MediaWiki master
testCompression.php
Go to the documentation of this file.
1<?php
27
28require_once __DIR__ . '/../Maintenance.php';
29
31 public function __construct() {
32 parent::__construct();
33
34 $this->addArg( 'title', 'The page to test' );
35
36 $this->addOption( 'type', 'The HistoryBlob subclass to use', false, true );
37 $this->addOption( 'start', 'The start date', false, true );
38 $this->addOption( 'limit', 'Maximum number of revisions to process', false, true );
39 }
40
41 public function execute() {
42 $lang = $this->getServiceContainer()->getLanguageFactory()->getLanguage( 'en' );
43 $title = Title::newFromText( $this->getArg( 0 ) );
44
45 if ( $this->hasOption( 'start' ) ) {
46 $start = wfTimestamp( TS_MW, strtotime( $this->getOption( 'start' ) ) );
47 echo "Starting from " . $lang->timeanddate( $start ) . "\n";
48 } else {
49 $start = '19700101000000';
50 }
51 if ( $this->hasOption( 'limit' ) ) {
52 $limit = $this->getOption( 'limit' );
53 $untilHappy = false;
54 } else {
55 $limit = 1000;
56 $untilHappy = true;
57 }
58 $type = $this->getOption( 'type', ConcatenatedGzipHistoryBlob::class );
59
60 $dbr = $this->getReplicaDB();
61
62 $revStore = $this->getServiceContainer()->getRevisionStore();
63 $res = $revStore->newSelectQueryBuilder( $dbr )
64 ->joinComment()
65 ->joinPage()
66 ->where( [
67 'page_namespace' => $title->getNamespace(),
68 'page_title' => $title->getDBkey(),
69 $dbr->expr( 'rev_timestamp', '>', $dbr->timestamp( $start ) ),
70 ] )
71 ->limit( $limit )
72 ->caller( __FILE__ )->fetchResultSet();
73
74 $blob = new $type;
75 $hashes = [];
76 $keys = [];
77 $uncompressedSize = 0;
78 $t = -microtime( true );
79 foreach ( $res as $row ) {
80 $revRecord = $revStore->newRevisionFromRow( $row );
81 $text = $revRecord->getSlot( SlotRecord::MAIN, RevisionRecord::RAW )
82 ->getContent()
83 ->serialize();
84 $uncompressedSize += strlen( $text );
85 $hashes[$row->rev_id] = md5( $text );
86 $keys[$row->rev_id] = $blob->addItem( $text );
87 if ( $untilHappy && !$blob->isHappy() ) {
88 break;
89 }
90 }
91
92 $serialized = serialize( $blob );
93 $t += microtime( true );
94 # print_r( $blob->mDiffMap );
95
96 printf( "%s\nCompression ratio for %d revisions: %5.2f, %s -> %d\n",
97 $type,
98 count( $hashes ),
99 $uncompressedSize / strlen( $serialized ),
100 $lang->formatSize( $uncompressedSize ),
101 strlen( $serialized )
102 );
103 printf( "Compression time: %5.2f ms\n", $t * 1000 );
104
105 $t = -microtime( true );
106 $blob = unserialize( $serialized );
107 foreach ( $keys as $id => $key ) {
108 $text = $blob->getItem( $key );
109 if ( md5( $text ) != $hashes[$id] ) {
110 echo "Content hash mismatch for rev_id $id\n";
111 # var_dump( $text );
112 }
113 }
114 $t += microtime( true );
115 printf( "Decompression time: %5.2f ms\n", $t * 1000 );
116 }
117}
118
119$maintClass = TestCompression::class;
120require_once RUN_MAINTENANCE_IF_MAIN;
wfTimestamp( $outputtype=TS_UNIX, $ts=0)
Get a timestamp string in one of various formats.
Abstract maintenance class for quickly writing and churning out maintenance scripts with minimal effo...
addArg( $arg, $description, $required=true, $multi=false)
Add some args that are needed.
hasOption( $name)
Checks to see if a particular option was set.
getServiceContainer()
Returns the main service container.
getArg( $argId=0, $default=null)
Get an argument.
addOption( $name, $description, $required=false, $withArg=false, $shortName=false, $multiOccurrence=false)
Add a parameter to the script.
getOption( $name, $default=null)
Get an option, or return the default.
Page revision base class.
Value object representing a content slot associated with a page revision.
Represents a title within MediaWiki.
Definition Title.php:78
execute()
Do the actual work.
__construct()
Default constructor.