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