includes/search/SearchSqlite.php

   1 <?php
   2 /**
   3  * SQLite search backend, based upon SearchMysql
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Search
  22  */
  23
  24 /**
  25  * Search engine hook for SQLite
  26  * @ingroup Search
  27  */
  28 class SearchSqlite extends SearchDatabase {
  29         /**
  30          * Whether fulltext search is supported by current schema
  31          * @return bool
  32          */
  33         function fulltextSearchSupported() {
  34                 return $this->db->checkForEnabledSearch();
  35         }
  36
  37         /**
  38          * Parse the user's query and transform it into an SQL fragment which will
  39          * become part of a WHERE clause
  40          *
  41          * @param string $filteredText
  42          * @param bool $fulltext
  43          * @return string
  44          */
  45         function parseQuery( $filteredText, $fulltext ) {
  46                 global $wgContLang;
  47                 $lc = $this->legalSearchChars(); // Minus format chars
  48                 $searchon = '';
  49                 $this->searchTerms = [];
  50
  51                 $m = [];
  52                 if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  53                                 $filteredText, $m, PREG_SET_ORDER ) ) {
  54                         foreach ( $m as $bits ) {
  55                                 MediaWiki\suppressWarnings();
  56                                 list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  57                                 MediaWiki\restoreWarnings();
  58
  59                                 if ( $nonQuoted != '' ) {
  60                                         $term = $nonQuoted;
  61                                         $quote = '';
  62                                 } else {
  63                                         $term = str_replace( '"', '', $term );
  64                                         $quote = '"';
  65                                 }
  66
  67                                 if ( $searchon !== '' ) {
  68                                         $searchon .= ' ';
  69                                 }
  70
  71                                 // Some languages such as Serbian store the input form in the search index,
  72                                 // so we may need to search for matches in multiple writing system variants.
  73                                 $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
  74                                 if ( is_array( $convertedVariants ) ) {
  75                                         $variants = array_unique( array_values( $convertedVariants ) );
  76                                 } else {
  77                                         $variants = [ $term ];
  78                                 }
  79
  80                                 // The low-level search index does some processing on input to work
  81                                 // around problems with minimum lengths and encoding in MySQL's
  82                                 // fulltext engine.
  83                                 // For Chinese this also inserts spaces between adjacent Han characters.
  84                                 $strippedVariants = array_map(
  85                                         [ $wgContLang, 'normalizeForSearch' ],
  86                                         $variants );
  87
  88                                 // Some languages such as Chinese force all variants to a canonical
  89                                 // form when stripping to the low-level search index, so to be sure
  90                                 // let's check our variants list for unique items after stripping.
  91                                 $strippedVariants = array_unique( $strippedVariants );
  92
  93                                 $searchon .= $modifier;
  94                                 if ( count( $strippedVariants ) > 1 ) {
  95                                         $searchon .= '(';
  96                                 }
  97                                 foreach ( $strippedVariants as $stripped ) {
  98                                         if ( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
  99                                                 // Hack for Chinese: we need to toss in quotes for
 100                                                 // multiple-character phrases since normalizeForSearch()
 101                                                 // added spaces between them to make word breaks.
 102                                                 $stripped = '"' . trim( $stripped ) . '"';
 103                                         }
 104                                         $searchon .= "$quote$stripped$quote$wildcard ";
 105                                 }
 106                                 if ( count( $strippedVariants ) > 1 ) {
 107                                         $searchon .= ')';
 108                                 }
 109
 110                                 // Match individual terms or quoted phrase in result highlighting...
 111                                 // Note that variants will be introduced in a later stage for highlighting!
 112                                 $regexp = $this->regexTerm( $term, $wildcard );
 113                                 $this->searchTerms[] = $regexp;
 114                         }
 115
 116                 } else {
 117                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 118                 }
 119
 120                 $searchon = $this->db->addQuotes( $searchon );
 121                 $field = $this->getIndexField( $fulltext );
 122                 return " $field MATCH $searchon ";
 123         }
 124
 125         function regexTerm( $string, $wildcard ) {
 126                 global $wgContLang;
 127
 128                 $regex = preg_quote( $string, '/' );
 129                 if ( $wgContLang->hasWordBreaks() ) {
 130                         if ( $wildcard ) {
 131                                 // Don't cut off the final bit!
 132                                 $regex = "\b$regex";
 133                         } else {
 134                                 $regex = "\b$regex\b";
 135                         }
 136                 } else {
 137                         // For Chinese, words may legitimately abut other words in the text literal.
 138                         // Don't add \b boundary checks... note this could cause false positives
 139                         // for latin chars.
 140                 }
 141                 return $regex;
 142         }
 143
 144         public function legalSearchCharsForUpdate() {
 145                 return "\"*" . parent::legalSearchCharsForUpdate();
 146         }
 147
 148         /**
 149          * Perform a full text search query and return a result set.
 150          *
 151          * @param string $term Raw search term
 152          * @return SqlSearchResultSet
 153          */
 154         function searchText( $term ) {
 155                 return $this->searchInternal( $term, true );
 156         }
 157
 158         /**
 159          * Perform a title-only search query and return a result set.
 160          *
 161          * @param string $term Raw search term
 162          * @return SqlSearchResultSet
 163          */
 164         function searchTitle( $term ) {
 165                 return $this->searchInternal( $term, false );
 166         }
 167
 168         protected function searchInternal( $term, $fulltext ) {
 169                 global $wgContLang;
 170
 171                 if ( !$this->fulltextSearchSupported() ) {
 172                         return null;
 173                 }
 174
 175                 $filteredTerm = $this->filter( $wgContLang->lc( $term ) );
 176                 $resultSet = $this->db->query( $this->getQuery( $filteredTerm, $fulltext ) );
 177
 178                 $total = null;
 179                 $totalResult = $this->db->query( $this->getCountQuery( $filteredTerm, $fulltext ) );
 180                 $row = $totalResult->fetchObject();
 181                 if ( $row ) {
 182                         $total = intval( $row->c );
 183                 }
 184                 $totalResult->free();
 185
 186                 return new SqlSearchResultSet( $resultSet, $this->searchTerms, $total );
 187         }
 188
 189         /**
 190          * Return a partial WHERE clause to limit the search to the given namespaces
 191          * @return string
 192          */
 193         function queryNamespaces() {
 194                 if ( is_null( $this->namespaces ) ) {
 195                         return '';  # search all
 196                 }
 197                 if ( !count( $this->namespaces ) ) {
 198                         $namespaces = '0';
 199                 } else {
 200                         $namespaces = $this->db->makeList( $this->namespaces );
 201                 }
 202                 return 'AND page_namespace IN (' . $namespaces . ')';
 203         }
 204
 205         /**
 206          * Returns a query with limit for number of results set.
 207          * @param string $sql
 208          * @return string
 209          */
 210         function limitResult( $sql ) {
 211                 return $this->db->limitResult( $sql, $this->limit, $this->offset );
 212         }
 213
 214         /**
 215          * Construct the full SQL query to do the search.
 216          * The guts shoulds be constructed in queryMain()
 217          * @param string $filteredTerm
 218          * @param bool $fulltext
 219          * @return string
 220          */
 221         function getQuery( $filteredTerm, $fulltext ) {
 222                 return $this->limitResult(
 223                         $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 224                         $this->queryNamespaces()
 225                 );
 226         }
 227
 228         /**
 229          * Picks which field to index on, depending on what type of query.
 230          * @param bool $fulltext
 231          * @return string
 232          */
 233         function getIndexField( $fulltext ) {
 234                 return $fulltext ? 'si_text' : 'si_title';
 235         }
 236
 237         /**
 238          * Get the base part of the search query.
 239          *
 240          * @param string $filteredTerm
 241          * @param bool $fulltext
 242          * @return string
 243          */
 244         function queryMain( $filteredTerm, $fulltext ) {
 245                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 246                 $page = $this->db->tableName( 'page' );
 247                 $searchindex = $this->db->tableName( 'searchindex' );
 248                 return "SELECT $searchindex.rowid, page_namespace, page_title " .
 249                         "FROM $page,$searchindex " .
 250                         "WHERE page_id=$searchindex.rowid AND $match";
 251         }
 252
 253         function getCountQuery( $filteredTerm, $fulltext ) {
 254                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 255                 $page = $this->db->tableName( 'page' );
 256                 $searchindex = $this->db->tableName( 'searchindex' );
 257                 return "SELECT COUNT(*) AS c " .
 258                         "FROM $page,$searchindex " .
 259                         "WHERE page_id=$searchindex.rowid AND $match " .
 260                         $this->queryNamespaces();
 261         }
 262
 263         /**
 264          * Create or update the search index record for the given page.
 265          * Title and text should be pre-processed.
 266          *
 267          * @param int $id
 268          * @param string $title
 269          * @param string $text
 270          */
 271         function update( $id, $title, $text ) {
 272                 if ( !$this->fulltextSearchSupported() ) {
 273                         return;
 274                 }
 275                 // @todo find a method to do it in a single request,
 276                 // couldn't do it so far due to typelessness of FTS3 tables.
 277                 $dbw = wfGetDB( DB_MASTER );
 278
 279                 $dbw->delete( 'searchindex', [ 'rowid' => $id ], __METHOD__ );
 280
 281                 $dbw->insert( 'searchindex',
 282                         [
 283                                 'rowid' => $id,
 284                                 'si_title' => $title,
 285                                 'si_text' => $text
 286                         ], __METHOD__ );
 287         }
 288
 289         /**
 290          * Update a search index record's title only.
 291          * Title should be pre-processed.
 292          *
 293          * @param int $id
 294          * @param string $title
 295          */
 296         function updateTitle( $id, $title ) {
 297                 if ( !$this->fulltextSearchSupported() ) {
 298                         return;
 299                 }
 300                 $dbw = wfGetDB( DB_MASTER );
 301
 302                 $dbw->update( 'searchindex',
 303                         [ 'si_title' => $title ],
 304                         [ 'rowid' => $id ],
 305                         __METHOD__ );
 306         }
 307 }