maintenance/convertLinks.php

   1 <?php
   2 /**
   3  * Convert from the old links schema (string->ID) to the new schema (ID->ID)
   4  * The wiki should be put into read-only mode while this script executes
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @ingroup Maintenance
  22  */
  23
  24 require_once( dirname( __FILE__ ) . '/Maintenance.php' );
  25
  26 class ConvertLinks extends Maintenance {
  27
  28         public function __construct() {
  29                 parent::__construct();
  30                 $this->mDescription = "Convert from the old links schema (string->ID) to the new schema (ID->ID)
  31 The wiki should be put into read-only mode while this script executes";
  32         }
  33
  34         public function execute() {
  35                 global $wgDBtype;
  36                 if ( $wgDBtype == 'postgres' ) {
  37                         $this->output( "Links table already ok on Postgres.\n" );
  38                         return;
  39                 }
  40
  41                 $this->output( "Converting links table to ID-ID...\n" );
  42
  43                 global $wgLang, $wgDBserver, $wgDBadminuser, $wgDBadminpassword, $wgDBname;
  44                 global $noKeys, $logPerformance, $fh;
  45
  46                 $tuplesAdded = $numBadLinks = $curRowsRead = 0; # counters etc
  47                 $totalTuplesInserted = 0; # total tuples INSERTed into links_temp
  48
  49                 $reportCurReadProgress = true; # whether or not to give progress reports while reading IDs from cur table
  50                 $curReadReportInterval = 1000; # number of rows between progress reports
  51
  52                 $reportLinksConvProgress = true; # whether or not to give progress reports during conversion
  53                 $linksConvInsertInterval = 1000; # number of rows per INSERT
  54
  55                 $initialRowOffset = 0;
  56                 # $finalRowOffset = 0; # not used yet; highest row number from links table to process
  57
  58                 # Overwrite the old links table with the new one.  If this is set to false,
  59                 # the new table will be left at links_temp.
  60                 $overwriteLinksTable = true;
  61
  62                 # Don't create keys, and so allow duplicates in the new links table.
  63                 # This gives a huge speed improvement for very large links tables which are MyISAM. (What about InnoDB?)
  64                 $noKeys = false;
  65
  66
  67                 $logPerformance = false; # output performance data to a file
  68                 $perfLogFilename = "convLinksPerf.txt";
  69                 # --------------------------------------------------------------------
  70
  71                 $dbw = wfGetDB( DB_MASTER );
  72                 list ( $cur, $links, $links_temp, $links_backup ) = $dbw->tableNamesN( 'cur', 'links', 'links_temp', 'links_backup' );
  73
  74                 $res = $dbw->query( "SELECT l_from FROM $links LIMIT 1" );
  75                 if ( $dbw->fieldType( $res, 0 ) == "int" ) {
  76                         $this->output( "Schema already converted\n" );
  77                         return;
  78                 }
  79
  80                 $res = $dbw->query( "SELECT COUNT(*) AS count FROM $links" );
  81                 $row = $dbw->fetchObject( $res );
  82                 $numRows = $row->count;
  83                 $dbw->freeResult( $res );
  84
  85                 if ( $numRows == 0 ) {
  86                         $this->output( "Updating schema (no rows to convert)...\n" );
  87                         $this->createTempTable();
  88                 } else {
  89                         if ( $logPerformance ) { $fh = fopen ( $perfLogFilename, "w" ); }
  90                         $baseTime = $startTime = $this->getMicroTime();
  91                         # Create a title -> cur_id map
  92                         $this->output( "Loading IDs from $cur table...\n" );
  93                         $this->performanceLog ( "Reading $numRows rows from cur table...\n" );
  94                         $this->performanceLog ( "rows read vs seconds elapsed:\n" );
  95
  96                         $dbw->bufferResults( false );
  97                         $res = $dbw->query( "SELECT cur_namespace,cur_title,cur_id FROM $cur" );
  98                         $ids = array();
  99
 100                         while ( $row = $dbw->fetchObject( $res ) ) {
 101                                 $title = $row->cur_title;
 102                                 if ( $row->cur_namespace ) {
 103                                         $title = $wgLang->getNsText( $row->cur_namespace ) . ":$title";
 104                                 }
 105                                 $ids[$title] = $row->cur_id;
 106                                 $curRowsRead++;
 107                                 if ( $reportCurReadProgress ) {
 108                                         if ( ( $curRowsRead % $curReadReportInterval ) == 0 ) {
 109                                                 $this->performanceLog( $curRowsRead . " " . ( $this->getMicroTime() - $baseTime ) . "\n" );
 110                                                 $this->output( "\t$curRowsRead rows of $cur table read.\n" );
 111                                         }
 112                                 }
 113                         }
 114                         $dbw->freeResult( $res );
 115                         $dbw->bufferResults( true );
 116                         $this->output( "Finished loading IDs.\n\n" );
 117                         $this->performanceLog( "Took " . ( $this->getMicroTime() - $baseTime ) . " seconds to load IDs.\n\n" );
 118                 # --------------------------------------------------------------------
 119
 120                         # Now, step through the links table (in chunks of $linksConvInsertInterval rows),
 121                         # convert, and write to the new table.
 122                         $this->createTempTable();
 123                         $this->performanceLog( "Resetting timer.\n\n" );
 124                         $baseTime = $this->getMicroTime();
 125                         $this->output( "Processing $numRows rows from $links table...\n" );
 126                         $this->performanceLog( "Processing $numRows rows from $links table...\n" );
 127                         $this->performanceLog( "rows inserted vs seconds elapsed:\n" );
 128
 129                         for ( $rowOffset = $initialRowOffset; $rowOffset < $numRows; $rowOffset += $linksConvInsertInterval ) {
 130                                 $sqlRead = "SELECT * FROM $links ";
 131                                 $sqlRead = $dbw->limitResult( $sqlRead, $linksConvInsertInterval, $rowOffset );
 132                                 $res = $dbw->query( $sqlRead );
 133                                 if ( $noKeys ) {
 134                                         $sqlWrite = array( "INSERT INTO $links_temp (l_from,l_to) VALUES " );
 135                                 } else {
 136                                         $sqlWrite = array( "INSERT IGNORE INTO $links_temp (l_from,l_to) VALUES " );
 137                                 }
 138
 139                                 $tuplesAdded = 0; # no tuples added to INSERT yet
 140                                 while ( $row = $dbw->fetchObject( $res ) ) {
 141                                         $fromTitle = $row->l_from;
 142                                         if ( array_key_exists( $fromTitle, $ids ) ) { # valid title
 143                                                 $from = $ids[$fromTitle];
 144                                                 $to = $row->l_to;
 145                                                 if ( $tuplesAdded != 0 ) {
 146                                                         $sqlWrite[] = ",";
 147                                                 }
 148                                                 $sqlWrite[] = "($from,$to)";
 149                                                 $tuplesAdded++;
 150                                         } else { # invalid title
 151                                                 $numBadLinks++;
 152                                         }
 153                                 }
 154                                 $dbw->freeResult( $res );
 155                                 # $this->output( "rowOffset: $rowOffset\ttuplesAdded: $tuplesAdded\tnumBadLinks: $numBadLinks\n" );
 156                                 if ( $tuplesAdded != 0  ) {
 157                                         if ( $reportLinksConvProgress ) {
 158                                                 $this->output( "Inserting $tuplesAdded tuples into $links_temp..." );
 159                                         }
 160                                         $dbw->query( implode( "", $sqlWrite ) );
 161                                         $totalTuplesInserted += $tuplesAdded;
 162                                         if ( $reportLinksConvProgress )
 163                                                 $this->output( " done. Total $totalTuplesInserted tuples inserted.\n" );
 164                                                 $this->performanceLog( $totalTuplesInserted . " " . ( $this->getMicroTime() - $baseTime ) . "\n"  );
 165                                 }
 166                         }
 167                         $this->output( "$totalTuplesInserted valid titles and $numBadLinks invalid titles were processed.\n\n" );
 168                         $this->performanceLog( "$totalTuplesInserted valid titles and $numBadLinks invalid titles were processed.\n" );
 169                         $this->performanceLog( "Total execution time: " . ( $this->getMicroTime() - $startTime ) . " seconds.\n" );
 170                         if ( $logPerformance ) { fclose ( $fh ); }
 171                 }
 172                 # --------------------------------------------------------------------
 173
 174                 if ( $overwriteLinksTable ) {
 175                         $dbConn = Database::newFromParams( $wgDBserver, $wgDBadminuser, $wgDBadminpassword, $wgDBname );
 176                         if ( !( $dbConn->isOpen() ) ) {
 177                                 $this->output( "Opening connection to database failed.\n" );
 178                                 return;
 179                         }
 180                         # Check for existing links_backup, and delete it if it exists.
 181                         $this->output( "Dropping backup links table if it exists..." );
 182                         $dbConn->query( "DROP TABLE IF EXISTS $links_backup", DB_MASTER );
 183                         $this->output( " done.\n" );
 184
 185                         # Swap in the new table, and move old links table to links_backup
 186                         $this->output( "Swapping tables '$links' to '$links_backup'; '$links_temp' to '$links'..." );
 187                         $dbConn->query( "RENAME TABLE links TO $links_backup, $links_temp TO $links", DB_MASTER );
 188                         $this->output( " done.\n\n" );
 189
 190                         $dbConn->close();
 191                         $this->output( "Conversion complete. The old table remains at $links_backup;\n" );
 192                         $this->output( "delete at your leisure.\n" );
 193                 } else {
 194                         $this->output( "Conversion complete.  The converted table is at $links_temp;\n" );
 195                         $this->output( "the original links table is unchanged.\n" );
 196                 }
 197         }
 198
 199         private function createTempTable() {
 200                 global $wgDBserver, $wgDBadminuser, $wgDBadminpassword, $wgDBname;
 201                 global $noKeys;
 202                 $dbConn = Database::newFromParams( $wgDBserver, $wgDBadminuser, $wgDBadminpassword, $wgDBname );
 203
 204                 if ( !( $dbConn->isOpen() ) ) {
 205                         $this->output( "Opening connection to database failed.\n" );
 206                         return;
 207                 }
 208                 $links_temp = $dbConn->tableName( 'links_temp' );
 209
 210                 $this->output( "Dropping temporary links table if it exists..." );
 211                 $dbConn->query( "DROP TABLE IF EXISTS $links_temp" );
 212                 $this->output( " done.\n" );
 213
 214                 $this->output( "Creating temporary links table..." );
 215                 if ( $noKeys ) {
 216                         $dbConn->query( "CREATE TABLE $links_temp ( " .
 217                         "l_from int(8) unsigned NOT NULL default '0', " .
 218                         "l_to int(8) unsigned NOT NULL default '0')" );
 219                 } else {
 220                         $dbConn->query( "CREATE TABLE $links_temp ( " .
 221                         "l_from int(8) unsigned NOT NULL default '0', " .
 222                         "l_to int(8) unsigned NOT NULL default '0', " .
 223                         "UNIQUE KEY l_from(l_from,l_to), " .
 224                         "KEY (l_to))" );
 225                 }
 226                 $this->output( " done.\n\n" );
 227         }
 228
 229         private function performanceLog( $text ) {
 230                 global $logPerformance, $fh;
 231                 if ( $logPerformance ) {
 232                         fwrite( $fh, $text );
 233                 }
 234         }
 235
 236         private function getMicroTime() { # return time in seconds, with microsecond accuracy
 237                 list( $usec, $sec ) = explode( " ", microtime() );
 238                 return ( (float)$usec + (float)$sec );
 239         }
 240 }
 241
 242 $maintClass = "ConvertLinks";
 243 require_once( DO_MAINTENANCE );