1 #!/usr/bin/env python 2 3 """ 4 Specific classes for storing term information. 5 6 Copyright (C) 2009 Paul Boddie <paul@boddie.org.uk> 7 8 This program is free software; you can redistribute it and/or modify it under 9 the terms of the GNU General Public License as published by the Free Software 10 Foundation; either version 3 of the License, or (at your option) any later 11 version. 12 13 This program is distributed in the hope that it will be useful, but WITHOUT ANY 14 WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A 15 PARTICULAR PURPOSE. See the GNU General Public License for more details. 16 17 You should have received a copy of the GNU General Public License along 18 with this program. If not, see <http://www.gnu.org/licenses/>. 19 """ 20 21 from iixr.files import * 22 from iixr.positions import * 23 from iixr.phrases import PhraseIterator 24 from os.path import commonprefix # to find common string prefixes 25 from bisect import bisect_right # to find terms in the dictionary index 26 27 class TermWriter(FileWriter): 28 29 "Writing term information to files." 30 31 def reset(self): 32 self.last_term = "" 33 self.last_offset = 0 34 35 def write_term(self, term, offset, frequency, doc_frequency): 36 37 """ 38 Write the given 'term', its position file 'offset', its 'frequency' and 39 its 'doc_frequency' (number of documents in which it appears) to the 40 term information file. 41 """ 42 43 # Write the prefix length and term suffix. 44 45 common = len(commonprefix([self.last_term, term])) 46 suffix = term[common:] 47 48 self.write_number(common) 49 self.write_string(suffix) 50 51 # Write the offset delta. 52 53 self.write_number(offset - self.last_offset) 54 55 # Write the frequency. 56 57 self.write_number(frequency) 58 59 # Write the document frequency. 60 61 self.write_number(doc_frequency) 62 63 self.last_term = term 64 self.last_offset = offset 65 66 class TermReader(FileReader): 67 68 "Reading term information from files." 69 70 def reset(self): 71 self.last_term = "" 72 self.last_offset = 0 73 74 def read_term(self): 75 76 """ 77 Read a term, its position file offset, its frequency and its document 78 frequency from the term information file. 79 """ 80 81 # Read the prefix length and term suffix. 82 83 common = self.read_number() 84 suffix = self.read_string() 85 86 self.last_term = self.last_term[:common] + suffix 87 88 # Read the offset delta. 89 90 self.last_offset += self.read_number() 91 92 # Read the frequency. 93 94 frequency = self.read_number() 95 96 # Read the document frequency. 97 98 doc_frequency = self.read_number() 99 100 return self.last_term, self.last_offset, frequency, doc_frequency 101 102 def go_to_term(self, term, offset, info_offset): 103 104 """ 105 Seek past the entry for 'term' having 'offset' to 'info_offset'. This 106 permits the scanning for later terms from the specified term. 107 """ 108 109 self.f.seek(info_offset) 110 self.last_term = term 111 self.last_offset = offset 112 113 class TermIndexWriter(TermWriter): 114 115 "Writing term dictionary index details to files." 116 117 def reset(self): 118 TermWriter.reset(self) 119 self.last_info_offset = 0 120 121 def write_term(self, term, offset, frequency, doc_frequency, info_offset): 122 123 """ 124 Write the given 'term', its position file 'offset', its 'frequency' and 125 its 'doc_frequency' to the term dictionary index file, along with the 126 'info_offset' in the term information file. 127 """ 128 129 TermWriter.write_term(self, term, offset, frequency, doc_frequency) 130 131 # Write the information file offset delta. 132 133 self.write_number(info_offset - self.last_info_offset) 134 self.last_info_offset = info_offset 135 136 class TermIndexReader(TermReader): 137 138 "Reading term dictionary index details from files." 139 140 def reset(self): 141 TermReader.reset(self) 142 self.last_info_offset = 0 143 144 def read_term(self): 145 146 """ 147 Read a term, its position file offset, its frequency, its document 148 frequency and a term information file offset from the term dictionary 149 index file. 150 """ 151 152 term, offset, frequency, doc_frequency = TermReader.read_term(self) 153 154 # Read the offset delta. 155 156 self.last_info_offset += self.read_number() 157 158 return term, offset, frequency, doc_frequency, self.last_info_offset 159 160 class TermDictionaryWriter: 161 162 "Writing term dictionaries." 163 164 def __init__(self, info_writer, index_writer, position_dict_writer, interval): 165 self.info_writer = info_writer 166 self.index_writer = index_writer 167 self.position_dict_writer = position_dict_writer 168 self.interval = interval 169 self.entry = 0 170 171 def _write_term(self, term, offset, frequency, doc_frequency): 172 173 """ 174 Write the given 'term', its position file 'offset', its 'frequency' and 175 its 'doc_frequency' (number of documents in which it appears) to the 176 term information file. Return the offset after the term information was 177 written to the file. 178 """ 179 180 self.info_writer.write_term(term, offset, frequency, doc_frequency) 181 182 if self.entry % self.interval == 0: 183 info_offset = self.info_writer.f.tell() 184 self.index_writer.write_term(term, offset, frequency, doc_frequency, info_offset) 185 186 self.entry += 1 187 188 def write_term_positions(self, term, doc_positions): 189 190 """ 191 Write the given 'term' and the 'doc_positions' recording the documents 192 and positions at which the term is found. 193 """ 194 195 offset, frequency, doc_frequency = self.position_dict_writer.write_term_positions(doc_positions) 196 self._write_term(term, offset, frequency, doc_frequency) 197 198 def close(self): 199 self.info_writer.close() 200 self.index_writer.close() 201 self.position_dict_writer.close() 202 203 class TermDictionaryReader: 204 205 "Reading term dictionaries." 206 207 def __init__(self, info_reader, index_reader, position_dict_reader): 208 self.info_reader = info_reader 209 self.index_reader = index_reader 210 self.position_dict_reader = position_dict_reader 211 212 self.terms = [] 213 try: 214 while 1: 215 self.terms.append(self.index_reader.read_term()) 216 except EOFError: 217 pass 218 219 # Large numbers for ordering purposes. 220 221 if self.terms: 222 self.max_offset = self.terms[-1][1] + 1 223 else: 224 self.max_offset = None 225 226 def _find_closest_entry(self, term): 227 228 """ 229 Find the offsets and frequencies of 'term' from the term dictionary or 230 the closest term starting with the value of 'term'. 231 232 Return the closest index entry consisting of a term, the position file 233 offset, the term frequency, the document frequency, and the term details 234 file offset. 235 """ 236 237 i = bisect_right(self.terms, (term, self.max_offset, 0, 0)) - 1 238 239 # Get the entry position providing the term or one preceding it. 240 # If no entry precedes the requested term, return the very first entry 241 # as the closest. 242 243 if i == -1: 244 return self.terms[0] 245 else: 246 return self.terms[i] 247 248 def _find_closest_term(self, term): 249 250 """ 251 Find the offsets and frequencies of 'term' from the term dictionary or 252 the closest term starting with the value of 'term'. 253 254 Return the closest term (or the term itself), the position file offset, 255 the term frequency, the document frequency, and the term details file 256 offset (or None if the reader is already positioned). 257 """ 258 259 found_term, offset, frequency, doc_frequency, info_offset = self._find_closest_entry(term) 260 261 # Where the term is found immediately, return the offset and 262 # frequencies. If the term does not appear, return the details of the 263 # closest entry. 264 265 if term <= found_term: 266 return found_term, offset, frequency, doc_frequency, info_offset 267 268 # Otherwise, seek past the index term's entry in the information file 269 # and scan for the desired term. 270 271 else: 272 self.info_reader.go_to_term(found_term, offset, info_offset) 273 try: 274 while term > found_term: 275 found_term, offset, frequency, doc_frequency = self.info_reader.read_term() 276 except EOFError: 277 pass 278 279 return found_term, offset, frequency, doc_frequency, None 280 281 def _find_term(self, term): 282 283 """ 284 Find the position file offset and frequency of 'term' from the term 285 dictionary. 286 """ 287 288 found_term, offset, frequency, doc_frequency, info_offset = self._find_closest_term(term) 289 290 # If the term is found, return the offset and frequencies. 291 292 if term == found_term: 293 return offset, frequency, doc_frequency 294 else: 295 return None 296 297 def _get_positions(self, offset, doc_frequency): 298 299 """ 300 Obtain positions from the position index 'offset' expecting a number of 301 documents equal to the given 'doc_frequency'. 302 """ 303 304 self.position_dict_reader.seek(offset, doc_frequency) 305 return self.position_dict_reader 306 307 # Iterator convenience methods. 308 309 def __iter__(self): 310 self.rewind() 311 return self 312 313 def next(self): 314 try: 315 return self.read_term() 316 except EOFError: 317 raise StopIteration 318 319 # Sequential access methods. 320 321 def rewind(self): 322 self.info_reader.rewind() 323 324 def read_term(self): 325 326 """ 327 Return the next term, its frequency, its document frequency, and the 328 documents and positions at which the term is found. 329 """ 330 331 term, offset, frequency, doc_frequency = self.info_reader.read_term() 332 333 self.position_dict_reader.seek(offset, doc_frequency) 334 return term, frequency, doc_frequency, self.position_dict_reader 335 336 # Query methods. 337 338 def find_terms(self, term): 339 340 "Return all terms whose values start with the value of 'term'." 341 342 terms = [] 343 344 found_term, offset, frequency, doc_frequency, info_offset = self._find_closest_term(term) 345 346 # Position the reader, if necessary. 347 348 if info_offset is not None: 349 self.info_reader.go_to_term(found_term, offset, info_offset) 350 351 # Read and record terms. 352 353 try: 354 # Add the found term if it starts with the specified term. 355 356 while found_term.startswith(term): 357 terms.append(found_term) 358 found_term, offset, frequency, doc_frequency = self.info_reader.read_term() 359 360 except EOFError: 361 pass 362 363 return terms 364 365 def find_positions(self, term): 366 367 "Return the documents and positions at which the given 'term' is found." 368 369 t = self._find_term(term) 370 if t is None: 371 return [] 372 else: 373 offset, frequency, doc_frequency = t 374 return self._get_positions(offset, doc_frequency) 375 376 def find_common_positions(self, terms): 377 378 """ 379 Return the documents and positions at which all the given 'terms' are 380 found, where only common documents are returned. 381 """ 382 383 return PhraseIterator([self.find_positions(term) for term in terms]) 384 385 def get_frequency(self, term): 386 387 "Return the frequency of the given 'term'." 388 389 t = self._find_term(term) 390 if t is None: 391 return None 392 else: 393 offset, frequency, doc_frequency = t 394 return frequency 395 396 def get_document_frequency(self, term): 397 398 "Return the document frequency of the given 'term'." 399 400 t = self._find_term(term) 401 if t is None: 402 return None 403 else: 404 offset, frequency, doc_frequency = t 405 return doc_frequency 406 407 def close(self): 408 self.info_reader.close() 409 self.index_reader.close() 410 self.position_dict_reader.close() 411 412 # vim: tabstop=4 expandtab shiftwidth=4