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