| 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@172 | 29 | quotes = "['" '"]' # ['"] |
| paul@172 | 30 | dotall = r"(.|\n)" # behave similarly to dot with DOTALL option |
| paul@172 | 31 | dotparagraph = r"(.|\n(?!\r?\n))" # match everything within paragraphs |
| paul@55 | 32 | |
| paul@121 | 33 | def choice(l): |
| paul@121 | 34 | |
| paul@121 | 35 | "Return a pattern matching a choice of patterns in 'l'." |
| paul@121 | 36 | |
| paul@121 | 37 | return "(%s)" % "|".join(l) |
| paul@121 | 38 | |
| paul@55 | 39 | def excl(s): |
| paul@55 | 40 | |
| paul@55 | 41 | "Return a non-matching pattern for 's'." |
| paul@55 | 42 | |
| paul@55 | 43 | return "(?!%s)" % s |
| paul@55 | 44 | |
| paul@55 | 45 | def expect(s): |
| paul@55 | 46 | |
| paul@55 | 47 | "Return a pattern expecting 's'." |
| paul@55 | 48 | |
| paul@55 | 49 | return "(?=%s)" % s |
| paul@55 | 50 | |
| paul@55 | 51 | def group(name, s): |
| paul@55 | 52 | |
| paul@55 | 53 | "Return a pattern group having 'name' and the pattern string 's'." |
| paul@55 | 54 | |
| paul@55 | 55 | return "(?P<%s>%s)" % (name, s) |
| paul@55 | 56 | |
| paul@55 | 57 | def optional(s): |
| paul@55 | 58 | |
| paul@55 | 59 | "Return an optional pattern." |
| paul@55 | 60 | |
| paul@55 | 61 | return "(?:%s)?" % s |
| paul@55 | 62 | |
| paul@55 | 63 | def recur(name): |
| paul@55 | 64 | |
| paul@55 | 65 | "Return a test for a recurrence of group 'name'." |
| paul@55 | 66 | |
| paul@55 | 67 | return "(?P=%s)" % name |
| paul@55 | 68 | |
| paul@55 | 69 | def repeat(s, min=None, max=None): |
| paul@55 | 70 | |
| paul@55 | 71 | "Return a pattern matching 's' for the given 'min' and 'max' limits." |
| paul@55 | 72 | |
| paul@55 | 73 | return "%s{%s,%s}" % (s, min is not None and min or "", |
| paul@55 | 74 | max is not None and max or "") |
| paul@36 | 75 | |
| paul@99 | 76 | def get_pattern(s): |
| paul@99 | 77 | |
| paul@99 | 78 | "Return a compiled regular expression for the given pattern 's'." |
| paul@99 | 79 | |
| paul@99 | 80 | return re.compile(s, re.UNICODE | re.MULTILINE) |
| paul@99 | 81 | |
| paul@33 | 82 | def get_patterns(syntax): |
| paul@33 | 83 | |
| paul@36 | 84 | """ |
| paul@36 | 85 | Define patterns for the regular expressions in the 'syntax' mapping. In each |
| paul@36 | 86 | pattern, replace \N with a pattern for matching whitespace excluding |
| paul@36 | 87 | newlines. |
| paul@36 | 88 | """ |
| paul@33 | 89 | |
| paul@33 | 90 | patterns = {} |
| paul@33 | 91 | for name, value in syntax.items(): |
| paul@36 | 92 | value = value.replace(r"\N", ws_excl_nl) |
| paul@55 | 93 | value = value.replace(r"\Q", quotes) |
| paul@155 | 94 | value = value.replace(r"\E", dotall) |
| paul@172 | 95 | value = value.replace(r"\P", dotparagraph) |
| paul@99 | 96 | patterns[name] = get_pattern(value) |
| paul@33 | 97 | return patterns |
| paul@33 | 98 | |
| paul@37 | 99 | def get_subset(d, keys): |
| paul@33 | 100 | |
| paul@37 | 101 | "Return a subset of 'd' having the given 'keys'." |
| paul@36 | 102 | |
| paul@37 | 103 | subset = {} |
| paul@37 | 104 | for key in keys: |
| paul@37 | 105 | subset[key] = d[key] |
| paul@37 | 106 | return subset |
| paul@36 | 107 | |
| paul@36 | 108 | |
| paul@32 | 109 | |
| paul@32 | 110 | # Tokenising functions. |
| paul@32 | 111 | |
| paul@32 | 112 | class TokenStream: |
| paul@32 | 113 | |
| paul@32 | 114 | "A stream of tokens taken from a string." |
| paul@32 | 115 | |
| paul@37 | 116 | def __init__(self, s, pos=0): |
| paul@32 | 117 | self.s = s |
| paul@36 | 118 | self.pos = pos |
| paul@45 | 119 | |
| paul@45 | 120 | # Match details. |
| paul@45 | 121 | |
| paul@32 | 122 | self.match = None |
| paul@45 | 123 | self.queued = None |
| paul@45 | 124 | self.match_start = None |
| paul@45 | 125 | |
| paul@45 | 126 | # Pattern name details. |
| paul@45 | 127 | |
| paul@32 | 128 | self.matching = None |
| paul@32 | 129 | |
| paul@32 | 130 | def rewind(self, length): |
| paul@32 | 131 | |
| paul@32 | 132 | "Rewind in the string by 'length'." |
| paul@32 | 133 | |
| paul@32 | 134 | self.pos -= min(length, self.pos) |
| paul@32 | 135 | |
| paul@45 | 136 | def queue_match(self): |
| paul@45 | 137 | |
| paul@45 | 138 | "Rewind in the string to the start of the last match." |
| paul@45 | 139 | |
| paul@45 | 140 | self.queued = self.match |
| paul@45 | 141 | |
| paul@37 | 142 | def read_until(self, patterns, remaining=True): |
| paul@32 | 143 | |
| paul@32 | 144 | """ |
| paul@37 | 145 | Find the first match for the given 'patterns'. Return the text preceding |
| paul@37 | 146 | any match, the remaining text if no match was found, or None if no match |
| paul@37 | 147 | was found and 'remaining' is given as a false value. |
| paul@32 | 148 | """ |
| paul@32 | 149 | |
| paul@45 | 150 | if self.queued: |
| paul@45 | 151 | self.match = self.queued |
| paul@45 | 152 | self.queued = None |
| paul@45 | 153 | else: |
| paul@45 | 154 | self.match_start = None |
| paul@45 | 155 | self.matching = None |
| paul@32 | 156 | |
| paul@45 | 157 | # Find the first matching pattern. |
| paul@32 | 158 | |
| paul@45 | 159 | for pattern_name, pattern in patterns.items(): |
| paul@45 | 160 | match = pattern.search(self.s, self.pos) |
| paul@45 | 161 | if match: |
| paul@45 | 162 | start, end = match.span() |
| paul@74 | 163 | if self.matching is None or start < self.start: |
| paul@45 | 164 | self.start = start |
| paul@45 | 165 | self.matching = pattern_name |
| paul@45 | 166 | self.match = match |
| paul@32 | 167 | |
| paul@32 | 168 | if self.matching is None: |
| paul@32 | 169 | if remaining: |
| paul@32 | 170 | return self.s[self.pos:] |
| paul@32 | 171 | else: |
| paul@32 | 172 | return None |
| paul@32 | 173 | else: |
| paul@45 | 174 | return self.s[self.pos:self.start] |
| paul@32 | 175 | |
| paul@54 | 176 | def match_group(self, group=1): |
| paul@32 | 177 | |
| paul@32 | 178 | """ |
| paul@32 | 179 | Return the matched text, updating the position in the stream. If 'group' |
| paul@32 | 180 | is specified, the indicated group in a match will be returned. |
| paul@32 | 181 | Typically, group 1 should contain all pertinent data, but groups defined |
| paul@32 | 182 | within group 1 can provide sections of the data. |
| paul@32 | 183 | """ |
| paul@32 | 184 | |
| paul@54 | 185 | self.update_pos() |
| paul@54 | 186 | |
| paul@32 | 187 | if self.match: |
| paul@32 | 188 | try: |
| paul@32 | 189 | return self.match.group(group) |
| paul@32 | 190 | except IndexError: |
| paul@32 | 191 | return "" |
| paul@32 | 192 | else: |
| paul@32 | 193 | return None |
| paul@32 | 194 | |
| paul@54 | 195 | def match_groups(self, groups=None): |
| paul@51 | 196 | |
| paul@54 | 197 | "Return the match 'groups', or all groups if unspecified." |
| paul@54 | 198 | |
| paul@54 | 199 | self.update_pos() |
| paul@51 | 200 | |
| paul@51 | 201 | if self.match: |
| paul@54 | 202 | if groups is None: |
| paul@54 | 203 | return self.match.groups() |
| paul@54 | 204 | else: |
| paul@54 | 205 | return self.match.groups(groups) |
| paul@51 | 206 | else: |
| paul@51 | 207 | return [] |
| paul@51 | 208 | |
| paul@54 | 209 | def update_pos(self): |
| paul@54 | 210 | |
| paul@54 | 211 | "Update the position in the stream." |
| paul@54 | 212 | |
| paul@54 | 213 | if self.match: |
| paul@54 | 214 | _start, self.pos = self.match.span() |
| paul@54 | 215 | else: |
| paul@54 | 216 | self.pos = len(self.s) |
| paul@54 | 217 | |
| paul@32 | 218 | |
| paul@32 | 219 | |
| paul@32 | 220 | # Parser abstractions. |
| paul@32 | 221 | |
| paul@32 | 222 | class ParserBase: |
| paul@32 | 223 | |
| paul@32 | 224 | "Common parsing methods." |
| paul@32 | 225 | |
| paul@37 | 226 | region_pattern_names = None |
| paul@37 | 227 | |
| paul@165 | 228 | def __init__(self, metadata, parsers=None, root=None): |
| paul@32 | 229 | |
| paul@32 | 230 | """ |
| paul@165 | 231 | Initialise the parser with the given 'metadata' and optional 'parsers'. |
| paul@165 | 232 | An optional 'root' indicates the document-level parser. |
| paul@32 | 233 | """ |
| paul@32 | 234 | |
| paul@165 | 235 | self.metadata = metadata |
| paul@165 | 236 | self.parsers = parsers |
| paul@87 | 237 | self.root = root |
| paul@37 | 238 | |
| paul@37 | 239 | def get_parser(self, format_type): |
| paul@37 | 240 | |
| paul@37 | 241 | """ |
| paul@37 | 242 | Return a parser for 'format_type' or None if no suitable parser is found. |
| paul@37 | 243 | """ |
| paul@37 | 244 | |
| paul@165 | 245 | cls = self.parsers and self.parsers.get(format_type) |
| paul@37 | 246 | if cls: |
| paul@165 | 247 | return cls(self.metadata, self.parsers, self.root or self) |
| paul@37 | 248 | else: |
| paul@37 | 249 | return None |
| paul@37 | 250 | |
| paul@37 | 251 | def get_patterns(self, pattern_names): |
| paul@37 | 252 | |
| paul@37 | 253 | "Return a mapping of the given 'pattern_names' to patterns." |
| paul@37 | 254 | |
| paul@37 | 255 | return get_subset(self.patterns, pattern_names) |
| paul@32 | 256 | |
| paul@36 | 257 | def get_items(self, s, pos=0): |
| paul@32 | 258 | |
| paul@36 | 259 | "Return a sequence of token items for 's' and 'pos'." |
| paul@32 | 260 | |
| paul@37 | 261 | return TokenStream(s, pos) |
| paul@37 | 262 | |
| paul@37 | 263 | def set_region(self, items, region): |
| paul@37 | 264 | |
| paul@37 | 265 | "Set the 'items' used to populate the given 'region'." |
| paul@32 | 266 | |
| paul@37 | 267 | self.items = items |
| paul@37 | 268 | self.region = region |
| paul@37 | 269 | |
| paul@37 | 270 | def read_until(self, pattern_names, remaining=True): |
| paul@36 | 271 | |
| paul@37 | 272 | """ |
| paul@37 | 273 | Read the next portion of input, matching using 'pattern_names'. Return |
| paul@37 | 274 | the text preceding any match, the remaining text if no match was found, |
| paul@37 | 275 | or None if no match was found and 'remaining' is given as a false value. |
| paul@37 | 276 | """ |
| paul@36 | 277 | |
| paul@37 | 278 | return self.items.read_until(self.get_patterns(pattern_names)) |
| paul@37 | 279 | |
| paul@54 | 280 | def match_group(self, group=1): |
| paul@37 | 281 | |
| paul@37 | 282 | """ |
| paul@37 | 283 | Return the group of the matching pattern with the given 'group' number. |
| paul@37 | 284 | """ |
| paul@36 | 285 | |
| paul@54 | 286 | return self.items.match_group(group) |
| paul@37 | 287 | |
| paul@54 | 288 | def matching_pattern(self): |
| paul@36 | 289 | |
| paul@37 | 290 | "Return the name of the matching pattern." |
| paul@36 | 291 | |
| paul@37 | 292 | return self.items.matching |
| paul@37 | 293 | |
| paul@51 | 294 | def match_groups(self): |
| paul@51 | 295 | |
| paul@51 | 296 | "Return the number of groups in the match." |
| paul@51 | 297 | |
| paul@51 | 298 | return self.items.match_groups() |
| paul@51 | 299 | |
| paul@37 | 300 | # Parser methods invoked from other objects. |
| paul@36 | 301 | |
| paul@32 | 302 | def parse(self, s): |
| paul@32 | 303 | |
| paul@32 | 304 | """ |
| paul@32 | 305 | Parse page text 's'. Pages consist of regions delimited by markers. |
| paul@32 | 306 | """ |
| paul@32 | 307 | |
| paul@37 | 308 | self.items = self.get_items(s) |
| paul@37 | 309 | self.region = self.parse_region() |
| paul@37 | 310 | return self.region |
| paul@37 | 311 | |
| paul@37 | 312 | def parse_region_content(self, items, region): |
| paul@37 | 313 | |
| paul@37 | 314 | "Parse the data provided by 'items' to populate a 'region'." |
| paul@37 | 315 | |
| paul@37 | 316 | self.set_region(items, region) |
| paul@32 | 317 | |
| paul@70 | 318 | # Parse inline and opaque regions. |
| paul@70 | 319 | |
| paul@70 | 320 | if not region.transparent: |
| paul@70 | 321 | pattern_names = ["regionend"] |
| paul@70 | 322 | |
| paul@70 | 323 | # Define a block to hold text. |
| paul@70 | 324 | |
| paul@70 | 325 | else: |
| paul@70 | 326 | self.new_block(region) |
| paul@70 | 327 | pattern_names = self.region_pattern_names |
| paul@37 | 328 | |
| paul@70 | 329 | # Start parsing. |
| paul@70 | 330 | |
| paul@70 | 331 | if pattern_names: |
| paul@70 | 332 | self.parse_region_details(region, pattern_names) |
| paul@37 | 333 | |
| paul@70 | 334 | # Reset the type if the region was not inline. |
| paul@70 | 335 | |
| paul@70 | 336 | if region.type == "inline": |
| paul@70 | 337 | first = region.nodes and region.nodes[0] |
| paul@70 | 338 | if first and isinstance(first, Text) and first.multiline(): |
| paul@70 | 339 | region.type = None |
| paul@37 | 340 | |
| paul@37 | 341 | # Top-level parser handler methods. |
| paul@37 | 342 | |
| paul@52 | 343 | def parse_region(self, level=0, indent=0, type=None): |
| paul@32 | 344 | |
| paul@32 | 345 | """ |
| paul@37 | 346 | Parse the data to populate a region with the given 'level' at the given |
| paul@52 | 347 | 'indent' having the given initial 'type'. |
| paul@32 | 348 | """ |
| paul@32 | 349 | |
| paul@52 | 350 | region = Region([], level, indent, type) |
| paul@32 | 351 | |
| paul@178 | 352 | # Parse section headers, then parse according to region type. |
| paul@32 | 353 | |
| paul@37 | 354 | self.parse_region_header(region) |
| paul@37 | 355 | self.parse_region_type(region) |
| paul@32 | 356 | |
| paul@32 | 357 | return region |
| paul@32 | 358 | |
| paul@37 | 359 | def parse_region_type(self, region): |
| paul@32 | 360 | |
| paul@32 | 361 | """ |
| paul@37 | 362 | Use configured parsers to parse 'region' based on its type. |
| paul@32 | 363 | """ |
| paul@32 | 364 | |
| paul@32 | 365 | # Find an appropriate parser given the type. |
| paul@32 | 366 | |
| paul@37 | 367 | parser = self.get_parser(region.type) |
| paul@70 | 368 | if not parser: |
| paul@70 | 369 | region.transparent = False |
| paul@70 | 370 | parser = parser or self.get_parser("moin") |
| paul@178 | 371 | |
| paul@178 | 372 | # Only parse directives if the region is transparent. |
| paul@178 | 373 | |
| paul@178 | 374 | if region.transparent: |
| paul@178 | 375 | self.parse_region_directives(region) |
| paul@178 | 376 | |
| paul@70 | 377 | parser.parse_region_content(self.items, region) |
| paul@32 | 378 | |
| paul@37 | 379 | def parse_region_header(self, region): |
| paul@32 | 380 | |
| paul@32 | 381 | """ |
| paul@37 | 382 | Parse the region header, setting it on the 'region' object. |
| paul@32 | 383 | """ |
| paul@32 | 384 | |
| paul@37 | 385 | if self.read_until(["header"], False) == "": # None means no header |
| paul@124 | 386 | region.args = self.match_group("args") |
| paul@124 | 387 | region.type = region.args.split(" ", 1)[0] |
| paul@32 | 388 | |
| paul@151 | 389 | def parse_region_directives(self, region): |
| paul@151 | 390 | |
| paul@151 | 391 | """ |
| paul@151 | 392 | Parse any directives immediately after the region header, adding them to |
| paul@151 | 393 | the 'region' object. |
| paul@151 | 394 | """ |
| paul@151 | 395 | |
| paul@151 | 396 | while True: |
| paul@151 | 397 | preceding = self.read_until(["directive"], False) |
| paul@151 | 398 | |
| paul@151 | 399 | # With an immediately-appearing directive, handle its details. |
| paul@151 | 400 | |
| paul@151 | 401 | if preceding == "": |
| paul@151 | 402 | handler = self.handlers.get(self.matching_pattern()) |
| paul@151 | 403 | if handler: |
| paul@151 | 404 | handler(self, region) |
| paul@151 | 405 | else: |
| paul@151 | 406 | break |
| paul@151 | 407 | |
| paul@151 | 408 | # Otherwise, with no immediate directive (or none at all), stop. |
| paul@151 | 409 | |
| paul@151 | 410 | else: |
| paul@151 | 411 | break |
| paul@151 | 412 | |
| paul@32 | 413 | # Parsing utilities. |
| paul@32 | 414 | |
| paul@43 | 415 | def parse_region_details(self, region, pattern_names, strict=False): |
| paul@32 | 416 | |
| paul@43 | 417 | """ |
| paul@43 | 418 | Search 'region' using the 'pattern_names'. If 'strict' is set to a true |
| paul@43 | 419 | value, forbid the accumulation of additional textual padding. |
| paul@43 | 420 | """ |
| paul@32 | 421 | |
| paul@32 | 422 | try: |
| paul@32 | 423 | while True: |
| paul@32 | 424 | |
| paul@32 | 425 | # Obtain text before any marker or the end of the input. |
| paul@32 | 426 | |
| paul@37 | 427 | preceding = self.read_until(pattern_names) |
| paul@32 | 428 | if preceding: |
| paul@43 | 429 | if not strict: |
| paul@43 | 430 | region.append_inline(Text(preceding)) |
| paul@43 | 431 | else: |
| paul@43 | 432 | break |
| paul@32 | 433 | |
| paul@32 | 434 | # End of input. |
| paul@32 | 435 | |
| paul@54 | 436 | if not self.matching_pattern(): |
| paul@32 | 437 | break |
| paul@32 | 438 | |
| paul@32 | 439 | # Obtain any feature. |
| paul@32 | 440 | |
| paul@67 | 441 | feature = self.match_group("feature") or self.match_group() |
| paul@54 | 442 | handler = self.handlers.get(self.matching_pattern()) |
| paul@32 | 443 | |
| paul@32 | 444 | # Handle each feature or add text to the region. |
| paul@32 | 445 | |
| paul@32 | 446 | if handler: |
| paul@37 | 447 | handler(self, region) |
| paul@43 | 448 | elif not strict: |
| paul@43 | 449 | region.append_inline(Text(feature)) |
| paul@32 | 450 | else: |
| paul@43 | 451 | break |
| paul@32 | 452 | |
| paul@32 | 453 | except StopIteration: |
| paul@32 | 454 | pass |
| paul@32 | 455 | |
| paul@32 | 456 | region.normalise() |
| paul@32 | 457 | |
| paul@43 | 458 | def add_node(self, region, node): |
| paul@43 | 459 | |
| paul@43 | 460 | "Add to 'region' the given 'node'." |
| paul@43 | 461 | |
| paul@43 | 462 | region.add(node) |
| paul@43 | 463 | |
| paul@43 | 464 | def append_node(self, region, node): |
| paul@43 | 465 | |
| paul@43 | 466 | "Append to 'region' the given 'node'." |
| paul@43 | 467 | |
| paul@43 | 468 | region.append(node) |
| paul@43 | 469 | |
| paul@37 | 470 | def end_region(self, region): |
| paul@32 | 471 | |
| paul@32 | 472 | "End the parsing of 'region', breaking out of the parsing loop." |
| paul@32 | 473 | |
| paul@32 | 474 | raise StopIteration |
| paul@32 | 475 | |
| paul@45 | 476 | def queue_match(self): |
| paul@43 | 477 | |
| paul@45 | 478 | "Queue the current match." |
| paul@43 | 479 | |
| paul@45 | 480 | self.items.queue_match() |
| paul@43 | 481 | |
| paul@43 | 482 | def new_block(self, region): |
| paul@43 | 483 | |
| paul@43 | 484 | "Start a new block in 'region'." |
| paul@43 | 485 | |
| paul@43 | 486 | self.add_node(region, Block([])) |
| paul@43 | 487 | |
| paul@98 | 488 | # Common handler methods. |
| paul@98 | 489 | |
| paul@98 | 490 | def parse_region_end(self, node): |
| paul@98 | 491 | |
| paul@98 | 492 | "Handle the end of a region occurring within 'node'." |
| paul@98 | 493 | |
| paul@98 | 494 | level = self.match_group("level") |
| paul@98 | 495 | feature = self.match_group("feature") |
| paul@98 | 496 | self.region.extra = self.match_group("extra") |
| paul@98 | 497 | |
| paul@98 | 498 | if self.region.have_end(level): |
| paul@98 | 499 | raise StopIteration |
| paul@98 | 500 | else: |
| paul@98 | 501 | node.append_inline(Text(feature)) |
| paul@98 | 502 | |
| paul@32 | 503 | # vim: tabstop=4 expandtab shiftwidth=4 |
