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          * Option to load each revision individually.
  54          */
  55         const LS_INDIVIDUAL = 0;
  56
  57         /**
  58          * Option to load revisions in chunks.
  59          */
  60         const LS_CHUNKED = 1;
  61
  62         public function __construct() {
  63                 parent::__construct();
  64                 $this->addDescription( 'Compress the text of a wiki' );
  65                 $this->addOption( 'type', 'Set compression type to either: gzip|concat', false, true, 't' );
  66                 $this->addOption(
  67                         'chunksize',
  68                         'Maximum number of revisions in a concat chunk',
  69                         false,
  70                         true,
  71                         'c'
  72                 );
  73                 $this->addOption(
  74                         'begin-date',
  75                         'Earliest date to check for uncompressed revisions',
  76                         false,
  77                         true,
  78                         'b'
  79                 );
  80                 $this->addOption( 'end-date', 'Latest revision date to compress', false, true, 'e' );
  81                 $this->addOption(
  82                         'startid',
  83                         'The id to start from (gzip -> text table, concat -> page table)',
  84                         false,
  85                         true,
  86                         's'
  87                 );
  88                 $this->addOption(
  89                         'extdb',
  90                         'Store specified revisions in an external cluster (untested)',
  91                         false,
  92                         true
  93                 );
  94                 $this->addOption(
  95                         'endid',
  96                         'The page_id to stop at (only when using concat compression type)',
  97                         false,
  98                         true,
  99                         'n'
 100                 );
 101         }
 102
 103         public function execute() {
 104                 global $wgDBname;
 105                 if ( !function_exists( "gzdeflate" ) ) {
 106                         $this->fatalError( "You must enable zlib support in PHP to compress old revisions!\n" .
 107                                 "Please see https://www.php.net/manual/en/ref.zlib.php\n" );
 108                 }
 109
 110                 $type = $this->getOption( 'type', 'concat' );
 111                 $chunkSize = $this->getOption( 'chunksize', 20 );
 112                 $startId = $this->getOption( 'startid', 0 );
 113                 $beginDate = $this->getOption( 'begin-date', '' );
 114                 $endDate = $this->getOption( 'end-date', '' );
 115                 $extDB = $this->getOption( 'extdb', '' );
 116                 $endId = $this->getOption( 'endid', false );
 117
 118                 if ( $type != 'concat' && $type != 'gzip' ) {
 119                         $this->error( "Type \"{$type}\" not supported" );
 120                 }
 121
 122                 if ( $extDB != '' ) {
 123                         $this->output( "Compressing database {$wgDBname} to external cluster {$extDB}\n"
 124                                 . str_repeat( '-', 76 ) . "\n\n" );
 125                 } else {
 126                         $this->output( "Compressing database {$wgDBname}\n"
 127                                 . str_repeat( '-', 76 ) . "\n\n" );
 128                 }
 129
 130                 $success = true;
 131                 if ( $type == 'concat' ) {
 132                         $success = $this->compressWithConcat( $startId, $chunkSize, $beginDate,
 133                                 $endDate, $extDB, $endId );
 134                 } else {
 135                         $this->compressOldPages( $startId, $extDB );
 136                 }
 137
 138                 if ( $success ) {
 139                         $this->output( "Done.\n" );
 140                 }
 141         }
 142
 143         /**
 144          * Fetch the text row-by-row to 'compressPage' function for compression.
 145          *
 146          * @param int $start
 147          * @param string $extdb
 148          */
 149         private function compressOldPages( $start = 0, $extdb = '' ) {
 150                 $chunksize = 50;
 151                 $this->output( "Starting from old_id $start...\n" );
 152                 $dbw = $this->getDB( DB_MASTER );
 153                 do {
 154                         $res = $dbw->select(
 155                                 'text',
 156                                 [ 'old_id', 'old_flags', 'old_text' ],
 157                                 "old_id>=$start",
 158                                 __METHOD__,
 159                                 [ 'ORDER BY' => 'old_id', 'LIMIT' => $chunksize, 'FOR UPDATE' ]
 160                         );
 161
 162                         if ( $res->numRows() == 0 ) {
 163                                 break;
 164                         }
 165
 166                         $last = $start;
 167
 168                         foreach ( $res as $row ) {
 169                                 # print "  {$row->old_id} - {$row->old_namespace}:{$row->old_title}\n";
 170                                 $this->compressPage( $row, $extdb );
 171                                 $last = $row->old_id;
 172                         }
 173
 174                         $start = $last + 1; # Deletion may leave long empty stretches
 175                         $this->output( "$start...\n" );
 176                 } while ( true );
 177         }
 178
 179         /**
 180          * Compress the text in gzip format.
 181          *
 182          * @param stdClass $row
 183          * @param string $extdb
 184          * @return bool
 185          */
 186         private function compressPage( $row, $extdb ) {
 187                 if ( strpos( $row->old_flags, 'gzip' ) !== false
 188                         || strpos( $row->old_flags, 'object' ) !== false
 189                 ) {
 190                         # print "Already compressed row {$row->old_id}\n";
 191                         return false;
 192                 }
 193                 $dbw = $this->getDB( DB_MASTER );
 194                 $flags = $row->old_flags ? "{$row->old_flags},gzip" : "gzip";
 195                 $compress = gzdeflate( $row->old_text );
 196
 197                 # Store in external storage if required
 198                 if ( $extdb !== '' ) {
 199                         $storeObj = new ExternalStoreDB;
 200                         $compress = $storeObj->store( $extdb, $compress );
 201                         if ( $compress === false ) {
 202                                 $this->error( "Unable to store object" );
 203
 204                                 return false;
 205                         }
 206                 }
 207
 208                 # Update text row
 209                 $dbw->update( 'text',
 210                         [ /* SET */
 211                                 'old_flags' => $flags,
 212                                 'old_text' => $compress
 213                         ], [ /* WHERE */
 214                                 'old_id' => $row->old_id
 215                         ], __METHOD__,
 216                         [ 'LIMIT' => 1 ]
 217                 );
 218
 219                 return true;
 220         }
 221
 222         /**
 223          * Compress the text in chunks after concatenating the revisions.
 224          *
 225          * @param int $startId
 226          * @param int $maxChunkSize
 227          * @param string $beginDate
 228          * @param string $endDate
 229          * @param string $extdb
 230          * @param bool|int $maxPageId
 231          * @return bool
 232          */
 233         private function compressWithConcat( $startId, $maxChunkSize, $beginDate,
 234                 $endDate, $extdb = "", $maxPageId = false
 235         ) {
 236                 $loadStyle = self::LS_CHUNKED;
 237
 238                 $dbr = $this->getDB( DB_REPLICA );
 239                 $dbw = $this->getDB( DB_MASTER );
 240
 241                 # Set up external storage
 242                 if ( $extdb != '' ) {
 243                         $storeObj = new ExternalStoreDB;
 244                 }
 245
 246                 # Get all articles by page_id
 247                 if ( !$maxPageId ) {
 248                         $maxPageId = $dbr->selectField( 'page', 'max(page_id)', '', __METHOD__ );
 249                 }
 250                 $this->output( "Starting from $startId of $maxPageId\n" );
 251                 $pageConds = [];
 252
 253                 /*
 254                 if ( $exclude_ns0 ) {
 255                         print "Excluding main namespace\n";
 256                         $pageConds[] = 'page_namespace<>0';
 257                 }
 258                 if ( $queryExtra ) {
 259                                         $pageConds[] = $queryExtra;
 260                 }
 261                  */
 262
 263                 # For each article, get a list of revisions which fit the criteria
 264
 265                 # No recompression, use a condition on old_flags
 266                 # Don't compress object type entities, because that might produce data loss when
 267                 # overwriting bulk storage concat rows. Don't compress external references, because
 268                 # the script doesn't yet delete rows from external storage.
 269                 $conds = [
 270                         'old_flags NOT ' . $dbr->buildLike( $dbr->anyString(), 'object', $dbr->anyString() )
 271                         . ' AND old_flags NOT '
 272                         . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() )
 273                 ];
 274
 275                 if ( $beginDate ) {
 276                         if ( !preg_match( '/^\d{14}$/', $beginDate ) ) {
 277                                 $this->error( "Invalid begin date \"$beginDate\"\n" );
 278
 279                                 return false;
 280                         }
 281                         $conds[] = "rev_timestamp>'" . $beginDate . "'";
 282                 }
 283                 if ( $endDate ) {
 284                         if ( !preg_match( '/^\d{14}$/', $endDate ) ) {
 285                                 $this->error( "Invalid end date \"$endDate\"\n" );
 286
 287                                 return false;
 288                         }
 289                         $conds[] = "rev_timestamp<'" . $endDate . "'";
 290                 }
 291                 if ( $loadStyle == self::LS_CHUNKED ) {
 292                         $tables = [ 'revision', 'text' ];
 293                         $fields = [ 'rev_id', 'rev_text_id', 'old_flags', 'old_text' ];
 294                         $conds[] = 'rev_text_id=old_id';
 295                         $revLoadOptions = 'FOR UPDATE';
 296                 } else {
 297                         $tables = [ 'revision' ];
 298                         $fields = [ 'rev_id', 'rev_text_id' ];
 299                         $revLoadOptions = [];
 300                 }
 301
 302                 # Don't work with current revisions
 303                 # Don't lock the page table for update either -- TS 2006-04-04
 304                 # $tables[] = 'page';
 305                 # $conds[] = 'page_id=rev_page AND rev_id != page_latest';
 306
 307                 for ( $pageId = $startId; $pageId <= $maxPageId; $pageId++ ) {
 308                         wfWaitForSlaves();
 309
 310                         # Wake up
 311                         $dbr->ping();
 312
 313                         # Get the page row
 314                         $pageRes = $dbr->select( 'page',
 315                                 [ 'page_id', 'page_namespace', 'page_title', 'page_latest' ],
 316                                 $pageConds + [ 'page_id' => $pageId ], __METHOD__ );
 317                         if ( $pageRes->numRows() == 0 ) {
 318                                 continue;
 319                         }
 320                         $pageRow = $dbr->fetchObject( $pageRes );
 321
 322                         # Display progress
 323                         $titleObj = Title::makeTitle( $pageRow->page_namespace, $pageRow->page_title );
 324                         $this->output( "$pageId\t" . $titleObj->getPrefixedDBkey() . " " );
 325
 326                         # Load revisions
 327                         $revRes = $dbw->select( $tables, $fields,
 328                                 array_merge( [
 329                                         'rev_page' => $pageRow->page_id,
 330                                         # Don't operate on the current revision
 331                                         # Use < instead of <> in case the current revision has changed
 332                                         # since the page select, which wasn't locking
 333                                         'rev_id < ' . $pageRow->page_latest
 334                                 ], $conds ),
 335                                 __METHOD__,
 336                                 $revLoadOptions
 337                         );
 338                         $revs = [];
 339                         foreach ( $revRes as $revRow ) {
 340                                 $revs[] = $revRow;
 341                         }
 342
 343                         if ( count( $revs ) < 2 ) {
 344                                 # No revisions matching, no further processing
 345                                 $this->output( "\n" );
 346                                 continue;
 347                         }
 348
 349                         # For each chunk
 350                         $i = 0;
 351                         while ( $i < count( $revs ) ) {
 352                                 if ( $i < count( $revs ) - $maxChunkSize ) {
 353                                         $thisChunkSize = $maxChunkSize;
 354                                 } else {
 355                                         $thisChunkSize = count( $revs ) - $i;
 356                                 }
 357
 358                                 $chunk = new ConcatenatedGzipHistoryBlob();
 359                                 $stubs = [];
 360                                 $this->beginTransaction( $dbw, __METHOD__ );
 361                                 $usedChunk = false;
 362                                 $primaryOldid = $revs[$i]->rev_text_id;
 363
 364                                 # Get the text of each revision and add it to the object
 365                                 for ( $j = 0; $j < $thisChunkSize && $chunk->isHappy(); $j++ ) {
 366                                         $oldid = $revs[$i + $j]->rev_text_id;
 367
 368                                         # Get text
 369                                         if ( $loadStyle == self::LS_INDIVIDUAL ) {
 370                                                 $textRow = $dbw->selectRow( 'text',
 371                                                         [ 'old_flags', 'old_text' ],
 372                                                         [ 'old_id' => $oldid ],
 373                                                         __METHOD__,
 374                                                         'FOR UPDATE'
 375                                                 );
 376                                                 $text = Revision::getRevisionText( $textRow );
 377                                         } else {
 378                                                 $text = Revision::getRevisionText( $revs[$i + $j] );
 379                                         }
 380
 381                                         if ( $text === false ) {
 382                                                 $this->error( "\nError, unable to get text in old_id $oldid" );
 383                                                 # $dbw->delete( 'old', [ 'old_id' => $oldid ] );
 384                                         }
 385
 386                                         if ( $extdb == "" && $j == 0 ) {
 387                                                 $chunk->setText( $text );
 388                                                 $this->output( '.' );
 389                                         } else {
 390                                                 # Don't make a stub if it's going to be longer than the article
 391                                                 # Stubs are typically about 100 bytes
 392                                                 if ( strlen( $text ) < 120 ) {
 393                                                         $stub = false;
 394                                                         $this->output( 'x' );
 395                                                 } else {
 396                                                         $stub = new HistoryBlobStub( $chunk->addItem( $text ) );
 397                                                         $stub->setLocation( $primaryOldid );
 398                                                         $stub->setReferrer( $oldid );
 399                                                         $this->output( '.' );
 400                                                         $usedChunk = true;
 401                                                 }
 402                                                 $stubs[$j] = $stub;
 403                                         }
 404                                 }
 405                                 $thisChunkSize = $j;
 406
 407                                 # If we couldn't actually use any stubs because the pages were too small, do nothing
 408                                 if ( $usedChunk ) {
 409                                         if ( $extdb != "" ) {
 410                                                 # Move blob objects to External Storage
 411                                                 $stored = $storeObj->store( $extdb, serialize( $chunk ) );
 412                                                 if ( $stored === false ) {
 413                                                         $this->error( "Unable to store object" );
 414
 415                                                         return false;
 416                                                 }
 417                                                 # Store External Storage URLs instead of Stub placeholders
 418                                                 foreach ( $stubs as $stub ) {
 419                                                         if ( $stub === false ) {
 420                                                                 continue;
 421                                                         }
 422                                                         # $stored should provide base path to a BLOB
 423                                                         $url = $stored . "/" . $stub->getHash();
 424                                                         $dbw->update( 'text',
 425                                                                 [ /* SET */
 426                                                                         'old_text' => $url,
 427                                                                         'old_flags' => 'external,utf-8',
 428                                                                 ], [ /* WHERE */
 429                                                                         'old_id' => $stub->getReferrer(),
 430                                                                 ]
 431                                                         );
 432                                                 }
 433                                         } else {
 434                                                 # Store the main object locally
 435                                                 $dbw->update( 'text',
 436                                                         [ /* SET */
 437                                                                 'old_text' => serialize( $chunk ),
 438                                                                 'old_flags' => 'object,utf-8',
 439                                                         ], [ /* WHERE */
 440                                                                 'old_id' => $primaryOldid
 441                                                         ]
 442                                                 );
 443
 444                                                 # Store the stub objects
 445                                                 for ( $j = 1; $j < $thisChunkSize; $j++ ) {
 446                                                         # Skip if not compressing and don't overwrite the first revision
 447                                                         if ( $stubs[$j] !== false && $revs[$i + $j]->rev_text_id != $primaryOldid ) {
 448                                                                 $dbw->update( 'text',
 449                                                                         [ /* SET */
 450                                                                                 'old_text' => serialize( $stubs[$j] ),
 451                                                                                 'old_flags' => 'object,utf-8',
 452                                                                         ], [ /* WHERE */
 453                                                                                 'old_id' => $revs[$i + $j]->rev_text_id
 454                                                                         ]
 455                                                                 );
 456                                                         }
 457                                                 }
 458                                         }
 459                                 }
 460                                 # Done, next
 461                                 $this->output( "/" );
 462                                 $this->commitTransaction( $dbw, __METHOD__ );
 463                                 $i += $thisChunkSize;
 464                         }
 465                         $this->output( "\n" );
 466                 }
 467
 468                 return true;
 469         }
 470 }
 471
 472 $maintClass = CompressOld::class;
 473 require_once RUN_MAINTENANCE_IF_MAIN;