maintenance/storage/compressOld.php

   1 <?php
   2 /**
   3  * Compress the text of a wiki.
   4  *
   5  * Usage:
   6  *
   7  * Non-wikimedia
   8  * php compressOld.php [options...]
   9  *
  10  * Wikimedia
  11  * php compressOld.php <database> [options...]
  12  *
  13  * Options are:
  14  *  -t <type>           set compression type to either:
  15  *                          gzip: compress revisions independently
  16  *                          concat: concatenate revisions and compress in chunks (default)
  17  *  -c <chunk-size>     maximum number of revisions in a concat chunk
  18  *  -b <begin-date>     earliest date to check for uncompressed revisions
  19  *  -e <end-date>       latest revision date to compress
  20  *  -s <start-id>       the old_id to start from
  21  *  --extdb <cluster>   store specified revisions in an external cluster (untested)
  22  *
  23  * This program is free software; you can redistribute it and/or modify
  24  * it under the terms of the GNU General Public License as published by
  25  * the Free Software Foundation; either version 2 of the License, or
  26  * (at your option) any later version.
  27  *
  28  * This program is distributed in the hope that it will be useful,
  29  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  30  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  31  * GNU General Public License for more details.
  32  *
  33  * You should have received a copy of the GNU General Public License along
  34  * with this program; if not, write to the Free Software Foundation, Inc.,
  35  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  36  * http://www.gnu.org/copyleft/gpl.html
  37  *
  38  * @file
  39  * @ingroup Maintenance ExternalStorage
  40  */
  41
  42 require_once( dirname( __FILE__ ) . '/../Maintenance.php' );
  43
  44 class CompressOld extends Maintenance {
  45         /**
  46          * @todo document
  47          */
  48         const LS_INDIVIDUAL = 0;
  49         const LS_CHUNKED    = 1;
  50
  51         public function __construct() {
  52                 parent::__construct();
  53                 $this->mDescription = 'Compress the text of a wiki';
  54                 $this->addOption( 'type', 'Set compression type to either: gzip|concat', false, true, 't' );
  55                 $this->addOption( 'chunksize', 'Maximum number of revisions in a concat chunk', false, true, 'c' );
  56                 $this->addOption( 'begin-date', 'Earliest date to check for uncompressed revisions', false, true, 'b' );
  57                 $this->addOption( 'end-date', 'Latest revision date to compress', false, true, 'e' );
  58                 $this->addOption( 'start-id', 'The old_id to start from', false, true, 's');
  59                 $this->addOption( 'extdb', 'Store specified revisions in an external cluster (untested)', false, true );
  60                 $this->addOption( 'endid', 'Stop at this old_id', false, true, 'n' );
  61         }
  62
  63         public function execute() {
  64                 global $wgDBname;
  65                 if ( !function_exists( "gzdeflate" ) ) {
  66                         $this->error( "You must enable zlib support in PHP to compress old revisions!\n" .
  67                                 "Please see http://www.php.net/manual/en/ref.zlib.php\n", true );
  68                 }
  69
  70                 $type = $this->getOption( 'type', 'concat' );
  71                 $chunkSize = $this->getOption( 'chunksize', 20 );
  72                 $startId = $this->getOption( 'start-id', 0 );
  73                 $beginDate = $this->getOption( 'begin-date', '' );
  74                 $endDate = $this->getOption( 'end-date', '' );
  75                 $extDB = $this->getOption( 'extdb', '' );
  76                 $endId = $this->getOption( 'endid', false );
  77
  78                 if ( $type != 'concat' && $type != 'gzip' ) {
  79                         $this->error( "Type \"{$type}\" not supported" );
  80                 }
  81
  82                 if ( $extDB != '' ) {
  83                         $this->output( "Compressing database {$wgDBname} to external cluster {$extDB}\n"
  84                                 . str_repeat( '-', 76 ) . "\n\n" );
  85                 } else {
  86                         $this->output( "Compressing database {$wgDBname}\n"
  87                                 . str_repeat( '-', 76 ) . "\n\n" );
  88                 }
  89
  90                 $success = true;
  91                 if ( $type == 'concat' ) {
  92                         $success = $this->compressWithConcat( $startId, $chunkSize, $beginDate,
  93                                 $endDate, $extDB, $endId );
  94                 } else {
  95                         $this->compressOldPages( $startId, $extDB );
  96                 }
  97
  98                 if ( $success ) {
  99                         $this->output( "Done.\n" );
 100                 }
 101         }
 102
 103         /** @todo document */
 104         private function compressOldPages( $start = 0, $extdb = '' ) {
 105                 $chunksize = 50;
 106                 $this->output( "Starting from old_id $start...\n" );
 107                 $dbw = wfGetDB( DB_MASTER );
 108                 do {
 109                         $res = $dbw->select( 'text', array( 'old_id','old_flags','old_text' ),
 110                                 "old_id>=$start", __METHOD__, array( 'ORDER BY' => 'old_id', 'LIMIT' => $chunksize, 'FOR UPDATE' ) );
 111                         if( $dbw->numRows( $res ) == 0 ) {
 112                                 break;
 113                         }
 114                         $last = $start;
 115                         foreach ( $res as $row ) {
 116                                 # print "  {$row->old_id} - {$row->old_namespace}:{$row->old_title}\n";
 117                                 $this->compressPage( $row, $extdb );
 118                                 $last = $row->old_id;
 119                         }
 120                         $start = $last + 1; # Deletion may leave long empty stretches
 121                         $this->output( "$start...\n" );
 122                 } while( true );
 123         }
 124
 125         /** @todo document */
 126         private function compressPage( $row, $extdb ) {
 127                 if ( false !== strpos( $row->old_flags, 'gzip' ) || false !== strpos( $row->old_flags, 'object' ) ) {
 128                         #print "Already compressed row {$row->old_id}\n";
 129                         return false;
 130                 }
 131                 $dbw = wfGetDB( DB_MASTER );
 132                 $flags = $row->old_flags ? "{$row->old_flags},gzip" : "gzip";
 133                 $compress = gzdeflate( $row->old_text );
 134
 135                 # Store in external storage if required
 136                 if ( $extdb !== '' ) {
 137                         $storeObj = new ExternalStoreDB;
 138                         $compress = $storeObj->store( $extdb, $compress );
 139                         if ( $compress === false ) {
 140                                 $this->error( "Unable to store object" );
 141                                 return false;
 142                         }
 143                 }
 144
 145                 # Update text row
 146                 $dbw->update( 'text',
 147                         array( /* SET */
 148                                 'old_flags' => $flags,
 149                                 'old_text' => $compress
 150                         ), array( /* WHERE */
 151                                 'old_id' => $row->old_id
 152                         ), __METHOD__,
 153                         array( 'LIMIT' => 1 )
 154                 );
 155                 return true;
 156         }
 157
 158         /** @todo document */
 159         private function compressWithConcat( $startId, $maxChunkSize, $beginDate,
 160                 $endDate, $extdb = "", $maxPageId = false )
 161         {
 162                 $loadStyle = self::LS_CHUNKED;
 163
 164                 $dbr = wfGetDB( DB_SLAVE );
 165                 $dbw = wfGetDB( DB_MASTER );
 166
 167                 # Set up external storage
 168                 if ( $extdb != '' ) {
 169                         $storeObj = new ExternalStoreDB;
 170                 }
 171
 172                 # Get all articles by page_id
 173                 if ( !$maxPageId ) {
 174                         $maxPageId = $dbr->selectField( 'page', 'max(page_id)', '', __METHOD__ );
 175                 }
 176                 $this->output( "Starting from $startId of $maxPageId\n" );
 177                 $pageConds = array();
 178
 179                 /*
 180                 if ( $exclude_ns0 ) {
 181                         print "Excluding main namespace\n";
 182                         $pageConds[] = 'page_namespace<>0';
 183                 }
 184                 if ( $queryExtra ) {
 185                                         $pageConds[] = $queryExtra;
 186                 }
 187                  */
 188
 189                 # For each article, get a list of revisions which fit the criteria
 190
 191                 # No recompression, use a condition on old_flags
 192                 # Don't compress object type entities, because that might produce data loss when
 193                 # overwriting bulk storage concat rows. Don't compress external references, because
 194                 # the script doesn't yet delete rows from external storage.
 195                 $conds = array(
 196                         'old_flags NOT ' . $dbr->buildLike( $dbr->anyString(), 'object', $dbr->anyString() ) . ' AND old_flags NOT '
 197                                 . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() ) );
 198
 199                 if ( $beginDate ) {
 200                         if ( !preg_match( '/^\d{14}$/', $beginDate ) ) {
 201                                 $this->error( "Invalid begin date \"$beginDate\"\n" );
 202                                 return false;
 203                         }
 204                         $conds[] = "rev_timestamp>'" . $beginDate . "'";
 205                 }
 206                 if ( $endDate )  {
 207                         if ( !preg_match( '/^\d{14}$/', $endDate ) ) {
 208                                 $this->error( "Invalid end date \"$endDate\"\n" );
 209                                 return false;
 210                         }
 211                         $conds[] = "rev_timestamp<'" . $endDate . "'";
 212                 }
 213                 if ( $loadStyle == self::LS_CHUNKED ) {
 214                         $tables = array( 'revision', 'text' );
 215                         $fields = array( 'rev_id', 'rev_text_id', 'old_flags', 'old_text' );
 216                         $conds[] = 'rev_text_id=old_id';
 217                         $revLoadOptions = 'FOR UPDATE';
 218                 } else {
 219                         $tables = array( 'revision' );
 220                         $fields = array( 'rev_id', 'rev_text_id' );
 221                         $revLoadOptions = array();
 222                 }
 223
 224                 # Don't work with current revisions
 225                 # Don't lock the page table for update either -- TS 2006-04-04
 226                 #$tables[] = 'page';
 227                 #$conds[] = 'page_id=rev_page AND rev_id != page_latest';
 228
 229                 for ( $pageId = $startId; $pageId <= $maxPageId; $pageId++ ) {
 230                         wfWaitForSlaves();
 231
 232                         # Wake up
 233                         $dbr->ping();
 234
 235                         # Get the page row
 236                         $pageRes = $dbr->select( 'page',
 237                                 array('page_id', 'page_namespace', 'page_title','page_latest'),
 238                                 $pageConds + array('page_id' => $pageId), __METHOD__ );
 239                         if ( $dbr->numRows( $pageRes ) == 0 ) {
 240                                 continue;
 241                         }
 242                         $pageRow = $dbr->fetchObject( $pageRes );
 243
 244                         # Display progress
 245                         $titleObj = Title::makeTitle( $pageRow->page_namespace, $pageRow->page_title );
 246                         $this->output( "$pageId\t" . $titleObj->getPrefixedDBkey() . " " );
 247
 248                         # Load revisions
 249                         $revRes = $dbw->select( $tables, $fields,
 250                                 array_merge( array(
 251                                         'rev_page' => $pageRow->page_id,
 252                                         # Don't operate on the current revision
 253                                         # Use < instead of <> in case the current revision has changed
 254                                         # since the page select, which wasn't locking
 255                                         'rev_id < ' . $pageRow->page_latest
 256                                 ), $conds ),
 257                                 __METHOD__,
 258                                 $revLoadOptions
 259                         );
 260                         $revs = array();
 261                         foreach ( $revRes as $revRow ) {
 262                                 $revs[] = $revRow;
 263                         }
 264
 265                         if ( count( $revs ) < 2) {
 266                                 # No revisions matching, no further processing
 267                                 $this->output( "\n" );
 268                                 continue;
 269                         }
 270
 271                         # For each chunk
 272                         $i = 0;
 273                         while ( $i < count( $revs ) ) {
 274                                 if ( $i < count( $revs ) - $maxChunkSize ) {
 275                                         $thisChunkSize = $maxChunkSize;
 276                                 } else {
 277                                         $thisChunkSize = count( $revs ) - $i;
 278                                 }
 279
 280                                 $chunk = new ConcatenatedGzipHistoryBlob();
 281                                 $stubs = array();
 282                                 $dbw->begin();
 283                                 $usedChunk = false;
 284                                 $primaryOldid = $revs[$i]->rev_text_id;
 285
 286                                 # Get the text of each revision and add it to the object
 287                                 for ( $j = 0; $j < $thisChunkSize && $chunk->isHappy(); $j++ ) {
 288                                         $oldid = $revs[$i + $j]->rev_text_id;
 289
 290                                         # Get text
 291                                         if ( $loadStyle == self::LS_INDIVIDUAL ) {
 292                                                 $textRow = $dbw->selectRow( 'text',
 293                                                         array( 'old_flags', 'old_text' ),
 294                                                         array( 'old_id' => $oldid ),
 295                                                         __METHOD__,
 296                                                         'FOR UPDATE'
 297                                                 );
 298                                                 $text = Revision::getRevisionText( $textRow );
 299                                         } else {
 300                                                 $text = Revision::getRevisionText( $revs[$i + $j] );
 301                                         }
 302
 303                                         if ( $text === false ) {
 304                                                 $this->error( "\nError, unable to get text in old_id $oldid" );
 305                                                 #$dbw->delete( 'old', array( 'old_id' => $oldid ) );
 306                                         }
 307
 308                                         if ( $extdb == "" && $j == 0 ) {
 309                                                 $chunk->setText( $text );
 310                                                 $this->output( '.' );
 311                                         } else {
 312                                                 # Don't make a stub if it's going to be longer than the article
 313                                                 # Stubs are typically about 100 bytes
 314                                                 if ( strlen( $text ) < 120 ) {
 315                                                         $stub = false;
 316                                                         $this->output( 'x' );
 317                                                 } else {
 318                                                         $stub = new HistoryBlobStub( $chunk->addItem( $text ) );
 319                                                         $stub->setLocation( $primaryOldid );
 320                                                         $stub->setReferrer( $oldid );
 321                                                         $this->output( '.' );
 322                                                         $usedChunk = true;
 323                                                 }
 324                                                 $stubs[$j] = $stub;
 325                                         }
 326                                 }
 327                                 $thisChunkSize = $j;
 328
 329                                 # If we couldn't actually use any stubs because the pages were too small, do nothing
 330                                 if ( $usedChunk ) {
 331                                         if ( $extdb != "" ) {
 332                                                 # Move blob objects to External Storage
 333                                                 $stored = $storeObj->store( $extdb, serialize( $chunk ));
 334                                                 if ($stored === false) {
 335                                                         $this->error(  "Unable to store object" );
 336                                                         return false;
 337                                                 }
 338                                                 # Store External Storage URLs instead of Stub placeholders
 339                                                 foreach ($stubs as $stub) {
 340                                                         if ($stub===false)
 341                                                                 continue;
 342                                                         # $stored should provide base path to a BLOB
 343                                                         $url = $stored."/".$stub->getHash();
 344                                                         $dbw->update( 'text',
 345                                                                 array( /* SET */
 346                                                                         'old_text' => $url,
 347                                                                         'old_flags' => 'external,utf-8',
 348                                                                 ), array ( /* WHERE */
 349                                                                         'old_id' => $stub->getReferrer(),
 350                                                                 )
 351                                                         );
 352                                                 }
 353                                         } else {
 354                                                 # Store the main object locally
 355                                                 $dbw->update( 'text',
 356                                                         array( /* SET */
 357                                                                 'old_text' => serialize( $chunk ),
 358                                                                 'old_flags' => 'object,utf-8',
 359                                                         ), array( /* WHERE */
 360                                                                 'old_id' => $primaryOldid
 361                                                         )
 362                                                 );
 363
 364                                                 # Store the stub objects
 365                                                 for ( $j = 1; $j < $thisChunkSize; $j++ ) {
 366                                                         # Skip if not compressing and don't overwrite the first revision
 367                                                         if ( $stubs[$j] !== false && $revs[$i + $j]->rev_text_id != $primaryOldid ) {
 368                                                                 $dbw->update( 'text',
 369                                                                         array( /* SET */
 370                                                                                 'old_text' => serialize($stubs[$j]),
 371                                                                                 'old_flags' => 'object,utf-8',
 372                                                                         ), array( /* WHERE */
 373                                                                                 'old_id' => $revs[$i + $j]->rev_text_id
 374                                                                         )
 375                                                                 );
 376                                                         }
 377                                                 }
 378                                         }
 379                                 }
 380                                 # Done, next
 381                                 $this->output( "/" );
 382                                 $dbw->commit();
 383                                 $i += $thisChunkSize;
 384                                 wfWaitForSlaves();
 385                         }
 386                         $this->output( "\n" );
 387                 }
 388                 return true;
 389         }
 390
 391 }
 392
 393 $maintClass = 'CompressOld';
 394 require_once( RUN_MAINTENANCE_IF_MAIN );