name = ['wR=' + WHITESPACE_KEYWORD] + name
            rx = rx[:-1]
        elif 'right' in self.directives['literalws']:
            name = ["wR=''"] + name
        try:
            arg = repr(self._check_rx(node, rx[1:-1].replace(r'\/', '/')))
        except AttributeError as error:
            errmsg = EBNFCompiler.AST_ERROR + " (" + str(error) + ")\n" + \
                     node.as_sexpr()
            node.add_error(errmsg)
            return '"' + errmsg + '"'
        return 'RE(' + ', '.join([arg] + name) + ')'

    def list_(self, node):
514
515
        assert node.children
        return set(item.result.strip() for item in node.result)
516
517


518
519
520
def grammar_changed(grammar_class, grammar_source):
    """Returns ``True`` if ``grammar_class`` does not reflect the latest
    changes of ``grammar_source``
521
522
523
524

    Parameters:
        grammar_class:  the parser class representing the grammar
            or the file name of a compiler suite containing the grammar
525
526
527
        grammar_source:  File name or string representation of the
            EBNF code of the grammar
            
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
    Returns (bool):
        True, if the source text of the grammar is different from the
        source from which the grammar class was generated
    """
    grammar = load_if_file(grammar_source)
    chksum = md5(grammar, __version__)
    if isinstance(grammar_class, str):
        # grammar_class = load_compiler_suite(grammar_class)[1]
        with open(grammar_class, 'r', encoding='utf8') as f:
            pycode = f.read()
        m = re.search('class \w*\(GrammarBase\)', pycode)
        if m:
            m = re.search('    source_hash__ *= *"([a-z0-9]*)"',
                          pycode[m.span()[1]:])
            return not (m and m.groups() and m.groups()[-1] == chksum)
        else:
            return True
    else:
        return chksum != grammar_class.source_hash__