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 use MediaWiki\MediaWikiServices;
  25 use Wikimedia\Rdbms\DatabaseSqlite;
  26
  27 /**
  28  * Search engine hook for SQLite
  29  * @ingroup Search
  30  */
  31 class SearchSqlite extends SearchDatabase {
  32         /**
  33          * Whether fulltext search is supported by current schema
  34          * @return bool
  35          */
  36         private function fulltextSearchSupported() {
  37                 // Avoid getConnectionRef() in order to get DatabaseSqlite specifically
  38                 /** @var DatabaseSqlite $dbr */
  39                 $dbr = $this->lb->getConnection( DB_REPLICA );
  40                 try {
  41                         return $dbr->checkForEnabledSearch();
  42                 } finally {
  43                         $this->lb->reuseConnection( $dbr );
  44                 }
  45         }
  46
  47         /**
  48          * Parse the user's query and transform it into an SQL fragment which will
  49          * become part of a WHERE clause
  50          *
  51          * @param string $filteredText
  52          * @param bool $fulltext
  53          * @return string
  54          */
  55         private function parseQuery( $filteredText, $fulltext ) {
  56                 $lc = $this->legalSearchChars( self::CHARS_NO_SYNTAX ); // Minus syntax chars (" and *)
  57                 $searchon = '';
  58                 $this->searchTerms = [];
  59
  60                 $m = [];
  61                 if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
  62                                 $filteredText, $m, PREG_SET_ORDER ) ) {
  63                         foreach ( $m as $bits ) {
  64                                 Wikimedia\suppressWarnings();
  65                                 list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
  66                                 Wikimedia\restoreWarnings();
  67
  68                                 if ( $nonQuoted != '' ) {
  69                                         $term = $nonQuoted;
  70                                         $quote = '';
  71                                 } else {
  72                                         $term = str_replace( '"', '', $term );
  73                                         $quote = '"';
  74                                 }
  75
  76                                 if ( $searchon !== '' ) {
  77                                         $searchon .= ' ';
  78                                 }
  79
  80                                 // Some languages such as Serbian store the input form in the search index,
  81                                 // so we may need to search for matches in multiple writing system variants.
  82                                 $convertedVariants = MediaWikiServices::getInstance()->getContentLanguage()->
  83                                         autoConvertToAllVariants( $term );
  84                                 if ( is_array( $convertedVariants ) ) {
  85                                         $variants = array_unique( array_values( $convertedVariants ) );
  86                                 } else {
  87                                         $variants = [ $term ];
  88                                 }
  89
  90                                 // The low-level search index does some processing on input to work
  91                                 // around problems with minimum lengths and encoding in MySQL's
  92                                 // fulltext engine.
  93                                 // For Chinese this also inserts spaces between adjacent Han characters.
  94                                 $strippedVariants = array_map(
  95                                         [ MediaWikiServices::getInstance()->getContentLanguage(),
  96                                                 'normalizeForSearch' ],
  97                                         $variants );
  98
  99                                 // Some languages such as Chinese force all variants to a canonical
 100                                 // form when stripping to the low-level search index, so to be sure
 101                                 // let's check our variants list for unique items after stripping.
 102                                 $strippedVariants = array_unique( $strippedVariants );
 103
 104                                 $searchon .= $modifier;
 105                                 if ( count( $strippedVariants ) > 1 ) {
 106                                         $searchon .= '(';
 107                                 }
 108                                 foreach ( $strippedVariants as $stripped ) {
 109                                         if ( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
 110                                                 // Hack for Chinese: we need to toss in quotes for
 111                                                 // multiple-character phrases since normalizeForSearch()
 112                                                 // added spaces between them to make word breaks.
 113                                                 $stripped = '"' . trim( $stripped ) . '"';
 114                                         }
 115                                         $searchon .= "$quote$stripped$quote$wildcard ";
 116                                 }
 117                                 if ( count( $strippedVariants ) > 1 ) {
 118                                         $searchon .= ')';
 119                                 }
 120
 121                                 // Match individual terms or quoted phrase in result highlighting...
 122                                 // Note that variants will be introduced in a later stage for highlighting!
 123                                 $regexp = $this->regexTerm( $term, $wildcard );
 124                                 $this->searchTerms[] = $regexp;
 125                         }
 126
 127                 } else {
 128                         wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
 129                 }
 130
 131                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 132                 $searchon = $dbr->addQuotes( $searchon );
 133                 $field = $this->getIndexField( $fulltext );
 134
 135                 return " $field MATCH $searchon ";
 136         }
 137
 138         private function regexTerm( $string, $wildcard ) {
 139                 $regex = preg_quote( $string, '/' );
 140                 if ( MediaWikiServices::getInstance()->getContentLanguage()->hasWordBreaks() ) {
 141                         if ( $wildcard ) {
 142                                 // Don't cut off the final bit!
 143                                 $regex = "\b$regex";
 144                         } else {
 145                                 $regex = "\b$regex\b";
 146                         }
 147                 } else {
 148                         // For Chinese, words may legitimately abut other words in the text literal.
 149                         // Don't add \b boundary checks... note this could cause false positives
 150                         // for Latin chars.
 151                 }
 152                 return $regex;
 153         }
 154
 155         public function legalSearchChars( $type = self::CHARS_ALL ) {
 156                 $searchChars = parent::legalSearchChars( $type );
 157                 if ( $type === self::CHARS_ALL ) {
 158                         // " for phrase, * for wildcard
 159                         $searchChars = "\"*" . $searchChars;
 160                 }
 161                 return $searchChars;
 162         }
 163
 164         /**
 165          * Perform a full text search query and return a result set.
 166          *
 167          * @param string $term Raw search term
 168          * @return SqlSearchResultSet|null
 169          */
 170         protected function doSearchTextInDB( $term ) {
 171                 return $this->searchInternal( $term, true );
 172         }
 173
 174         /**
 175          * Perform a title-only search query and return a result set.
 176          *
 177          * @param string $term Raw search term
 178          * @return SqlSearchResultSet|null
 179          */
 180         protected function doSearchTitleInDB( $term ) {
 181                 return $this->searchInternal( $term, false );
 182         }
 183
 184         protected function searchInternal( $term, $fulltext ) {
 185                 if ( !$this->fulltextSearchSupported() ) {
 186                         return null;
 187                 }
 188
 189                 $filteredTerm =
 190                         $this->filter( MediaWikiServices::getInstance()->getContentLanguage()->lc( $term ) );
 191                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 192                 $resultSet = $dbr->query( $this->getQuery( $filteredTerm, $fulltext ) );
 193
 194                 $total = null;
 195                 $totalResult = $dbr->query( $this->getCountQuery( $filteredTerm, $fulltext ) );
 196                 $row = $totalResult->fetchObject();
 197                 if ( $row ) {
 198                         $total = intval( $row->c );
 199                 }
 200                 $totalResult->free();
 201
 202                 return new SqlSearchResultSet( $resultSet, $this->searchTerms, $total );
 203         }
 204
 205         /**
 206          * Return a partial WHERE clause to limit the search to the given namespaces
 207          * @return string
 208          */
 209         private function queryNamespaces() {
 210                 if ( is_null( $this->namespaces ) ) {
 211                         return '';  # search all
 212                 }
 213                 if ( $this->namespaces === [] ) {
 214                         $namespaces = '0';
 215                 } else {
 216                         $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 217                         $namespaces = $dbr->makeList( $this->namespaces );
 218                 }
 219                 return 'AND page_namespace IN (' . $namespaces . ')';
 220         }
 221
 222         /**
 223          * Returns a query with limit for number of results set.
 224          * @param string $sql
 225          * @return string
 226          */
 227         private function limitResult( $sql ) {
 228                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 229
 230                 return $dbr->limitResult( $sql, $this->limit, $this->offset );
 231         }
 232
 233         /**
 234          * Construct the full SQL query to do the search.
 235          * The guts shoulds be constructed in queryMain()
 236          * @param string $filteredTerm
 237          * @param bool $fulltext
 238          * @return string
 239          */
 240         private function getQuery( $filteredTerm, $fulltext ) {
 241                 return $this->limitResult(
 242                         $this->queryMain( $filteredTerm, $fulltext ) . ' ' .
 243                         $this->queryNamespaces()
 244                 );
 245         }
 246
 247         /**
 248          * Picks which field to index on, depending on what type of query.
 249          * @param bool $fulltext
 250          * @return string
 251          */
 252         private function getIndexField( $fulltext ) {
 253                 return $fulltext ? 'si_text' : 'si_title';
 254         }
 255
 256         /**
 257          * Get the base part of the search query.
 258          *
 259          * @param string $filteredTerm
 260          * @param bool $fulltext
 261          * @return string
 262          */
 263         private function queryMain( $filteredTerm, $fulltext ) {
 264                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 265                 $dbr = $this->lb->getMaintenanceConnectionRef( DB_REPLICA );
 266                 $page = $dbr->tableName( 'page' );
 267                 $searchindex = $dbr->tableName( 'searchindex' );
 268                 return "SELECT $searchindex.rowid, page_namespace, page_title " .
 269                         "FROM $page,$searchindex " .
 270                         "WHERE page_id=$searchindex.rowid AND $match";
 271         }
 272
 273         private function getCountQuery( $filteredTerm, $fulltext ) {
 274                 $match = $this->parseQuery( $filteredTerm, $fulltext );
 275                 $dbr = $this->lb->getMaintenanceConnectionRef( DB_REPLICA );
 276                 $page = $dbr->tableName( 'page' );
 277                 $searchindex = $dbr->tableName( 'searchindex' );
 278                 return "SELECT COUNT(*) AS c " .
 279                         "FROM $page,$searchindex " .
 280                         "WHERE page_id=$searchindex.rowid AND $match " .
 281                         $this->queryNamespaces();
 282         }
 283
 284         /**
 285          * Create or update the search index record for the given page.
 286          * Title and text should be pre-processed.
 287          *
 288          * @param int $id
 289          * @param string $title
 290          * @param string $text
 291          */
 292         public function update( $id, $title, $text ) {
 293                 if ( !$this->fulltextSearchSupported() ) {
 294                         return;
 295                 }
 296                 // @todo find a method to do it in a single request,
 297                 // couldn't do it so far due to typelessness of FTS3 tables.
 298                 $dbw = $this->lb->getConnectionRef( DB_MASTER );
 299                 $dbw->delete( 'searchindex', [ 'rowid' => $id ], __METHOD__ );
 300                 $dbw->insert( 'searchindex',
 301                         [
 302                                 'rowid' => $id,
 303                                 'si_title' => $title,
 304                                 'si_text' => $text
 305                         ], __METHOD__ );
 306         }
 307
 308         /**
 309          * Update a search index record's title only.
 310          * Title should be pre-processed.
 311          *
 312          * @param int $id
 313          * @param string $title
 314          */
 315         public function updateTitle( $id, $title ) {
 316                 if ( !$this->fulltextSearchSupported() ) {
 317                         return;
 318                 }
 319
 320                 $dbw = $this->lb->getConnectionRef( DB_MASTER );
 321                 $dbw->update( 'searchindex',
 322                         [ 'si_title' => $title ],
 323                         [ 'rowid' => $id ],
 324                         __METHOD__ );
 325         }
 326 }