3 * Compress the text of a wiki.
8 * php compressOld.php [options...]
11 * php compressOld.php <database> [options...]
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)
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.
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.
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
41 * @ingroup Maintenance ExternalStorage
43 use MediaWiki\MediaWikiServices
;
44 use MediaWiki\Revision\SlotRecord
;
46 require_once __DIR__
. '/../Maintenance.php';
49 * Maintenance script that compress the text of a wiki.
51 * @ingroup Maintenance ExternalStorage
53 class CompressOld
extends Maintenance
{
54 public function __construct() {
55 parent
::__construct();
56 $this->addDescription( 'Compress the text of a wiki' );
57 $this->addOption( 'type', 'Set compression type to either: gzip|concat', false, true, 't' );
60 'Maximum number of revisions in a concat chunk',
67 'Earliest date to check for uncompressed revisions',
72 $this->addOption( 'end-date', 'Latest revision date to compress', false, true, 'e' );
75 'The id to start from (gzip -> text table, concat -> page table)',
82 'Store specified revisions in an external cluster (untested)',
88 'The page_id to stop at (only when using concat compression type)',
95 public function execute() {
97 if ( !function_exists( "gzdeflate" ) ) {
98 $this->fatalError( "You must enable zlib support in PHP to compress old revisions!\n" .
99 "Please see https://www.php.net/manual/en/ref.zlib.php\n" );
102 $type = $this->getOption( 'type', 'concat' );
103 $chunkSize = $this->getOption( 'chunksize', 20 );
104 $startId = $this->getOption( 'startid', 0 );
105 $beginDate = $this->getOption( 'begin-date', '' );
106 $endDate = $this->getOption( 'end-date', '' );
107 $extDB = $this->getOption( 'extdb', '' );
108 $endId = $this->getOption( 'endid', false );
110 if ( $type != 'concat' && $type != 'gzip' ) {
111 $this->error( "Type \"{$type}\" not supported" );
114 if ( $extDB != '' ) {
115 $this->output( "Compressing database {$wgDBname} to external cluster {$extDB}\n"
116 . str_repeat( '-', 76 ) . "\n\n" );
118 $this->output( "Compressing database {$wgDBname}\n"
119 . str_repeat( '-', 76 ) . "\n\n" );
123 if ( $type == 'concat' ) {
124 $success = $this->compressWithConcat( $startId, $chunkSize, $beginDate,
125 $endDate, $extDB, $endId );
127 $this->compressOldPages( $startId, $extDB );
131 $this->output( "Done.\n" );
136 * Fetch the text row-by-row to 'compressPage' function for compression.
139 * @param string $extdb
141 private function compressOldPages( $start = 0, $extdb = '' ) {
143 $this->output( "Starting from old_id $start...\n" );
144 $dbw = $this->getDB( DB_MASTER
);
148 [ 'old_id', 'old_flags', 'old_text' ],
151 [ 'ORDER BY' => 'old_id', 'LIMIT' => $chunksize, 'FOR UPDATE' ]
154 if ( $res->numRows() == 0 ) {
160 foreach ( $res as $row ) {
161 # print " {$row->old_id} - {$row->old_namespace}:{$row->old_title}\n";
162 $this->compressPage( $row, $extdb );
163 $last = $row->old_id
;
166 $start = $last +
1; # Deletion may leave long empty stretches
167 $this->output( "$start...\n" );
172 * Compress the text in gzip format.
174 * @param stdClass $row
175 * @param string $extdb
178 private function compressPage( $row, $extdb ) {
179 if ( strpos( $row->old_flags
, 'gzip' ) !== false
180 ||
strpos( $row->old_flags
, 'object' ) !== false
182 # print "Already compressed row {$row->old_id}\n";
185 $dbw = $this->getDB( DB_MASTER
);
186 $flags = $row->old_flags ?
"{$row->old_flags},gzip" : "gzip";
187 $compress = gzdeflate( $row->old_text
);
189 # Store in external storage if required
190 if ( $extdb !== '' ) {
191 $esFactory = MediaWikiServices
::getInstance()->getExternalStoreFactory();
192 /** @var ExternalStoreDB $storeObj */
193 $storeObj = $esFactory->getStore( 'DB' );
194 $compress = $storeObj->store( $extdb, $compress );
195 if ( $compress === false ) {
196 $this->error( "Unable to store object" );
203 $dbw->update( 'text',
205 'old_flags' => $flags,
206 'old_text' => $compress
208 'old_id' => $row->old_id
217 * Compress the text in chunks after concatenating the revisions.
219 * @param int $startId
220 * @param int $maxChunkSize
221 * @param string $beginDate
222 * @param string $endDate
223 * @param string $extdb
224 * @param bool|int $maxPageId
227 private function compressWithConcat( $startId, $maxChunkSize, $beginDate,
228 $endDate, $extdb = "", $maxPageId = false
230 global $wgMultiContentRevisionSchemaMigrationStage;
232 $dbr = $this->getDB( DB_REPLICA
);
233 $dbw = $this->getDB( DB_MASTER
);
235 # Set up external storage
236 if ( $extdb != '' ) {
237 $esFactory = MediaWikiServices
::getInstance()->getExternalStoreFactory();
238 /** @var ExternalStoreDB $storeObj */
239 $storeObj = $esFactory->getStore( 'DB' );
242 # Get all articles by page_id
244 $maxPageId = $dbr->selectField( 'page', 'max(page_id)', '', __METHOD__
);
246 $this->output( "Starting from $startId of $maxPageId\n" );
250 if ( $exclude_ns0 ) {
251 print "Excluding main namespace\n";
252 $pageConds[] = 'page_namespace<>0';
255 $pageConds[] = $queryExtra;
259 # For each article, get a list of revisions which fit the criteria
261 # No recompression, use a condition on old_flags
262 # Don't compress object type entities, because that might produce data loss when
263 # overwriting bulk storage concat rows. Don't compress external references, because
264 # the script doesn't yet delete rows from external storage.
266 'old_flags NOT ' . $dbr->buildLike( $dbr->anyString(), 'object', $dbr->anyString() )
267 . ' AND old_flags NOT '
268 . $dbr->buildLike( $dbr->anyString(), 'external', $dbr->anyString() )
272 if ( !preg_match( '/^\d{14}$/', $beginDate ) ) {
273 $this->error( "Invalid begin date \"$beginDate\"\n" );
277 $conds[] = "rev_timestamp>'" . $beginDate . "'";
280 if ( !preg_match( '/^\d{14}$/', $endDate ) ) {
281 $this->error( "Invalid end date \"$endDate\"\n" );
285 $conds[] = "rev_timestamp<'" . $endDate . "'";
288 if ( $wgMultiContentRevisionSchemaMigrationStage & SCHEMA_COMPAT_READ_OLD
) {
289 $tables = [ 'revision', 'text' ];
290 $conds[] = 'rev_text_id=old_id';
292 $slotRoleStore = MediaWikiServices
::getInstance()->getSlotRoleStore();
293 $tables = [ 'revision', 'slots', 'content', 'text' ];
294 $conds = array_merge( [
295 'rev_id=slot_revision_id',
296 'slot_role_id=' . $slotRoleStore->getId( SlotRecord
::MAIN
),
297 'content_id=slot_content_id',
298 'SUBSTRING(content_address, 1, 3)=' . $dbr->addQuotes( 'tt:' ),
299 'SUBSTRING(content_address, 4)=old_id',
303 $fields = [ 'rev_id', 'old_id', 'old_flags', 'old_text' ];
304 $revLoadOptions = 'FOR UPDATE';
306 # Don't work with current revisions
307 # Don't lock the page table for update either -- TS 2006-04-04
308 # $tables[] = 'page';
309 # $conds[] = 'page_id=rev_page AND rev_id != page_latest';
311 for ( $pageId = $startId; $pageId <= $maxPageId; $pageId++
) {
318 $pageRes = $dbr->select( 'page',
319 [ 'page_id', 'page_namespace', 'page_title', 'page_latest' ],
320 $pageConds +
[ 'page_id' => $pageId ], __METHOD__
);
321 if ( $pageRes->numRows() == 0 ) {
324 $pageRow = $dbr->fetchObject( $pageRes );
327 $titleObj = Title
::makeTitle( $pageRow->page_namespace
, $pageRow->page_title
);
328 $this->output( "$pageId\t" . $titleObj->getPrefixedDBkey() . " " );
331 $revRes = $dbw->select( $tables, $fields,
333 'rev_page' => $pageRow->page_id
,
334 # Don't operate on the current revision
335 # Use < instead of <> in case the current revision has changed
336 # since the page select, which wasn't locking
337 'rev_id < ' . $pageRow->page_latest
343 foreach ( $revRes as $revRow ) {
347 if ( count( $revs ) < 2 ) {
348 # No revisions matching, no further processing
349 $this->output( "\n" );
355 while ( $i < count( $revs ) ) {
356 if ( $i < count( $revs ) - $maxChunkSize ) {
357 $thisChunkSize = $maxChunkSize;
359 $thisChunkSize = count( $revs ) - $i;
362 $chunk = new ConcatenatedGzipHistoryBlob();
364 $this->beginTransaction( $dbw, __METHOD__
);
366 $primaryOldid = $revs[$i]->old_id
;
368 # Get the text of each revision and add it to the object
369 for ( $j = 0; $j < $thisChunkSize && $chunk->isHappy(); $j++
) {
370 $oldid = $revs[$i +
$j]->old_id
;
373 $text = Revision
::getRevisionText( $revs[$i +
$j] );
375 if ( $text === false ) {
376 $this->error( "\nError, unable to get text in old_id $oldid" );
377 # $dbw->delete( 'old', [ 'old_id' => $oldid ] );
380 if ( $extdb == "" && $j == 0 ) {
381 $chunk->setText( $text );
382 $this->output( '.' );
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 ) {
388 $this->output( 'x' );
390 $stub = new HistoryBlobStub( $chunk->addItem( $text ) );
391 $stub->setLocation( $primaryOldid );
392 $stub->setReferrer( $oldid );
393 $this->output( '.' );
401 # If we couldn't actually use any stubs because the pages were too small, do nothing
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" );
411 # Store External Storage URLs instead of Stub placeholders
412 foreach ( $stubs as $stub ) {
413 if ( $stub === false ) {
416 # $stored should provide base path to a BLOB
417 $url = $stored . "/" . $stub->getHash();
418 $dbw->update( 'text',
421 'old_flags' => 'external,utf-8',
423 'old_id' => $stub->getReferrer(),
428 # Store the main object locally
429 $dbw->update( 'text',
431 'old_text' => serialize( $chunk ),
432 'old_flags' => 'object,utf-8',
434 'old_id' => $primaryOldid
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]->old_id
!= $primaryOldid ) {
442 $dbw->update( 'text',
444 'old_text' => serialize( $stubs[$j] ),
445 'old_flags' => 'object,utf-8',
447 'old_id' => $revs[$i +
$j]->old_id
455 $this->output( "/" );
456 $this->commitTransaction( $dbw, __METHOD__
);
457 $i +
= $thisChunkSize;
459 $this->output( "\n" );
466 $maintClass = CompressOld
::class;
467 require_once RUN_MAINTENANCE_IF_MAIN
;