includes/SearchPostgres.php

   1 <?php
   2 # Copyright (C) 2006-2007 Greg Sabino Mullane <greg@turnstep.com>
   3 # http://www.mediawiki.org/
   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 /**
  21  * @file
  22  * @ingroup Search
  23  */
  24
  25 /**
  26  * Search engine hook base class for Postgres
  27  * @ingroup Search
  28  */
  29 class SearchPostgres extends SearchEngine {
  30
  31         function __construct( $db ) {
  32                 $this->db = $db;
  33         }
  34
  35         /**
  36          * Perform a full text search query via tsearch2 and return a result set.
  37          * Currently searches a page's current title (page.page_title) and
  38          * latest revision article text (pagecontent.old_text)
  39          *
  40          * @param string $term - Raw search term
  41          * @return PostgresSearchResultSet
  42          * @access public
  43          */
  44         function searchTitle( $term ) {
  45                 $q = $this->searchQuery( $term , 'titlevector', 'page_title' );
  46                 $olderror = error_reporting(E_ERROR);
  47                 $resultSet = $this->db->resultObject( $this->db->query( $q, 'SearchPostgres', true ) );
  48                 error_reporting($olderror);
  49                 if (!$resultSet) {
  50                         // Needed for "Query requires full scan, GIN doesn't support it"
  51                         return new SearchResultTooMany();
  52                 }
  53                 return new PostgresSearchResultSet( $resultSet, $this->searchTerms );
  54         }
  55         function searchText( $term ) {
  56                 $q = $this->searchQuery( $term, 'textvector', 'old_text' );
  57                 $olderror = error_reporting(E_ERROR);
  58                 $resultSet = $this->db->resultObject( $this->db->query( $q, 'SearchPostgres', true ) );
  59                 error_reporting($olderror);
  60                 if (!$resultSet) {
  61                         return new SearchResultTooMany();
  62                 }
  63                 return new PostgresSearchResultSet( $resultSet, $this->searchTerms );
  64         }
  65
  66
  67         /*
  68          * Transform the user's search string into a better form for tsearch2
  69          * Returns an SQL fragment consisting of quoted text to search for.
  70         */
  71         function parseQuery( $term ) {
  72
  73                 wfDebug( "parseQuery received: $term \n" );
  74
  75                 ## No backslashes allowed
  76                 $term = preg_replace('/\\\/', '', $term);
  77
  78                 ## Collapse parens into nearby words:
  79                 $term = preg_replace('/\s*\(\s*/', ' (', $term);
  80                 $term = preg_replace('/\s*\)\s*/', ') ', $term);
  81
  82                 ## Treat colons as word separators:
  83                 $term = preg_replace('/:/', ' ', $term);
  84
  85                 $searchstring = '';
  86                 $m = array();
  87                 if( preg_match_all('/([-!]?)(\S+)\s*/', $term, $m, PREG_SET_ORDER ) ) {
  88                         foreach( $m as $terms ) {
  89                                 if (strlen($terms[1])) {
  90                                         $searchstring .= ' & !';
  91                                 }
  92                                 if (strtolower($terms[2]) === 'and') {
  93                                         $searchstring .= ' & ';
  94                                 }
  95                                 else if (strtolower($terms[2]) === 'or' or $terms[2] === '|') {
  96                                         $searchstring .= ' | ';
  97                                 }
  98                                 else if (strtolower($terms[2]) === 'not') {
  99                                         $searchstring .= ' & !';
 100                                 }
 101                                 else {
 102                                         $searchstring .= " & $terms[2]";
 103                                 }
 104                         }
 105                 }
 106
 107                 ## Strip out leading junk
 108                 $searchstring = preg_replace('/^[\s\&\|]+/', '', $searchstring);
 109
 110                 ## Remove any doubled-up operators
 111                 $searchstring = preg_replace('/([\!\&\|]) +(?:[\&\|] +)+/', "$1 ", $searchstring);
 112
 113                 ## Remove any non-spaced operators (e.g. "Zounds!")
 114                 $searchstring = preg_replace('/([^ ])[\!\&\|]/', "$1", $searchstring);
 115
 116                 ## Remove any trailing whitespace or operators
 117                 $searchstring = preg_replace('/[\s\!\&\|]+$/', '', $searchstring);
 118
 119                 ## Remove unnecessary quotes around everything
 120                 $searchstring = preg_replace('/^[\'"](.*)[\'"]$/', "$1", $searchstring);
 121
 122                 ## Quote the whole thing
 123                 $searchstring = $this->db->addQuotes($searchstring);
 124
 125                 wfDebug( "parseQuery returned: $searchstring \n" );
 126
 127                 return $searchstring;
 128
 129         }
 130
 131         /**
 132          * Construct the full SQL query to do the search.
 133          * @param string $filteredTerm
 134          * @param string $fulltext
 135          * @private
 136          */
 137         function searchQuery( $term, $fulltext, $colname ) {
 138                 global $wgDBversion;
 139
 140                 if ( !isset( $wgDBversion ) ) {
 141                         $this->db->getServerVersion();
 142                         $wgDBversion = $this->db->numeric_version;
 143                 }
 144                 $prefix = $wgDBversion < 8.3 ? "'default'," : '';
 145
 146                 # Get the SQL fragment for the given term
 147                 $searchstring = $this->parseQuery( $term );
 148
 149                 ## We need a separate query here so gin does not complain about empty searches
 150                 $SQL = "SELECT to_tsquery($prefix $searchstring)";
 151                 $res = $this->db->doQuery($SQL);
 152                 if (!$res) {
 153                         ## TODO: Better output (example to catch: one 'two)
 154                         die ("Sorry, that was not a valid search string. Please go back and try again");
 155                 }
 156                 $top = pg_fetch_result($res,0,0);
 157
 158                 if ($top === "") { ## e.g. if only stopwords are used XXX return something better
 159                         $query = "SELECT page_id, page_namespace, page_title, 0 AS score ".
 160                                 "FROM page p, revision r, pagecontent c WHERE p.page_latest = r.rev_id " .
 161                                 "AND r.rev_text_id = c.old_id AND 1=0";
 162                 }
 163                 else {
 164                         $m = array();
 165                         if( preg_match_all("/'([^']+)'/", $top, $m, PREG_SET_ORDER ) ) {
 166                                 foreach( $m as $terms ) {
 167                                         $this->searchTerms[$terms[1]] = $terms[1];
 168                                 }
 169                         }
 170
 171                         $rankscore = $wgDBversion > 8.2 ? 5 : 1;
 172                         $rank = $wgDBversion < 8.3 ? 'rank' : 'ts_rank';
 173                         $query = "SELECT page_id, page_namespace, page_title, ".
 174                         "$rank($fulltext, to_tsquery($prefix $searchstring), $rankscore) AS score ".
 175                         "FROM page p, revision r, pagecontent c WHERE p.page_latest = r.rev_id " .
 176                         "AND r.rev_text_id = c.old_id AND $fulltext @@ to_tsquery($prefix $searchstring)";
 177                 }
 178
 179                 ## Redirects
 180                 if (! $this->showRedirects)
 181                         $query .= ' AND page_is_redirect = 0';
 182
 183                 ## Namespaces - defaults to 0
 184                 if( !is_null($this->namespaces) ){ // null -> search all
 185                         if ( count($this->namespaces) < 1)
 186                                 $query .= ' AND page_namespace = 0';
 187                         else {
 188                                 $namespaces = $this->db->makeList( $this->namespaces );
 189                                 $query .= " AND page_namespace IN ($namespaces)";
 190                         }
 191                 }
 192
 193                 $query .= " ORDER BY score DESC, page_id DESC";
 194
 195                 $query .= $this->db->limitResult( '', $this->limit, $this->offset );
 196
 197                 wfDebug( "searchQuery returned: $query \n" );
 198
 199                 return $query;
 200         }
 201
 202         ## Most of the work of these two functions are done automatically via triggers
 203
 204         function update( $pageid, $title, $text ) {
 205                 ## We don't want to index older revisions
 206                 $SQL = "UPDATE pagecontent SET textvector = NULL WHERE old_id IN ".
 207                                 "(SELECT rev_text_id FROM revision WHERE rev_page = " . intval( $pageid ) .
 208                                 " ORDER BY rev_text_id DESC OFFSET 1)";
 209                 $this->db->doQuery($SQL);
 210                 return true;
 211         }
 212
 213         function updateTitle( $id, $title ) {
 214                 return true;
 215         }
 216
 217 } ## end of the SearchPostgres class
 218
 219 /**
 220  * @ingroup Search
 221  */
 222 class PostgresSearchResult extends SearchResult {
 223         function __construct( $row ) {
 224                 parent::__construct($row);
 225                 $this->score = $row->score;
 226         }
 227         function getScore() {
 228                 return $this->score;
 229         }
 230 }
 231
 232 /**
 233  * @ingroup Search
 234  */
 235 class PostgresSearchResultSet extends SearchResultSet {
 236         function __construct( $resultSet, $terms ) {
 237                 $this->mResultSet = $resultSet;
 238                 $this->mTerms = $terms;
 239         }
 240
 241         function termMatches() {
 242                 return $this->mTerms;
 243         }
 244
 245         function numRows() {
 246                 return $this->mResultSet->numRows();
 247         }
 248
 249         function next() {
 250                 $row = $this->mResultSet->fetchObject();
 251                 if( $row === false ) {
 252                         return false;
 253                 } else {
 254                         return new PostgresSearchResult( $row );
 255                 }
 256         }
 257 }