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