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 use MediaWiki\MediaWikiServices;
  25
  26 require_once __DIR__ . '/Maintenance.php';
  27
  28 /**
  29  * Maintenance script to convert from the old links schema (string->ID)
  30  * to the new schema (ID->ID).
  31  *
  32  * The wiki should be put into read-only mode while this script executes.
  33  *
  34  * @ingroup Maintenance
  35  */
  36 class ConvertLinks extends Maintenance {
  37         private $logPerformance;
  38
  39         public function __construct() {
  40                 parent::__construct();
  41                 $this->addDescription(
  42                         'Convert from the old links schema (string->ID) to the new schema (ID->ID). '
  43                                 . 'The wiki should be put into read-only mode while this script executes' );
  44
  45                 $this->addArg( 'logperformance', "Log performance to perfLogFilename.", false );
  46                 $this->addArg(
  47                         'perfLogFilename',
  48                         "Filename where performance is logged if --logperformance was set "
  49                                 . "(defaults to 'convLinksPerf.txt').",
  50                         false
  51                 );
  52                 $this->addArg(
  53                         'keep-links-table',
  54                         "Don't overwrite the old links table with the new one, leave the new table at links_temp.",
  55                         false
  56                 );
  57                 $this->addArg(
  58                         'nokeys',
  59                         /* (What about InnoDB?) */
  60                         "Don't create keys, and so allow duplicates in the new links table.\n"
  61                                 . "This gives a huge speed improvement for very large links tables which are MyISAM.",
  62                         false
  63                 );
  64         }
  65
  66         public function getDbType() {
  67                 return Maintenance::DB_ADMIN;
  68         }
  69
  70         public function execute() {
  71                 $dbw = $this->getDB( DB_MASTER );
  72
  73                 $type = $dbw->getType();
  74                 if ( $type != 'mysql' ) {
  75                         $this->output( "Link table conversion not necessary for $type\n" );
  76
  77                         return;
  78                 }
  79
  80                 # counters etc
  81                 $numBadLinks = $curRowsRead = 0;
  82
  83                 # total tuples INSERTed into links_temp
  84                 $totalTuplesInserted = 0;
  85
  86                 # whether or not to give progress reports while reading IDs from cur table
  87                 $reportCurReadProgress = true;
  88
  89                 # number of rows between progress reports
  90                 $curReadReportInterval = 1000;
  91
  92                 # whether or not to give progress reports during conversion
  93                 $reportLinksConvProgress = true;
  94
  95                 # number of rows per INSERT
  96                 $linksConvInsertInterval = 1000;
  97
  98                 $initialRowOffset = 0;
  99
 100                 # not used yet; highest row number from links table to process
 101                 # $finalRowOffset = 0;
 102
 103                 $overwriteLinksTable = !$this->hasOption( 'keep-links-table' );
 104                 $noKeys = $this->hasOption( 'noKeys' );
 105                 $this->logPerformance = $this->hasOption( 'logperformance' );
 106                 $perfLogFilename = $this->getArg( 'perfLogFilename', "convLinksPerf.txt" );
 107
 108                 # --------------------------------------------------------------------
 109
 110                 list( $cur, $links, $links_temp, $links_backup ) =
 111                         $dbw->tableNamesN( 'cur', 'links', 'links_temp', 'links_backup' );
 112
 113                 if ( $dbw->tableExists( 'pagelinks' ) ) {
 114                         $this->output( "...have pagelinks; skipping old links table updates\n" );
 115
 116                         return;
 117                 }
 118
 119                 $res = $dbw->query( "SELECT l_from FROM $links LIMIT 1" );
 120                 if ( $dbw->fieldType( $res, 0 ) == "int" ) {
 121                         $this->output( "Schema already converted\n" );
 122
 123                         return;
 124                 }
 125
 126                 $res = $dbw->query( "SELECT COUNT(*) AS count FROM $links" );
 127                 $row = $dbw->fetchObject( $res );
 128                 $numRows = $row->count;
 129
 130                 if ( $numRows == 0 ) {
 131                         $this->output( "Updating schema (no rows to convert)...\n" );
 132                         $this->createTempTable();
 133                 } else {
 134                         $fh = false;
 135                         if ( $this->logPerformance ) {
 136                                 $fh = fopen( $perfLogFilename, "w" );
 137                                 if ( !$fh ) {
 138                                         $this->error( "Couldn't open $perfLogFilename" );
 139                                         $this->logPerformance = false;
 140                                 }
 141                         }
 142                         $baseTime = $startTime = microtime( true );
 143                         # Create a title -> cur_id map
 144                         $this->output( "Loading IDs from $cur table...\n" );
 145                         $this->performanceLog( $fh, "Reading $numRows rows from cur table...\n" );
 146                         $this->performanceLog( $fh, "rows read vs seconds elapsed:\n" );
 147                         $contentLang = MediaWikiServices::getInstance()->getContentLanguage();
 148
 149                         $ids = [];
 150                         $lastId = 0;
 151                         do {
 152                                 $res = $dbw->query(
 153                                         "SELECT cur_namespace,cur_title,cur_id FROM $cur " .
 154                                         "WHERE cur_id > $lastId ORDER BY cur_id LIMIT 10000"
 155                                 );
 156                                 foreach ( $res as $row ) {
 157                                         $title = $row->cur_title;
 158                                         if ( $row->cur_namespace ) {
 159                                                 $title = $contentLang->getNsText( $row->cur_namespace ) . ":$title";
 160                                         }
 161                                         $ids[$title] = $row->cur_id;
 162                                         $curRowsRead++;
 163                                         if ( $reportCurReadProgress ) {
 164                                                 if ( ( $curRowsRead % $curReadReportInterval ) == 0 ) {
 165                                                         $this->performanceLog(
 166                                                                 $fh,
 167                                                                 $curRowsRead . " " . ( microtime( true ) - $baseTime ) . "\n"
 168                                                         );
 169                                                         $this->output( "\t$curRowsRead rows of $cur table read.\n" );
 170                                                 }
 171                                         }
 172                                         $lastId = $row->cur_id;
 173                                 }
 174                         } while ( $res->numRows() > 0 );
 175                         $this->output( "Finished loading IDs.\n\n" );
 176                         $this->performanceLog(
 177                                 $fh,
 178                                 "Took " . ( microtime( true ) - $baseTime ) . " seconds to load IDs.\n\n"
 179                         );
 180
 181                         # --------------------------------------------------------------------
 182
 183                         # Now, step through the links table (in chunks of $linksConvInsertInterval rows),
 184                         # convert, and write to the new table.
 185                         $this->createTempTable();
 186                         $this->performanceLog( $fh, "Resetting timer.\n\n" );
 187                         $baseTime = microtime( true );
 188                         $this->output( "Processing $numRows rows from $links table...\n" );
 189                         $this->performanceLog( $fh, "Processing $numRows rows from $links table...\n" );
 190                         $this->performanceLog( $fh, "rows inserted vs seconds elapsed:\n" );
 191
 192                         for ( $rowOffset = $initialRowOffset; $rowOffset < $numRows;
 193                                 $rowOffset += $linksConvInsertInterval
 194                         ) {
 195                                 $sqlRead = "SELECT * FROM $links ";
 196                                 $sqlRead = $dbw->limitResult( $sqlRead, $linksConvInsertInterval, $rowOffset );
 197                                 $res = $dbw->query( $sqlRead );
 198                                 if ( $noKeys ) {
 199                                         $sqlWrite = [ "INSERT INTO $links_temp (l_from,l_to) VALUES " ];
 200                                 } else {
 201                                         $sqlWrite = [ "INSERT IGNORE INTO $links_temp (l_from,l_to) VALUES " ];
 202                                 }
 203
 204                                 $tuplesAdded = 0; # no tuples added to INSERT yet
 205                                 foreach ( $res as $row ) {
 206                                         $fromTitle = $row->l_from;
 207                                         if ( array_key_exists( $fromTitle, $ids ) ) { # valid title
 208                                                 $from = $ids[$fromTitle];
 209                                                 $to = $row->l_to;
 210                                                 if ( $tuplesAdded != 0 ) {
 211                                                         $sqlWrite[] = ",";
 212                                                 }
 213                                                 $sqlWrite[] = "($from,$to)";
 214                                                 $tuplesAdded++;
 215                                         } else { # invalid title
 216                                                 $numBadLinks++;
 217                                         }
 218                                 }
 219                                 # $this->output( "rowOffset: $rowOffset\ttuplesAdded: "
 220                                 #       . "$tuplesAdded\tnumBadLinks: $numBadLinks\n" );
 221                                 if ( $tuplesAdded != 0 ) {
 222                                         if ( $reportLinksConvProgress ) {
 223                                                 $this->output( "Inserting $tuplesAdded tuples into $links_temp..." );
 224                                         }
 225                                         $dbw->query( implode( "", $sqlWrite ) );
 226                                         $totalTuplesInserted += $tuplesAdded;
 227                                         if ( $reportLinksConvProgress ) {
 228                                                 $this->output( " done. Total $totalTuplesInserted tuples inserted.\n" );
 229                                                 $this->performanceLog(
 230                                                         $fh,
 231                                                         $totalTuplesInserted . " " . ( microtime( true ) - $baseTime ) . "\n"
 232                                                 );
 233                                         }
 234                                 }
 235                         }
 236                         $this->output( "$totalTuplesInserted valid titles and "
 237                                 . "$numBadLinks invalid titles were processed.\n\n" );
 238                         $this->performanceLog(
 239                                 $fh,
 240                                 "$totalTuplesInserted valid titles and $numBadLinks invalid titles were processed.\n"
 241                         );
 242                         $this->performanceLog(
 243                                 $fh,
 244                                 "Total execution time: " . ( microtime( true ) - $startTime ) . " seconds.\n"
 245                         );
 246                         if ( $this->logPerformance ) {
 247                                 fclose( $fh );
 248                         }
 249                 }
 250                 # --------------------------------------------------------------------
 251
 252                 if ( $overwriteLinksTable ) {
 253                         # Check for existing links_backup, and delete it if it exists.
 254                         $this->output( "Dropping backup links table if it exists..." );
 255                         $dbw->query( "DROP TABLE IF EXISTS $links_backup", __METHOD__ );
 256                         $this->output( " done.\n" );
 257
 258                         # Swap in the new table, and move old links table to links_backup
 259                         $this->output( "Swapping tables '$links' to '$links_backup'; '$links_temp' to '$links'..." );
 260                         $dbw->query( "RENAME TABLE links TO $links_backup, $links_temp TO $links", __METHOD__ );
 261                         $this->output( " done.\n\n" );
 262
 263                         $this->output( "Conversion complete. The old table remains at $links_backup;\n" );
 264                         $this->output( "delete at your leisure.\n" );
 265                 } else {
 266                         $this->output( "Conversion complete.  The converted table is at $links_temp;\n" );
 267                         $this->output( "the original links table is unchanged.\n" );
 268                 }
 269         }
 270
 271         private function createTempTable() {
 272                 $dbConn = $this->getDB( DB_MASTER );
 273
 274                 if ( !( $dbConn->isOpen() ) ) {
 275                         $this->output( "Opening connection to database failed.\n" );
 276
 277                         return;
 278                 }
 279                 $links_temp = $dbConn->tableName( 'links_temp' );
 280
 281                 $this->output( "Dropping temporary links table if it exists..." );
 282                 $dbConn->query( "DROP TABLE IF EXISTS $links_temp" );
 283                 $this->output( " done.\n" );
 284
 285                 $this->output( "Creating temporary links table..." );
 286                 if ( $this->hasOption( 'noKeys' ) ) {
 287                         $dbConn->query( "CREATE TABLE $links_temp ( " .
 288                                 "l_from int(8) unsigned NOT NULL default '0', " .
 289                                 "l_to int(8) unsigned NOT NULL default '0')" );
 290                 } else {
 291                         $dbConn->query( "CREATE TABLE $links_temp ( " .
 292                                 "l_from int(8) unsigned NOT NULL default '0', " .
 293                                 "l_to int(8) unsigned NOT NULL default '0', " .
 294                                 "UNIQUE KEY l_from(l_from,l_to), " .
 295                                 "KEY (l_to))" );
 296                 }
 297                 $this->output( " done.\n\n" );
 298         }
 299
 300         private function performanceLog( $fh, $text ) {
 301                 if ( $this->logPerformance ) {
 302                         fwrite( $fh, $text );
 303                 }
 304         }
 305 }
 306
 307 $maintClass = ConvertLinks::class;
 308 require_once RUN_MAINTENANCE_IF_MAIN;