includes/search/SearchPostgres.php

   1 <?php
   2 /**
   3  * PostgreSQL search engine
   4  *
   5  * Copyright © 2006-2007 Greg Sabino Mullane <greg@turnstep.com>
   6  * http://www.mediawiki.org/
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Search
  25  */
  26
  27 /**
  28  * Search engine hook base class for Postgres
  29  * @ingroup Search
  30  */
  31 class SearchPostgres extends SearchEngine {
  32         /**
  33          * Creates an instance of this class
  34          * @param $db DatabaseSqlite: database object
  35          */
  36         function __construct( $db ) {
  37                 parent::__construct( $db );
  38         }
  39
  40         /**
  41          * Perform a full text search query via tsearch2 and return a result set.
  42          * Currently searches a page's current title (page.page_title) and
  43          * latest revision article text (pagecontent.old_text)
  44          *
  45          * @param $term String: raw search term
  46          * @return PostgresSearchResultSet
  47          */
  48         function searchTitle( $term ) {
  49                 $q = $this->searchQuery( $term , 'titlevector', 'page_title' );
  50                 $olderror = error_reporting(E_ERROR);
  51                 $resultSet = $this->db->resultObject( $this->db->query( $q, 'SearchPostgres', true ) );
  52                 error_reporting($olderror);
  53                 if (!$resultSet) {
  54                         // Needed for "Query requires full scan, GIN doesn't support it"
  55                         return new SearchResultTooMany();
  56                 }
  57                 return new PostgresSearchResultSet( $resultSet, $this->searchTerms );
  58         }
  59         function searchText( $term ) {
  60                 $q = $this->searchQuery( $term, 'textvector', 'old_text' );
  61                 $olderror = error_reporting(E_ERROR);
  62                 $resultSet = $this->db->resultObject( $this->db->query( $q, 'SearchPostgres', true ) );
  63                 error_reporting($olderror);
  64                 if (!$resultSet) {
  65                         return new SearchResultTooMany();
  66                 }
  67                 return new PostgresSearchResultSet( $resultSet, $this->searchTerms );
  68         }
  69
  70
  71         /*
  72          * Transform the user's search string into a better form for tsearch2
  73          * Returns an SQL fragment consisting of quoted text to search for.
  74         */
  75         function parseQuery( $term ) {
  76
  77                 wfDebug( "parseQuery received: $term \n" );
  78
  79                 ## No backslashes allowed
  80                 $term = preg_replace('/\\\/', '', $term);
  81
  82                 ## Collapse parens into nearby words:
  83                 $term = preg_replace('/\s*\(\s*/', ' (', $term);
  84                 $term = preg_replace('/\s*\)\s*/', ') ', $term);
  85
  86                 ## Treat colons as word separators:
  87                 $term = preg_replace('/:/', ' ', $term);
  88
  89                 $searchstring = '';
  90                 $m = array();
  91                 if( preg_match_all('/([-!]?)(\S+)\s*/', $term, $m, PREG_SET_ORDER ) ) {
  92                         foreach( $m as $terms ) {
  93                                 if (strlen($terms[1])) {
  94                                         $searchstring .= ' & !';
  95                                 }
  96                                 if (strtolower($terms[2]) === 'and') {
  97                                         $searchstring .= ' & ';
  98                                 }
  99                                 else if (strtolower($terms[2]) === 'or' or $terms[2] === '|') {
 100                                         $searchstring .= ' | ';
 101                                 }
 102                                 else if (strtolower($terms[2]) === 'not') {
 103                                         $searchstring .= ' & !';
 104                                 }
 105                                 else {
 106                                         $searchstring .= " & $terms[2]";
 107                                 }
 108                         }
 109                 }
 110
 111                 ## Strip out leading junk
 112                 $searchstring = preg_replace('/^[\s\&\|]+/', '', $searchstring);
 113
 114                 ## Remove any doubled-up operators
 115                 $searchstring = preg_replace('/([\!\&\|]) +(?:[\&\|] +)+/', "$1 ", $searchstring);
 116
 117                 ## Remove any non-spaced operators (e.g. "Zounds!")
 118                 $searchstring = preg_replace('/([^ ])[\!\&\|]/', "$1", $searchstring);
 119
 120                 ## Remove any trailing whitespace or operators
 121                 $searchstring = preg_replace('/[\s\!\&\|]+$/', '', $searchstring);
 122
 123                 ## Remove unnecessary quotes around everything
 124                 $searchstring = preg_replace('/^[\'"](.*)[\'"]$/', "$1", $searchstring);
 125
 126                 ## Quote the whole thing
 127                 $searchstring = $this->db->addQuotes($searchstring);
 128
 129                 wfDebug( "parseQuery returned: $searchstring \n" );
 130
 131                 return $searchstring;
 132
 133         }
 134
 135         /**
 136          * Construct the full SQL query to do the search.
 137          * @param $term String
 138          * @param $fulltext String
 139          * @param $colname
 140          */
 141         function searchQuery( $term, $fulltext, $colname ) {
 142                 $postgresVersion = $this->db->getServerVersion();
 143
 144                 $prefix = $postgresVersion < 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 = $postgresVersion > 8.2 ? 5 : 1;
 172                         $rank = $postgresVersion < 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 SqlSearchResultSet {
 236         function __construct( $resultSet, $terms ) {
 237                 parent::__construct( $resultSet, $terms );
 238         }
 239
 240         function next() {
 241                 $row = $this->mResultSet->fetchObject();
 242                 if( $row === false ) {
 243                         return false;
 244                 } else {
 245                         return new PostgresSearchResult( $row );
 246                 }
 247         }
 248 }