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 /**
  47  * Maintenance script that compress the text of a wiki.
  48  *
  49  * @ingroup Maintenance ExternalStorage
  50  */
  51 class CompressOld extends Maintenance {
  52         /**
  53          * @todo document
  54          */
  55         const LS_INDIVIDUAL = 0;
  56         const LS_CHUNKED = 1;
  57
  58         public function __construct() {
  59                 parent::__construct();
  60                 $this->mDescription = 'Compress the text of a wiki';
  61                 $this->addOption( 'type', 'Set compression type to either: gzip|concat', false, true, 't' );
  62                 $this->addOption( 'chunksize', 'Maximum number of revisions in a concat chunk', false, true, 'c' );
  63                 $this->addOption( 'begin-date', 'Earliest date to check for uncompressed revisions', false, true, 'b' );
  64                 $this->addOption( 'end-date', 'Latest revision date to compress', false, true, 'e' );
  65                 $this->addOption( 'startid', 'The id to start from (gzip -> text table, concat -> page table)', false, true, 's' );
  66                 $this->addOption( 'extdb', 'Store specified revisions in an external cluster (untested)', false, true );
  67                 $this->addOption( 'endid', 'The page_id to stop at (only when using concat compression type)', false, true, 'n' );
  68         }
  69
  70         public function execute() {
  71                 global $wgDBname;
  72                 if ( !function_exists( "gzdeflate" ) ) {
  73                         $this->error( "You must enable zlib support in PHP to compress old revisions!\n" .
  74                                 "Please see http://www.php.net/manual/en/ref.zlib.php\n", true );
  75                 }
  76
  77                 $type = $this->getOption( 'type', 'concat' );
  78                 $chunkSize = $this->getOption( 'chunksize', 20 );
  79                 $startId = $this->getOption( 'startid', 0 );
  80                 $beginDate = $this->getOption( 'begin-date', '' );
  81                 $endDate = $this->getOption( 'end-date', '' );
  82                 $extDB = $this->getOption( 'extdb', '' );
  83                 $endId = $this->getOption( 'endid', false );
  84
  85                 if ( $type != 'concat' && $type != 'gzip' ) {
  86                         $this->error( "Type \"{$type}\" not supported" );
  87                 }
  88
  89                 if ( $extDB != '' ) {
  90                         $this->output( "Compressing database {$wgDBname} to external cluster {$extDB}\n"
  91                                 . str_repeat( '-', 76 ) . "\n\n" );
  92                 } else {
  93                         $this->output( "Compressing database {$wgDBname}\n"
  94                                 . str_repeat( '-', 76 ) . "\n\n" );
  95                 }
  96
  97                 $success = true;
  98                 if ( $type == 'concat' ) {
  99                         $success = $this->compressWithConcat( $startId, $chunkSize, $beginDate,
 100                                 $endDate, $extDB, $endId );
 101                 } else {
 102                         $this->compressOldPages( $startId, $extDB );
 103                 }
 104
 105                 if ( $success ) {
 106                         $this->output( "Done.\n" );
 107                 }
 108         }
 109
 110         /** @todo document */
 111         private function compressOldPages( $start = 0, $extdb = '' ) {
 112                 $chunksize = 50;
 113                 $this->output( "Starting from old_id $start...\n" );
 114                 $dbw = wfGetDB( DB_MASTER );
 115                 do {
 116                         $res = $dbw->select( 'text', array( 'old_id', 'old_flags', 'old_text' ),
 117                                 "old_id>=$start", __METHOD__, array( 'ORDER BY' => 'old_id', 'LIMIT' => $chunksize, 'FOR UPDATE' ) );
 118                         if ( $res->numRows() == 0 ) {
 119                                 break;
 120                         }
 121                         $last = $start;
 122                         foreach ( $res as $row ) {
 123                                 # print "  {$row->old_id} - {$row->old_namespace}:{$row->old_title}\n";
 124                                 $this->compressPage( $row, $extdb );
 125                                 $last = $row->old_id;
 126                         }
 127                         $start = $last + 1; # Deletion may leave long empty stretches
 128                         $this->output( "$start...\n" );
 129                 } while ( true );
 130         }
 131
 132         /**
 133          * @todo document
 134          * @param $row
 135          * @param $extdb
 136          * @return bool
 137          */
 138         private function compressPage( $row, $extdb ) {
 139                 if ( false !== strpos( $row->old_flags, 'gzip' ) || false !== strpos( $row->old_flags, 'object' ) ) {
 140                         #print "Already compressed row {$row->old_id}\n";
 141                         return false;
 142                 }
 143                 $dbw = wfGetDB( DB_MASTER );
 144                 $flags = $row->old_flags ? "{$row->old_flags},gzip" : "gzip";
 145                 $compress = gzdeflate( $row->old_text );
 146
 147                 # Store in external storage if required
 148                 if ( $extdb !== '' ) {
 149                         $storeObj = new ExternalStoreDB;
 150                         $compress = $storeObj->store( $extdb, $compress );
 151                         if ( $compress === false ) {
 152                                 $this->error( "Unable to store object" );
 153                                 return false;
 154                         }
 155                 }
 156
 157                 # Update text row
 158                 $dbw->update( 'text',
 159                         array( /* SET */
 160                                 'old_flags' => $flags,
 161                                 'old_text' => $compress
 162                         ), array( /* WHERE */
 163                                 'old_id' => $row->old_id
 164                         ), __METHOD__,
 165                         array( 'LIMIT' => 1 )
 166                 );
 167                 return true;
 168         }
 169
 170         /**
 171          * @param $startId
 172          * @param $maxChunkSize
 173          * @param $beginDate
 174          * @param $endDate
 175          * @param $extdb string
 176          * @param $maxPageId bool|int
 177          * @return bool
 178          */
 179         private function compressWithConcat( $startId, $maxChunkSize, $beginDate,
 180                 $endDate, $extdb = "", $maxPageId = false )
 181         {
 182                 $loadStyle = self::LS_CHUNKED;
 183
 184                 $dbr = wfGetDB( DB_SLAVE );
 185                 $dbw = wfGetDB( DB_MASTER );
 186
 187                 # Set up external storage
 188                 if ( $extdb != '' ) {
 189                         $storeObj = new ExternalStoreDB;
 190                 }
 191
 192                 # Get all articles by page_id
 193                 if ( !$maxPageId ) {
 194                         $maxPageId = $dbr->selectField( 'page', 'max(page_id)', '', __METHOD__ );
 195                 }
 196                 $this->output( "Starting from $startId of $maxPageId\n" );
 197                 $pageConds = array();
 198
 199                 /*
 200                 if ( $exclude_ns0 ) {
 201                         print "Excluding main namespace\n";
 202                         $pageConds[] = 'page_namespace<>0';
 203                 }
 204                 if ( $queryExtra ) {
 205                                         $pageConds[] = $queryExtra;
 206                 }
 207                  */
 208
 209                 # For each article, get a list of revisions which fit the criteria
 210
 211                 # No recompression, use a condition on old_flags
 212                 # Don't compress object type entities, because that might produce data loss when
 213                 # overwriting bulk storage concat rows. Don't compress external references, because
 214                 # the script doesn't yet delete rows from external storage.
 215                 $conds = array(
 216                         'old_flags NOT ' . $dbr->buildLike( $dbr->anyString(), 'object', $dbr->anyString() ) . ' AND old_flags NOT '
 217                                 . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() ) );
 218
 219                 if ( $beginDate ) {
 220                         if ( !preg_match( '/^\d{14}$/', $beginDate ) ) {
 221                                 $this->error( "Invalid begin date \"$beginDate\"\n" );
 222                                 return false;
 223                         }
 224                         $conds[] = "rev_timestamp>'" . $beginDate . "'";
 225                 }
 226                 if ( $endDate ) {
 227                         if ( !preg_match( '/^\d{14}$/', $endDate ) ) {
 228                                 $this->error( "Invalid end date \"$endDate\"\n" );
 229                                 return false;
 230                         }
 231                         $conds[] = "rev_timestamp<'" . $endDate . "'";
 232                 }
 233                 if ( $loadStyle == self::LS_CHUNKED ) {
 234                         $tables = array( 'revision', 'text' );
 235                         $fields = array( 'rev_id', 'rev_text_id', 'old_flags', 'old_text' );
 236                         $conds[] = 'rev_text_id=old_id';
 237                         $revLoadOptions = 'FOR UPDATE';
 238                 } else {
 239                         $tables = array( 'revision' );
 240                         $fields = array( 'rev_id', 'rev_text_id' );
 241                         $revLoadOptions = array();
 242                 }
 243
 244                 # Don't work with current revisions
 245                 # Don't lock the page table for update either -- TS 2006-04-04
 246                 #$tables[] = 'page';
 247                 #$conds[] = 'page_id=rev_page AND rev_id != page_latest';
 248
 249                 for ( $pageId = $startId; $pageId <= $maxPageId; $pageId++ ) {
 250                         wfWaitForSlaves();
 251
 252                         # Wake up
 253                         $dbr->ping();
 254
 255                         # Get the page row
 256                         $pageRes = $dbr->select( 'page',
 257                                 array( 'page_id', 'page_namespace', 'page_title', 'page_latest' ),
 258                                 $pageConds + array( 'page_id' => $pageId ), __METHOD__ );
 259                         if ( $pageRes->numRows() == 0 ) {
 260                                 continue;
 261                         }
 262                         $pageRow = $dbr->fetchObject( $pageRes );
 263
 264                         # Display progress
 265                         $titleObj = Title::makeTitle( $pageRow->page_namespace, $pageRow->page_title );
 266                         $this->output( "$pageId\t" . $titleObj->getPrefixedDBkey() . " " );
 267
 268                         # Load revisions
 269                         $revRes = $dbw->select( $tables, $fields,
 270                                 array_merge( array(
 271                                         'rev_page' => $pageRow->page_id,
 272                                         # Don't operate on the current revision
 273                                         # Use < instead of <> in case the current revision has changed
 274                                         # since the page select, which wasn't locking
 275                                         'rev_id < ' . $pageRow->page_latest
 276                                 ), $conds ),
 277                                 __METHOD__,
 278                                 $revLoadOptions
 279                         );
 280                         $revs = array();
 281                         foreach ( $revRes as $revRow ) {
 282                                 $revs[] = $revRow;
 283                         }
 284
 285                         if ( count( $revs ) < 2 ) {
 286                                 # No revisions matching, no further processing
 287                                 $this->output( "\n" );
 288                                 continue;
 289                         }
 290
 291                         # For each chunk
 292                         $i = 0;
 293                         while ( $i < count( $revs ) ) {
 294                                 if ( $i < count( $revs ) - $maxChunkSize ) {
 295                                         $thisChunkSize = $maxChunkSize;
 296                                 } else {
 297                                         $thisChunkSize = count( $revs ) - $i;
 298                                 }
 299
 300                                 $chunk = new ConcatenatedGzipHistoryBlob();
 301                                 $stubs = array();
 302                                 $dbw->begin( __METHOD__ );
 303                                 $usedChunk = false;
 304                                 $primaryOldid = $revs[$i]->rev_text_id;
 305
 306                                 # Get the text of each revision and add it to the object
 307                                 for ( $j = 0; $j < $thisChunkSize && $chunk->isHappy(); $j++ ) {
 308                                         $oldid = $revs[$i + $j]->rev_text_id;
 309
 310                                         # Get text
 311                                         if ( $loadStyle == self::LS_INDIVIDUAL ) {
 312                                                 $textRow = $dbw->selectRow( 'text',
 313                                                         array( 'old_flags', 'old_text' ),
 314                                                         array( 'old_id' => $oldid ),
 315                                                         __METHOD__,
 316                                                         'FOR UPDATE'
 317                                                 );
 318                                                 $text = Revision::getRevisionText( $textRow );
 319                                         } else {
 320                                                 $text = Revision::getRevisionText( $revs[$i + $j] );
 321                                         }
 322
 323                                         if ( $text === false ) {
 324                                                 $this->error( "\nError, unable to get text in old_id $oldid" );
 325                                                 #$dbw->delete( 'old', array( 'old_id' => $oldid ) );
 326                                         }
 327
 328                                         if ( $extdb == "" && $j == 0 ) {
 329                                                 $chunk->setText( $text );
 330                                                 $this->output( '.' );
 331                                         } else {
 332                                                 # Don't make a stub if it's going to be longer than the article
 333                                                 # Stubs are typically about 100 bytes
 334                                                 if ( strlen( $text ) < 120 ) {
 335                                                         $stub = false;
 336                                                         $this->output( 'x' );
 337                                                 } else {
 338                                                         $stub = new HistoryBlobStub( $chunk->addItem( $text ) );
 339                                                         $stub->setLocation( $primaryOldid );
 340                                                         $stub->setReferrer( $oldid );
 341                                                         $this->output( '.' );
 342                                                         $usedChunk = true;
 343                                                 }
 344                                                 $stubs[$j] = $stub;
 345                                         }
 346                                 }
 347                                 $thisChunkSize = $j;
 348
 349                                 # If we couldn't actually use any stubs because the pages were too small, do nothing
 350                                 if ( $usedChunk ) {
 351                                         if ( $extdb != "" ) {
 352                                                 # Move blob objects to External Storage
 353                                                 $stored = $storeObj->store( $extdb, serialize( $chunk ));
 354                                                 if ( $stored === false ) {
 355                                                         $this->error( "Unable to store object" );
 356                                                         return false;
 357                                                 }
 358                                                 # Store External Storage URLs instead of Stub placeholders
 359                                                 foreach ( $stubs as $stub ) {
 360                                                         if ( $stub === false ) {
 361                                                                 continue;
 362                                                         }
 363                                                         # $stored should provide base path to a BLOB
 364                                                         $url = $stored . "/" . $stub->getHash();
 365                                                         $dbw->update( 'text',
 366                                                                 array( /* SET */
 367                                                                         'old_text' => $url,
 368                                                                         'old_flags' => 'external,utf-8',
 369                                                                 ), array( /* WHERE */
 370                                                                         'old_id' => $stub->getReferrer(),
 371                                                                 )
 372                                                         );
 373                                                 }
 374                                         } else {
 375                                                 # Store the main object locally
 376                                                 $dbw->update( 'text',
 377                                                         array( /* SET */
 378                                                                 'old_text' => serialize( $chunk ),
 379                                                                 'old_flags' => 'object,utf-8',
 380                                                         ), array( /* WHERE */
 381                                                                 'old_id' => $primaryOldid
 382                                                         )
 383                                                 );
 384
 385                                                 # Store the stub objects
 386                                                 for ( $j = 1; $j < $thisChunkSize; $j++ ) {
 387                                                         # Skip if not compressing and don't overwrite the first revision
 388                                                         if ( $stubs[$j] !== false && $revs[$i + $j]->rev_text_id != $primaryOldid ) {
 389                                                                 $dbw->update( 'text',
 390                                                                         array( /* SET */
 391                                                                                 'old_text' => serialize( $stubs[$j] ),
 392                                                                                 'old_flags' => 'object,utf-8',
 393                                                                         ), array( /* WHERE */
 394                                                                                 'old_id' => $revs[$i + $j]->rev_text_id
 395                                                                         )
 396                                                                 );
 397                                                         }
 398                                                 }
 399                                         }
 400                                 }
 401                                 # Done, next
 402                                 $this->output( "/" );
 403                                 $dbw->commit( __METHOD__ );
 404                                 $i += $thisChunkSize;
 405                                 wfWaitForSlaves();
 406                         }
 407                         $this->output( "\n" );
 408                 }
 409                 return true;
 410         }
 411
 412 }
 413
 414 $maintClass = 'CompressOld';
 415 require_once RUN_MAINTENANCE_IF_MAIN;