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