# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
# http://www.gnu.org/copyleft/gpl.html
-## XXX Better catching of SELECT to_tsquery('the')
+/**
+ * @file
+ * @ingroup Search
+ */
/**
* Search engine hook base class for Postgres
- * @addtogroup Search
+ * @ingroup Search
*/
-
class SearchPostgres extends SearchEngine {
function SearchPostgres( $db ) {
/**
* Perform a full text search query via tsearch2 and return a result set.
- * Currently searches a page's current title (p.page_title) and text (t.old_text)
+ * Currently searches a page's current title (page.page_title) and
+ * latest revision article text (pagecontent.old_text)
*
* @param string $term - Raw search term
* @return PostgresSearchResultSet
* @access public
*/
- function searchText( $term ) {
- $resultSet = $this->db->resultObject( $this->db->query( $this->searchQuery( $term, 'textvector' ) ) );
+ function searchTitle( $term ) {
+ $q = $this->searchQuery( $term , 'titlevector', 'page_title' );
+ $olderror = error_reporting(E_ERROR);
+ $resultSet = $this->db->resultObject( $this->db->query( $q, 'SearchPostgres', true ) );
+ error_reporting($olderror);
+ if (!$resultSet) {
+ // Needed for "Query requires full scan, GIN doesn't support it"
+ return new SearchResultTooMany();
+ }
return new PostgresSearchResultSet( $resultSet, $this->searchTerms );
}
- function searchTitle( $term ) {
- $resultSet = $this->db->resultObject( $this->db->query( $this->searchQuery( $term , 'titlevector' ) ) );
+ function searchText( $term ) {
+ $q = $this->searchQuery( $term, 'textvector', 'old_text' );
+ $olderror = error_reporting(E_ERROR);
+ $resultSet = $this->db->resultObject( $this->db->query( $q, 'SearchPostgres', true ) );
+ error_reporting($olderror);
+ if (!$resultSet) {
+ return new SearchResultTooMany();
+ }
return new PostgresSearchResultSet( $resultSet, $this->searchTerms );
}
/*
* Transform the user's search string into a better form for tsearch2
*/
- function parseQuery( $filteredText, $fulltext ) {
- global $wgContLang;
- $lc = SearchEngine::legalSearchChars();
- $searchon = '';
- $searchconst = '';
- $this->searchTerms = array();
-
- # FIXME: This doesn't handle parenthetical expressions.
+ function parseQuery( $term ) {
+
+ wfDebug( "parseQuery received: $term" );
+
+ ## No backslashes allowed
+ $term = preg_replace('/\\\/', '', $term);
+
+ ## Collapse parens into nearby words:
+ $term = preg_replace('/\s*\(\s*/', ' (', $term);
+ $term = preg_replace('/\s*\)\s*/', ') ', $term);
+
+ ## Treat colons as word separators:
+ $term = preg_replace('/:/', ' ', $term);
+
+ $searchstring = '';
$m = array();
- if( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
- $filteredText, $m, PREG_SET_ORDER ) ) {
+ if( preg_match_all('/([-!]?)(\S+)\s*/', $term, $m, PREG_SET_ORDER ) ) {
foreach( $m as $terms ) {
- switch ($terms[1]) {
- case '~': // negate but do not exclude like "-" for now, simply negate
- case '-': $searchconst.= '&!' . $wgContLang->stripForSearch( $terms[2] );
- break;
- case '+': $searchconst.= '&' . $wgContLang->stripForSearch( $terms[2] );
- break;
- case '<': // decrease priority of word - not implemented
- case '>': // increase priority of word - not implemented
- default : $searchon.= '|' . $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( '"', '', $terms[2] ), '/' );
+ if (strlen($terms[1])) {
+ $searchstring .= ' & !';
+ }
+ if (strtolower($terms[2]) === 'and') {
+ $searchstring .= ' & ';
+ }
+ else if (strtolower($terms[2]) === 'or' or $terms[2] === '|') {
+ $searchstring .= ' | ';
+ }
+ else if (strtolower($terms[2]) === 'not') {
+ $searchstring .= ' & !';
+ }
+ else {
+ $searchstring .= " & $terms[2]";
}
- $this->searchTerms[] = $regexp;
}
- wfDebug( "Would search with '$searchon'\n" );
- wfDebug( 'Match with /\b' . implode( '\b|\b', $this->searchTerms ) . "\b/\n" );
- } else {
- wfDebug( "Can't understand search query '{$this->filteredText}'\n" );
}
- if (substr_count($searchon,'|')==1) {
- $searchon = str_replace ('|','&',$searchon);
- }
- $searchon = substr($searchconst . $searchon, 1) ;
- $searchon = preg_replace('/(\s+)/','&',$searchon);
- $searchon = $this->db->strencode( $searchon );
- return $searchon;
+ ## Strip out leading junk
+ $searchstring = preg_replace('/^[\s\&\|]+/', '', $searchstring);
+
+ ## Remove any doubled-up operators
+ $searchstring = preg_replace('/([\!\&\|]) +(?:[\&\|] +)+/', "$1 ", $searchstring);
+
+ ## Remove any non-spaced operators (e.g. "Zounds!")
+ $searchstring = preg_replace('/([^ ])[\!\&\|]/', "$1", $searchstring);
+
+ ## Remove any trailing whitespace or operators
+ $searchstring = preg_replace('/[\s\!\&\|]+$/', '', $searchstring);
+
+ ## Remove unnecessary quotes around everything
+ $searchstring = preg_replace('/^[\'"](.*)[\'"]$/', "$1", $searchstring);
+
+ ## Quote the whole thing
+ $searchstring = $this->db->addQuotes($searchstring);
+
+ wfDebug( "parseQuery returned: $searchstring" );
+
+ return $searchstring;
+
}
/**
* @param string $fulltext
* @private
*/
- function searchQuery( $filteredTerm, $fulltext ) {
+ function searchQuery( $term, $fulltext, $colname ) {
+ global $wgDBversion;
+
+ if ( !isset( $wgDBversion ) ) {
+ $this->db->getServerVersion();
+ $wgDBversion = $this->db->numeric_version;
+ }
+ $prefix = $wgDBversion < 8.3 ? "'default'," : '';
- $match = $this->parseQuery( $filteredTerm, $fulltext );
+ $searchstring = $this->parseQuery( $term );
- $query = "SELECT page_id, page_namespace, page_title, old_text AS page_text, ".
- "rank(titlevector, to_tsquery('default','$match')) AS rnk ".
+ ## We need a separate query here so gin does not complain about empty searches
+ $SQL = "SELECT to_tsquery($prefix $searchstring)";
+ $res = $this->db->doQuery($SQL);
+ if (!$res) {
+ ## TODO: Better output (example to catch: one 'two)
+ die ("Sorry, that was not a valid search string. Please go back and try again");
+ }
+ $top = pg_fetch_result($res,0,0);
+
+ if ($top === "") { ## e.g. if only stopwords are used XXX return something better
+ $query = "SELECT page_id, page_namespace, page_title, 0 AS score ".
+ "FROM page p, revision r, pagecontent c WHERE p.page_latest = r.rev_id " .
+ "AND r.rev_text_id = c.old_id AND 1=0";
+ }
+ else {
+ $m = array();
+ if( preg_match_all("/'([^']+)'/", $top, $m, PREG_SET_ORDER ) ) {
+ foreach( $m as $terms ) {
+ $this->searchTerms[$terms[1]] = $terms[1];
+ }
+ }
+
+ $rankscore = $wgDBversion > 8.2 ? 5 : 1;
+ $rank = $wgDBversion < 8.3 ? 'rank' : 'ts_rank';
+ $query = "SELECT page_id, page_namespace, page_title, ".
+ "$rank($fulltext, to_tsquery($prefix $searchstring), $rankscore) AS score ".
"FROM page p, revision r, pagecontent c WHERE p.page_latest = r.rev_id " .
- "AND r.rev_text_id = c.old_id AND $fulltext @@ to_tsquery('default','$match')";
+ "AND r.rev_text_id = c.old_id AND $fulltext @@ to_tsquery($prefix $searchstring)";
+ }
## Redirects
if (! $this->showRedirects)
- $query .= ' AND page_is_redirect = 0'; ## IS FALSE
+ $query .= ' AND page_is_redirect = 0';
## Namespaces - defaults to 0
- if ( count($this->namespaces) < 1)
- $query .= ' AND page_namespace = 0';
- else {
- $namespaces = implode( ',', $this->namespaces );
- $query .= " AND page_namespace IN ($namespaces)";
+ if( !is_null($this->namespaces) ){ // null -> search all
+ if ( count($this->namespaces) < 1)
+ $query .= ' AND page_namespace = 0';
+ else {
+ $namespaces = implode( ',', $this->namespaces );
+ $query .= " AND page_namespace IN ($namespaces)";
+ }
}
- $query .= " ORDER BY rnk DESC, page_id DESC";
+ $query .= " ORDER BY score DESC, page_id DESC";
$query .= $this->db->limitResult( '', $this->limit, $this->offset );
+ wfDebug( "searchQuery returned: $query" );
+
return $query;
}
- ## These two functions are done automatically via triggers
+ ## Most of the work of these two functions are done automatically via triggers
function update( $pageid, $title, $text ) {
- $dbw = wfGetDB( DB_MASTER );
## We don't want to index older revisions
$SQL = "UPDATE pagecontent SET textvector = NULL WHERE old_id = ".
"(SELECT rev_text_id FROM revision WHERE rev_page = $pageid ".
"ORDER BY rev_text_id DESC LIMIT 1 OFFSET 1)";
- $dbw->doQuery($SQL);
+ $this->db->doQuery($SQL);
+ return true;
+ }
+
+ function updateTitle( $id, $title ) {
return true;
}
- function updateTitle( $id, $title ) { return true; }
} ## end of the SearchPostgres class
+/**
+ * @ingroup Search
+ */
+class PostgresSearchResult extends SearchResult {
+ function PostgresSearchResult( $row ) {
+ $this->mTitle = Title::makeTitle( $row->page_namespace, $row->page_title );
+ $this->score = $row->score;
+ }
+ function getScore() {
+ return $this->score;
+ }
+}
+/**
+ * @ingroup Search
+ */
class PostgresSearchResultSet extends SearchResultSet {
function PostgresSearchResultSet( $resultSet, $terms ) {
$this->mResultSet = $resultSet;
if( $row === false ) {
return false;
} else {
- return new SearchResult( $row );
+ return new PostgresSearchResult( $row );
}
}
}
-
-?>
dépôts / lhc / web / wiklou.git / blobdiff