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 <startid>        the id to start from (referring to the text table for
  21  *                      type gzip, and to the page table for type concat)
  22  *  -n <endid>          the page_id to stop at (only when using concat compression type)
  23  *  --extdb <cluster>   store specified revisions in an external cluster (untested)
  24  *
  25  * This program is free software; you can redistribute it and/or modify
  26  * it under the terms of the GNU General Public License as published by
  27  * the Free Software Foundation; either version 2 of the License, or
  28  * (at your option) any later version.
  29  *
  30  * This program is distributed in the hope that it will be useful,
  31  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  32  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  33  * GNU General Public License for more details.
  34  *
  35  * You should have received a copy of the GNU General Public License along
  36  * with this program; if not, write to the Free Software Foundation, Inc.,
  37  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  38  * http://www.gnu.org/copyleft/gpl.html
  39  *
  40  * @file
  41  * @ingroup Maintenance ExternalStorage
  42  */
  43
  44 require_once( __DIR__ . '/../Maintenance.php' );
  45
  46 class CompressOld extends Maintenance {
  47         /**
  48          * @todo document
  49          */
  50         const LS_INDIVIDUAL = 0;
  51         const LS_CHUNKED    = 1;
  52
  53         public function __construct() {
  54                 parent::__construct();
  55                 $this->mDescription = 'Compress the text of a wiki';
  56                 $this->addOption( 'type', 'Set compression type to either: gzip|concat', false, true, 't' );
  57                 $this->addOption( 'chunksize', 'Maximum number of revisions in a concat chunk', false, true, 'c' );
  58                 $this->addOption( 'begin-date', 'Earliest date to check for uncompressed revisions', false, true, 'b' );
  59                 $this->addOption( 'end-date', 'Latest revision date to compress', false, true, 'e' );
  60                 $this->addOption( 'startid', 'The id to start from (gzip -> text table, concat -> page table)', false, true, 's' );
  61                 $this->addOption( 'extdb', 'Store specified revisions in an external cluster (untested)', false, true );
  62                 $this->addOption( 'endid', 'The page_id to stop at (only when using concat compression type)', false, true, 'n' );
  63         }
  64
  65         public function execute() {
  66                 global $wgDBname;
  67                 if ( !function_exists( "gzdeflate" ) ) {
  68                         $this->error( "You must enable zlib support in PHP to compress old revisions!\n" .
  69                                 "Please see http://www.php.net/manual/en/ref.zlib.php\n", true );
  70                 }
  71
  72                 $type = $this->getOption( 'type', 'concat' );
  73                 $chunkSize = $this->getOption( 'chunksize', 20 );
  74                 $startId = $this->getOption( 'startid', 0 );
  75                 $beginDate = $this->getOption( 'begin-date', '' );
  76                 $endDate = $this->getOption( 'end-date', '' );
  77                 $extDB = $this->getOption( 'extdb', '' );
  78                 $endId = $this->getOption( 'endid', false );
  79
  80                 if ( $type != 'concat' && $type != 'gzip' ) {
  81                         $this->error( "Type \"{$type}\" not supported" );
  82                 }
  83
  84                 if ( $extDB != '' ) {
  85                         $this->output( "Compressing database {$wgDBname} to external cluster {$extDB}\n"
  86                                 . str_repeat( '-', 76 ) . "\n\n" );
  87                 } else {
  88                         $this->output( "Compressing database {$wgDBname}\n"
  89                                 . str_repeat( '-', 76 ) . "\n\n" );
  90                 }
  91
  92                 $success = true;
  93                 if ( $type == 'concat' ) {
  94                         $success = $this->compressWithConcat( $startId, $chunkSize, $beginDate,
  95                                 $endDate, $extDB, $endId );
  96                 } else {
  97                         $this->compressOldPages( $startId, $extDB );
  98                 }
  99
 100                 if ( $success ) {
 101                         $this->output( "Done.\n" );
 102                 }
 103         }
 104
 105         /** @todo document */
 106         private function compressOldPages( $start = 0, $extdb = '' ) {
 107                 $chunksize = 50;
 108                 $this->output( "Starting from old_id $start...\n" );
 109                 $dbw = wfGetDB( DB_MASTER );
 110                 do {
 111                         $res = $dbw->select( 'text', array( 'old_id','old_flags','old_text' ),
 112                                 "old_id>=$start", __METHOD__, array( 'ORDER BY' => 'old_id', 'LIMIT' => $chunksize, 'FOR UPDATE' ) );
 113                         if( $dbw->numRows( $res ) == 0 ) {
 114                                 break;
 115                         }
 116                         $last = $start;
 117                         foreach ( $res as $row ) {
 118                                 # print "  {$row->old_id} - {$row->old_namespace}:{$row->old_title}\n";
 119                                 $this->compressPage( $row, $extdb );
 120                                 $last = $row->old_id;
 121                         }
 122                         $start = $last + 1; # Deletion may leave long empty stretches
 123                         $this->output( "$start...\n" );
 124                 } while( true );
 125         }
 126
 127         /**
 128          * @todo document
 129          * @param $row
 130          * @param $extdb
 131          * @return bool
 132          */
 133         private function compressPage( $row, $extdb ) {
 134                 if ( false !== strpos( $row->old_flags, 'gzip' ) || false !== strpos( $row->old_flags, 'object' ) ) {
 135                         #print "Already compressed row {$row->old_id}\n";
 136                         return false;
 137                 }
 138                 $dbw = wfGetDB( DB_MASTER );
 139                 $flags = $row->old_flags ? "{$row->old_flags},gzip" : "gzip";
 140                 $compress = gzdeflate( $row->old_text );
 141
 142                 # Store in external storage if required
 143                 if ( $extdb !== '' ) {
 144                         $storeObj = new ExternalStoreDB;
 145                         $compress = $storeObj->store( $extdb, $compress );
 146                         if ( $compress === false ) {
 147                                 $this->error( "Unable to store object" );
 148                                 return false;
 149                         }
 150                 }
 151
 152                 # Update text row
 153                 $dbw->update( 'text',
 154                         array( /* SET */
 155                                 'old_flags' => $flags,
 156                                 'old_text' => $compress
 157                         ), array( /* WHERE */
 158                                 'old_id' => $row->old_id
 159                         ), __METHOD__,
 160                         array( 'LIMIT' => 1 )
 161                 );
 162                 return true;
 163         }
 164
 165         /**
 166          * @param $startId
 167          * @param $maxChunkSize
 168          * @param $beginDate
 169          * @param $endDate
 170          * @param $extdb string
 171          * @param $maxPageId bool|int
 172          * @return bool
 173          */
 174         private function compressWithConcat( $startId, $maxChunkSize, $beginDate,
 175                 $endDate, $extdb = "", $maxPageId = false )
 176         {
 177                 $loadStyle = self::LS_CHUNKED;
 178
 179                 $dbr = wfGetDB( DB_SLAVE );
 180                 $dbw = wfGetDB( DB_MASTER );
 181
 182                 # Set up external storage
 183                 if ( $extdb != '' ) {
 184                         $storeObj = new ExternalStoreDB;
 185                 }
 186
 187                 # Get all articles by page_id
 188                 if ( !$maxPageId ) {
 189                         $maxPageId = $dbr->selectField( 'page', 'max(page_id)', '', __METHOD__ );
 190                 }
 191                 $this->output( "Starting from $startId of $maxPageId\n" );
 192                 $pageConds = array();
 193
 194                 /*
 195                 if ( $exclude_ns0 ) {
 196                         print "Excluding main namespace\n";
 197                         $pageConds[] = 'page_namespace<>0';
 198                 }
 199                 if ( $queryExtra ) {
 200                                         $pageConds[] = $queryExtra;
 201                 }
 202                  */
 203
 204                 # For each article, get a list of revisions which fit the criteria
 205
 206                 # No recompression, use a condition on old_flags
 207                 # Don't compress object type entities, because that might produce data loss when
 208                 # overwriting bulk storage concat rows. Don't compress external references, because
 209                 # the script doesn't yet delete rows from external storage.
 210                 $conds = array(
 211                         'old_flags NOT ' . $dbr->buildLike( $dbr->anyString(), 'object', $dbr->anyString() ) . ' AND old_flags NOT '
 212                                 . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() ) );
 213
 214                 if ( $beginDate ) {
 215                         if ( !preg_match( '/^\d{14}$/', $beginDate ) ) {
 216                                 $this->error( "Invalid begin date \"$beginDate\"\n" );
 217                                 return false;
 218                         }
 219                         $conds[] = "rev_timestamp>'" . $beginDate . "'";
 220                 }
 221                 if ( $endDate )  {
 222                         if ( !preg_match( '/^\d{14}$/', $endDate ) ) {
 223                                 $this->error( "Invalid end date \"$endDate\"\n" );
 224                                 return false;
 225                         }
 226                         $conds[] = "rev_timestamp<'" . $endDate . "'";
 227                 }
 228                 if ( $loadStyle == self::LS_CHUNKED ) {
 229                         $tables = array( 'revision', 'text' );
 230                         $fields = array( 'rev_id', 'rev_text_id', 'old_flags', 'old_text' );
 231                         $conds[] = 'rev_text_id=old_id';
 232                         $revLoadOptions = 'FOR UPDATE';
 233                 } else {
 234                         $tables = array( 'revision' );
 235                         $fields = array( 'rev_id', 'rev_text_id' );
 236                         $revLoadOptions = array();
 237                 }
 238
 239                 # Don't work with current revisions
 240                 # Don't lock the page table for update either -- TS 2006-04-04
 241                 #$tables[] = 'page';
 242                 #$conds[] = 'page_id=rev_page AND rev_id != page_latest';
 243
 244                 for ( $pageId = $startId; $pageId <= $maxPageId; $pageId++ ) {
 245                         wfWaitForSlaves();
 246
 247                         # Wake up
 248                         $dbr->ping();
 249
 250                         # Get the page row
 251                         $pageRes = $dbr->select( 'page',
 252                                 array('page_id', 'page_namespace', 'page_title','page_latest'),
 253                                 $pageConds + array('page_id' => $pageId), __METHOD__ );
 254                         if ( $dbr->numRows( $pageRes ) == 0 ) {
 255                                 continue;
 256                         }
 257                         $pageRow = $dbr->fetchObject( $pageRes );
 258
 259                         # Display progress
 260                         $titleObj = Title::makeTitle( $pageRow->page_namespace, $pageRow->page_title );
 261                         $this->output( "$pageId\t" . $titleObj->getPrefixedDBkey() . " " );
 262
 263                         # Load revisions
 264                         $revRes = $dbw->select( $tables, $fields,
 265                                 array_merge( array(
 266                                         'rev_page' => $pageRow->page_id,
 267                                         # Don't operate on the current revision
 268                                         # Use < instead of <> in case the current revision has changed
 269                                         # since the page select, which wasn't locking
 270                                         'rev_id < ' . $pageRow->page_latest
 271                                 ), $conds ),
 272                                 __METHOD__,
 273                                 $revLoadOptions
 274                         );
 275                         $revs = array();
 276                         foreach ( $revRes as $revRow ) {
 277                                 $revs[] = $revRow;
 278                         }
 279
 280                         if ( count( $revs ) < 2) {
 281                                 # No revisions matching, no further processing
 282                                 $this->output( "\n" );
 283                                 continue;
 284                         }
 285
 286                         # For each chunk
 287                         $i = 0;
 288                         while ( $i < count( $revs ) ) {
 289                                 if ( $i < count( $revs ) - $maxChunkSize ) {
 290                                         $thisChunkSize = $maxChunkSize;
 291                                 } else {
 292                                         $thisChunkSize = count( $revs ) - $i;
 293                                 }
 294
 295                                 $chunk = new ConcatenatedGzipHistoryBlob();
 296                                 $stubs = array();
 297                                 $dbw->begin( __METHOD__ );
 298                                 $usedChunk = false;
 299                                 $primaryOldid = $revs[$i]->rev_text_id;
 300
 301                                 # Get the text of each revision and add it to the object
 302                                 for ( $j = 0; $j < $thisChunkSize && $chunk->isHappy(); $j++ ) {
 303                                         $oldid = $revs[$i + $j]->rev_text_id;
 304
 305                                         # Get text
 306                                         if ( $loadStyle == self::LS_INDIVIDUAL ) {
 307                                                 $textRow = $dbw->selectRow( 'text',
 308                                                         array( 'old_flags', 'old_text' ),
 309                                                         array( 'old_id' => $oldid ),
 310                                                         __METHOD__,
 311                                                         'FOR UPDATE'
 312                                                 );
 313                                                 $text = Revision::getRevisionText( $textRow );
 314                                         } else {
 315                                                 $text = Revision::getRevisionText( $revs[$i + $j] );
 316                                         }
 317
 318                                         if ( $text === false ) {
 319                                                 $this->error( "\nError, unable to get text in old_id $oldid" );
 320                                                 #$dbw->delete( 'old', array( 'old_id' => $oldid ) );
 321                                         }
 322
 323                                         if ( $extdb == "" && $j == 0 ) {
 324                                                 $chunk->setText( $text );
 325                                                 $this->output( '.' );
 326                                         } else {
 327                                                 # Don't make a stub if it's going to be longer than the article
 328                                                 # Stubs are typically about 100 bytes
 329                                                 if ( strlen( $text ) < 120 ) {
 330                                                         $stub = false;
 331                                                         $this->output( 'x' );
 332                                                 } else {
 333                                                         $stub = new HistoryBlobStub( $chunk->addItem( $text ) );
 334                                                         $stub->setLocation( $primaryOldid );
 335                                                         $stub->setReferrer( $oldid );
 336                                                         $this->output( '.' );
 337                                                         $usedChunk = true;
 338                                                 }
 339                                                 $stubs[$j] = $stub;
 340                                         }
 341                                 }
 342                                 $thisChunkSize = $j;
 343
 344                                 # If we couldn't actually use any stubs because the pages were too small, do nothing
 345                                 if ( $usedChunk ) {
 346                                         if ( $extdb != "" ) {
 347                                                 # Move blob objects to External Storage
 348                                                 $stored = $storeObj->store( $extdb, serialize( $chunk ));
 349                                                 if ($stored === false) {
 350                                                         $this->error(  "Unable to store object" );
 351                                                         return false;
 352                                                 }
 353                                                 # Store External Storage URLs instead of Stub placeholders
 354                                                 foreach ($stubs as $stub) {
 355                                                         if ($stub===false)
 356                                                                 continue;
 357                                                         # $stored should provide base path to a BLOB
 358                                                         $url = $stored."/".$stub->getHash();
 359                                                         $dbw->update( 'text',
 360                                                                 array( /* SET */
 361                                                                         'old_text' => $url,
 362                                                                         'old_flags' => 'external,utf-8',
 363                                                                 ), array ( /* WHERE */
 364                                                                         'old_id' => $stub->getReferrer(),
 365                                                                 )
 366                                                         );
 367                                                 }
 368                                         } else {
 369                                                 # Store the main object locally
 370                                                 $dbw->update( 'text',
 371                                                         array( /* SET */
 372                                                                 'old_text' => serialize( $chunk ),
 373                                                                 'old_flags' => 'object,utf-8',
 374                                                         ), array( /* WHERE */
 375                                                                 'old_id' => $primaryOldid
 376                                                         )
 377                                                 );
 378
 379                                                 # Store the stub objects
 380                                                 for ( $j = 1; $j < $thisChunkSize; $j++ ) {
 381                                                         # Skip if not compressing and don't overwrite the first revision
 382                                                         if ( $stubs[$j] !== false && $revs[$i + $j]->rev_text_id != $primaryOldid ) {
 383                                                                 $dbw->update( 'text',
 384                                                                         array( /* SET */
 385                                                                                 'old_text' => serialize($stubs[$j]),
 386                                                                                 'old_flags' => 'object,utf-8',
 387                                                                         ), array( /* WHERE */
 388                                                                                 'old_id' => $revs[$i + $j]->rev_text_id
 389                                                                         )
 390                                                                 );
 391                                                         }
 392                                                 }
 393                                         }
 394                                 }
 395                                 # Done, next
 396                                 $this->output( "/" );
 397                                 $dbw->commit( __METHOD__ );
 398                                 $i += $thisChunkSize;
 399                                 wfWaitForSlaves();
 400                         }
 401                         $this->output( "\n" );
 402                 }
 403                 return true;
 404         }
 405
 406 }
 407
 408 $maintClass = 'CompressOld';
 409 require_once( RUN_MAINTENANCE_IF_MAIN );