X-Git-Url: https://git.heureux-cyclage.org/?a=blobdiff_plain;f=includes%2FSearchEngine.php;h=c3b38519a7f3e1085fe52df29a13c8ce57a39e75;hb=cdbbe0ad4aaf40c1872d204a68a014d77981e8b7;hp=930ab83160f2a8f10439308cb8dcb1de580c02d9;hpb=bdb843ade7d2c98425982fc912ecce46eb530a89;p=lhc%2Fweb%2Fwiklou.git
diff --git a/includes/SearchEngine.php b/includes/SearchEngine.php
index 930ab83160..c3b38519a7 100644
--- a/includes/SearchEngine.php
+++ b/includes/SearchEngine.php
@@ -1,645 +1,345 @@
rawText = trim( $text );
-
- # We display the query, so let's strip it for safety
- #
- global $wgDBmysql4;
- $lc = SearchEngine::legalSearchChars() . '()';
- if( $wgDBmysql4 ) {
- $lc .= "\"~<>*+-";
- }
- $this->filteredText = trim( preg_replace( "/[^{$lc}]/", " ", $text ) );
- $this->searchTerms = array();
- $this->strictMatching = true; # Google-style, add '+' on all terms
-
- $this->db =& wfGetDB( DB_SLAVE );
- }
+ var $limit = 10;
+ var $offset = 0;
+ var $searchTerms = array();
+ var $namespaces = array( NS_MAIN );
+ var $showRedirects = false;
/**
- * Return a partial WHERE clause to limit the search to the given namespaces
+ * Perform a full text search query and return a result set.
+ * If title searches are not supported or disabled, return null.
+ *
+ * @param string $term - Raw search term
+ * @return SearchResultSet
+ * @access public
+ * @abstract
*/
- function queryNamespaces() {
- $namespaces = implode( ',', $this->namespacesToSearch );
- if ($namespaces == '') {
- $namespaces = '0';
- }
- return "AND cur_namespace IN (" . $namespaces . ')';
+ function searchText( $term ) {
+ return null;
}
/**
- * Return a partial WHERE clause to include or exclude redirects from results
+ * Perform a title-only search query and return a result set.
+ * If title searches are not supported or disabled, return null.
+ *
+ * @param string $term - Raw search term
+ * @return SearchResultSet
+ * @access public
+ * @abstract
*/
- function searchRedirects() {
- if ( $this->doSearchRedirects ) {
- return '';
- } else {
- return 'AND cur_is_redirect=0 ';
- }
+ function searchTitle( $term ) {
+ return null;
}
/**
- * @access private
- */ function initNamespaceCheckbox( $i ) {
- global $wgUser, $wgNamespacesToBeSearchedDefault;
-
- if ($wgUser->getID()) {
- // User is logged in so we retrieve his default namespaces
- return $wgUser->getOption( 'searchNs'.$i );
- } else {
- // User is not logged in so we give him the global default namespaces
- return !empty($wgNamespacesToBeSearchedDefault[ $i ]);
- }
- }
-
- /**
- * Display the "power search" footer. Does not actually perform the search,
- * that is done by showResults()
+ * If an exact title match can be find, or a very slightly close match,
+ * return the title. If no match, returns NULL.
+ *
+ * @static
+ * @param string $term
+ * @return Title
+ * @private
*/
- function powersearch() {
- global $wgUser, $wgOut, $wgContLang, $wgTitle, $wgRequest;
- $sk =& $wgUser->getSkin();
-
- $search = $this->rawText;
- $searchx = $wgRequest->getVal( 'searchx' );
- $listredirs = $wgRequest->getVal( 'redirs' );
-
- $ret = wfMsg('powersearchtext'); # Text to be returned
- $tempText = ''; # Temporary text, for substitution into $ret
-
- if( isset( $_REQUEST['searchx'] ) ) {
- $this->addToQuery['searchx'] = '1';
- }
-
- # Do namespace checkboxes
- $namespaces = $wgContLang->getNamespaces();
- foreach ( $namespaces as $i => $namespace ) {
- # Skip virtual namespaces
- if ( $i < 0 ) {
- continue;
- }
-
- $formVar = 'ns'.$i;
-
- # Initialise checkboxValues, either from defaults or from
- # a previous invocation
- if ( !isset( $searchx ) ) {
- $checkboxValue = $this->initNamespaceCheckbox( $i );
- } else {
- $checkboxValue = $wgRequest->getVal( $formVar );
- }
-
- $checked = '';
- if ( $checkboxValue == 1 ) {
- $checked = ' checked="checked"';
- $this->addToQuery['ns'.$i] = 1;
- array_push( $this->namespacesToSearch, $i );
- }
- $name = str_replace( '_', ' ', $namespaces[$i] );
- if ( '' == $name ) {
- $name = wfMsg( 'blanknamespace' );
- }
+ function getNearMatch( $term ) {
+ # Exact match? No need to look further.
+ $title = Title::newFromText( $term );
+ if (is_null($title))
+ return NULL;
- if ( $tempText !== '' ) {
- $tempText .= ' ';
- }
- $tempText .= "{$name}\n";
+ if ( $title->getNamespace() == NS_SPECIAL || $title->exists() ) {
+ return $title;
}
- $ret = str_replace ( '$1', $tempText, $ret );
-
- # List redirects checkbox
- $checked = '';
- if ( $listredirs == 1 ) {
- $this->addToQuery['redirs'] = 1;
- $checked = ' checked="checked"';
+ # Now try all lower case (i.e. first letter capitalized)
+ #
+ $title = Title::newFromText( strtolower( $term ) );
+ if ( $title->exists() ) {
+ return $title;
}
- $tempText = "\n";
- $ret = str_replace( '$2', $tempText, $ret );
-
- # Search field
-
- $tempText = "\n";
- $ret = str_replace( "$3", $tempText, $ret );
- # Searchx button
+ # Now try capitalized string
+ #
+ $title = Title::newFromText( ucwords( strtolower( $term ) ) );
+ if ( $title->exists() ) {
+ return $title;
+ }
- $tempText = '\n";
- $ret = str_replace( '$9', $tempText, $ret );
+ # Now try all upper case
+ #
+ $title = Title::newFromText( strtoupper( $term ) );
+ if ( $title->exists() ) {
+ return $title;
+ }
- $action = $sk->escapeSearchLink();
- $ret = "
\n
{$top}
\n" ); - # For powersearch - $a2l = ''; - $akk = array_keys( $this->addToQuery ); - foreach ( $akk AS $ak ) { - $a2l .= "&{$ak}={$this->addToQuery[$ak]}" ; + # Quoted term? Try without the quotes... + if( preg_match( '/^"([^"]+)"$/', $term, $matches ) ) { + return SearchEngine::getNearMatch( $matches[1] ); } - $prevnext = wfViewPrevNext( $offset, $limit, '', - 'search=' . wfUrlencode( $this->filteredText ) . $a2l ); - $wgOut->addHTML( "{$prevnext}
\n" ); - $wgOut->addHTML( $powersearch ); + return NULL; } function legalSearchChars() { - $lc = "A-Za-z_'0-9\\x80-\\xFF\\-"; - return $lc; + return "A-Za-z_'0-9\\x80-\\xFF\\-"; } - function parseQuery() { - global $wgDBmysql4; - if (strlen($this->filteredText) < 1) - return MW_SEARCH_BAD_QUERY; - - if( $wgDBmysql4 ) { - # Use cleaner boolean search if available - return $this->parseQuery4(); - } else { - # Fall back to ugly hack with multiple search clauses - return $this->parseQuery3(); - } + /** + * Set the maximum number of results to return + * and how many to skip before returning the first. + * + * @param int $limit + * @param int $offset + * @access public + */ + function setLimitOffset( $limit, $offset = 0 ) { + $this->limit = intval( $limit ); + $this->offset = intval( $offset ); } - - function parseQuery3() { - global $wgDBminWordLen, $wgContLang; - - # on non mysql4 database: get list of words we don't want to search for - require_once( 'FulltextStoplist.php' ); - - $lc = SearchEngine::legalSearchChars() . '()'; - $q = preg_replace( "/([()])/", " \\1 ", $this->filteredText ); - $q = preg_replace( "/\\s+/", " ", $q ); - $w = explode( ' ', trim( $q ) ); - - $last = $cond = ''; - foreach ( $w as $word ) { - $word = $wgContLang->stripForSearch( $word ); - if ( 'and' == $word || 'or' == $word || 'not' == $word - || '(' == $word || ')' == $word ) { - $cond .= ' ' . strtoupper( $word ); - $last = ''; - } else if ( strlen( $word ) < $wgDBminWordLen ) { - continue; - } else if ( FulltextStoplist::inList( $word ) ) { - continue; - } else { - if ( '' != $last ) { $cond .= ' AND'; } - $cond .= " (MATCH (##field##) AGAINST ('" . - $this->db->strencode( $word ). "'))"; - $last = $word; - array_push( $this->searchTerms, "\\b" . $word . "\\b" ); - } - } - if ( 0 == count( $this->searchTerms ) ) { - return MW_SEARCH_BAD_QUERY; - } - $this->titleCond = '(' . str_replace( '##field##', - 'si_title', $cond ) . ' )'; - - $this->textCond = '(' . str_replace( '##field##', - 'si_text', $cond ) . ' AND (cur_is_redirect=0) )'; - - return MW_SEARCH_OK; + /** + * Set which namespaces the search should include. + * Give an array of namespace index numbers. + * + * @param array $namespaces + * @access public + */ + function setNamespaces( $namespaces ) { + $this->namespaces = $namespaces; } - - function parseQuery4() { + + /** + * Make a list of searchable namespaces and their canonical names. + * @return array + * @access public + */ + function searchableNamespaces() { global $wgContLang; - $lc = SearchEngine::legalSearchChars(); - $searchon = ''; - $this->searchTerms = array(); - - # FIXME: This doesn't handle parenthetical expressions. - if( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/', - $this->filteredText, $m, PREG_SET_ORDER ) ) { - foreach( $m as $terms ) { - if( $searchon !== '' ) $searchon .= ' '; - if( $this->strictMatching && ($terms[1] == '') ) { - $terms[1] = '+'; - } - $searchon .= $terms[1] . $wgContLang->stripForSearch( $terms[2] ); - if( !empty( $terms[3] ) ) { - $regexp = preg_quote( $terms[3] ); - if( $terms[4] ) $regexp .= "[0-9A-Za-z_]+"; - } else { - $regexp = preg_quote( str_replace( '"', '', $terms[2] ) ); - } - $this->searchTerms[] = $regexp; + $arr = array(); + foreach( $wgContLang->getNamespaces() as $ns => $name ) { + if( $ns >= NS_MAIN ) { + $arr[$ns] = $name; } - wfDebug( "Would search with '$searchon'\n" ); - wfDebug( "Match with /\b" . implode( '\b|\b', $this->searchTerms ) . "\b/\n" ); - } else { - wfDebug( "Can't understand search query '{$this->filteredText}'\n" ); } - - $searchon = $this->db->strencode( $searchon ); - $this->titleCond = " MATCH(si_title) AGAINST('$searchon' IN BOOLEAN MODE)"; - $this->textCond = " (MATCH(si_text) AGAINST('$searchon' IN BOOLEAN MODE) AND cur_is_redirect=0)"; - return MW_SEARCH_OK; + return $arr; } - function &getMatches( $cond, $limit, $offset = 0 ) { - $searchindex = $this->db->tableName( 'searchindex' ); - $cur = $this->db->tableName( 'cur' ); - $searchnamespaces = $this->queryNamespaces(); - $redircond = $this->searchRedirects(); - - $sql = "SELECT cur_id,cur_namespace,cur_title," . - "cur_text FROM $cur,$searchindex " . - "WHERE cur_id=si_page AND {$cond} " . - "{$searchnamespaces} {$redircond} " . - $this->db->limitResult( $limit, $offset ); - - $res = $this->db->query( $sql, 'SearchEngine::getMatches' ); - $matches = array(); - while ( $row = $this->db->fetchObject( $res ) ) { - $matches[] = $row; - } - $this->db->freeResult( $res ); - - return $matches; + /** + * Return a 'cleaned up' search string + * + * @return string + * @access public + */ + function filter( $text ) { + $lc = $this->legalSearchChars(); + return trim( preg_replace( "/[^{$lc}]/", " ", $text ) ); } - - function showMatches( &$matches, $offset, $msgEmpty, $msgFound ) { - global $wgOut; - if ( 0 == count( $matches ) ) { - $wgOut->addHTML( "' . wfMsg('nogomatch', $editurl ) . "
\n" ); - - # Try a fuzzy title search - $anyhit = false; - global $wgDisableFuzzySearch; - if(! $wgDisableFuzzySearch ){ - foreach( array(NS_MAIN, NS_PROJECT, NS_USER, NS_IMAGE, NS_MEDIAWIKI) as $namespace){ - $anyhit |= SearchEngine::doFuzzyTitleSearch( $this->rawText, $namespace ); - } - } - - if( ! $anyhit ){ - return $this->showResults(); - } + /** + * Some search modes return a total hit count for the query + * in the entire article database. This may include pages + * in namespaces that would not be matched on the given + * settings. + * + * Return null if no total hits number is supported. + * + * @return int + * @access public + */ + function getTotalHits() { + return null; } /** - * @static + * Some search modes return a suggested alternate term if there are + * no exact hits. Returns true if there is one on this set. + * + * @return bool + * @access public */ - function doFuzzyTitleSearch( $search, $namespace ){ - global $wgContLang, $wgOut; - - $this->setupPage(); - - $sstr = ucfirst($search); - $sstr = str_replace(' ', '_', $sstr); - $fuzzymatches = SearchEngine::fuzzyTitles( $sstr, $namespace ); - $fuzzymatches = array_slice($fuzzymatches, 0, 10); - $slen = strlen( $search ); - $wikitext = ''; - foreach($fuzzymatches as $res){ - $t = str_replace('_', ' ', $res[1]); - $tfull = $wgContLang->getNsText( $namespace ) . ":$t|$t"; - if( $namespace == NS_MAIN ) - $tfull = $t; - $distance = $res[0]; - $closeness = (strlen( $search ) - $distance) / strlen( $search ); - $percent = intval( $closeness * 100 ) . '%'; - $stars = str_repeat('*', ceil(5 * $closeness) ); - $wikitext .= "* [[$tfull]] $percent ($stars)\n"; - } - if( $wikitext ){ - if( $namespace != NS_MAIN ) - $wikitext = '=== ' . $wgContLang->getNsText( $namespace ) . " ===\n" . $wikitext; - $wgOut->addWikiText( $wikitext ); - return true; - } + function hasSuggestion() { return false; } /** - * @static + * Some search modes return a suggested alternate term if there are + * no exact hits. Check hasSuggestion() first. + * + * @return string + * @access public */ - function fuzzyTitles( $sstr, $namespace = NS_MAIN ){ - $span = 0.10; // weed on title length before doing levenshtein. - $tolerance = 0.35; // allowed percentage of erronous characters - $slen = strlen($sstr); - $tolerance_count = ceil($tolerance * $slen); - $spanabs = ceil($slen * (1 + $span)) - $slen; - # print "Word: $sstr, len = $slen, range = [$min, $max], tolerance_count = $tolerance_count