Coverage for /pythoncovmergedfiles/medio/medio/usr/local/lib/python3.8/site-packages/lark/lark.py: 51%

330 statements  

« prev     ^ index     » next       coverage.py v7.4.1, created at 2024-02-14 06:19 +0000

1from abc import ABC, abstractmethod 

2import getpass 

3import sys, os, pickle 

4import tempfile 

5import types 

6import re 

7from typing import ( 

8 TypeVar, Type, List, Dict, Iterator, Callable, Union, Optional, Sequence, 

9 Tuple, Iterable, IO, Any, TYPE_CHECKING, Collection 

10) 

11if TYPE_CHECKING: 

12 from .parsers.lalr_interactive_parser import InteractiveParser 

13 from .tree import ParseTree 

14 from .visitors import Transformer 

15 if sys.version_info >= (3, 8): 

16 from typing import Literal 

17 else: 

18 from typing_extensions import Literal 

19 from .parser_frontends import ParsingFrontend 

20 

21from .exceptions import ConfigurationError, assert_config, UnexpectedInput 

22from .utils import Serialize, SerializeMemoizer, FS, isascii, logger 

23from .load_grammar import load_grammar, FromPackageLoader, Grammar, verify_used_files, PackageResource, sha256_digest 

24from .tree import Tree 

25from .common import LexerConf, ParserConf, _ParserArgType, _LexerArgType 

26 

27from .lexer import Lexer, BasicLexer, TerminalDef, LexerThread, Token 

28from .parse_tree_builder import ParseTreeBuilder 

29from .parser_frontends import _validate_frontend_args, _get_lexer_callbacks, _deserialize_parsing_frontend, _construct_parsing_frontend 

30from .grammar import Rule 

31 

32 

33try: 

34 import regex 

35 _has_regex = True 

36except ImportError: 

37 _has_regex = False 

38 

39 

40###{standalone 

41 

42 

43class PostLex(ABC): 

44 @abstractmethod 

45 def process(self, stream: Iterator[Token]) -> Iterator[Token]: 

46 return stream 

47 

48 always_accept: Iterable[str] = () 

49 

50class LarkOptions(Serialize): 

51 """Specifies the options for Lark 

52 

53 """ 

54 

55 start: List[str] 

56 debug: bool 

57 strict: bool 

58 transformer: 'Optional[Transformer]' 

59 propagate_positions: Union[bool, str] 

60 maybe_placeholders: bool 

61 cache: Union[bool, str] 

62 regex: bool 

63 g_regex_flags: int 

64 keep_all_tokens: bool 

65 tree_class: Optional[Callable[[str, List], Any]] 

66 parser: _ParserArgType 

67 lexer: _LexerArgType 

68 ambiguity: 'Literal["auto", "resolve", "explicit", "forest"]' 

69 postlex: Optional[PostLex] 

70 priority: 'Optional[Literal["auto", "normal", "invert"]]' 

71 lexer_callbacks: Dict[str, Callable[[Token], Token]] 

72 use_bytes: bool 

73 ordered_sets: bool 

74 edit_terminals: Optional[Callable[[TerminalDef], TerminalDef]] 

75 import_paths: 'List[Union[str, Callable[[Union[None, str, PackageResource], str], Tuple[str, str]]]]' 

76 source_path: Optional[str] 

77 

78 OPTIONS_DOC = r""" 

79 **=== General Options ===** 

80 

81 start 

82 The start symbol. Either a string, or a list of strings for multiple possible starts (Default: "start") 

83 debug 

84 Display debug information and extra warnings. Use only when debugging (Default: ``False``) 

85 When used with Earley, it generates a forest graph as "sppf.png", if 'dot' is installed. 

86 strict 

87 Throw an exception on any potential ambiguity, including shift/reduce conflicts, and regex collisions. 

88 transformer 

89 Applies the transformer to every parse tree (equivalent to applying it after the parse, but faster) 

90 propagate_positions 

91 Propagates positional attributes into the 'meta' attribute of all tree branches. 

92 Sets attributes: (line, column, end_line, end_column, start_pos, end_pos, 

93 container_line, container_column, container_end_line, container_end_column) 

94 Accepts ``False``, ``True``, or a callable, which will filter which nodes to ignore when propagating. 

95 maybe_placeholders 

96 When ``True``, the ``[]`` operator returns ``None`` when not matched. 

97 When ``False``, ``[]`` behaves like the ``?`` operator, and returns no value at all. 

98 (default= ``True``) 

99 cache 

100 Cache the results of the Lark grammar analysis, for x2 to x3 faster loading. LALR only for now. 

101 

102 - When ``False``, does nothing (default) 

103 - When ``True``, caches to a temporary file in the local directory 

104 - When given a string, caches to the path pointed by the string 

105 regex 

106 When True, uses the ``regex`` module instead of the stdlib ``re``. 

107 g_regex_flags 

108 Flags that are applied to all terminals (both regex and strings) 

109 keep_all_tokens 

110 Prevent the tree builder from automagically removing "punctuation" tokens (Default: ``False``) 

111 tree_class 

112 Lark will produce trees comprised of instances of this class instead of the default ``lark.Tree``. 

113 

114 **=== Algorithm Options ===** 

115 

116 parser 

117 Decides which parser engine to use. Accepts "earley" or "lalr". (Default: "earley"). 

118 (there is also a "cyk" option for legacy) 

119 lexer 

120 Decides whether or not to use a lexer stage 

121 

122 - "auto" (default): Choose for me based on the parser 

123 - "basic": Use a basic lexer 

124 - "contextual": Stronger lexer (only works with parser="lalr") 

125 - "dynamic": Flexible and powerful (only with parser="earley") 

126 - "dynamic_complete": Same as dynamic, but tries *every* variation of tokenizing possible. 

127 ambiguity 

128 Decides how to handle ambiguity in the parse. Only relevant if parser="earley" 

129 

130 - "resolve": The parser will automatically choose the simplest derivation 

131 (it chooses consistently: greedy for tokens, non-greedy for rules) 

132 - "explicit": The parser will return all derivations wrapped in "_ambig" tree nodes (i.e. a forest). 

133 - "forest": The parser will return the root of the shared packed parse forest. 

134 

135 **=== Misc. / Domain Specific Options ===** 

136 

137 postlex 

138 Lexer post-processing (Default: ``None``) Only works with the basic and contextual lexers. 

139 priority 

140 How priorities should be evaluated - "auto", ``None``, "normal", "invert" (Default: "auto") 

141 lexer_callbacks 

142 Dictionary of callbacks for the lexer. May alter tokens during lexing. Use with caution. 

143 use_bytes 

144 Accept an input of type ``bytes`` instead of ``str``. 

145 ordered_sets 

146 Should Earley use ordered-sets to achieve stable output (~10% slower than regular sets. Default: True) 

147 edit_terminals 

148 A callback for editing the terminals before parse. 

149 import_paths 

150 A List of either paths or loader functions to specify from where grammars are imported 

151 source_path 

152 Override the source of from where the grammar was loaded. Useful for relative imports and unconventional grammar loading 

153 **=== End of Options ===** 

154 """ 

155 if __doc__: 

156 __doc__ += OPTIONS_DOC 

157 

158 

159 # Adding a new option needs to be done in multiple places: 

160 # - In the dictionary below. This is the primary truth of which options `Lark.__init__` accepts 

161 # - In the docstring above. It is used both for the docstring of `LarkOptions` and `Lark`, and in readthedocs 

162 # - As an attribute of `LarkOptions` above 

163 # - Potentially in `_LOAD_ALLOWED_OPTIONS` below this class, when the option doesn't change how the grammar is loaded 

164 # - Potentially in `lark.tools.__init__`, if it makes sense, and it can easily be passed as a cmd argument 

165 _defaults: Dict[str, Any] = { 

166 'debug': False, 

167 'strict': False, 

168 'keep_all_tokens': False, 

169 'tree_class': None, 

170 'cache': False, 

171 'postlex': None, 

172 'parser': 'earley', 

173 'lexer': 'auto', 

174 'transformer': None, 

175 'start': 'start', 

176 'priority': 'auto', 

177 'ambiguity': 'auto', 

178 'regex': False, 

179 'propagate_positions': False, 

180 'lexer_callbacks': {}, 

181 'maybe_placeholders': True, 

182 'edit_terminals': None, 

183 'g_regex_flags': 0, 

184 'use_bytes': False, 

185 'ordered_sets': True, 

186 'import_paths': [], 

187 'source_path': None, 

188 '_plugins': {}, 

189 } 

190 

191 def __init__(self, options_dict: Dict[str, Any]) -> None: 

192 o = dict(options_dict) 

193 

194 options = {} 

195 for name, default in self._defaults.items(): 

196 if name in o: 

197 value = o.pop(name) 

198 if isinstance(default, bool) and name not in ('cache', 'use_bytes', 'propagate_positions'): 

199 value = bool(value) 

200 else: 

201 value = default 

202 

203 options[name] = value 

204 

205 if isinstance(options['start'], str): 

206 options['start'] = [options['start']] 

207 

208 self.__dict__['options'] = options 

209 

210 

211 assert_config(self.parser, ('earley', 'lalr', 'cyk', None)) 

212 

213 if self.parser == 'earley' and self.transformer: 

214 raise ConfigurationError('Cannot specify an embedded transformer when using the Earley algorithm. ' 

215 'Please use your transformer on the resulting parse tree, or use a different algorithm (i.e. LALR)') 

216 

217 if o: 

218 raise ConfigurationError("Unknown options: %s" % o.keys()) 

219 

220 def __getattr__(self, name: str) -> Any: 

221 try: 

222 return self.__dict__['options'][name] 

223 except KeyError as e: 

224 raise AttributeError(e) 

225 

226 def __setattr__(self, name: str, value: str) -> None: 

227 assert_config(name, self.options.keys(), "%r isn't a valid option. Expected one of: %s") 

228 self.options[name] = value 

229 

230 def serialize(self, memo = None) -> Dict[str, Any]: 

231 return self.options 

232 

233 @classmethod 

234 def deserialize(cls, data: Dict[str, Any], memo: Dict[int, Union[TerminalDef, Rule]]) -> "LarkOptions": 

235 return cls(data) 

236 

237 

238# Options that can be passed to the Lark parser, even when it was loaded from cache/standalone. 

239# These options are only used outside of `load_grammar`. 

240_LOAD_ALLOWED_OPTIONS = {'postlex', 'transformer', 'lexer_callbacks', 'use_bytes', 'debug', 'g_regex_flags', 'regex', 'propagate_positions', 'tree_class', '_plugins'} 

241 

242_VALID_PRIORITY_OPTIONS = ('auto', 'normal', 'invert', None) 

243_VALID_AMBIGUITY_OPTIONS = ('auto', 'resolve', 'explicit', 'forest') 

244 

245 

246_T = TypeVar('_T', bound="Lark") 

247 

248class Lark(Serialize): 

249 """Main interface for the library. 

250 

251 It's mostly a thin wrapper for the many different parsers, and for the tree constructor. 

252 

253 Parameters: 

254 grammar: a string or file-object containing the grammar spec (using Lark's ebnf syntax) 

255 options: a dictionary controlling various aspects of Lark. 

256 

257 Example: 

258 >>> Lark(r'''start: "foo" ''') 

259 Lark(...) 

260 """ 

261 

262 source_path: str 

263 source_grammar: str 

264 grammar: 'Grammar' 

265 options: LarkOptions 

266 lexer: Lexer 

267 parser: 'ParsingFrontend' 

268 terminals: Collection[TerminalDef] 

269 

270 def __init__(self, grammar: 'Union[Grammar, str, IO[str]]', **options) -> None: 

271 self.options = LarkOptions(options) 

272 re_module: types.ModuleType 

273 

274 # Set regex or re module 

275 use_regex = self.options.regex 

276 if use_regex: 

277 if _has_regex: 

278 re_module = regex 

279 else: 

280 raise ImportError('`regex` module must be installed if calling `Lark(regex=True)`.') 

281 else: 

282 re_module = re 

283 

284 # Some, but not all file-like objects have a 'name' attribute 

285 if self.options.source_path is None: 

286 try: 

287 self.source_path = grammar.name # type: ignore[union-attr] 

288 except AttributeError: 

289 self.source_path = '<string>' 

290 else: 

291 self.source_path = self.options.source_path 

292 

293 # Drain file-like objects to get their contents 

294 try: 

295 read = grammar.read # type: ignore[union-attr] 

296 except AttributeError: 

297 pass 

298 else: 

299 grammar = read() 

300 

301 cache_fn = None 

302 cache_sha256 = None 

303 if isinstance(grammar, str): 

304 self.source_grammar = grammar 

305 if self.options.use_bytes: 

306 if not isascii(grammar): 

307 raise ConfigurationError("Grammar must be ascii only, when use_bytes=True") 

308 

309 if self.options.cache: 

310 if self.options.parser != 'lalr': 

311 raise ConfigurationError("cache only works with parser='lalr' for now") 

312 

313 unhashable = ('transformer', 'postlex', 'lexer_callbacks', 'edit_terminals', '_plugins') 

314 options_str = ''.join(k+str(v) for k, v in options.items() if k not in unhashable) 

315 from . import __version__ 

316 s = grammar + options_str + __version__ + str(sys.version_info[:2]) 

317 cache_sha256 = sha256_digest(s) 

318 

319 if isinstance(self.options.cache, str): 

320 cache_fn = self.options.cache 

321 else: 

322 if self.options.cache is not True: 

323 raise ConfigurationError("cache argument must be bool or str") 

324 

325 try: 

326 username = getpass.getuser() 

327 except Exception: 

328 # The exception raised may be ImportError or OSError in 

329 # the future. For the cache, we don't care about the 

330 # specific reason - we just want a username. 

331 username = "unknown" 

332 

333 cache_fn = tempfile.gettempdir() + "/.lark_cache_%s_%s_%s_%s.tmp" % (username, cache_sha256, *sys.version_info[:2]) 

334 

335 old_options = self.options 

336 try: 

337 with FS.open(cache_fn, 'rb') as f: 

338 logger.debug('Loading grammar from cache: %s', cache_fn) 

339 # Remove options that aren't relevant for loading from cache 

340 for name in (set(options) - _LOAD_ALLOWED_OPTIONS): 

341 del options[name] 

342 file_sha256 = f.readline().rstrip(b'\n') 

343 cached_used_files = pickle.load(f) 

344 if file_sha256 == cache_sha256.encode('utf8') and verify_used_files(cached_used_files): 

345 cached_parser_data = pickle.load(f) 

346 self._load(cached_parser_data, **options) 

347 return 

348 except FileNotFoundError: 

349 # The cache file doesn't exist; parse and compose the grammar as normal 

350 pass 

351 except Exception: # We should probably narrow done which errors we catch here. 

352 logger.exception("Failed to load Lark from cache: %r. We will try to carry on.", cache_fn) 

353 

354 # In theory, the Lark instance might have been messed up by the call to `_load`. 

355 # In practice the only relevant thing that might have been overwritten should be `options` 

356 self.options = old_options 

357 

358 

359 # Parse the grammar file and compose the grammars 

360 self.grammar, used_files = load_grammar(grammar, self.source_path, self.options.import_paths, self.options.keep_all_tokens) 

361 else: 

362 assert isinstance(grammar, Grammar) 

363 self.grammar = grammar 

364 

365 

366 if self.options.lexer == 'auto': 

367 if self.options.parser == 'lalr': 

368 self.options.lexer = 'contextual' 

369 elif self.options.parser == 'earley': 

370 if self.options.postlex is not None: 

371 logger.info("postlex can't be used with the dynamic lexer, so we use 'basic' instead. " 

372 "Consider using lalr with contextual instead of earley") 

373 self.options.lexer = 'basic' 

374 else: 

375 self.options.lexer = 'dynamic' 

376 elif self.options.parser == 'cyk': 

377 self.options.lexer = 'basic' 

378 else: 

379 assert False, self.options.parser 

380 lexer = self.options.lexer 

381 if isinstance(lexer, type): 

382 assert issubclass(lexer, Lexer) # XXX Is this really important? Maybe just ensure interface compliance 

383 else: 

384 assert_config(lexer, ('basic', 'contextual', 'dynamic', 'dynamic_complete')) 

385 if self.options.postlex is not None and 'dynamic' in lexer: 

386 raise ConfigurationError("Can't use postlex with a dynamic lexer. Use basic or contextual instead") 

387 

388 if self.options.ambiguity == 'auto': 

389 if self.options.parser == 'earley': 

390 self.options.ambiguity = 'resolve' 

391 else: 

392 assert_config(self.options.parser, ('earley', 'cyk'), "%r doesn't support disambiguation. Use one of these parsers instead: %s") 

393 

394 if self.options.priority == 'auto': 

395 self.options.priority = 'normal' 

396 

397 if self.options.priority not in _VALID_PRIORITY_OPTIONS: 

398 raise ConfigurationError("invalid priority option: %r. Must be one of %r" % (self.options.priority, _VALID_PRIORITY_OPTIONS)) 

399 if self.options.ambiguity not in _VALID_AMBIGUITY_OPTIONS: 

400 raise ConfigurationError("invalid ambiguity option: %r. Must be one of %r" % (self.options.ambiguity, _VALID_AMBIGUITY_OPTIONS)) 

401 

402 if self.options.parser is None: 

403 terminals_to_keep = '*' 

404 elif self.options.postlex is not None: 

405 terminals_to_keep = set(self.options.postlex.always_accept) 

406 else: 

407 terminals_to_keep = set() 

408 

409 # Compile the EBNF grammar into BNF 

410 self.terminals, self.rules, self.ignore_tokens = self.grammar.compile(self.options.start, terminals_to_keep) 

411 

412 if self.options.edit_terminals: 

413 for t in self.terminals: 

414 self.options.edit_terminals(t) 

415 

416 self._terminals_dict = {t.name: t for t in self.terminals} 

417 

418 # If the user asked to invert the priorities, negate them all here. 

419 if self.options.priority == 'invert': 

420 for rule in self.rules: 

421 if rule.options.priority is not None: 

422 rule.options.priority = -rule.options.priority 

423 for term in self.terminals: 

424 term.priority = -term.priority 

425 # Else, if the user asked to disable priorities, strip them from the 

426 # rules and terminals. This allows the Earley parsers to skip an extra forest walk 

427 # for improved performance, if you don't need them (or didn't specify any). 

428 elif self.options.priority is None: 

429 for rule in self.rules: 

430 if rule.options.priority is not None: 

431 rule.options.priority = None 

432 for term in self.terminals: 

433 term.priority = 0 

434 

435 # TODO Deprecate lexer_callbacks? 

436 self.lexer_conf = LexerConf( 

437 self.terminals, re_module, self.ignore_tokens, self.options.postlex, 

438 self.options.lexer_callbacks, self.options.g_regex_flags, use_bytes=self.options.use_bytes, strict=self.options.strict 

439 ) 

440 

441 if self.options.parser: 

442 self.parser = self._build_parser() 

443 elif lexer: 

444 self.lexer = self._build_lexer() 

445 

446 if cache_fn: 

447 logger.debug('Saving grammar to cache: %s', cache_fn) 

448 try: 

449 with FS.open(cache_fn, 'wb') as f: 

450 assert cache_sha256 is not None 

451 f.write(cache_sha256.encode('utf8') + b'\n') 

452 pickle.dump(used_files, f) 

453 self.save(f, _LOAD_ALLOWED_OPTIONS) 

454 except IOError as e: 

455 logger.exception("Failed to save Lark to cache: %r.", cache_fn, e) 

456 

457 if __doc__: 

458 __doc__ += "\n\n" + LarkOptions.OPTIONS_DOC 

459 

460 __serialize_fields__ = 'parser', 'rules', 'options' 

461 

462 def _build_lexer(self, dont_ignore: bool=False) -> BasicLexer: 

463 lexer_conf = self.lexer_conf 

464 if dont_ignore: 

465 from copy import copy 

466 lexer_conf = copy(lexer_conf) 

467 lexer_conf.ignore = () 

468 return BasicLexer(lexer_conf) 

469 

470 def _prepare_callbacks(self) -> None: 

471 self._callbacks = {} 

472 # we don't need these callbacks if we aren't building a tree 

473 if self.options.ambiguity != 'forest': 

474 self._parse_tree_builder = ParseTreeBuilder( 

475 self.rules, 

476 self.options.tree_class or Tree, 

477 self.options.propagate_positions, 

478 self.options.parser != 'lalr' and self.options.ambiguity == 'explicit', 

479 self.options.maybe_placeholders 

480 ) 

481 self._callbacks = self._parse_tree_builder.create_callback(self.options.transformer) 

482 self._callbacks.update(_get_lexer_callbacks(self.options.transformer, self.terminals)) 

483 

484 def _build_parser(self) -> "ParsingFrontend": 

485 self._prepare_callbacks() 

486 _validate_frontend_args(self.options.parser, self.options.lexer) 

487 parser_conf = ParserConf(self.rules, self._callbacks, self.options.start) 

488 return _construct_parsing_frontend( 

489 self.options.parser, 

490 self.options.lexer, 

491 self.lexer_conf, 

492 parser_conf, 

493 options=self.options 

494 ) 

495 

496 def save(self, f, exclude_options: Collection[str] = ()) -> None: 

497 """Saves the instance into the given file object 

498 

499 Useful for caching and multiprocessing. 

500 """ 

501 if self.options.parser != 'lalr': 

502 raise NotImplementedError("Lark.save() is only implemented for the LALR(1) parser.") 

503 data, m = self.memo_serialize([TerminalDef, Rule]) 

504 if exclude_options: 

505 data["options"] = {n: v for n, v in data["options"].items() if n not in exclude_options} 

506 pickle.dump({'data': data, 'memo': m}, f, protocol=pickle.HIGHEST_PROTOCOL) 

507 

508 @classmethod 

509 def load(cls: Type[_T], f) -> _T: 

510 """Loads an instance from the given file object 

511 

512 Useful for caching and multiprocessing. 

513 """ 

514 inst = cls.__new__(cls) 

515 return inst._load(f) 

516 

517 def _deserialize_lexer_conf(self, data: Dict[str, Any], memo: Dict[int, Union[TerminalDef, Rule]], options: LarkOptions) -> LexerConf: 

518 lexer_conf = LexerConf.deserialize(data['lexer_conf'], memo) 

519 lexer_conf.callbacks = options.lexer_callbacks or {} 

520 lexer_conf.re_module = regex if options.regex else re 

521 lexer_conf.use_bytes = options.use_bytes 

522 lexer_conf.g_regex_flags = options.g_regex_flags 

523 lexer_conf.skip_validation = True 

524 lexer_conf.postlex = options.postlex 

525 return lexer_conf 

526 

527 def _load(self: _T, f: Any, **kwargs) -> _T: 

528 if isinstance(f, dict): 

529 d = f 

530 else: 

531 d = pickle.load(f) 

532 memo_json = d['memo'] 

533 data = d['data'] 

534 

535 assert memo_json 

536 memo = SerializeMemoizer.deserialize(memo_json, {'Rule': Rule, 'TerminalDef': TerminalDef}, {}) 

537 options = dict(data['options']) 

538 if (set(kwargs) - _LOAD_ALLOWED_OPTIONS) & set(LarkOptions._defaults): 

539 raise ConfigurationError("Some options are not allowed when loading a Parser: {}" 

540 .format(set(kwargs) - _LOAD_ALLOWED_OPTIONS)) 

541 options.update(kwargs) 

542 self.options = LarkOptions.deserialize(options, memo) 

543 self.rules = [Rule.deserialize(r, memo) for r in data['rules']] 

544 self.source_path = '<deserialized>' 

545 _validate_frontend_args(self.options.parser, self.options.lexer) 

546 self.lexer_conf = self._deserialize_lexer_conf(data['parser'], memo, self.options) 

547 self.terminals = self.lexer_conf.terminals 

548 self._prepare_callbacks() 

549 self._terminals_dict = {t.name: t for t in self.terminals} 

550 self.parser = _deserialize_parsing_frontend( 

551 data['parser'], 

552 memo, 

553 self.lexer_conf, 

554 self._callbacks, 

555 self.options, # Not all, but multiple attributes are used 

556 ) 

557 return self 

558 

559 @classmethod 

560 def _load_from_dict(cls, data, memo, **kwargs): 

561 inst = cls.__new__(cls) 

562 return inst._load({'data': data, 'memo': memo}, **kwargs) 

563 

564 @classmethod 

565 def open(cls: Type[_T], grammar_filename: str, rel_to: Optional[str]=None, **options) -> _T: 

566 """Create an instance of Lark with the grammar given by its filename 

567 

568 If ``rel_to`` is provided, the function will find the grammar filename in relation to it. 

569 

570 Example: 

571 

572 >>> Lark.open("grammar_file.lark", rel_to=__file__, parser="lalr") 

573 Lark(...) 

574 

575 """ 

576 if rel_to: 

577 basepath = os.path.dirname(rel_to) 

578 grammar_filename = os.path.join(basepath, grammar_filename) 

579 with open(grammar_filename, encoding='utf8') as f: 

580 return cls(f, **options) 

581 

582 @classmethod 

583 def open_from_package(cls: Type[_T], package: str, grammar_path: str, search_paths: 'Sequence[str]'=[""], **options) -> _T: 

584 """Create an instance of Lark with the grammar loaded from within the package `package`. 

585 This allows grammar loading from zipapps. 

586 

587 Imports in the grammar will use the `package` and `search_paths` provided, through `FromPackageLoader` 

588 

589 Example: 

590 

591 Lark.open_from_package(__name__, "example.lark", ("grammars",), parser=...) 

592 """ 

593 package_loader = FromPackageLoader(package, search_paths) 

594 full_path, text = package_loader(None, grammar_path) 

595 options.setdefault('source_path', full_path) 

596 options.setdefault('import_paths', []) 

597 options['import_paths'].append(package_loader) 

598 return cls(text, **options) 

599 

600 def __repr__(self): 

601 return 'Lark(open(%r), parser=%r, lexer=%r, ...)' % (self.source_path, self.options.parser, self.options.lexer) 

602 

603 

604 def lex(self, text: str, dont_ignore: bool=False) -> Iterator[Token]: 

605 """Only lex (and postlex) the text, without parsing it. Only relevant when lexer='basic' 

606 

607 When dont_ignore=True, the lexer will return all tokens, even those marked for %ignore. 

608 

609 :raises UnexpectedCharacters: In case the lexer cannot find a suitable match. 

610 """ 

611 lexer: Lexer 

612 if not hasattr(self, 'lexer') or dont_ignore: 

613 lexer = self._build_lexer(dont_ignore) 

614 else: 

615 lexer = self.lexer 

616 lexer_thread = LexerThread.from_text(lexer, text) 

617 stream = lexer_thread.lex(None) 

618 if self.options.postlex: 

619 return self.options.postlex.process(stream) 

620 return stream 

621 

622 def get_terminal(self, name: str) -> TerminalDef: 

623 """Get information about a terminal""" 

624 return self._terminals_dict[name] 

625 

626 def parse_interactive(self, text: Optional[str]=None, start: Optional[str]=None) -> 'InteractiveParser': 

627 """Start an interactive parsing session. 

628 

629 Parameters: 

630 text (str, optional): Text to be parsed. Required for ``resume_parse()``. 

631 start (str, optional): Start symbol 

632 

633 Returns: 

634 A new InteractiveParser instance. 

635 

636 See Also: ``Lark.parse()`` 

637 """ 

638 return self.parser.parse_interactive(text, start=start) 

639 

640 def parse(self, text: str, start: Optional[str]=None, on_error: 'Optional[Callable[[UnexpectedInput], bool]]'=None) -> 'ParseTree': 

641 """Parse the given text, according to the options provided. 

642 

643 Parameters: 

644 text (str): Text to be parsed. 

645 start (str, optional): Required if Lark was given multiple possible start symbols (using the start option). 

646 on_error (function, optional): if provided, will be called on UnexpectedToken error. Return true to resume parsing. 

647 LALR only. See examples/advanced/error_handling.py for an example of how to use on_error. 

648 

649 Returns: 

650 If a transformer is supplied to ``__init__``, returns whatever is the 

651 result of the transformation. Otherwise, returns a Tree instance. 

652 

653 :raises UnexpectedInput: On a parse error, one of these sub-exceptions will rise: 

654 ``UnexpectedCharacters``, ``UnexpectedToken``, or ``UnexpectedEOF``. 

655 For convenience, these sub-exceptions also inherit from ``ParserError`` and ``LexerError``. 

656 

657 """ 

658 return self.parser.parse(text, start=start, on_error=on_error) 

659 

660 

661###}