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(
  63                         'chunksize',
  64                         'Maximum number of revisions in a concat chunk',
  65                         false,
  66                         true,
  67                         'c'
  68                 );
  69                 $this->addOption(
  70                         'begin-date',
  71                         'Earliest date to check for uncompressed revisions',
  72                         false,
  73                         true,
  74                         'b'
  75                 );
  76                 $this->addOption( 'end-date', 'Latest revision date to compress', false, true, 'e' );
  77                 $this->addOption(
  78                         'startid',
  79                         'The id to start from (gzip -> text table, concat -> page table)',
  80                         false,
  81                         true,
  82                         's'
  83                 );
  84                 $this->addOption(
  85                         'extdb',
  86                         'Store specified revisions in an external cluster (untested)',
  87                         false,
  88                         true
  89                 );
  90                 $this->addOption(
  91                         'endid',
  92                         'The page_id to stop at (only when using concat compression type)',
  93                         false,
  94                         true,
  95                         'n'
  96                 );
  97         }
  98
  99         public function execute() {
 100                 global $wgDBname;
 101                 if ( !function_exists( "gzdeflate" ) ) {
 102                         $this->error( "You must enable zlib support in PHP to compress old revisions!\n" .
 103                                 "Please see http://www.php.net/manual/en/ref.zlib.php\n", true );
 104                 }
 105
 106                 $type = $this->getOption( 'type', 'concat' );
 107                 $chunkSize = $this->getOption( 'chunksize', 20 );
 108                 $startId = $this->getOption( 'startid', 0 );
 109                 $beginDate = $this->getOption( 'begin-date', '' );
 110                 $endDate = $this->getOption( 'end-date', '' );
 111                 $extDB = $this->getOption( 'extdb', '' );
 112                 $endId = $this->getOption( 'endid', false );
 113
 114                 if ( $type != 'concat' && $type != 'gzip' ) {
 115                         $this->error( "Type \"{$type}\" not supported" );
 116                 }
 117
 118                 if ( $extDB != '' ) {
 119                         $this->output( "Compressing database {$wgDBname} to external cluster {$extDB}\n"
 120                                 . str_repeat( '-', 76 ) . "\n\n" );
 121                 } else {
 122                         $this->output( "Compressing database {$wgDBname}\n"
 123                                 . str_repeat( '-', 76 ) . "\n\n" );
 124                 }
 125
 126                 $success = true;
 127                 if ( $type == 'concat' ) {
 128                         $success = $this->compressWithConcat( $startId, $chunkSize, $beginDate,
 129                                 $endDate, $extDB, $endId );
 130                 } else {
 131                         $this->compressOldPages( $startId, $extDB );
 132                 }
 133
 134                 if ( $success ) {
 135                         $this->output( "Done.\n" );
 136                 }
 137         }
 138
 139         /**
 140          * @todo document
 141          * @param int $start
 142          * @param string $extdb
 143          */
 144         private function compressOldPages( $start = 0, $extdb = '' ) {
 145                 $chunksize = 50;
 146                 $this->output( "Starting from old_id $start...\n" );
 147                 $dbw = wfGetDB( DB_MASTER );
 148                 do {
 149                         $res = $dbw->select(
 150                                 'text',
 151                                 array( 'old_id', 'old_flags', 'old_text' ),
 152                                 "old_id>=$start",
 153                                 __METHOD__,
 154                                 array( 'ORDER BY' => 'old_id', 'LIMIT' => $chunksize, 'FOR UPDATE' )
 155                         );
 156
 157                         if ( $res->numRows() == 0 ) {
 158                                 break;
 159                         }
 160
 161                         $last = $start;
 162
 163                         foreach ( $res as $row ) {
 164                                 # print "  {$row->old_id} - {$row->old_namespace}:{$row->old_title}\n";
 165                                 $this->compressPage( $row, $extdb );
 166                                 $last = $row->old_id;
 167                         }
 168
 169                         $start = $last + 1; # Deletion may leave long empty stretches
 170                         $this->output( "$start...\n" );
 171                 } while ( true );
 172         }
 173
 174         /**
 175          * @todo document
 176          * @param stdClass $row
 177          * @param string $extdb
 178          * @return bool
 179          */
 180         private function compressPage( $row, $extdb ) {
 181                 if ( false !== strpos( $row->old_flags, 'gzip' )
 182                         || false !== strpos( $row->old_flags, 'object' )
 183                 ) {
 184                         #print "Already compressed row {$row->old_id}\n";
 185                         return false;
 186                 }
 187                 $dbw = wfGetDB( DB_MASTER );
 188                 $flags = $row->old_flags ? "{$row->old_flags},gzip" : "gzip";
 189                 $compress = gzdeflate( $row->old_text );
 190
 191                 # Store in external storage if required
 192                 if ( $extdb !== '' ) {
 193                         $storeObj = new ExternalStoreDB;
 194                         $compress = $storeObj->store( $extdb, $compress );
 195                         if ( $compress === false ) {
 196                                 $this->error( "Unable to store object" );
 197
 198                                 return false;
 199                         }
 200                 }
 201
 202                 # Update text row
 203                 $dbw->update( 'text',
 204                         array( /* SET */
 205                                 'old_flags' => $flags,
 206                                 'old_text' => $compress
 207                         ), array( /* WHERE */
 208                                 'old_id' => $row->old_id
 209                         ), __METHOD__,
 210                         array( 'LIMIT' => 1 )
 211                 );
 212
 213                 return true;
 214         }
 215
 216         /**
 217          * @param int $startId
 218          * @param int $maxChunkSize
 219          * @param string $beginDate
 220          * @param string $endDate
 221          * @param string $extdb
 222          * @param bool|int $maxPageId
 223          * @return bool
 224          */
 225         private function compressWithConcat( $startId, $maxChunkSize, $beginDate,
 226                 $endDate, $extdb = "", $maxPageId = false
 227         ) {
 228                 $loadStyle = self::LS_CHUNKED;
 229
 230                 $dbr = wfGetDB( DB_SLAVE );
 231                 $dbw = wfGetDB( DB_MASTER );
 232
 233                 # Set up external storage
 234                 if ( $extdb != '' ) {
 235                         $storeObj = new ExternalStoreDB;
 236                 }
 237
 238                 # Get all articles by page_id
 239                 if ( !$maxPageId ) {
 240                         $maxPageId = $dbr->selectField( 'page', 'max(page_id)', '', __METHOD__ );
 241                 }
 242                 $this->output( "Starting from $startId of $maxPageId\n" );
 243                 $pageConds = array();
 244
 245                 /*
 246                 if ( $exclude_ns0 ) {
 247                         print "Excluding main namespace\n";
 248                         $pageConds[] = 'page_namespace<>0';
 249                 }
 250                 if ( $queryExtra ) {
 251                                         $pageConds[] = $queryExtra;
 252                 }
 253                  */
 254
 255                 # For each article, get a list of revisions which fit the criteria
 256
 257                 # No recompression, use a condition on old_flags
 258                 # Don't compress object type entities, because that might produce data loss when
 259                 # overwriting bulk storage concat rows. Don't compress external references, because
 260                 # the script doesn't yet delete rows from external storage.
 261                 $conds = array(
 262                         'old_flags NOT ' . $dbr->buildLike( $dbr->anyString(), 'object', $dbr->anyString() )
 263                         . ' AND old_flags NOT '
 264                         . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() )
 265                 );
 266
 267                 if ( $beginDate ) {
 268                         if ( !preg_match( '/^\d{14}$/', $beginDate ) ) {
 269                                 $this->error( "Invalid begin date \"$beginDate\"\n" );
 270
 271                                 return false;
 272                         }
 273                         $conds[] = "rev_timestamp>'" . $beginDate . "'";
 274                 }
 275                 if ( $endDate ) {
 276                         if ( !preg_match( '/^\d{14}$/', $endDate ) ) {
 277                                 $this->error( "Invalid end date \"$endDate\"\n" );
 278
 279                                 return false;
 280                         }
 281                         $conds[] = "rev_timestamp<'" . $endDate . "'";
 282                 }
 283                 if ( $loadStyle == self::LS_CHUNKED ) {
 284                         $tables = array( 'revision', 'text' );
 285                         $fields = array( 'rev_id', 'rev_text_id', 'old_flags', 'old_text' );
 286                         $conds[] = 'rev_text_id=old_id';
 287                         $revLoadOptions = 'FOR UPDATE';
 288                 } else {
 289                         $tables = array( 'revision' );
 290                         $fields = array( 'rev_id', 'rev_text_id' );
 291                         $revLoadOptions = array();
 292                 }
 293
 294                 # Don't work with current revisions
 295                 # Don't lock the page table for update either -- TS 2006-04-04
 296                 #$tables[] = 'page';
 297                 #$conds[] = 'page_id=rev_page AND rev_id != page_latest';
 298
 299                 for ( $pageId = $startId; $pageId <= $maxPageId; $pageId++ ) {
 300                         wfWaitForSlaves();
 301
 302                         # Wake up
 303                         $dbr->ping();
 304
 305                         # Get the page row
 306                         $pageRes = $dbr->select( 'page',
 307                                 array( 'page_id', 'page_namespace', 'page_title', 'page_latest' ),
 308                                 $pageConds + array( 'page_id' => $pageId ), __METHOD__ );
 309                         if ( $pageRes->numRows() == 0 ) {
 310                                 continue;
 311                         }
 312                         $pageRow = $dbr->fetchObject( $pageRes );
 313
 314                         # Display progress
 315                         $titleObj = Title::makeTitle( $pageRow->page_namespace, $pageRow->page_title );
 316                         $this->output( "$pageId\t" . $titleObj->getPrefixedDBkey() . " " );
 317
 318                         # Load revisions
 319                         $revRes = $dbw->select( $tables, $fields,
 320                                 array_merge( array(
 321                                         'rev_page' => $pageRow->page_id,
 322                                         # Don't operate on the current revision
 323                                         # Use < instead of <> in case the current revision has changed
 324                                         # since the page select, which wasn't locking
 325                                         'rev_id < ' . $pageRow->page_latest
 326                                 ), $conds ),
 327                                 __METHOD__,
 328                                 $revLoadOptions
 329                         );
 330                         $revs = array();
 331                         foreach ( $revRes as $revRow ) {
 332                                 $revs[] = $revRow;
 333                         }
 334
 335                         if ( count( $revs ) < 2 ) {
 336                                 # No revisions matching, no further processing
 337                                 $this->output( "\n" );
 338                                 continue;
 339                         }
 340
 341                         # For each chunk
 342                         $i = 0;
 343                         while ( $i < count( $revs ) ) {
 344                                 if ( $i < count( $revs ) - $maxChunkSize ) {
 345                                         $thisChunkSize = $maxChunkSize;
 346                                 } else {
 347                                         $thisChunkSize = count( $revs ) - $i;
 348                                 }
 349
 350                                 $chunk = new ConcatenatedGzipHistoryBlob();
 351                                 $stubs = array();
 352                                 $dbw->begin( __METHOD__ );
 353                                 $usedChunk = false;
 354                                 $primaryOldid = $revs[$i]->rev_text_id;
 355
 356                                 // @codingStandardsIgnoreStart Ignore avoid function calls in a FOR loop test part warning
 357                                 # Get the text of each revision and add it to the object
 358                                 for ( $j = 0; $j < $thisChunkSize && $chunk->isHappy(); $j++ ) {
 359                                         // @codingStandardsIgnoreEnd
 360                                         $oldid = $revs[$i + $j]->rev_text_id;
 361
 362                                         # Get text
 363                                         if ( $loadStyle == self::LS_INDIVIDUAL ) {
 364                                                 $textRow = $dbw->selectRow( 'text',
 365                                                         array( 'old_flags', 'old_text' ),
 366                                                         array( 'old_id' => $oldid ),
 367                                                         __METHOD__,
 368                                                         'FOR UPDATE'
 369                                                 );
 370                                                 $text = Revision::getRevisionText( $textRow );
 371                                         } else {
 372                                                 $text = Revision::getRevisionText( $revs[$i + $j] );
 373                                         }
 374
 375                                         if ( $text === false ) {
 376                                                 $this->error( "\nError, unable to get text in old_id $oldid" );
 377                                                 #$dbw->delete( 'old', array( 'old_id' => $oldid ) );
 378                                         }
 379
 380                                         if ( $extdb == "" && $j == 0 ) {
 381                                                 $chunk->setText( $text );
 382                                                 $this->output( '.' );
 383                                         } else {
 384                                                 # Don't make a stub if it's going to be longer than the article
 385                                                 # Stubs are typically about 100 bytes
 386                                                 if ( strlen( $text ) < 120 ) {
 387                                                         $stub = false;
 388                                                         $this->output( 'x' );
 389                                                 } else {
 390                                                         $stub = new HistoryBlobStub( $chunk->addItem( $text ) );
 391                                                         $stub->setLocation( $primaryOldid );
 392                                                         $stub->setReferrer( $oldid );
 393                                                         $this->output( '.' );
 394                                                         $usedChunk = true;
 395                                                 }
 396                                                 $stubs[$j] = $stub;
 397                                         }
 398                                 }
 399                                 $thisChunkSize = $j;
 400
 401                                 # If we couldn't actually use any stubs because the pages were too small, do nothing
 402                                 if ( $usedChunk ) {
 403                                         if ( $extdb != "" ) {
 404                                                 # Move blob objects to External Storage
 405                                                 $stored = $storeObj->store( $extdb, serialize( $chunk ) );
 406                                                 if ( $stored === false ) {
 407                                                         $this->error( "Unable to store object" );
 408
 409                                                         return false;
 410                                                 }
 411                                                 # Store External Storage URLs instead of Stub placeholders
 412                                                 foreach ( $stubs as $stub ) {
 413                                                         if ( $stub === false ) {
 414                                                                 continue;
 415                                                         }
 416                                                         # $stored should provide base path to a BLOB
 417                                                         $url = $stored . "/" . $stub->getHash();
 418                                                         $dbw->update( 'text',
 419                                                                 array( /* SET */
 420                                                                         'old_text' => $url,
 421                                                                         'old_flags' => 'external,utf-8',
 422                                                                 ), array( /* WHERE */
 423                                                                         'old_id' => $stub->getReferrer(),
 424                                                                 )
 425                                                         );
 426                                                 }
 427                                         } else {
 428                                                 # Store the main object locally
 429                                                 $dbw->update( 'text',
 430                                                         array( /* SET */
 431                                                                 'old_text' => serialize( $chunk ),
 432                                                                 'old_flags' => 'object,utf-8',
 433                                                         ), array( /* WHERE */
 434                                                                 'old_id' => $primaryOldid
 435                                                         )
 436                                                 );
 437
 438                                                 # Store the stub objects
 439                                                 for ( $j = 1; $j < $thisChunkSize; $j++ ) {
 440                                                         # Skip if not compressing and don't overwrite the first revision
 441                                                         if ( $stubs[$j] !== false && $revs[$i + $j]->rev_text_id != $primaryOldid ) {
 442                                                                 $dbw->update( 'text',
 443                                                                         array( /* SET */
 444                                                                                 'old_text' => serialize( $stubs[$j] ),
 445                                                                                 'old_flags' => 'object,utf-8',
 446                                                                         ), array( /* WHERE */
 447                                                                                 'old_id' => $revs[$i + $j]->rev_text_id
 448                                                                         )
 449                                                                 );
 450                                                         }
 451                                                 }
 452                                         }
 453                                 }
 454                                 # Done, next
 455                                 $this->output( "/" );
 456                                 $dbw->commit( __METHOD__ );
 457                                 $i += $thisChunkSize;
 458                                 wfWaitForSlaves();
 459                         }
 460                         $this->output( "\n" );
 461                 }
 462
 463                 return true;
 464         }
 465 }
 466
 467 $maintClass = 'CompressOld';
 468 require_once RUN_MAINTENANCE_IF_MAIN;