maintenance/convertLinks.php

   1 <?php
   2 /**
   3  * Convert from the old links schema (string->ID) to the new schema (ID->ID).
   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  * @file
  21  * @ingroup Maintenance
  22  */
  23
  24 require_once( dirname( __FILE__ ) . '/Maintenance.php' );
  25
  26 /**
  27  * Maintenance script to convert from the old links schema (string->ID)
  28  * to the new schema (ID->ID).
  29  *
  30  * The wiki should be put into read-only mode while this script executes.
  31  *
  32  * @ingroup Maintenance
  33  */
  34 class ConvertLinks extends Maintenance {
  35         private $logPerformance;
  36
  37         public function __construct() {
  38                 parent::__construct();
  39                 $this->mDescription = "Convert from the old links schema (string->ID) to the new schema (ID->ID).
  40 The wiki should be put into read-only mode while this script executes";
  41
  42                 $this->addArg( 'logperformance', "Log performance to perfLogFilename.", false );
  43                 $this->addArg( 'perfLogFilename', "Filename where performance is logged if --logperformance was set (defaults to 'convLinksPerf.txt').", false );
  44                 $this->addArg( 'keep-links-table', "Don't overwrite the old links table with the new one, leave the new table at links_temp.", false );
  45                 $this->addArg( 'nokeys', "Don't create keys, and so allow duplicates in the new links table.\n
  46 This gives a huge speed improvement for very large links tables which are MyISAM." /* (What about InnoDB?) */, false );
  47         }
  48
  49         public function getDbType() {
  50                 return Maintenance::DB_ADMIN;
  51         }
  52
  53         public function execute() {
  54                 $dbw = wfGetDB( DB_MASTER );
  55
  56                 $type = $dbw->getType();
  57                 if ( $type != 'mysql' ) {
  58                         $this->output( "Link table conversion not necessary for $type\n" );
  59                         return;
  60                 }
  61
  62                 global $wgContLang;
  63
  64                 $numBadLinks = $curRowsRead = 0; # counters etc
  65                 $totalTuplesInserted = 0; # total tuples INSERTed into links_temp
  66
  67                 $reportCurReadProgress = true; # whether or not to give progress reports while reading IDs from cur table
  68                 $curReadReportInterval = 1000; # number of rows between progress reports
  69
  70                 $reportLinksConvProgress = true; # whether or not to give progress reports during conversion
  71                 $linksConvInsertInterval = 1000; # number of rows per INSERT
  72
  73                 $initialRowOffset = 0;
  74                 # $finalRowOffset = 0; # not used yet; highest row number from links table to process
  75
  76                 $overwriteLinksTable = !$this->hasOption( 'keep-links-table' );
  77                 $noKeys = $this->hasOption( 'noKeys' );
  78                 $this->logPerformance = $this->hasOption( 'logperformance' );
  79                 $perfLogFilename = $this->getArg( 'perfLogFilename', "convLinksPerf.txt" );
  80
  81                 # --------------------------------------------------------------------
  82
  83                 list ( $cur, $links, $links_temp, $links_backup ) = $dbw->tableNamesN( 'cur', 'links', 'links_temp', 'links_backup' );
  84
  85                 if( $dbw->tableExists( 'pagelinks' ) ) {
  86                         $this->output( "...have pagelinks; skipping old links table updates\n" );
  87                         return;
  88                 }
  89
  90                 $res = $dbw->query( "SELECT l_from FROM $links LIMIT 1" );
  91                 if ( $dbw->fieldType( $res, 0 ) == "int" ) {
  92                         $this->output( "Schema already converted\n" );
  93                         return;
  94                 }
  95
  96                 $res = $dbw->query( "SELECT COUNT(*) AS count FROM $links" );
  97                 $row = $dbw->fetchObject( $res );
  98                 $numRows = $row->count;
  99                 $dbw->freeResult( $res );
 100
 101                 if ( $numRows == 0 ) {
 102                         $this->output( "Updating schema (no rows to convert)...\n" );
 103                         $this->createTempTable();
 104                 } else {
 105                         $fh = false;
 106                         if ( $this->logPerformance ) {
 107                                 $fh = fopen ( $perfLogFilename, "w" );
 108                                 if ( !$fh ) {
 109                                         $this->error( "Couldn't open $perfLogFilename" );
 110                                         $this->logPerformance = false;
 111                                 }
 112                         }
 113                         $baseTime = $startTime = $this->getMicroTime();
 114                         # Create a title -> cur_id map
 115                         $this->output( "Loading IDs from $cur table...\n" );
 116                         $this->performanceLog ( $fh, "Reading $numRows rows from cur table...\n" );
 117                         $this->performanceLog ( $fh, "rows read vs seconds elapsed:\n" );
 118
 119                         $dbw->bufferResults( false );
 120                         $res = $dbw->query( "SELECT cur_namespace,cur_title,cur_id FROM $cur" );
 121                         $ids = array();
 122
 123                         foreach ( $res as $row ) {
 124                                 $title = $row->cur_title;
 125                                 if ( $row->cur_namespace ) {
 126                                         $title = $wgContLang->getNsText( $row->cur_namespace ) . ":$title";
 127                                 }
 128                                 $ids[$title] = $row->cur_id;
 129                                 $curRowsRead++;
 130                                 if ( $reportCurReadProgress ) {
 131                                         if ( ( $curRowsRead % $curReadReportInterval ) == 0 ) {
 132                                                 $this->performanceLog( $fh, $curRowsRead . " " . ( $this->getMicroTime() - $baseTime ) . "\n" );
 133                                                 $this->output( "\t$curRowsRead rows of $cur table read.\n" );
 134                                         }
 135                                 }
 136                         }
 137                         $dbw->freeResult( $res );
 138                         $dbw->bufferResults( true );
 139                         $this->output( "Finished loading IDs.\n\n" );
 140                         $this->performanceLog( $fh, "Took " . ( $this->getMicroTime() - $baseTime ) . " seconds to load IDs.\n\n" );
 141
 142                         # --------------------------------------------------------------------
 143
 144                         # Now, step through the links table (in chunks of $linksConvInsertInterval rows),
 145                         # convert, and write to the new table.
 146                         $this->createTempTable();
 147                         $this->performanceLog( $fh, "Resetting timer.\n\n" );
 148                         $baseTime = $this->getMicroTime();
 149                         $this->output( "Processing $numRows rows from $links table...\n" );
 150                         $this->performanceLog( $fh, "Processing $numRows rows from $links table...\n" );
 151                         $this->performanceLog( $fh, "rows inserted vs seconds elapsed:\n" );
 152
 153                         for ( $rowOffset = $initialRowOffset; $rowOffset < $numRows; $rowOffset += $linksConvInsertInterval ) {
 154                                 $sqlRead = "SELECT * FROM $links ";
 155                                 $sqlRead = $dbw->limitResult( $sqlRead, $linksConvInsertInterval, $rowOffset );
 156                                 $res = $dbw->query( $sqlRead );
 157                                 if ( $noKeys ) {
 158                                         $sqlWrite = array( "INSERT INTO $links_temp (l_from,l_to) VALUES " );
 159                                 } else {
 160                                         $sqlWrite = array( "INSERT IGNORE INTO $links_temp (l_from,l_to) VALUES " );
 161                                 }
 162
 163                                 $tuplesAdded = 0; # no tuples added to INSERT yet
 164                                 foreach ( $res as $row ) {
 165                                         $fromTitle = $row->l_from;
 166                                         if ( array_key_exists( $fromTitle, $ids ) ) { # valid title
 167                                                 $from = $ids[$fromTitle];
 168                                                 $to = $row->l_to;
 169                                                 if ( $tuplesAdded != 0 ) {
 170                                                         $sqlWrite[] = ",";
 171                                                 }
 172                                                 $sqlWrite[] = "($from,$to)";
 173                                                 $tuplesAdded++;
 174                                         } else { # invalid title
 175                                                 $numBadLinks++;
 176                                         }
 177                                 }
 178                                 $dbw->freeResult( $res );
 179                                 # $this->output( "rowOffset: $rowOffset\ttuplesAdded: $tuplesAdded\tnumBadLinks: $numBadLinks\n" );
 180                                 if ( $tuplesAdded != 0  ) {
 181                                         if ( $reportLinksConvProgress ) {
 182                                                 $this->output( "Inserting $tuplesAdded tuples into $links_temp..." );
 183                                         }
 184                                         $dbw->query( implode( "", $sqlWrite ) );
 185                                         $totalTuplesInserted += $tuplesAdded;
 186                                         if ( $reportLinksConvProgress )
 187                                                 $this->output( " done. Total $totalTuplesInserted tuples inserted.\n" );
 188                                                 $this->performanceLog( $fh, $totalTuplesInserted . " " . ( $this->getMicroTime() - $baseTime ) . "\n"  );
 189                                 }
 190                         }
 191                         $this->output( "$totalTuplesInserted valid titles and $numBadLinks invalid titles were processed.\n\n" );
 192                         $this->performanceLog( $fh, "$totalTuplesInserted valid titles and $numBadLinks invalid titles were processed.\n" );
 193                         $this->performanceLog( $fh, "Total execution time: " . ( $this->getMicroTime() - $startTime ) . " seconds.\n" );
 194                         if ( $this->logPerformance ) {
 195                                 fclose ( $fh );
 196                         }
 197                 }
 198                 # --------------------------------------------------------------------
 199
 200                 if ( $overwriteLinksTable ) {
 201                         # Check for existing links_backup, and delete it if it exists.
 202                         $this->output( "Dropping backup links table if it exists..." );
 203                         $dbw->query( "DROP TABLE IF EXISTS $links_backup", __METHOD__ );
 204                         $this->output( " done.\n" );
 205
 206                         # Swap in the new table, and move old links table to links_backup
 207                         $this->output( "Swapping tables '$links' to '$links_backup'; '$links_temp' to '$links'..." );
 208                         $dbw->query( "RENAME TABLE links TO $links_backup, $links_temp TO $links", __METHOD__ );
 209                         $this->output( " done.\n\n" );
 210
 211                         $dbw->close();
 212                         $this->output( "Conversion complete. The old table remains at $links_backup;\n" );
 213                         $this->output( "delete at your leisure.\n" );
 214                 } else {
 215                         $this->output( "Conversion complete.  The converted table is at $links_temp;\n" );
 216                         $this->output( "the original links table is unchanged.\n" );
 217                 }
 218         }
 219
 220         private function createTempTable() {
 221                 $dbConn = wfGetDB( DB_MASTER );
 222
 223                 if ( !( $dbConn->isOpen() ) ) {
 224                         $this->output( "Opening connection to database failed.\n" );
 225                         return;
 226                 }
 227                 $links_temp = $dbConn->tableName( 'links_temp' );
 228
 229                 $this->output( "Dropping temporary links table if it exists..." );
 230                 $dbConn->query( "DROP TABLE IF EXISTS $links_temp" );
 231                 $this->output( " done.\n" );
 232
 233                 $this->output( "Creating temporary links table..." );
 234                 if ( $this->hasOption( 'noKeys' ) ) {
 235                         $dbConn->query( "CREATE TABLE $links_temp ( " .
 236                         "l_from int(8) unsigned NOT NULL default '0', " .
 237                         "l_to int(8) unsigned NOT NULL default '0')" );
 238                 } else {
 239                         $dbConn->query( "CREATE TABLE $links_temp ( " .
 240                         "l_from int(8) unsigned NOT NULL default '0', " .
 241                         "l_to int(8) unsigned NOT NULL default '0', " .
 242                         "UNIQUE KEY l_from(l_from,l_to), " .
 243                         "KEY (l_to))" );
 244                 }
 245                 $this->output( " done.\n\n" );
 246         }
 247
 248         private function performanceLog( $fh, $text ) {
 249                 if ( $this->logPerformance ) {
 250                         fwrite( $fh, $text );
 251                 }
 252         }
 253
 254         private function getMicroTime() { # return time in seconds, with microsecond accuracy
 255                 list( $usec, $sec ) = explode( " ", microtime() );
 256                 return ( (float)$usec + (float)$sec );
 257         }
 258 }
 259
 260 $maintClass = "ConvertLinks";
 261 require_once( RUN_MAINTENANCE_IF_MAIN );