paul@32 | 1 | #!/usr/bin/env python |
paul@32 | 2 | |
paul@32 | 3 | """ |
paul@32 | 4 | Moin wiki parsing functionality. |
paul@32 | 5 | |
paul@45 | 6 | Copyright (C) 2017, 2018 Paul Boddie <paul@boddie.org.uk> |
paul@32 | 7 | |
paul@32 | 8 | This program is free software; you can redistribute it and/or modify it under |
paul@32 | 9 | the terms of the GNU General Public License as published by the Free Software |
paul@32 | 10 | Foundation; either version 3 of the License, or (at your option) any later |
paul@32 | 11 | version. |
paul@32 | 12 | |
paul@32 | 13 | This program is distributed in the hope that it will be useful, but WITHOUT |
paul@32 | 14 | ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS |
paul@32 | 15 | FOR A PARTICULAR PURPOSE. See the GNU General Public License for more |
paul@32 | 16 | details. |
paul@32 | 17 | |
paul@32 | 18 | You should have received a copy of the GNU General Public License along with |
paul@32 | 19 | this program. If not, see <http://www.gnu.org/licenses/>. |
paul@32 | 20 | """ |
paul@32 | 21 | |
paul@43 | 22 | from collections import defaultdict |
paul@83 | 23 | from moinformat.tree.moin import Block, Region, Text |
paul@33 | 24 | import re |
paul@33 | 25 | |
paul@33 | 26 | # Pattern management. |
paul@33 | 27 | |
paul@36 | 28 | ws_excl_nl = r"[ \f\r\t\v]" |
paul@55 | 29 | quotes = "['" '"]' # ['"] |
paul@155 | 30 | dotall = r"(.|\n)" |
paul@55 | 31 | |
paul@121 | 32 | def choice(l): |
paul@121 | 33 | |
paul@121 | 34 | "Return a pattern matching a choice of patterns in 'l'." |
paul@121 | 35 | |
paul@121 | 36 | return "(%s)" % "|".join(l) |
paul@121 | 37 | |
paul@55 | 38 | def excl(s): |
paul@55 | 39 | |
paul@55 | 40 | "Return a non-matching pattern for 's'." |
paul@55 | 41 | |
paul@55 | 42 | return "(?!%s)" % s |
paul@55 | 43 | |
paul@55 | 44 | def expect(s): |
paul@55 | 45 | |
paul@55 | 46 | "Return a pattern expecting 's'." |
paul@55 | 47 | |
paul@55 | 48 | return "(?=%s)" % s |
paul@55 | 49 | |
paul@55 | 50 | def group(name, s): |
paul@55 | 51 | |
paul@55 | 52 | "Return a pattern group having 'name' and the pattern string 's'." |
paul@55 | 53 | |
paul@55 | 54 | return "(?P<%s>%s)" % (name, s) |
paul@55 | 55 | |
paul@55 | 56 | def optional(s): |
paul@55 | 57 | |
paul@55 | 58 | "Return an optional pattern." |
paul@55 | 59 | |
paul@55 | 60 | return "(?:%s)?" % s |
paul@55 | 61 | |
paul@55 | 62 | def recur(name): |
paul@55 | 63 | |
paul@55 | 64 | "Return a test for a recurrence of group 'name'." |
paul@55 | 65 | |
paul@55 | 66 | return "(?P=%s)" % name |
paul@55 | 67 | |
paul@55 | 68 | def repeat(s, min=None, max=None): |
paul@55 | 69 | |
paul@55 | 70 | "Return a pattern matching 's' for the given 'min' and 'max' limits." |
paul@55 | 71 | |
paul@55 | 72 | return "%s{%s,%s}" % (s, min is not None and min or "", |
paul@55 | 73 | max is not None and max or "") |
paul@36 | 74 | |
paul@99 | 75 | def get_pattern(s): |
paul@99 | 76 | |
paul@99 | 77 | "Return a compiled regular expression for the given pattern 's'." |
paul@99 | 78 | |
paul@99 | 79 | return re.compile(s, re.UNICODE | re.MULTILINE) |
paul@99 | 80 | |
paul@33 | 81 | def get_patterns(syntax): |
paul@33 | 82 | |
paul@36 | 83 | """ |
paul@36 | 84 | Define patterns for the regular expressions in the 'syntax' mapping. In each |
paul@36 | 85 | pattern, replace \N with a pattern for matching whitespace excluding |
paul@36 | 86 | newlines. |
paul@36 | 87 | """ |
paul@33 | 88 | |
paul@33 | 89 | patterns = {} |
paul@33 | 90 | for name, value in syntax.items(): |
paul@36 | 91 | value = value.replace(r"\N", ws_excl_nl) |
paul@55 | 92 | value = value.replace(r"\Q", quotes) |
paul@155 | 93 | value = value.replace(r"\E", dotall) |
paul@99 | 94 | patterns[name] = get_pattern(value) |
paul@33 | 95 | return patterns |
paul@33 | 96 | |
paul@37 | 97 | def get_subset(d, keys): |
paul@33 | 98 | |
paul@37 | 99 | "Return a subset of 'd' having the given 'keys'." |
paul@36 | 100 | |
paul@37 | 101 | subset = {} |
paul@37 | 102 | for key in keys: |
paul@37 | 103 | subset[key] = d[key] |
paul@37 | 104 | return subset |
paul@36 | 105 | |
paul@36 | 106 | |
paul@32 | 107 | |
paul@32 | 108 | # Tokenising functions. |
paul@32 | 109 | |
paul@32 | 110 | class TokenStream: |
paul@32 | 111 | |
paul@32 | 112 | "A stream of tokens taken from a string." |
paul@32 | 113 | |
paul@37 | 114 | def __init__(self, s, pos=0): |
paul@32 | 115 | self.s = s |
paul@36 | 116 | self.pos = pos |
paul@45 | 117 | |
paul@45 | 118 | # Match details. |
paul@45 | 119 | |
paul@32 | 120 | self.match = None |
paul@45 | 121 | self.queued = None |
paul@45 | 122 | self.match_start = None |
paul@45 | 123 | |
paul@45 | 124 | # Pattern name details. |
paul@45 | 125 | |
paul@32 | 126 | self.matching = None |
paul@32 | 127 | |
paul@32 | 128 | def rewind(self, length): |
paul@32 | 129 | |
paul@32 | 130 | "Rewind in the string by 'length'." |
paul@32 | 131 | |
paul@32 | 132 | self.pos -= min(length, self.pos) |
paul@32 | 133 | |
paul@45 | 134 | def queue_match(self): |
paul@45 | 135 | |
paul@45 | 136 | "Rewind in the string to the start of the last match." |
paul@45 | 137 | |
paul@45 | 138 | self.queued = self.match |
paul@45 | 139 | |
paul@37 | 140 | def read_until(self, patterns, remaining=True): |
paul@32 | 141 | |
paul@32 | 142 | """ |
paul@37 | 143 | Find the first match for the given 'patterns'. Return the text preceding |
paul@37 | 144 | any match, the remaining text if no match was found, or None if no match |
paul@37 | 145 | was found and 'remaining' is given as a false value. |
paul@32 | 146 | """ |
paul@32 | 147 | |
paul@45 | 148 | if self.queued: |
paul@45 | 149 | self.match = self.queued |
paul@45 | 150 | self.queued = None |
paul@45 | 151 | else: |
paul@45 | 152 | self.match_start = None |
paul@45 | 153 | self.matching = None |
paul@32 | 154 | |
paul@45 | 155 | # Find the first matching pattern. |
paul@32 | 156 | |
paul@45 | 157 | for pattern_name, pattern in patterns.items(): |
paul@45 | 158 | match = pattern.search(self.s, self.pos) |
paul@45 | 159 | if match: |
paul@45 | 160 | start, end = match.span() |
paul@74 | 161 | if self.matching is None or start < self.start: |
paul@45 | 162 | self.start = start |
paul@45 | 163 | self.matching = pattern_name |
paul@45 | 164 | self.match = match |
paul@32 | 165 | |
paul@32 | 166 | if self.matching is None: |
paul@32 | 167 | if remaining: |
paul@32 | 168 | return self.s[self.pos:] |
paul@32 | 169 | else: |
paul@32 | 170 | return None |
paul@32 | 171 | else: |
paul@45 | 172 | return self.s[self.pos:self.start] |
paul@32 | 173 | |
paul@54 | 174 | def match_group(self, group=1): |
paul@32 | 175 | |
paul@32 | 176 | """ |
paul@32 | 177 | Return the matched text, updating the position in the stream. If 'group' |
paul@32 | 178 | is specified, the indicated group in a match will be returned. |
paul@32 | 179 | Typically, group 1 should contain all pertinent data, but groups defined |
paul@32 | 180 | within group 1 can provide sections of the data. |
paul@32 | 181 | """ |
paul@32 | 182 | |
paul@54 | 183 | self.update_pos() |
paul@54 | 184 | |
paul@32 | 185 | if self.match: |
paul@32 | 186 | try: |
paul@32 | 187 | return self.match.group(group) |
paul@32 | 188 | except IndexError: |
paul@32 | 189 | return "" |
paul@32 | 190 | else: |
paul@32 | 191 | return None |
paul@32 | 192 | |
paul@54 | 193 | def match_groups(self, groups=None): |
paul@51 | 194 | |
paul@54 | 195 | "Return the match 'groups', or all groups if unspecified." |
paul@54 | 196 | |
paul@54 | 197 | self.update_pos() |
paul@51 | 198 | |
paul@51 | 199 | if self.match: |
paul@54 | 200 | if groups is None: |
paul@54 | 201 | return self.match.groups() |
paul@54 | 202 | else: |
paul@54 | 203 | return self.match.groups(groups) |
paul@51 | 204 | else: |
paul@51 | 205 | return [] |
paul@51 | 206 | |
paul@54 | 207 | def update_pos(self): |
paul@54 | 208 | |
paul@54 | 209 | "Update the position in the stream." |
paul@54 | 210 | |
paul@54 | 211 | if self.match: |
paul@54 | 212 | _start, self.pos = self.match.span() |
paul@54 | 213 | else: |
paul@54 | 214 | self.pos = len(self.s) |
paul@54 | 215 | |
paul@32 | 216 | |
paul@32 | 217 | |
paul@32 | 218 | # Parser abstractions. |
paul@32 | 219 | |
paul@32 | 220 | class ParserBase: |
paul@32 | 221 | |
paul@32 | 222 | "Common parsing methods." |
paul@32 | 223 | |
paul@37 | 224 | region_pattern_names = None |
paul@37 | 225 | |
paul@87 | 226 | def __init__(self, formats=None, root=None): |
paul@32 | 227 | |
paul@32 | 228 | """ |
paul@32 | 229 | Initialise the parser with any given 'formats' mapping from region type |
paul@87 | 230 | names to parser objects. An optional 'root' indicates the document-level |
paul@87 | 231 | parser. |
paul@32 | 232 | """ |
paul@32 | 233 | |
paul@32 | 234 | self.formats = formats |
paul@87 | 235 | self.root = root |
paul@37 | 236 | |
paul@37 | 237 | def get_parser(self, format_type): |
paul@37 | 238 | |
paul@37 | 239 | """ |
paul@37 | 240 | Return a parser for 'format_type' or None if no suitable parser is found. |
paul@37 | 241 | """ |
paul@37 | 242 | |
paul@37 | 243 | if not self.formats: |
paul@37 | 244 | return None |
paul@37 | 245 | |
paul@37 | 246 | cls = self.formats.get(format_type) |
paul@37 | 247 | if cls: |
paul@87 | 248 | return cls(self.formats, self.root or self) |
paul@37 | 249 | else: |
paul@37 | 250 | return None |
paul@37 | 251 | |
paul@37 | 252 | def get_patterns(self, pattern_names): |
paul@37 | 253 | |
paul@37 | 254 | "Return a mapping of the given 'pattern_names' to patterns." |
paul@37 | 255 | |
paul@37 | 256 | return get_subset(self.patterns, pattern_names) |
paul@32 | 257 | |
paul@36 | 258 | def get_items(self, s, pos=0): |
paul@32 | 259 | |
paul@36 | 260 | "Return a sequence of token items for 's' and 'pos'." |
paul@32 | 261 | |
paul@37 | 262 | return TokenStream(s, pos) |
paul@37 | 263 | |
paul@37 | 264 | def set_region(self, items, region): |
paul@37 | 265 | |
paul@37 | 266 | "Set the 'items' used to populate the given 'region'." |
paul@32 | 267 | |
paul@37 | 268 | self.items = items |
paul@37 | 269 | self.region = region |
paul@37 | 270 | |
paul@37 | 271 | def read_until(self, pattern_names, remaining=True): |
paul@36 | 272 | |
paul@37 | 273 | """ |
paul@37 | 274 | Read the next portion of input, matching using 'pattern_names'. Return |
paul@37 | 275 | the text preceding any match, the remaining text if no match was found, |
paul@37 | 276 | or None if no match was found and 'remaining' is given as a false value. |
paul@37 | 277 | """ |
paul@36 | 278 | |
paul@37 | 279 | return self.items.read_until(self.get_patterns(pattern_names)) |
paul@37 | 280 | |
paul@54 | 281 | def match_group(self, group=1): |
paul@37 | 282 | |
paul@37 | 283 | """ |
paul@37 | 284 | Return the group of the matching pattern with the given 'group' number. |
paul@37 | 285 | """ |
paul@36 | 286 | |
paul@54 | 287 | return self.items.match_group(group) |
paul@37 | 288 | |
paul@54 | 289 | def matching_pattern(self): |
paul@36 | 290 | |
paul@37 | 291 | "Return the name of the matching pattern." |
paul@36 | 292 | |
paul@37 | 293 | return self.items.matching |
paul@37 | 294 | |
paul@51 | 295 | def match_groups(self): |
paul@51 | 296 | |
paul@51 | 297 | "Return the number of groups in the match." |
paul@51 | 298 | |
paul@51 | 299 | return self.items.match_groups() |
paul@51 | 300 | |
paul@37 | 301 | # Parser methods invoked from other objects. |
paul@36 | 302 | |
paul@32 | 303 | def parse(self, s): |
paul@32 | 304 | |
paul@32 | 305 | """ |
paul@32 | 306 | Parse page text 's'. Pages consist of regions delimited by markers. |
paul@32 | 307 | """ |
paul@32 | 308 | |
paul@37 | 309 | self.items = self.get_items(s) |
paul@37 | 310 | self.region = self.parse_region() |
paul@37 | 311 | return self.region |
paul@37 | 312 | |
paul@37 | 313 | def parse_region_content(self, items, region): |
paul@37 | 314 | |
paul@37 | 315 | "Parse the data provided by 'items' to populate a 'region'." |
paul@37 | 316 | |
paul@37 | 317 | self.set_region(items, region) |
paul@32 | 318 | |
paul@70 | 319 | # Parse inline and opaque regions. |
paul@70 | 320 | |
paul@70 | 321 | if not region.transparent: |
paul@70 | 322 | pattern_names = ["regionend"] |
paul@70 | 323 | |
paul@70 | 324 | # Define a block to hold text. |
paul@70 | 325 | |
paul@70 | 326 | else: |
paul@70 | 327 | self.new_block(region) |
paul@70 | 328 | pattern_names = self.region_pattern_names |
paul@37 | 329 | |
paul@70 | 330 | # Start parsing. |
paul@70 | 331 | |
paul@70 | 332 | if pattern_names: |
paul@70 | 333 | self.parse_region_details(region, pattern_names) |
paul@37 | 334 | |
paul@70 | 335 | # Reset the type if the region was not inline. |
paul@70 | 336 | |
paul@70 | 337 | if region.type == "inline": |
paul@70 | 338 | first = region.nodes and region.nodes[0] |
paul@70 | 339 | if first and isinstance(first, Text) and first.multiline(): |
paul@70 | 340 | region.type = None |
paul@37 | 341 | |
paul@37 | 342 | # Top-level parser handler methods. |
paul@37 | 343 | |
paul@52 | 344 | def parse_region(self, level=0, indent=0, type=None): |
paul@32 | 345 | |
paul@32 | 346 | """ |
paul@37 | 347 | Parse the data to populate a region with the given 'level' at the given |
paul@52 | 348 | 'indent' having the given initial 'type'. |
paul@32 | 349 | """ |
paul@32 | 350 | |
paul@52 | 351 | region = Region([], level, indent, type) |
paul@32 | 352 | |
paul@151 | 353 | # Parse section headers and directives, then parse according to region |
paul@151 | 354 | # type. |
paul@32 | 355 | |
paul@37 | 356 | self.parse_region_header(region) |
paul@151 | 357 | self.parse_region_directives(region) |
paul@37 | 358 | self.parse_region_type(region) |
paul@32 | 359 | |
paul@32 | 360 | return region |
paul@32 | 361 | |
paul@37 | 362 | def parse_region_type(self, region): |
paul@32 | 363 | |
paul@32 | 364 | """ |
paul@37 | 365 | Use configured parsers to parse 'region' based on its type. |
paul@32 | 366 | """ |
paul@32 | 367 | |
paul@32 | 368 | # Find an appropriate parser given the type. |
paul@32 | 369 | |
paul@37 | 370 | parser = self.get_parser(region.type) |
paul@70 | 371 | if not parser: |
paul@70 | 372 | region.transparent = False |
paul@70 | 373 | parser = parser or self.get_parser("moin") |
paul@70 | 374 | parser.parse_region_content(self.items, region) |
paul@32 | 375 | |
paul@37 | 376 | def parse_region_header(self, region): |
paul@32 | 377 | |
paul@32 | 378 | """ |
paul@37 | 379 | Parse the region header, setting it on the 'region' object. |
paul@32 | 380 | """ |
paul@32 | 381 | |
paul@37 | 382 | if self.read_until(["header"], False) == "": # None means no header |
paul@124 | 383 | region.args = self.match_group("args") |
paul@124 | 384 | region.type = region.args.split(" ", 1)[0] |
paul@32 | 385 | |
paul@151 | 386 | def parse_region_directives(self, region): |
paul@151 | 387 | |
paul@151 | 388 | """ |
paul@151 | 389 | Parse any directives immediately after the region header, adding them to |
paul@151 | 390 | the 'region' object. |
paul@151 | 391 | """ |
paul@151 | 392 | |
paul@151 | 393 | while True: |
paul@151 | 394 | preceding = self.read_until(["directive"], False) |
paul@151 | 395 | |
paul@151 | 396 | # With an immediately-appearing directive, handle its details. |
paul@151 | 397 | |
paul@151 | 398 | if preceding == "": |
paul@151 | 399 | handler = self.handlers.get(self.matching_pattern()) |
paul@151 | 400 | if handler: |
paul@151 | 401 | handler(self, region) |
paul@151 | 402 | else: |
paul@151 | 403 | break |
paul@151 | 404 | |
paul@151 | 405 | # Otherwise, with no immediate directive (or none at all), stop. |
paul@151 | 406 | |
paul@151 | 407 | else: |
paul@151 | 408 | break |
paul@151 | 409 | |
paul@32 | 410 | # Parsing utilities. |
paul@32 | 411 | |
paul@43 | 412 | def parse_region_details(self, region, pattern_names, strict=False): |
paul@32 | 413 | |
paul@43 | 414 | """ |
paul@43 | 415 | Search 'region' using the 'pattern_names'. If 'strict' is set to a true |
paul@43 | 416 | value, forbid the accumulation of additional textual padding. |
paul@43 | 417 | """ |
paul@32 | 418 | |
paul@32 | 419 | try: |
paul@32 | 420 | while True: |
paul@32 | 421 | |
paul@32 | 422 | # Obtain text before any marker or the end of the input. |
paul@32 | 423 | |
paul@37 | 424 | preceding = self.read_until(pattern_names) |
paul@32 | 425 | if preceding: |
paul@43 | 426 | if not strict: |
paul@43 | 427 | region.append_inline(Text(preceding)) |
paul@43 | 428 | else: |
paul@43 | 429 | break |
paul@32 | 430 | |
paul@32 | 431 | # End of input. |
paul@32 | 432 | |
paul@54 | 433 | if not self.matching_pattern(): |
paul@32 | 434 | break |
paul@32 | 435 | |
paul@32 | 436 | # Obtain any feature. |
paul@32 | 437 | |
paul@67 | 438 | feature = self.match_group("feature") or self.match_group() |
paul@54 | 439 | handler = self.handlers.get(self.matching_pattern()) |
paul@32 | 440 | |
paul@32 | 441 | # Handle each feature or add text to the region. |
paul@32 | 442 | |
paul@32 | 443 | if handler: |
paul@37 | 444 | handler(self, region) |
paul@43 | 445 | elif not strict: |
paul@43 | 446 | region.append_inline(Text(feature)) |
paul@32 | 447 | else: |
paul@43 | 448 | break |
paul@32 | 449 | |
paul@32 | 450 | except StopIteration: |
paul@32 | 451 | pass |
paul@32 | 452 | |
paul@32 | 453 | region.normalise() |
paul@32 | 454 | |
paul@43 | 455 | def add_node(self, region, node): |
paul@43 | 456 | |
paul@43 | 457 | "Add to 'region' the given 'node'." |
paul@43 | 458 | |
paul@43 | 459 | region.add(node) |
paul@43 | 460 | |
paul@43 | 461 | def append_node(self, region, node): |
paul@43 | 462 | |
paul@43 | 463 | "Append to 'region' the given 'node'." |
paul@43 | 464 | |
paul@43 | 465 | region.append(node) |
paul@43 | 466 | |
paul@37 | 467 | def end_region(self, region): |
paul@32 | 468 | |
paul@32 | 469 | "End the parsing of 'region', breaking out of the parsing loop." |
paul@32 | 470 | |
paul@32 | 471 | raise StopIteration |
paul@32 | 472 | |
paul@45 | 473 | def queue_match(self): |
paul@43 | 474 | |
paul@45 | 475 | "Queue the current match." |
paul@43 | 476 | |
paul@45 | 477 | self.items.queue_match() |
paul@43 | 478 | |
paul@43 | 479 | def new_block(self, region): |
paul@43 | 480 | |
paul@43 | 481 | "Start a new block in 'region'." |
paul@43 | 482 | |
paul@43 | 483 | self.add_node(region, Block([])) |
paul@43 | 484 | |
paul@98 | 485 | # Common handler methods. |
paul@98 | 486 | |
paul@98 | 487 | def parse_region_end(self, node): |
paul@98 | 488 | |
paul@98 | 489 | "Handle the end of a region occurring within 'node'." |
paul@98 | 490 | |
paul@98 | 491 | level = self.match_group("level") |
paul@98 | 492 | feature = self.match_group("feature") |
paul@98 | 493 | self.region.extra = self.match_group("extra") |
paul@98 | 494 | |
paul@98 | 495 | if self.region.have_end(level): |
paul@98 | 496 | raise StopIteration |
paul@98 | 497 | else: |
paul@98 | 498 | node.append_inline(Text(feature)) |
paul@98 | 499 | |
paul@32 | 500 | # vim: tabstop=4 expandtab shiftwidth=4 |