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 use MediaWiki\MediaWikiServices;
  28
  29 /**
  30  * Search engine hook base class for Oracle (ConText).
  31  * @ingroup Search
  32  */
  33 class SearchOracle extends SearchDatabase {
  34         private $reservedWords = [
  35                 'ABOUT' => 1,
  36                 'ACCUM' => 1,
  37                 'AND' => 1,
  38                 'BT' => 1,
  39                 'BTG' => 1,
  40                 'BTI' => 1,
  41                 'BTP' => 1,
  42                 'FUZZY' => 1,
  43                 'HASPATH' => 1,
  44                 'INPATH' => 1,
  45                 'MINUS' => 1,
  46                 'NEAR' => 1,
  47                 'NOT' => 1,
  48                 'NT' => 1,
  49                 'NTG' => 1,
  50                 'NTI' => 1,
  51                 'NTP' => 1,
  52                 'OR' => 1,
  53                 'PT' => 1,
  54                 'RT' => 1,
  55                 'SQE' => 1,
  56                 'SYN' => 1,
  57                 'TR' => 1,
  58                 'TRSYN' => 1,
  59                 'TT' => 1,
  60                 'WITHIN' => 1,
  61         ];
  62
  63         /**
  64          * Perform a full text search query and return a result set.
  65          *
  66          * @param string $term Raw search term
  67          * @return SqlSearchResultSet
  68          */
  69         protected function doSearchTextInDB( $term ) {
  70                 if ( $term == '' ) {
  71                         return new SqlSearchResultSet( false, '' );
  72                 }
  73
  74                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
  75                 $resultSet = $dbr->query( $this->getQuery( $this->filter( $term ), true ) );
  76                 return new SqlSearchResultSet( $resultSet, $this->searchTerms );
  77         }
  78
  79         /**
  80          * Perform a title-only search query and return a result set.
  81          *
  82          * @param string $term Raw search term
  83          * @return SqlSearchResultSet
  84          */
  85         protected function doSearchTitleInDB( $term ) {
  86                 if ( $term == '' ) {
  87                         return new SqlSearchResultSet( false, '' );
  88                 }
  89
  90                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
  91                 $resultSet = $dbr->query( $this->getQuery( $this->filter( $term ), false ) );
  92                 return new SqlSearchResultSet( $resultSet, $this->searchTerms );
  93         }
  94
  95         /**
  96          * Return a partial WHERE clause to limit the search to the given namespaces
  97          * @return string
  98          */
  99         private function queryNamespaces() {
 100                 if ( is_null( $this->namespaces ) ) {
 101                         return '';
 102                 }
 103                 if ( $this->namespaces === [] ) {
 104                         $namespaces = '0';
 105                 } else {
 106                         $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 107                         $namespaces = $dbr->makeList( $this->namespaces );
 108                 }
 109                 return 'AND page_namespace IN (' . $namespaces . ')';
 110         }
 111
 112         /**
 113          * Return a LIMIT clause to limit results on the query.
 114          *
 115          * @param string $sql
 116          *
 117          * @return string
 118          */
 119         private function queryLimit( $sql ) {
 120                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 121
 122                 return $dbr->limitResult( $sql, $this->limit, $this->offset );
 123         }
 124
 125         /**
 126          * Does not do anything for generic search engine
 127          * subclasses may define this though
 128          *
 129          * @param string $filteredTerm
 130          * @param bool $fulltext
 131          * @return string
 132          */
 133         function queryRanking( $filteredTerm, $fulltext ) {
 134                 return ' ORDER BY score(1)';
 135         }
 136
 137         /**
 138          * Construct the full SQL query to do the search.
 139          * The guts shoulds be constructed in queryMain()
 140          * @param string $filteredTerm
 141          * @param bool $fulltext
 142          * @return string
 143          */
 144         private function getQuery( $filteredTerm, $fulltext ) {
 145                 return $this->queryLimit( $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 146                         $this->queryNamespaces() . ' ' .
 147                         $this->queryRanking( $filteredTerm, $fulltext ) . ' ' );
 148         }
 149
 150         /**
 151          * Picks which field to index on, depending on what type of query.
 152          * @param bool $fulltext
 153          * @return string
 154          */
 155         private function getIndexField( $fulltext ) {
 156                 return $fulltext ? 'si_text' : 'si_title';
 157         }
 158
 159         /**
 160          * Get the base part of the search query.
 161          *
 162          * @param string $filteredTerm
 163          * @param bool $fulltext
 164          * @return string
 165          */
 166         function queryMain( $filteredTerm, $fulltext ) {
 167                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 168
 169                 $dbr = $this->lb->getMaintenanceConnectionRef( DB_REPLICA );
 170                 $page = $dbr->tableName( 'page' );
 171                 $searchindex = $dbr->tableName( 'searchindex' );
 172
 173                 return 'SELECT page_id, page_namespace, page_title ' .
 174                         "FROM $page,$searchindex " .
 175                         'WHERE page_id=si_page AND ' . $match;
 176         }
 177
 178         /**
 179          * Parse a user input search string, and return an SQL fragment to be used
 180          * as part of a WHERE clause
 181          * @param string $filteredText
 182          * @param bool $fulltext
 183          * @return string
 184          */
 185         private function parseQuery( $filteredText, $fulltext ) {
 186                 $lc = $this->legalSearchChars( self::CHARS_NO_SYNTAX );
 187                 $this->searchTerms = [];
 188
 189                 # @todo FIXME: This doesn't handle parenthetical expressions.
 190                 $m = [];
 191                 $searchon = '';
 192                 if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
 193                                 $filteredText, $m, PREG_SET_ORDER ) ) {
 194                         foreach ( $m as $terms ) {
 195                                 // Search terms in all variant forms, only
 196                                 // apply on wiki with LanguageConverter
 197                                 $temp_terms = MediaWikiServices::getInstance()->getContentLanguage()->
 198                                         autoConvertToAllVariants( $terms[2] );
 199                                 if ( is_array( $temp_terms ) ) {
 200                                         $temp_terms = array_unique( array_values( $temp_terms ) );
 201                                         foreach ( $temp_terms as $t ) {
 202                                                 $searchon .= ( $terms[1] == '-' ? ' ~' : ' & ' ) . $this->escapeTerm( $t );
 203                                         }
 204                                 } else {
 205                                         $searchon .= ( $terms[1] == '-' ? ' ~' : ' & ' ) . $this->escapeTerm( $terms[2] );
 206                                 }
 207                                 if ( !empty( $terms[3] ) ) {
 208                                         $regexp = preg_quote( $terms[3], '/' );
 209                                         if ( $terms[4] ) {
 210                                                 $regexp .= "[0-9A-Za-z_]+";
 211                                         }
 212                                 } else {
 213                                         $regexp = preg_quote( str_replace( '"', '', $terms[2] ), '/' );
 214                                 }
 215                                 $this->searchTerms[] = $regexp;
 216                         }
 217                 }
 218
 219                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 220                 $searchon = $dbr->addQuotes( ltrim( $searchon, ' &' ) );
 221                 $field = $this->getIndexField( $fulltext );
 222
 223                 return " CONTAINS($field, $searchon, 1) > 0 ";
 224         }
 225
 226         private function escapeTerm( $t ) {
 227                 $t = MediaWikiServices::getInstance()->getContentLanguage()->normalizeForSearch( $t );
 228                 $t = isset( $this->reservedWords[strtoupper( $t )] ) ? '{' . $t . '}' : $t;
 229                 $t = preg_replace( '/^"(.*)"$/', '($1)', $t );
 230                 $t = preg_replace( '/([-&|])/', '\\\\$1', $t );
 231                 return $t;
 232         }
 233
 234         /**
 235          * Create or update the search index record for the given page.
 236          * Title and text should be pre-processed.
 237          *
 238          * @param int $id
 239          * @param string $title
 240          * @param string $text
 241          */
 242         function update( $id, $title, $text ) {
 243                 $dbw = $this->lb->getConnection( DB_MASTER );
 244                 $dbw->replace( 'searchindex',
 245                         [ 'si_page' ],
 246                         [
 247                                 'si_page' => $id,
 248                                 'si_title' => $title,
 249                                 'si_text' => $text
 250                         ], 'SearchOracle::update' );
 251
 252                 // Sync the index
 253                 // We need to specify the DB name (i.e. user/schema) here so that
 254                 // it can work from the installer, where
 255                 //     ALTER SESSION SET CURRENT_SCHEMA = ...
 256                 // was used.
 257                 $dbw->query( "CALL ctx_ddl.sync_index(" .
 258                         $dbw->addQuotes( $dbw->getDBname() . '.' . $dbw->tableName( 'si_text_idx', 'raw' ) ) . ")" );
 259                 $dbw->query( "CALL ctx_ddl.sync_index(" .
 260                         $dbw->addQuotes( $dbw->getDBname() . '.' . $dbw->tableName( 'si_title_idx', 'raw' ) ) . ")" );
 261         }
 262
 263         /**
 264          * Update a search index record's title only.
 265          * Title should be pre-processed.
 266          *
 267          * @param int $id
 268          * @param string $title
 269          */
 270         function updateTitle( $id, $title ) {
 271                 $dbw = $this->lb->getConnectionRef( DB_MASTER );
 272                 $dbw->update( 'searchindex',
 273                         [ 'si_title' => $title ],
 274                         [ 'si_page' => $id ],
 275                         'SearchOracle::updateTitle',
 276                         [] );
 277         }
 278
 279         public function legalSearchChars( $type = self::CHARS_ALL ) {
 280                 $searchChars = parent::legalSearchChars( $type );
 281                 if ( $type === self::CHARS_ALL ) {
 282                         $searchChars = "\"" . $searchChars;
 283                 }
 284                 return $searchChars;
 285         }
 286 }