includes/search/SearchPostgres.php

   1 <?php
   2 /**
   3  * PostgreSQL search engine
   4  *
   5  * Copyright © 2006-2007 Greg Sabino Mullane <greg@turnstep.com>
   6  * https://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 use MediaWiki\MediaWikiServices;
  27 use MediaWiki\Revision\SlotRecord;
  28
  29 /**
  30  * Search engine hook base class for Postgres
  31  * @ingroup Search
  32  */
  33 class SearchPostgres extends SearchDatabase {
  34         /**
  35          * Perform a full text search query via tsearch2 and return a result set.
  36          * Currently searches a page's current title (page.page_title) and
  37          * latest revision article text (pagecontent.old_text)
  38          *
  39          * @param string $term Raw search term
  40          * @return SqlSearchResultSet
  41          */
  42         protected function doSearchTitleInDB( $term ) {
  43                 $q = $this->searchQuery( $term, 'titlevector', 'page_title' );
  44                 $olderror = error_reporting( E_ERROR );
  45                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
  46                 $resultSet = $dbr->query( $q, 'SearchPostgres', true );
  47                 error_reporting( $olderror );
  48                 return new SqlSearchResultSet( $resultSet, $this->searchTerms );
  49         }
  50
  51         protected function doSearchTextInDB( $term ) {
  52                 $q = $this->searchQuery( $term, 'textvector', 'old_text' );
  53                 $olderror = error_reporting( E_ERROR );
  54                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
  55                 $resultSet = $dbr->query( $q, 'SearchPostgres', true );
  56                 error_reporting( $olderror );
  57                 return new SqlSearchResultSet( $resultSet, $this->searchTerms );
  58         }
  59
  60         /**
  61          * Transform the user's search string into a better form for tsearch2
  62          * Returns an SQL fragment consisting of quoted text to search for.
  63          *
  64          * @param string $term
  65          *
  66          * @return string
  67          */
  68         private function parseQuery( $term ) {
  69                 wfDebug( "parseQuery received: $term \n" );
  70
  71                 # # No backslashes allowed
  72                 $term = preg_replace( '/\\\/', '', $term );
  73
  74                 # # Collapse parens into nearby words:
  75                 $term = preg_replace( '/\s*\(\s*/', ' (', $term );
  76                 $term = preg_replace( '/\s*\)\s*/', ') ', $term );
  77
  78                 # # Treat colons as word separators:
  79                 $term = preg_replace( '/:/', ' ', $term );
  80
  81                 $searchstring = '';
  82                 $m = [];
  83                 if ( preg_match_all( '/([-!]?)(\S+)\s*/', $term, $m, PREG_SET_ORDER ) ) {
  84                         foreach ( $m as $terms ) {
  85                                 if ( strlen( $terms[1] ) ) {
  86                                         $searchstring .= ' & !';
  87                                 }
  88                                 if ( strtolower( $terms[2] ) === 'and' ) {
  89                                         $searchstring .= ' & ';
  90                                 } elseif ( strtolower( $terms[2] ) === 'or' || $terms[2] === '|' ) {
  91                                         $searchstring .= ' | ';
  92                                 } elseif ( strtolower( $terms[2] ) === 'not' ) {
  93                                         $searchstring .= ' & !';
  94                                 } else {
  95                                         $searchstring .= " & $terms[2]";
  96                                 }
  97                         }
  98                 }
  99
 100                 # # Strip out leading junk
 101                 $searchstring = preg_replace( '/^[\s\&\|]+/', '', $searchstring );
 102
 103                 # # Remove any doubled-up operators
 104                 $searchstring = preg_replace( '/([\!\&\|]) +(?:[\&\|] +)+/', "$1 ", $searchstring );
 105
 106                 # # Remove any non-spaced operators (e.g. "Zounds!")
 107                 $searchstring = preg_replace( '/([^ ])[\!\&\|]/', "$1", $searchstring );
 108
 109                 # # Remove any trailing whitespace or operators
 110                 $searchstring = preg_replace( '/[\s\!\&\|]+$/', '', $searchstring );
 111
 112                 # # Remove unnecessary quotes around everything
 113                 $searchstring = preg_replace( '/^[\'"](.*)[\'"]$/', "$1", $searchstring );
 114
 115                 # # Quote the whole thing
 116                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 117                 $searchstring = $dbr->addQuotes( $searchstring );
 118
 119                 wfDebug( "parseQuery returned: $searchstring \n" );
 120
 121                 return $searchstring;
 122         }
 123
 124         /**
 125          * Construct the full SQL query to do the search.
 126          * @param string $term
 127          * @param string $fulltext
 128          * @param string $colname
 129          * @return string
 130          */
 131         private function searchQuery( $term, $fulltext, $colname ) {
 132                 # Get the SQL fragment for the given term
 133                 $searchstring = $this->parseQuery( $term );
 134
 135                 # # We need a separate query here so gin does not complain about empty searches
 136                 $sql = "SELECT to_tsquery($searchstring)";
 137                 $dbr = $this->lb->getConnectionRef( DB_REPLICA );
 138                 $res = $dbr->query( $sql );
 139                 if ( !$res ) {
 140                         # # TODO: Better output (example to catch: one 'two)
 141                         die( "Sorry, that was not a valid search string. Please go back and try again" );
 142                 }
 143                 $top = $res->fetchRow()[0];
 144
 145                 $this->searchTerms = [];
 146                 $slotRoleStore = MediaWikiServices::getInstance()->getSlotRoleStore();
 147                 if ( $top === "" ) { # # e.g. if only stopwords are used XXX return something better
 148                         $query = "SELECT page_id, page_namespace, page_title, 0 AS score " .
 149                                 "FROM page p, revision r, slots s, content c, pagecontent pc " .
 150                                 "WHERE p.page_latest = r.rev_id " .
 151                                 "AND s.slot_revision_id = r.rev_id " .
 152                                 "AND s.slot_role_id = " . $slotRoleStore->getId( SlotRecord::MAIN ) . " " .
 153                                 "AND c.content_id = s.slot_content_id " .
 154                                 "AND pc.old_id = substring( c.content_address from '^tt:([0-9]+)$' )::int " .
 155                                 "AND 1=0";
 156                 } else {
 157                         $m = [];
 158                         if ( preg_match_all( "/'([^']+)'/", $top, $m, PREG_SET_ORDER ) ) {
 159                                 foreach ( $m as $terms ) {
 160                                         $this->searchTerms[$terms[1]] = $terms[1];
 161                                 }
 162                         }
 163
 164                         $query = "SELECT page_id, page_namespace, page_title, " .
 165                                 "ts_rank($fulltext, to_tsquery($searchstring), 5) AS score " .
 166                                 "FROM page p, revision r, slots s, content c, pagecontent pc " .
 167                                 "WHERE p.page_latest = r.rev_id " .
 168                                 "AND s.slot_revision_id = r.rev_id " .
 169                                 "AND s.slot_role_id = " . $slotRoleStore->getId( SlotRecord::MAIN ) . " " .
 170                                 "AND c.content_id = s.slot_content_id " .
 171                                 "AND pc.old_id = substring( c.content_address from '^tt:([0-9]+)$' )::int " .
 172                                 "AND $fulltext @@ to_tsquery($searchstring)";
 173                 }
 174                 # # Namespaces - defaults to 0
 175                 if ( !is_null( $this->namespaces ) ) { // null -> search all
 176                         if ( count( $this->namespaces ) < 1 ) {
 177                                 $query .= ' AND page_namespace = 0';
 178                         } else {
 179                                 $namespaces = $dbr->makeList( $this->namespaces );
 180                                 $query .= " AND page_namespace IN ($namespaces)";
 181                         }
 182                 }
 183
 184                 $query .= " ORDER BY score DESC, page_id DESC";
 185
 186                 $query .= $dbr->limitResult( '', $this->limit, $this->offset );
 187
 188                 wfDebug( "searchQuery returned: $query \n" );
 189
 190                 return $query;
 191         }
 192
 193         # # Most of the work of these two functions are done automatically via triggers
 194
 195         function update( $pageid, $title, $text ) {
 196                 # # We don't want to index older revisions
 197                 $slotRoleStore = MediaWikiServices::getInstance()->getSlotRoleStore();
 198                 $sql = "UPDATE pagecontent SET textvector = NULL " .
 199                         "WHERE textvector IS NOT NULL " .
 200                         "AND old_id IN " .
 201                         "(SELECT DISTINCT substring( c.content_address from '^tt:([0-9]+)$' )::int AS old_rev_text_id " .
 202                         " FROM content c, slots s, revision r " .
 203                         " WHERE r.rev_page = $pageid " .
 204                         " AND s.slot_revision_id = r.rev_id " .
 205                         " AND s.slot_role_id = " . $slotRoleStore->getId( SlotRecord::MAIN ) . " " .
 206                         " AND c.content_id = s.slot_content_id " .
 207                         " ORDER BY old_rev_text_id DESC OFFSET 1)";
 208
 209                 $dbw = $this->lb->getConnectionRef( DB_MASTER );
 210                 $dbw->query( $sql );
 211
 212                 return true;
 213         }
 214
 215         function updateTitle( $id, $title ) {
 216                 return true;
 217         }
 218 }