Root/
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 | <?php/* -*- tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- *//*# ***** BEGIN LICENSE BLOCK *****# This file is part of Plume Framework, a simple PHP Application Framework.# Copyright (C) 2001-2007 Loic d'Anterroches and contributors.## Plume Framework is free software; you can redistribute it and/or modify# it under the terms of the GNU Lesser General Public License as published by# the Free Software Foundation; either version 2.1 of the License, or# (at your option) any later version.## Plume Framework is distributed in the hope that it will be useful,# but WITHOUT ANY WARRANTY; without even the implied warranty of# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the# GNU Lesser General Public License for more details.## You should have received a copy of the GNU Lesser General Public License# along with this program; if not, write to the Free Software# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA## ***** END LICENSE BLOCK ***** *//** * Class implementing a small search engine. * * Ideal for a small website with up to 100,000 documents. */class Pluf_Search{ /** * Search. * * Returns an array of array with model_class, model_id and * score. The list is already sorted by score descending. * * You can then filter the list as you wish with another set of * weights. * * @param string Query string. * @return array Results. */ public static function search($query, $stemmer='Pluf_Text_Stemmer_Porter') { $query = Pluf_Text::cleanString(html_entity_decode($query, ENT_QUOTES, 'UTF-8')); $words = Pluf_Text::tokenize($query); if ($stemmer != null) { $words = self::stem($words, $stemmer); } $words_flat = array(); foreach ($words as $word=>$c) { $words_flat[] = $word; } $word_ids = self::getWordIds($words_flat); if (in_array(null, $word_ids)) { return array(); } return self::searchDocuments($word_ids); } /** * Stem the words with the given stemmer. */ public static function stem($words, $stemmer) { $nwords = array(); foreach ($words as $word => $occ) { $word = call_user_func(array($stemmer, 'stem'), $word); if (isset($nwords[$word])) { $nwords[$word] += $occ; } else { $nwords[$word] = $occ; } } return $nwords; } /** * Search documents. * * Only the total of the ponderated occurences is used to sort the * results. * * @param array Ids. * @return array Sorted by score, returns model_class, model_id and score. */ public static function searchDocuments($wids) { $db =& Pluf::db(); $gocc = new Pluf_Search_Occ(); $where = array(); foreach ($wids as $id) { $where[] = $db->qn('word').'='.(int)$id; } $select = 'SELECT model_class, model_id, SUM(pondocc) AS score FROM '.$gocc->getSqlTable().' WHERE '.implode(' OR ', $where).' GROUP BY model_class, model_id HAVING COUNT(*)='.count($wids).' ORDER BY score DESC'; return $db->select($select); } /** * Get the id of each word. * * @param array Words * @return array Ids, null if no matching word. */ public static function getWordIds($words) { $ids = array(); $gword = new Pluf_Search_Word(); foreach ($words as $word) { $sql = new Pluf_SQL('word=%s', array($word)); $l = $gword->getList(array('filter' => $sql->gen())); if ($l->count() > 0) { $ids[] = $l[0]->id; } else { $ids[] = null; } } return $ids; } /** * Index a document. * * The document must provide a method _toIndex() returning the * document as a string for indexation. The string must be clean * and will simply be tokenized by Pluf_Text::tokenize(). * * So a recommended way to clean it at the end is to remove all * the HTML tags and then run the following on it: * * return Pluf_Text::cleanString(html_entity_decode($string, * ENT_QUOTES, 'UTF-8')); * * Indexing is resource intensive so it is recommanded to run the * indexing in an asynchronous way. When you save a resource to be * indexed, just write a log "need to index resource x" and then * you can every few minutes index the resources. Nobody care if * your index is not perfectly fresh, but your end users care if * it takes 0.6s to get back the page instead of 0.1s. * * Take 500 average documents, index them while counting the total * time it takes to index. Divide by 500 and if the result is more * than 0.1s, use a log/queue. * * FIXME: Concurrency problem if you index at the same time the same doc. * * @param Pluf_Model Document to index. * @param Stemmer used. ('Pluf_Text_Stemmer_Porter') * @return array Statistics. */ public static function index($doc, $stemmer='Pluf_Text_Stemmer_Porter') { $words = Pluf_Text::tokenize($doc->_toIndex()); if ($stemmer != null) { $words = self::stem($words, $stemmer); } // Get the total number of words. $total = 0.0; $words_flat = array(); foreach ($words as $word => $occ) { $total += (float) $occ; $words_flat[] = $word; } // Drop the last indexation. $gocc = new Pluf_Search_Occ(); $sql = new Pluf_SQL('DELETE FROM '.$gocc->getSqlTable().' WHERE model_class=%s AND model_id=%s', array($doc->_model, $doc->id)); $db =& Pluf::db(); $db->execute($sql->gen()); // Get the ids for each word. $ids = self::getWordIds($words_flat); // Insert a new word for the missing words and add the occ. $n = count($ids); $new_words = 0; $done = array(); for ($i=0;$i<$n;$i++) { if ($ids[$i] === null) { $word = new Pluf_Search_Word(); $word->word = $words_flat[$i]; try { $word->create(); $ids[$i] = $word->id; } catch (Exception $e) { // most likely concurrent addition of a word, try // to read it. $_ids = self::getWordIds(array($words_flat[$i])); if ($_ids[0] !== null) { // if we miss it here, just forget about it $ids[$i] = $_ids[0]; } } $new_words++; } if (isset($done[$ids[$i]])) { continue; } $done[$ids[$i]] = true; $occ = new Pluf_Search_Occ(); $occ->word = new Pluf_Search_Word($ids[$i]); $occ->model_class = $doc->_model; $occ->model_id = $doc->id; $occ->occ = $words[$words_flat[$i]]; $occ->pondocc = $words[$words_flat[$i]]/$total; $occ->create(); } // update the stats $sql = new Pluf_SQL('model_class=%s AND model_id=%s', array($doc->_model, $doc->id)); $last_index = Pluf::factory('Pluf_Search_Stats')->getList(array('filter' => $sql->gen())); if ($last_index->count() == 0) { $stats = new Pluf_Search_Stats(); $stats->model_class = $doc->_model; $stats->model_id = $doc->id; $stats->indexations = 1; $stats->create(); } else { $last_index[0]->indexations += 1; $last_index[0]->update(); } return array('total' => $total, 'new' => $new_words, 'unique'=>$n); }} |
