includes/search/SearchOracle.php

   1 <?php
   2 /**
   3  * Oracle search engine
   4  *
   5  * Copyright © 2004 Brion Vibber <brion@pobox.com>
   6  * http://www.mediawiki.org/
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Search
  25  */
  26
  27 /**
  28  * Search engine hook base class for Oracle (ConText).
  29  * @ingroup Search
  30  */
  31 class SearchOracle extends SearchEngine {
  32
  33         private $reservedWords = array(
  34                 'ABOUT' => 1,
  35                 'ACCUM' => 1,
  36                 'AND' => 1,
  37                 'BT' => 1,
  38                 'BTG' => 1,
  39                 'BTI' => 1,
  40                 'BTP' => 1,
  41                 'FUZZY' => 1,
  42                 'HASPATH' => 1,
  43                 'INPATH' => 1,
  44                 'MINUS' => 1,
  45                 'NEAR' => 1,
  46                 'NOT' => 1,
  47                 'NT' => 1,
  48                 'NTG' => 1,
  49                 'NTI' => 1,
  50                 'NTP' => 1,
  51                 'OR' => 1,
  52                 'PT' => 1,
  53                 'RT' => 1,
  54                 'SQE' => 1,
  55                 'SYN' => 1,
  56                 'TR' => 1,
  57                 'TRSYN' => 1,
  58                 'TT' => 1,
  59                 'WITHIN' => 1,
  60         );
  61
  62         /**
  63          * Creates an instance of this class
  64          * @param $db DatabasePostgres: database object
  65          */
  66         function __construct( $db ) {
  67                 parent::__construct( $db );
  68         }
  69
  70         /**
  71          * Perform a full text search query and return a result set.
  72          *
  73          * @param string $term raw search term
  74          * @return SqlSearchResultSet
  75          */
  76         function searchText( $term ) {
  77                 if ( $term == '' ) {
  78                         return new SqlSearchResultSet( false, '' );
  79                 }
  80
  81                 $resultSet = $this->db->resultObject( $this->db->query( $this->getQuery( $this->filter( $term ), true ) ) );
  82                 return new SqlSearchResultSet( $resultSet, $this->searchTerms );
  83         }
  84
  85         /**
  86          * Perform a title-only search query and return a result set.
  87          *
  88          * @param string $term raw search term
  89          * @return SqlSearchResultSet
  90          */
  91         function searchTitle( $term ) {
  92                 if ( $term == '' ) {
  93                         return new SqlSearchResultSet( false, '' );
  94                 }
  95
  96                 $resultSet = $this->db->resultObject( $this->db->query( $this->getQuery( $this->filter( $term ), false ) ) );
  97                 return new MySQLSearchResultSet( $resultSet, $this->searchTerms );
  98         }
  99
 100         /**
 101          * Return a partial WHERE clause to exclude redirects, if so set
 102          * @return String
 103          */
 104         function queryRedirect() {
 105                 if ( $this->showRedirects ) {
 106                         return '';
 107                 } else {
 108                         return 'AND page_is_redirect=0';
 109                 }
 110         }
 111
 112         /**
 113          * Return a partial WHERE clause to limit the search to the given namespaces
 114          * @return String
 115          */
 116         function queryNamespaces() {
 117                 if ( is_null( $this->namespaces ) ) {
 118                         return '';
 119                 }
 120                 if ( !count( $this->namespaces ) ) {
 121                         $namespaces = '0';
 122                 } else {
 123                         $namespaces = $this->db->makeList( $this->namespaces );
 124                 }
 125                 return 'AND page_namespace IN (' . $namespaces . ')';
 126         }
 127
 128         /**
 129          * Return a LIMIT clause to limit results on the query.
 130          *
 131          * @param $sql string
 132          *
 133          * @return String
 134          */
 135         function queryLimit( $sql ) {
 136                 return $this->db->limitResult( $sql, $this->limit, $this->offset );
 137         }
 138
 139         /**
 140          * Does not do anything for generic search engine
 141          * subclasses may define this though
 142          *
 143          * @return String
 144          */
 145         function queryRanking( $filteredTerm, $fulltext ) {
 146                 return ' ORDER BY score(1)';
 147         }
 148
 149         /**
 150          * Construct the full SQL query to do the search.
 151          * The guts shoulds be constructed in queryMain()
 152          * @param $filteredTerm String
 153          * @param $fulltext Boolean
 154          * @return String
 155          */
 156         function getQuery( $filteredTerm, $fulltext ) {
 157                 return $this->queryLimit( $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 158                         $this->queryRedirect() . ' ' .
 159                         $this->queryNamespaces() . ' ' .
 160                         $this->queryRanking( $filteredTerm, $fulltext ) . ' ' );
 161         }
 162
 163         /**
 164          * Picks which field to index on, depending on what type of query.
 165          * @param $fulltext Boolean
 166          * @return String
 167          */
 168         function getIndexField( $fulltext ) {
 169                 return $fulltext ? 'si_text' : 'si_title';
 170         }
 171
 172         /**
 173          * Get the base part of the search query.
 174          *
 175          * @param $filteredTerm String
 176          * @param $fulltext Boolean
 177          * @return String
 178          */
 179         function queryMain( $filteredTerm, $fulltext ) {
 180                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 181                 $page = $this->db->tableName( 'page' );
 182                 $searchindex = $this->db->tableName( 'searchindex' );
 183                 return 'SELECT page_id, page_namespace, page_title ' .
 184                         "FROM $page,$searchindex " .
 185                         'WHERE page_id=si_page AND ' . $match;
 186         }
 187
 188         /**
 189          * Parse a user input search string, and return an SQL fragment to be used
 190          * as part of a WHERE clause
 191          * @return string
 192          */
 193         function parseQuery( $filteredText, $fulltext ) {
 194                 global $wgContLang;
 195                 $lc = SearchEngine::legalSearchChars();
 196                 $this->searchTerms = array();
 197
 198                 # @todo FIXME: This doesn't handle parenthetical expressions.
 199                 $m = array();
 200                 $searchon = '';
 201                 if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
 202                                 $filteredText, $m, PREG_SET_ORDER ) ) {
 203                         foreach ( $m as $terms ) {
 204                                 // Search terms in all variant forms, only
 205                                 // apply on wiki with LanguageConverter
 206                                 $temp_terms = $wgContLang->autoConvertToAllVariants( $terms[2] );
 207                                 if ( is_array( $temp_terms ) ) {
 208                                         $temp_terms = array_unique( array_values( $temp_terms ) );
 209                                         foreach ( $temp_terms as $t ) {
 210                                                 $searchon .= ( $terms[1] == '-' ? ' ~' : ' & ' ) . $this->escapeTerm( $t );
 211                                         }
 212                                 }
 213                                 else {
 214                                         $searchon .= ( $terms[1] == '-' ? ' ~' : ' & ' ) . $this->escapeTerm( $terms[2] );
 215                                 }
 216                                 if ( !empty( $terms[3] ) ) {
 217                                         $regexp = preg_quote( $terms[3], '/' );
 218                                         if ( $terms[4] ) {
 219                                                 $regexp .= "[0-9A-Za-z_]+";
 220                                         }
 221                                 } else {
 222                                         $regexp = preg_quote( str_replace( '"', '', $terms[2] ), '/' );
 223                                 }
 224                                 $this->searchTerms[] = $regexp;
 225                         }
 226                 }
 227
 228                 $searchon = $this->db->addQuotes( ltrim( $searchon, ' &' ) );
 229                 $field = $this->getIndexField( $fulltext );
 230                 return " CONTAINS($field, $searchon, 1) > 0 ";
 231         }
 232
 233         private function escapeTerm( $t ) {
 234                 global $wgContLang;
 235                 $t = $wgContLang->normalizeForSearch( $t );
 236                 $t = isset( $this->reservedWords[strtoupper( $t )] ) ? '{' . $t . '}' : $t;
 237                 $t = preg_replace( '/^"(.*)"$/', '($1)', $t );
 238                 $t = preg_replace( '/([-&|])/', '\\\\$1', $t );
 239                 return $t;
 240         }
 241         /**
 242          * Create or update the search index record for the given page.
 243          * Title and text should be pre-processed.
 244          *
 245          * @param $id Integer
 246          * @param $title String
 247          * @param $text String
 248          */
 249         function update( $id, $title, $text ) {
 250                 $dbw = wfGetDB( DB_MASTER );
 251                 $dbw->replace( 'searchindex',
 252                         array( 'si_page' ),
 253                         array(
 254                                 'si_page' => $id,
 255                                 'si_title' => $title,
 256                                 'si_text' => $text
 257                         ), 'SearchOracle::update' );
 258
 259                 // Sync the index
 260                 // We need to specify the DB name (i.e. user/schema) here so that
 261                 // it can work from the installer, where
 262                 //     ALTER SESSION SET CURRENT_SCHEMA = ...
 263                 // was used.
 264                 $dbw->query( "CALL ctx_ddl.sync_index(" .
 265                         $dbw->addQuotes( $dbw->getDBname() . '.' . $dbw->tableName( 'si_text_idx', 'raw' ) ) . ")" );
 266                 $dbw->query( "CALL ctx_ddl.sync_index(" .
 267                         $dbw->addQuotes( $dbw->getDBname() . '.' . $dbw->tableName( 'si_title_idx', 'raw' ) ) . ")" );
 268         }
 269
 270         /**
 271          * Update a search index record's title only.
 272          * Title should be pre-processed.
 273          *
 274          * @param $id Integer
 275          * @param $title String
 276          */
 277         function updateTitle( $id, $title ) {
 278                 $dbw = wfGetDB( DB_MASTER );
 279
 280                 $dbw->update( 'searchindex',
 281                         array( 'si_title' => $title ),
 282                         array( 'si_page' => $id ),
 283                         'SearchOracle::updateTitle',
 284                         array() );
 285         }
 286
 287         public static function legalSearchChars() {
 288                 return "\"" . parent::legalSearchChars();
 289         }
 290 }