includes/search/SearchOracle.php

   1 <?php
   2 /**
   3  * Oracle search engine
   4  *
   5  * Copyright © 2004 Brion Vibber <brion@pobox.com>
   6  * https://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 SearchDatabase {
  32         private $reservedWords = [
  33                 'ABOUT' => 1,
  34                 'ACCUM' => 1,
  35                 'AND' => 1,
  36                 'BT' => 1,
  37                 'BTG' => 1,
  38                 'BTI' => 1,
  39                 'BTP' => 1,
  40                 'FUZZY' => 1,
  41                 'HASPATH' => 1,
  42                 'INPATH' => 1,
  43                 'MINUS' => 1,
  44                 'NEAR' => 1,
  45                 'NOT' => 1,
  46                 'NT' => 1,
  47                 'NTG' => 1,
  48                 'NTI' => 1,
  49                 'NTP' => 1,
  50                 'OR' => 1,
  51                 'PT' => 1,
  52                 'RT' => 1,
  53                 'SQE' => 1,
  54                 'SYN' => 1,
  55                 'TR' => 1,
  56                 'TRSYN' => 1,
  57                 'TT' => 1,
  58                 'WITHIN' => 1,
  59         ];
  60
  61         /**
  62          * Perform a full text search query and return a result set.
  63          *
  64          * @param string $term Raw search term
  65          * @return SqlSearchResultSet
  66          */
  67         function searchText( $term ) {
  68                 if ( $term == '' ) {
  69                         return new SqlSearchResultSet( false, '' );
  70                 }
  71
  72                 $resultSet = $this->db->query( $this->getQuery( $this->filter( $term ), true ) );
  73                 return new SqlSearchResultSet( $resultSet, $this->searchTerms );
  74         }
  75
  76         /**
  77          * Perform a title-only search query and return a result set.
  78          *
  79          * @param string $term Raw search term
  80          * @return SqlSearchResultSet
  81          */
  82         function searchTitle( $term ) {
  83                 if ( $term == '' ) {
  84                         return new SqlSearchResultSet( false, '' );
  85                 }
  86
  87                 $resultSet = $this->db->query( $this->getQuery( $this->filter( $term ), false ) );
  88                 return new SqlSearchResultSet( $resultSet, $this->searchTerms );
  89         }
  90
  91         /**
  92          * Return a partial WHERE clause to limit the search to the given namespaces
  93          * @return string
  94          */
  95         function queryNamespaces() {
  96                 if ( is_null( $this->namespaces ) ) {
  97                         return '';
  98                 }
  99                 if ( !count( $this->namespaces ) ) {
 100                         $namespaces = '0';
 101                 } else {
 102                         $namespaces = $this->db->makeList( $this->namespaces );
 103                 }
 104                 return 'AND page_namespace IN (' . $namespaces . ')';
 105         }
 106
 107         /**
 108          * Return a LIMIT clause to limit results on the query.
 109          *
 110          * @param string $sql
 111          *
 112          * @return string
 113          */
 114         function queryLimit( $sql ) {
 115                 return $this->db->limitResult( $sql, $this->limit, $this->offset );
 116         }
 117
 118         /**
 119          * Does not do anything for generic search engine
 120          * subclasses may define this though
 121          *
 122          * @param string $filteredTerm
 123          * @param bool $fulltext
 124          * @return string
 125          */
 126         function queryRanking( $filteredTerm, $fulltext ) {
 127                 return ' ORDER BY score(1)';
 128         }
 129
 130         /**
 131          * Construct the full SQL query to do the search.
 132          * The guts shoulds be constructed in queryMain()
 133          * @param string $filteredTerm
 134          * @param bool $fulltext
 135          * @return string
 136          */
 137         function getQuery( $filteredTerm, $fulltext ) {
 138                 return $this->queryLimit( $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 139                         $this->queryNamespaces() . ' ' .
 140                         $this->queryRanking( $filteredTerm, $fulltext ) . ' ' );
 141         }
 142
 143         /**
 144          * Picks which field to index on, depending on what type of query.
 145          * @param bool $fulltext
 146          * @return string
 147          */
 148         function getIndexField( $fulltext ) {
 149                 return $fulltext ? 'si_text' : 'si_title';
 150         }
 151
 152         /**
 153          * Get the base part of the search query.
 154          *
 155          * @param string $filteredTerm
 156          * @param bool $fulltext
 157          * @return string
 158          */
 159         function queryMain( $filteredTerm, $fulltext ) {
 160                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 161                 $page = $this->db->tableName( 'page' );
 162                 $searchindex = $this->db->tableName( 'searchindex' );
 163                 return 'SELECT page_id, page_namespace, page_title ' .
 164                         "FROM $page,$searchindex " .
 165                         'WHERE page_id=si_page AND ' . $match;
 166         }
 167
 168         /**
 169          * Parse a user input search string, and return an SQL fragment to be used
 170          * as part of a WHERE clause
 171          * @param string $filteredText
 172          * @param bool $fulltext
 173          * @return string
 174          */
 175         function parseQuery( $filteredText, $fulltext ) {
 176                 global $wgContLang;
 177                 $lc = $this->legalSearchChars( self::CHARS_NO_SYNTAX );
 178                 $this->searchTerms = [];
 179
 180                 # @todo FIXME: This doesn't handle parenthetical expressions.
 181                 $m = [];
 182                 $searchon = '';
 183                 if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
 184                                 $filteredText, $m, PREG_SET_ORDER ) ) {
 185                         foreach ( $m as $terms ) {
 186                                 // Search terms in all variant forms, only
 187                                 // apply on wiki with LanguageConverter
 188                                 $temp_terms = $wgContLang->autoConvertToAllVariants( $terms[2] );
 189                                 if ( is_array( $temp_terms ) ) {
 190                                         $temp_terms = array_unique( array_values( $temp_terms ) );
 191                                         foreach ( $temp_terms as $t ) {
 192                                                 $searchon .= ( $terms[1] == '-' ? ' ~' : ' & ' ) . $this->escapeTerm( $t );
 193                                         }
 194                                 } else {
 195                                         $searchon .= ( $terms[1] == '-' ? ' ~' : ' & ' ) . $this->escapeTerm( $terms[2] );
 196                                 }
 197                                 if ( !empty( $terms[3] ) ) {
 198                                         $regexp = preg_quote( $terms[3], '/' );
 199                                         if ( $terms[4] ) {
 200                                                 $regexp .= "[0-9A-Za-z_]+";
 201                                         }
 202                                 } else {
 203                                         $regexp = preg_quote( str_replace( '"', '', $terms[2] ), '/' );
 204                                 }
 205                                 $this->searchTerms[] = $regexp;
 206                         }
 207                 }
 208
 209                 $searchon = $this->db->addQuotes( ltrim( $searchon, ' &' ) );
 210                 $field = $this->getIndexField( $fulltext );
 211                 return " CONTAINS($field, $searchon, 1) > 0 ";
 212         }
 213
 214         private function escapeTerm( $t ) {
 215                 global $wgContLang;
 216                 $t = $wgContLang->normalizeForSearch( $t );
 217                 $t = isset( $this->reservedWords[strtoupper( $t )] ) ? '{' . $t . '}' : $t;
 218                 $t = preg_replace( '/^"(.*)"$/', '($1)', $t );
 219                 $t = preg_replace( '/([-&|])/', '\\\\$1', $t );
 220                 return $t;
 221         }
 222
 223         /**
 224          * Create or update the search index record for the given page.
 225          * Title and text should be pre-processed.
 226          *
 227          * @param int $id
 228          * @param string $title
 229          * @param string $text
 230          */
 231         function update( $id, $title, $text ) {
 232                 $dbw = wfGetDB( DB_MASTER );
 233                 $dbw->replace( 'searchindex',
 234                         [ 'si_page' ],
 235                         [
 236                                 'si_page' => $id,
 237                                 'si_title' => $title,
 238                                 'si_text' => $text
 239                         ], 'SearchOracle::update' );
 240
 241                 // Sync the index
 242                 // We need to specify the DB name (i.e. user/schema) here so that
 243                 // it can work from the installer, where
 244                 //     ALTER SESSION SET CURRENT_SCHEMA = ...
 245                 // was used.
 246                 $dbw->query( "CALL ctx_ddl.sync_index(" .
 247                         $dbw->addQuotes( $dbw->getDBname() . '.' . $dbw->tableName( 'si_text_idx', 'raw' ) ) . ")" );
 248                 $dbw->query( "CALL ctx_ddl.sync_index(" .
 249                         $dbw->addQuotes( $dbw->getDBname() . '.' . $dbw->tableName( 'si_title_idx', 'raw' ) ) . ")" );
 250         }
 251
 252         /**
 253          * Update a search index record's title only.
 254          * Title should be pre-processed.
 255          *
 256          * @param int $id
 257          * @param string $title
 258          */
 259         function updateTitle( $id, $title ) {
 260                 $dbw = wfGetDB( DB_MASTER );
 261
 262                 $dbw->update( 'searchindex',
 263                         [ 'si_title' => $title ],
 264                         [ 'si_page' => $id ],
 265                         'SearchOracle::updateTitle',
 266                         [] );
 267         }
 268
 269         public static function legalSearchChars( $type = self::CHARS_ALL ) {
 270                 $searchChars = parent::legalSearchChars( $type );
 271                 if ( $type === self::CHARS_ALL ) {
 272                         $searchChars = "\"" . $searchChars;
 273                 }
 274                 return $searchChars;
 275         }
 276 }