<?php
-# See search.doc
-
+/**
+ * Contain site class
+ * See search.doc
+ * @package MediaWiki
+ */
+
+/**
+ *
+ */
+define( 'MW_SEARCH_OK', true );
+define( 'MW_SEARCH_BAD_QUERY', false );
+
+/**
+ * @todo document
+ * @package MediaWiki
+ */
class SearchEngine {
- /* private */ var $mUsertext, $mSearchterms;
- /* private */ var $mTitlecond, $mTextcond;
+ /* private */ var $rawText, $filteredText, $searchTerms;
+ /* private */ var $titleCond, $textCond;
var $doSearchRedirects = true;
- var $addtoquery = array();
+ var $addToQuery = array();
var $namespacesToSearch = array();
var $alternateTitle;
- var $all_titles = false;
+ var $allTitles = false;
+
+ function SearchEngine( $text ) {
+ $this->rawText = trim( $text );
- function SearchEngine( $text )
- {
# We display the query, so let's strip it for safety
#
global $wgDBmysql4;
- $lc = SearchEngine::legalSearchChars() . "()";
- if( $wgDBmysql4 ) $lc .= "\"~<>*+-";
- $this->mUsertext = trim( preg_replace( "/[^{$lc}]/", " ", $text ) );
- $this->mSearchterms = array();
- $this->mStrictMatching = true; # Google-style, add '+' on all terms
+ $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 );
}
- function queryNamespaces()
- {
- $namespaces = implode( ",", $this->namespacesToSearch );
- if ($namespaces == "") {
- $namespaces = "0";
+ /**
+ * Return a partial WHERE clause to limit the search to the given namespaces
+ */
+ function queryNamespaces() {
+ $namespaces = implode( ',', $this->namespacesToSearch );
+ if ($namespaces == '') {
+ $namespaces = '0';
}
- return "AND cur_namespace IN (" . $namespaces . ")";
+ return "AND cur_namespace IN (" . $namespaces . ')';
}
- function searchRedirects()
- {
+ /**
+ * Return a partial WHERE clause to include or exclude redirects from results
+ */
+ function searchRedirects() {
if ( $this->doSearchRedirects ) {
- return "";
+ return '';
} else {
- return "AND cur_is_redirect=0 ";
+ return 'AND cur_is_redirect=0 ';
}
}
- /* private */ function initNamespaceCheckbox( $i )
- {
+ /**
+ * @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 );
+ 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()
- function powersearch()
- {
- global $wgUser, $wgOut, $wgLang, $wgTitle, $wgRequest;
+ /**
+ * Display the "power search" footer. Does not actually perform the search,
+ * that is done by showResults()
+ */
+ function powersearch() {
+ global $wgUser, $wgOut, $wgContLang, $wgTitle, $wgRequest;
$sk =& $wgUser->getSkin();
- $search = $wgRequest->getText( 'search' );
+ $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
+ $ret = wfMsg('powersearchtext'); # Text to be returned
+ $tempText = ''; # Temporary text, for substitution into $ret
- if( isset( $_REQUEST["searchx"] ) ) {
- $this->addtoquery["searchx"] = "1";
+ if( isset( $_REQUEST['searchx'] ) ) {
+ $this->addToQuery['searchx'] = '1';
}
# Do namespace checkboxes
- $namespaces = $wgLang->getNamespaces();
+ $namespaces = $wgContLang->getNamespaces();
foreach ( $namespaces as $i => $namespace ) {
# Skip virtual namespaces
if ( $i < 0 ) {
continue;
}
- $formVar = "ns$i";
+ $formVar = 'ns'.$i;
# Initialise checkboxValues, either from defaults or from
# a previous invocation
$checkboxValue = $wgRequest->getVal( $formVar );
}
- $checked = "";
+ $checked = '';
if ( $checkboxValue == 1 ) {
- $checked = " checked='checked'";
- $this->addtoquery["ns{$i}"] = 1;
+ $checked = ' checked="checked"';
+ $this->addToQuery['ns'.$i] = 1;
array_push( $this->namespacesToSearch, $i );
}
- $name = str_replace( "_", " ", $namespaces[$i] );
- if ( "" == $name ) {
- $name = wfMsg( "blanknamespace" );
+ $name = str_replace( '_', ' ', $namespaces[$i] );
+ if ( '' == $name ) {
+ $name = wfMsg( 'blanknamespace' );
}
- if ( $tempText !== "" ) {
- $tempText .= " ";
+ if ( $tempText !== '' ) {
+ $tempText .= ' ';
}
$tempText .= "<input type='checkbox' value=\"1\" name=\"" .
"ns{$i}\"{$checked} />{$name}\n";
}
- $ret = str_replace ( "$1", $tempText, $ret );
+ $ret = str_replace ( '$1', $tempText, $ret );
# List redirects checkbox
- $checked = "";
+ $checked = '';
if ( $listredirs == 1 ) {
- $this->addtoquery["redirs"] = 1;
- $checked = " checked='checked'";
+ $this->addToQuery['redirs'] = 1;
+ $checked = ' checked="checked"';
}
$tempText = "<input type='checkbox' value='1' name=\"redirs\"{$checked} />\n";
- $ret = str_replace( "$2", $tempText, $ret );
+ $ret = str_replace( '$2', $tempText, $ret );
# Search field
$tempText = "<input type='text' name=\"search\" value=\"" .
- htmlspecialchars( $search ) ."\" width='80' />\n";
+ htmlspecialchars( $search ) ."\" width=\"80\" />\n";
$ret = str_replace( "$3", $tempText, $ret );
# Searchx button
- $tempText = "<input type='submit' name=\"searchx\" value=\"" .
- wfMsg("powersearch") . "\" />\n";
- $ret = str_replace( "$9", $tempText, $ret );
+ $tempText = '<input type="submit" name="searchx" value="' .
+ wfMsg('powersearch') . "\" />\n";
+ $ret = str_replace( '$9', $tempText, $ret );
$action = $sk->escapeSearchLink();
$ret = "<br /><br />\n<form id=\"powersearch\" method=\"get\" " .
function setupPage() {
global $wgOut;
- $wgOut->setPageTitle( wfMsg( "searchresults" ) );
- $q = wfMsg( "searchquery", htmlspecialchars( $this->mUsertext ) );
- $wgOut->setSubtitle( $q );
+ $wgOut->setPageTitle( wfMsg( 'searchresults' ) );
+ $wgOut->setSubtitle( wfMsg( 'searchquery', htmlspecialchars( $this->rawText ) ) );
$wgOut->setArticleRelated( false );
- $wgOut->setRobotpolicy( "noindex,nofollow" );
+ $wgOut->setRobotpolicy( 'noindex,nofollow' );
}
- # Perform the search and construct the results page
- function showResults()
- {
- global $wgUser, $wgTitle, $wgOut, $wgLang, $wgRequest;
+ /**
+ * Perform the search and construct the results page
+ */
+ function showResults() {
+ global $wgUser, $wgTitle, $wgOut, $wgLang;
global $wgDisableTextSearch, $wgInputEncoding;
- $fname = "SearchEngine::showResults";
+ $fname = 'SearchEngine::showResults';
- $search = $wgRequest->getText( 'search' );
+ $search = $this->rawText;
$powersearch = $this->powersearch(); /* Need side-effects here? */
$this->setupPage();
$sk = $wgUser->getSkin();
- $header = wfMsg( "searchresulttext", $sk->makeKnownLink(
- wfMsg( "searchhelppage" ), wfMsg( "searchingwikipedia" ) ) );
- $wgOut->addHTML( $header );
-
- $this->parseQuery();
- if ( "" == $this->mTitlecond || "" == $this->mTextcond ) {
- $wgOut->addHTML( "<h2>" . wfMsg( "badquery" ) . "</h2>\n" .
- "<p>" . wfMsg( "badquerytext" ) . "</p>\n" );
+ $wgOut->addWikiText( wfMsg( 'searchresulttext' ) );
+
+ if ( !$this->parseQuery() ) {
+ $wgOut->addWikiText(
+ '==' . wfMsg( 'badquery' ) . "==\n" .
+ wfMsg( 'badquerytext' ) );
return;
}
- list( $limit, $offset ) = wfCheckLimits( 20, "searchlimit" );
-
- $searchnamespaces = $this->queryNamespaces();
- $redircond = $this->searchRedirects();
- $dbr =& wfGetDB( DB_SLAVE );
- $searchindex = $dbr->tableName( 'searchindex' );
- $cur = $dbr->tableName( 'cur' );
+ list( $limit, $offset ) = wfCheckLimits( 20, 'searchlimit' );
if ( $wgDisableTextSearch ) {
- $wgOut->addHTML( wfMsg( "searchdisabled" ) );
- $wgOut->addHTML( wfMsg( "googlesearch", htmlspecialchars( $search ), $GLOBALS['wgInputEncoding'] ) );
+ $wgOut->addHTML( wfMsg( 'searchdisabled' ) );
+ $wgOut->addHTML( wfMsg( 'googlesearch',
+ htmlspecialchars( $this->rawText ),
+ htmlspecialchars( $wgInputEncoding ) ) );
+ return;
+ }
+
+ $titleMatches = $this->getMatches( $this->titleCond, $limit, $offset );
+ $textMatches = $this->getMatches( $this->textCond, $limit, $offset );
+
+ $sk = $wgUser->getSkin();
+
+ $num = count( $titleMatches ) + count( $textMatches );
+ if ( $num >= $limit ) {
+ $top = wfShowingResults( $offset, $limit );
} else {
- $sql = "SELECT cur_id,cur_namespace,cur_title," .
- "cur_text FROM $cur,$searchindex " .
- "WHERE cur_id=si_page AND {$this->mTitlecond} " .
- "{$searchnamespaces} {$redircond}" .
- "LIMIT {$offset}, {$limit}";
- $res1 = $dbr->query( $sql, $fname );
- $num = $dbr->numRows($res1);
-
- $sk = $wgUser->getSkin();
- $text = "";
-
- $this->parseQuery( $dbr );
- if ( "" == $this->mTitlecond || "" == $this->mTextcond ) {
- $wgOut->addHTML( "<h2>" . wfMsg( "badquery" ) . "</h2>\n" .
- "<p>" . wfMsg( "badquerytext" ) . "</p>\n" );
- return;
- }
- list( $limit, $offset ) = wfCheckLimits( 20, "searchlimit" );
-
- $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 {$this->mTitlecond} " .
- "{$searchnamespaces} {$redircond}" .
- "LIMIT {$offset}, {$limit}";
- $res1 = $dbr->query( $sql, $fname );
- $num = $dbr->numRows($res1);
-
- $sql = "SELECT cur_id,cur_namespace,cur_title," .
- "cur_text FROM $cur,$searchindex " .
- "WHERE cur_id=si_page AND {$this->mTextcond} " .
- "{$searchnamespaces} {$redircond} " .
- "LIMIT {$offset}, {$limit}";
- $res2 = $dbr->query( $sql, $fname );
- $num = $num + $dbr->numRows($res2);
-
- if ( $num == $limit ) {
- $top = wfShowingResults( $offset, $limit);
- } else {
- $top = wfShowingResultsNum( $offset, $limit, $num );
- }
- $wgOut->addHTML( "<p>{$top}</p>\n" );
-
- # For powersearch
-
- $a2l = "" ;
- $akk = array_keys( $this->addtoquery ) ;
- foreach ( $akk AS $ak ) {
- $a2l .= "&{$ak}={$this->addtoquery[$ak]}" ;
- }
-
- $sl = wfViewPrevNext( $offset, $limit, "",
- "search=" . wfUrlencode( $this->mUsertext ) . $a2l );
- $wgOut->addHTML( "<br />{$sl}\n" );
-
- $foundsome = false;
-
- if ( 0 == $dbr->numRows( $res1 ) ) {
- $wgOut->addHTML( "<h2>" . wfMsg( "notitlematches" ) .
- "</h2>\n" );
- } else {
- $foundsome = true;
- $off = $offset + 1;
- $wgOut->addHTML( "<h2>" . wfMsg( "titlematches" ) .
- "</h2>\n<ol start='{$off}'>" );
-
- while ( $row = $dbr->fetchObject( $res1 ) ) {
- $this->showHit( $row );
- }
- $dbr->freeResult( $res1 );
- $wgOut->addHTML( "</ol>\n" );
- }
+ $top = wfShowingResultsNum( $offset, $limit, $num );
+ }
+ $wgOut->addHTML( "<p>{$top}</p>\n" );
- if ( 0 == $dbr->numRows( $res2 ) ) {
- $wgOut->addHTML( "<h2>" . wfMsg( "notextmatches" ) .
- "</h2>\n" );
- } else {
- $foundsome = true;
- $off = $offset + 1;
- $wgOut->addHTML( "<h2>" . wfMsg( "textmatches" ) . "</h2>\n" .
- "<ol start='{$off}'>" );
- while ( $row = $dbr->fetchObject( $res2 ) ) {
- $this->showHit( $row );
- }
- $dbr->freeResult( $res2 );
- $wgOut->addHTML( "</ol>\n" );
- }
- if ( ! $foundsome ) {
- $wgOut->addHTML( "<p>" . wfMsg( "nonefound" ) . "</p>\n" );
- }
- $wgOut->addHTML( "<p>{$sl}</p>\n" );
- $wgOut->addHTML( $powersearch );
+ # For powersearch
+ $a2l = '';
+ $akk = array_keys( $this->addToQuery );
+ foreach ( $akk AS $ak ) {
+ $a2l .= "&{$ak}={$this->addToQuery[$ak]}" ;
+ }
+
+ $prevnext = wfViewPrevNext( $offset, $limit, '',
+ 'search=' . wfUrlencode( $this->filteredText ) . $a2l );
+ $wgOut->addHTML( "<br />{$prevnext}\n" );
+
+ $foundsome = $this->showMatches( $titleMatches, $offset, 'notitlematches', 'titlematches' )
+ || $this->showMatches( $textMatches, $offset, 'notextmatches', 'textmatches' );
+
+ if ( !$foundsome ) {
+ $wgOut->addWikiText( wfMsg( 'nonefound' ) );
}
+ $wgOut->addHTML( "<p>{$prevnext}</p>\n" );
+ $wgOut->addHTML( $powersearch );
}
- function legalSearchChars()
- {
+ function legalSearchChars() {
$lc = "A-Za-z_'0-9\\x80-\\xFF\\-";
return $lc;
}
- function parseQuery( &$db )
- {
- global $wgDBminWordLen, $wgLang, $wgDBmysql4;
+ 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( $db );
+ return $this->parseQuery4();
+ } else {
+ # Fall back to ugly hack with multiple search clauses
+ return $this->parseQuery3();
}
+ }
+
+ function parseQuery3() {
+ global $wgDBminWordLen, $wgContLang;
+
# on non mysql4 database: get list of words we don't want to search for
- require_once( "FulltextStoplist.php" );
+ require_once( 'FulltextStoplist.php' );
- $lc = SearchEngine::legalSearchChars() . "()";
- $q = preg_replace( "/([()])/", " \\1 ", $this->mUsertext );
+ $lc = SearchEngine::legalSearchChars() . '()';
+ $q = preg_replace( "/([()])/", " \\1 ", $this->filteredText );
$q = preg_replace( "/\\s+/", " ", $q );
- $w = explode( " ", strtolower( trim( $q ) ) );
+ $w = explode( ' ', trim( $q ) );
- $last = $cond = "";
+ $last = $cond = '';
foreach ( $w as $word ) {
- $word = $wgLang->stripForSearch( $word );
- if ( "and" == $word || "or" == $word || "not" == $word
- || "(" == $word || ")" == $word ) {
- $cond .= " " . strtoupper( $word );
- $last = "";
+ $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"; }
+ if ( '' != $last ) { $cond .= ' AND'; }
$cond .= " (MATCH (##field##) AGAINST ('" .
- $db->strencode( $word ). "'))";
+ $this->db->strencode( $word ). "'))";
$last = $word;
- array_push( $this->mSearchterms, "\\b" . $word . "\\b" );
+ array_push( $this->searchTerms, "\\b" . $word . "\\b" );
}
}
- if ( 0 == count( $this->mSearchterms ) ) { return; }
+ if ( 0 == count( $this->searchTerms ) ) {
+ return MW_SEARCH_BAD_QUERY;
+ }
- $this->mTitlecond = "(" . str_replace( "##field##",
- "si_title", $cond ) . " )";
+ $this->titleCond = '(' . str_replace( '##field##',
+ 'si_title', $cond ) . ' )';
- $this->mTextcond = "(" . str_replace( "##field##",
- "si_text", $cond ) . " AND (cur_is_redirect=0) )";
+ $this->textCond = '(' . str_replace( '##field##',
+ 'si_text', $cond ) . ' AND (cur_is_redirect=0) )';
+
+ return MW_SEARCH_OK;
}
- function parseQuery4( &$db )
- {
- global $wgLang;
+ function parseQuery4() {
+ global $wgContLang;
$lc = SearchEngine::legalSearchChars();
- $searchon = "";
- $this->mSearchterms = array();
+ $searchon = '';
+ $this->searchTerms = array();
# FIXME: This doesn't handle parenthetical expressions.
if( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
- $this->mUsertext, $m, PREG_SET_ORDER ) ) {
+ $this->filteredText, $m, PREG_SET_ORDER ) ) {
foreach( $m as $terms ) {
- if( $searchon !== "" ) $searchon .= " ";
- if( $this->mStrictMatching && ($terms[1] == "") ) {
- $terms[1] = "+";
+ if( $searchon !== '' ) $searchon .= ' ';
+ if( $this->strictMatching && ($terms[1] == '') ) {
+ $terms[1] = '+';
}
- $searchon .= $terms[1] . $wgLang->stripForSearch( $terms[2] );
- if( $terms[3] ) {
+ $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->mSearchterms[] = $regexp;
+ $this->searchTerms[] = $regexp;
}
wfDebug( "Would search with '$searchon'\n" );
- wfDebug( "Match with /\b" . implode( '\b|\b', $this->mSearchterms ) . "\b/\n" );
+ wfDebug( "Match with /\b" . implode( '\b|\b', $this->searchTerms ) . "\b/\n" );
} else {
- wfDebug( "Can't understand search query '$this->mUsertext'\n" );
+ 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;
+ }
+
+ 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 );
- $searchon = $db->strencode( $searchon );
- $this->mTitlecond = " MATCH(si_title) AGAINST('$searchon' IN BOOLEAN MODE)";
- $this->mTextcond = " (MATCH(si_text) AGAINST('$searchon' IN BOOLEAN MODE) AND cur_is_redirect=0)";
+ return $matches;
+ }
+
+ function showMatches( &$matches, $offset, $msgEmpty, $msgFound ) {
+ global $wgOut;
+ if ( 0 == count( $matches ) ) {
+ $wgOut->addHTML( "<h2>" . wfMsg( $msgEmpty ) .
+ "</h2>\n" );
+ return false;
+ } else {
+ $off = $offset + 1;
+ $wgOut->addHTML( "<h2>" . wfMsg( $msgFound ) .
+ "</h2>\n<ol start='{$off}'>" );
+
+ foreach( $matches as $row ) {
+ $this->showHit( $row );
+ }
+ $wgOut->addHTML( "</ol>\n" );
+ return true;
+ }
}
- function showHit( $row )
- {
- global $wgUser, $wgOut;
+ function showHit( $row ) {
+ global $wgUser, $wgOut, $wgContLang;
$t = Title::makeName( $row->cur_namespace, $row->cur_title );
+ if( is_null( $t ) ) {
+ $wgOut->addHTML( "<!-- Broken link in search result -->\n" );
+ return;
+ }
$sk = $wgUser->getSkin();
- $contextlines = $wgUser->getOption( "contextlines" );
- if ( "" == $contextlines ) { $contextlines = 5; }
- $contextchars = $wgUser->getOption( "contextchars" );
- if ( "" == $contextchars ) { $contextchars = 50; }
+ $contextlines = $wgUser->getOption( 'contextlines' );
+ if ( '' == $contextlines ) { $contextlines = 5; }
+ $contextchars = $wgUser->getOption( 'contextchars' );
+ if ( '' == $contextchars ) { $contextchars = 50; }
- $link = $sk->makeKnownLink( $t, "" );
- $size = wfMsg( "nbytes", strlen( $row->cur_text ) );
+ $link = $sk->makeKnownLink( $t, '' );
+ $size = wfMsg( 'nbytes', strlen( $row->cur_text ) );
$wgOut->addHTML( "<li>{$link} ({$size})" );
$lines = explode( "\n", $row->cur_text );
- $pat1 = "/(.*)(" . implode( "|", $this->mSearchterms ) . ")(.*)/i";
+ $pat1 = "/(.*)(" . implode( "|", $this->searchTerms ) . ")(.*)/i";
$lineno = 0;
foreach ( $lines as $line ) {
- if ( 0 == $contextlines ) { break; }
+ if ( 0 == $contextlines ) {
+ break;
+ }
--$contextlines;
++$lineno;
- if ( ! preg_match( $pat1, $line, $m ) ) { continue; }
-
- $pre = $m[1];
- if ( 0 == $contextchars ) { $pre = "..."; }
- else {
- if ( strlen( $pre ) > $contextchars ) {
- $pre = "..." . substr( $pre, -$contextchars );
- }
+ if ( ! preg_match( $pat1, $line, $m ) ) {
+ continue;
}
- $pre = wfEscapeHTML( $pre );
- if ( count( $m ) < 3 ) { $post = ""; }
- else { $post = $m[3]; }
+ $pre = $wgContLang->truncate( $m[1], -$contextchars, '...' );
- if ( 0 == $contextchars ) { $post = "..."; }
- else {
- if ( strlen( $post ) > $contextchars ) {
- $post = substr( $post, 0, $contextchars ) . "...";
- }
+ if ( count( $m ) < 3 ) {
+ $post = '';
+ } else {
+ $post = $wgContLang->truncate( $m[3], $contextchars, '...' );
}
- $post = wfEscapeHTML( $post );
- $found = wfEscapeHTML( $m[2] );
- $line = "{$pre}{$found}{$post}";
- $pat2 = "/(" . implode( "|", $this->mSearchterms ) . ")/i";
+ $found = $m[2];
+
+ $line = htmlspecialchars( $pre . $found . $post );
+ $pat2 = '/(' . implode( '|', $this->searchTerms ) . ")/i";
$line = preg_replace( $pat2,
- "<font color='red'>\\1</font>", $line );
+ "<span class='searchmatch'>\\1</span>", $line );
$wgOut->addHTML( "<br /><small>{$lineno}: {$line}</small>\n" );
}
$wgOut->addHTML( "</li>\n" );
}
- function goResult()
- {
- global $wgOut, $wgRequest, $wgGoToEdit;
- global $wgDisableTextSearch;
- $fname = "SearchEngine::goResult";
-
- $search = trim( $wgRequest->getText( "search" ) );
-
- # Try to go to page as entered.
- #
- $t = Title::newFromText( $search );
-
- # If the string cannot be used to create a title
- if( false == $t ){
- $this->showResults();
- return;
- }
-
+ function getNearMatch() {
# Exact match? No need to look further.
- if ( $t->getNamespace() == NS_SPECIAL || 0 != $t->getArticleID() ) {
- $wgOut->redirect( $t->getFullURL() );
- return;
+ $title = Title::newFromText( $this->rawText );
+ if ( $title->getNamespace() == NS_SPECIAL || 0 != $title->getArticleID() ) {
+ return $title;
}
# Now try all lower case (i.e. first letter capitalized)
#
- $t = Title::newFromText( strtolower( $search ) );
- if ( 0 != $t->getArticleID() ) {
- $wgOut->redirect( $t->getFullURL() );
- return;
+ $title = Title::newFromText( strtolower( $this->rawText ) );
+ if ( 0 != $title->getArticleID() ) {
+ return $title;
}
# Now try capitalized string
#
- $t = Title::newFromText( ucwords( strtolower( $search ) ) );
- if ( 0 != $t->getArticleID() ) {
- $wgOut->redirect( $t->getFullURL() );
- return;
+ $title = Title::newFromText( ucwords( strtolower( $this->rawText ) ) );
+ if ( 0 != $title->getArticleID() ) {
+ return $title;
}
# Now try all upper case
#
- $t = Title::newFromText( strtoupper( $search ) );
- if ( 0 != $t->getArticleID() ) {
- $wgOut->redirect( $t->getFullURL() );
- return;
+ $title = Title::newFromText( strtoupper( $this->rawText ) );
+ if ( 0 != $title->getArticleID() ) {
+ return $title;
}
# Entering an IP address goes to the contributions page
- if ( preg_match( '/^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}$/', $search ) ) {
- $title = Title::makeTitle( NS_SPECIAL, "Contributions" );
- $wgOut->redirect( $title->getFullUrl( "target=$search" ) );
+ if ( preg_match( '/^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}$/', $this->rawText ) ) {
+ $title = Title::makeTitle( NS_SPECIAL, "Contributions/" . $this->rawText );
+ return $title;
+ }
+
+ return NULL;
+ }
+
+ function goResult() {
+ global $wgOut, $wgGoToEdit;
+ global $wgDisableTextSearch;
+ $fname = 'SearchEngine::goResult';
+
+ # Try to go to page as entered.
+ #
+ $t = Title::newFromText( $this->rawText );
+
+ # If the string cannot be used to create a title
+ if( is_null( $t ) ){
+ $this->showResults();
return;
}
+ # If there's an exact or very near match, jump right there.
+ $t = $this->getNearMatch();
+ if( !is_null( $t ) ) {
+ $wgOut->redirect( $t->getFullURL() );
+ return;
+ }
+
# No match, generate an edit URL
- $t = Title::newFromText( $this->mUsertext );
+ $t = Title::newFromText( $this->rawText );
# If the feature is enabled, go straight to the edit page
if ( $wgGoToEdit ) {
- $wgOut->redirect( $t->getFullURL( "action=edit" ) );
+ $wgOut->redirect( $t->getFullURL( 'action=edit' ) );
return;
}
if( $t ) {
- $editurl = $t->escapeLocalURL( "action=edit" );
+ $editurl = $t->escapeLocalURL( 'action=edit' );
} else {
- $editurl = ""; # ??
+ $editurl = ''; # ??
}
- $wgOut->addHTML( "<p>" . wfMsg("nogomatch", $editurl ) . "</p>\n" );
+ $wgOut->addHTML( '<p>' . wfMsg('nogomatch', $editurl, htmlspecialchars( $search ) ) . "</p>\n" );
# Try a fuzzy title search
$anyhit = false;
global $wgDisableFuzzySearch;
if(! $wgDisableFuzzySearch ){
- foreach( array(NS_MAIN, NS_WP, NS_USER, NS_IMAGE, NS_MEDIAWIKI) as $namespace){
- $anyhit |= SearchEngine::doFuzzyTitleSearch( $search, $namespace );
+ foreach( array(NS_MAIN, NS_PROJECT, NS_USER, NS_IMAGE, NS_MEDIAWIKI) as $namespace){
+ $anyhit |= SearchEngine::doFuzzyTitleSearch( $this->rawText, $namespace );
}
}
}
}
- /* static */ function doFuzzyTitleSearch( $search, $namespace ){
- global $wgLang, $wgOut;
+ /**
+ * @static
+ */
+ function doFuzzyTitleSearch( $search, $namespace ){
+ global $wgContLang, $wgOut;
$this->setupPage();
$sstr = ucfirst($search);
- $sstr = str_replace(" ", "_", $sstr);
+ $sstr = str_replace(' ', '_', $sstr);
$fuzzymatches = SearchEngine::fuzzyTitles( $sstr, $namespace );
$fuzzymatches = array_slice($fuzzymatches, 0, 10);
$slen = strlen( $search );
- $wikitext = "";
+ $wikitext = '';
foreach($fuzzymatches as $res){
- $t = str_replace("_", " ", $res[1]);
- $tfull = $wgLang->getNsText( $namespace ) . ":$t|$t";
+ $t = str_replace('_', ' ', $res[1]);
+ $tfull = $wgContLang->getNsText( $namespace ) . ":$t|$t";
if( $namespace == NS_MAIN )
- $tfull = "$t";
+ $tfull = $t;
$distance = $res[0];
$closeness = (strlen( $search ) - $distance) / strlen( $search );
- $percent = intval( $closeness * 100 ) . "%";
- $stars = str_repeat("*", ceil(5 * $closeness) );
+ $percent = intval( $closeness * 100 ) . '%';
+ $stars = str_repeat('*', ceil(5 * $closeness) );
$wikitext .= "* [[$tfull]] $percent ($stars)\n";
}
if( $wikitext ){
if( $namespace != NS_MAIN )
- $wikitext = "=== " . $wgLang->getNsText( $namespace ) . " ===\n" . $wikitext;
+ $wikitext = '=== ' . $wgContLang->getNsText( $namespace ) . " ===\n" . $wikitext;
$wgOut->addWikiText( $wikitext );
return true;
}
return false;
}
- /* static */ function fuzzyTitles( $sstr, $namespace = NS_MAIN ){
+ /**
+ * @static
+ */
+ 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);
$cnt++;
}
}
- usort($result, "SearchEngine_pcmp");
+ usort($result, 'SearchEngine_pcmp');
return $result;
}
- /* static */ function getTitlesByLength($aLength, $aNamespace = 0){
+ /**
+ * static
+ */
+ function getTitlesByLength($aLength, $aNamespace = 0){
global $wgMemc, $wgDBname;
$fname = 'SearchEngin::getTitlesByLength';
// to avoid multiple costly SELECTs in case of no memcached
- if( $this->all_titles ){
- if( isset( $this->all_titles[$aLength][$aNamespace] ) ){
- return $this->all_titles[$aLength][$aNamespace];
+ if( $this->allTitles ){
+ if( isset( $this->allTitles[$aLength][$aNamespace] ) ){
+ return $this->allTitles[$aLength][$aNamespace];
} else {
return array();
}
$wgMemc->set( $mkeyts, time() );
- $dbr =& wfGetDB( DB_SLAVE );
- $res = $dbr->select( 'cur', array( 'cur_title', 'cur_namespace' ), false, $fname );
+ $res = $this->db->select( 'cur', array( 'cur_title', 'cur_namespace' ), false, $fname );
$titles = array(); // length, ns, [titles]
- while( $obj = $dbr->fetchObject( $res ) ){
+ while( $obj = $this->db->fetchObject( $res ) ){
$title = $obj->cur_title;
$ns = $obj->cur_namespace;
$len = strlen( $title );
$wgMemc->set( $mkey, $title_arr, 3600 * 24 );
}
}
- $this->all_titles = $titles;
+ $this->allTitles = $titles;
if( isset( $titles[$aLength][$aNamespace] ) )
return $titles[$aLength][$aNamespace];
else
}
}
-/* private static */ function SearchEngine_pcmp($a, $b){ return $a[0] - $b[0]; }
+/**
+ * @access private
+ * @static
+ */
+function SearchEngine_pcmp($a, $b){ return $a[0] - $b[0]; }
?>