MediaWiki  REL1_31
benchmarkParse.php
Go to the documentation of this file.
1 <?php
25 require __DIR__ . '/../Maintenance.php';
26 
28 
35 class BenchmarkParse extends Maintenance {
37  private $templateTimestamp = null;
38 
39  private $clearLinkCache = false;
40 
44  private $linkCache;
45 
47  private $idCache = [];
48 
49  function __construct() {
50  parent::__construct();
51  $this->addDescription( 'Benchmark parse operation' );
52  $this->addArg( 'title', 'The name of the page to parse' );
53  $this->addOption( 'warmup', 'Repeat the parse operation this number of times to warm the cache',
54  false, true );
55  $this->addOption( 'loops', 'Number of times to repeat parse operation post-warmup',
56  false, true );
57  $this->addOption( 'page-time',
58  'Use the version of the page which was current at the given time',
59  false, true );
60  $this->addOption( 'tpl-time',
61  'Use templates which were current at the given time (except that moves and ' .
62  'deletes are not handled properly)',
63  false, true );
64  $this->addOption( 'reset-linkcache', 'Reset the LinkCache after every parse.',
65  false, false );
66  }
67 
68  function execute() {
69  if ( $this->hasOption( 'tpl-time' ) ) {
70  $this->templateTimestamp = wfTimestamp( TS_MW, strtotime( $this->getOption( 'tpl-time' ) ) );
71  Hooks::register( 'BeforeParserFetchTemplateAndtitle', [ $this, 'onFetchTemplate' ] );
72  }
73 
74  $this->clearLinkCache = $this->hasOption( 'reset-linkcache' );
75  // Set as a member variable to avoid function calls when we're timing the parse
76  $this->linkCache = MediaWikiServices::getInstance()->getLinkCache();
77 
78  $title = Title::newFromText( $this->getArg() );
79  if ( !$title ) {
80  $this->error( "Invalid title" );
81  exit( 1 );
82  }
83 
84  if ( $this->hasOption( 'page-time' ) ) {
85  $pageTimestamp = wfTimestamp( TS_MW, strtotime( $this->getOption( 'page-time' ) ) );
86  $id = $this->getRevIdForTime( $title, $pageTimestamp );
87  if ( !$id ) {
88  $this->error( "The page did not exist at that time" );
89  exit( 1 );
90  }
91 
92  $revision = Revision::newFromId( $id );
93  } else {
94  $revision = Revision::newFromTitle( $title );
95  }
96 
97  if ( !$revision ) {
98  $this->error( "Unable to load revision, incorrect title?" );
99  exit( 1 );
100  }
101 
102  $warmup = $this->getOption( 'warmup', 1 );
103  for ( $i = 0; $i < $warmup; $i++ ) {
104  $this->runParser( $revision );
105  }
106 
107  $loops = $this->getOption( 'loops', 1 );
108  if ( $loops < 1 ) {
109  $this->fatalError( 'Invalid number of loops specified' );
110  }
111  $startUsage = getrusage();
112  $startTime = microtime( true );
113  for ( $i = 0; $i < $loops; $i++ ) {
114  $this->runParser( $revision );
115  }
116  $endUsage = getrusage();
117  $endTime = microtime( true );
118 
119  printf( "CPU time = %.3f s, wall clock time = %.3f s\n",
120  // CPU time
121  ( $endUsage['ru_utime.tv_sec'] + $endUsage['ru_utime.tv_usec'] * 1e-6
122  - $startUsage['ru_utime.tv_sec'] - $startUsage['ru_utime.tv_usec'] * 1e-6 ) / $loops,
123  // Wall clock time
124  ( $endTime - $startTime ) / $loops
125  );
126  }
127 
135  function getRevIdForTime( Title $title, $timestamp ) {
136  $dbr = $this->getDB( DB_REPLICA );
137 
138  $id = $dbr->selectField(
139  [ 'revision', 'page' ],
140  'rev_id',
141  [
142  'page_namespace' => $title->getNamespace(),
143  'page_title' => $title->getDBkey(),
144  'rev_timestamp <= ' . $dbr->addQuotes( $timestamp )
145  ],
146  __METHOD__,
147  [ 'ORDER BY' => 'rev_timestamp DESC', 'LIMIT' => 1 ],
148  [ 'revision' => [ 'INNER JOIN', 'rev_page=page_id' ] ]
149  );
150 
151  return $id;
152  }
153 
159  function runParser( Revision $revision ) {
160  $content = $revision->getContent();
161  $content->getParserOutput( $revision->getTitle(), $revision->getId() );
162  if ( $this->clearLinkCache ) {
163  $this->linkCache->clear();
164  }
165  }
166 
177  function onFetchTemplate( Parser $parser, Title $title, &$skip, &$id ) {
178  $pdbk = $title->getPrefixedDBkey();
179  if ( !isset( $this->idCache[$pdbk] ) ) {
180  $proposedId = $this->getRevIdForTime( $title, $this->templateTimestamp );
181  $this->idCache[$pdbk] = $proposedId;
182  }
183  if ( $this->idCache[$pdbk] !== false ) {
184  $id = $this->idCache[$pdbk];
185  }
186 
187  return true;
188  }
189 }
190 
LinkCache
Cache for article titles (prefixed DB keys) and ids linked from one source.
Definition: LinkCache.php:34
Title\newFromText
static newFromText( $text, $defaultNamespace=NS_MAIN)
Create a new Title from text, such as what one would find in a link.
Definition: Title.php:273
use
Apache License January AND DISTRIBUTION Definitions License shall mean the terms and conditions for use
Definition: APACHE-LICENSE-2.0.txt:10
Revision\newFromId
static newFromId( $id, $flags=0)
Load a page revision from a given revision ID number.
Definition: Revision.php:114
array
the array() calling protocol came about after MediaWiki 1.4rc1.
Maintenance\fatalError
fatalError( $msg, $exitCode=1)
Output a message and terminate the current script.
Definition: Maintenance.php:439
Maintenance\addDescription
addDescription( $text)
Set the description text.
Definition: Maintenance.php:291
Revision\getContent
getContent( $audience=self::FOR_PUBLIC, User $user=null)
Fetch revision content if it's available to the specified audience.
Definition: Revision.php:929
wfTimestamp
wfTimestamp( $outputtype=TS_UNIX, $ts=0)
Get a timestamp string in one of various formats.
Definition: GlobalFunctions.php:1980
RUN_MAINTENANCE_IF_MAIN
require_once RUN_MAINTENANCE_IF_MAIN
Definition: maintenance.txt:50
e
in this case you re responsible for computing and outputting the entire conflict i e
Definition: hooks.txt:1421
Maintenance
Abstract maintenance class for quickly writing and churning out maintenance scripts with minimal effo...
Definition: maintenance.txt:39
Revision\getId
getId()
Get revision ID.
Definition: Revision.php:617
php
injection txt This is an overview of how MediaWiki makes use of dependency injection The design described here grew from the discussion of RFC T384 The term dependency this means that anything an object needs to operate should be injected from the the object itself should only know narrow no concrete implementation of the logic it relies on The requirement to inject everything typically results in an architecture that based on two main types of and essentially stateless service objects that use other service objects to operate on the value objects As of the beginning MediaWiki is only starting to use the DI approach Much of the code still relies on global state or direct resulting in a highly cyclical dependency which acts as the top level factory for services in MediaWiki which can be used to gain access to default instances of various services MediaWikiServices however also allows new services to be defined and default services to be redefined Services are defined or redefined by providing a callback the instantiator that will return a new instance of the service When it will create an instance of MediaWikiServices and populate it with the services defined in the files listed by thereby bootstrapping the DI framework Per $wgServiceWiringFiles lists includes ServiceWiring php
Definition: injection.txt:37
$dbr
$dbr
Definition: testCompression.php:50
Revision
Definition: Revision.php:41
Revision\newFromTitle
static newFromTitle(LinkTarget $linkTarget, $id=0, $flags=0)
Load either the current, or a specified, revision that's attached to a given link target.
Definition: Revision.php:133
BenchmarkParse
Maintenance script to benchmark how long it takes to parse a given title at an optionally specified t...
Definition: benchmarkParse.php:35
$maintClass
$maintClass
Definition: benchmarkParse.php:191
$parser
do that in ParserLimitReportFormat instead $parser
Definition: hooks.txt:2603
Maintenance\addOption
addOption( $name, $description, $required=false, $withArg=false, $shortName=false, $multiOccurrence=false)
Add a parameter to the script.
Definition: Maintenance.php:219
BenchmarkParse\$templateTimestamp
string $templateTimestamp
MediaWiki concatenated string timestamp (YYYYMMDDHHMMSS)
Definition: benchmarkParse.php:37
$title
namespace and then decline to actually register it file or subcat img or subcat $title
Definition: hooks.txt:964
DB_REPLICA
const DB_REPLICA
Definition: defines.php:25
BenchmarkParse\$linkCache
LinkCache $linkCache
Definition: benchmarkParse.php:44
Revision\getTitle
getTitle()
Returns the title of the page associated with this entry.
Definition: Revision.php:734
BenchmarkParse\onFetchTemplate
onFetchTemplate(Parser $parser, Title $title, &$skip, &$id)
Hook into the parser's revision ID fetcher.
Definition: benchmarkParse.php:177
BenchmarkParse\runParser
runParser(Revision $revision)
Parse the text from a given Revision.
Definition: benchmarkParse.php:159
Hooks\register
static register( $name, $callback)
Attach an event handler to a given hook.
Definition: Hooks.php:49
Parser
PHP Parser - Processes wiki markup (which uses a more user-friendly syntax, such as "[[link]]" for ma...
Definition: Parser.php:70
BenchmarkParse\execute
execute()
Do the actual work.
Definition: benchmarkParse.php:68
Title
Represents a title within MediaWiki.
Definition: Title.php:39
Maintenance\getOption
getOption( $name, $default=null)
Get an option, or return the default.
Definition: Maintenance.php:254
Maintenance\addArg
addArg( $arg, $description, $required=true)
Add some args that are needed.
Definition: Maintenance.php:271
BenchmarkParse\$clearLinkCache
$clearLinkCache
Definition: benchmarkParse.php:39
BenchmarkParse\__construct
__construct()
Default constructor.
Definition: benchmarkParse.php:49
Maintenance\getDB
getDB( $db, $groups=[], $wiki=false)
Returns a database to be used by current maintenance script.
Definition: Maintenance.php:1309
Maintenance\error
error( $err, $die=0)
Throw an error to the user.
Definition: Maintenance.php:416
class
you have access to all of the normal MediaWiki so you can get a DB use the etc For full docs on the Maintenance class
Definition: maintenance.txt:56
MediaWikiServices
injection txt This is an overview of how MediaWiki makes use of dependency injection The design described here grew from the discussion of RFC T384 The term dependency this means that anything an object needs to operate should be injected from the the object itself should only know narrow no concrete implementation of the logic it relies on The requirement to inject everything typically results in an architecture that based on two main types of and essentially stateless service objects that use other service objects to operate on the value objects As of the beginning MediaWiki is only starting to use the DI approach Much of the code still relies on global state or direct resulting in a highly cyclical dependency MediaWikiServices
Definition: injection.txt:25
Maintenance\hasOption
hasOption( $name)
Checks to see if a particular param exists.
Definition: Maintenance.php:240
Maintenance\getArg
getArg( $argId=0, $default=null)
Get an argument.
Definition: Maintenance.php:310
BenchmarkParse\$idCache
array $idCache
Cache that maps a Title DB key to revision ID for the requested timestamp.
Definition: benchmarkParse.php:47
BenchmarkParse\getRevIdForTime
getRevIdForTime(Title $title, $timestamp)
Fetch the ID of the revision of a Title that occurred.
Definition: benchmarkParse.php:135