MediaWiki  master
grep.php
Go to the documentation of this file.
1 <?php
2 // phpcs:disable MediaWiki.Files.ClassMatchesFilename.NotMatch
8 
9 require_once __DIR__ . '/Maintenance.php';
10 
16 class GrepPages extends Maintenance {
18  private $contLang;
19 
21  private $wikiPageFactory;
22 
23  public function __construct() {
24  parent::__construct();
25  $this->addDescription( 'Search the source text of pages for lines matching ' .
26  'a given regex, and print the lines.' );
27  $this->addOption( 'prefix',
28  'Title prefix. Can be specified more than once. ' .
29  'Use e.g. --prefix=Talk: to search an entire namespace.',
30  false, true, false, true );
31  $this->addOption( 'show-wiki', 'Add the wiki ID to the output' );
32  $this->addOption( 'pages-with-matches',
33  'Suppress normal output; instead print the title of each page ' .
34  'from which output would normally have been printed.',
35  false, false, 'l' );
36  $this->addArg( 'regex', 'The regex to search for' );
37  }
38 
39  private function init() {
40  $services = MediaWikiServices::getInstance();
41  $this->contLang = $services->getContentLanguage();
42  $this->wikiPageFactory = $services->getWikiPageFactory();
43  }
44 
45  public function execute() {
46  $this->init();
47 
48  $showWiki = $this->getOption( 'show-wiki' );
49  $wikiId = WikiMap::getCurrentWikiId();
50  $prefix = $this->getOption( 'prefix' );
51  $regex = $this->getArg( 0 );
52  $titleOnly = $this->hasOption( 'pages-with-matches' );
53 
54  if ( ( $regex[0] ?? '' ) === '/' ) {
55  $delimRegex = $regex[0];
56  } else {
57  $delimRegex = '{' . $regex . '}';
58  }
59 
60  foreach ( $this->findPages( $prefix ) as $page ) {
61  $content = $page->getContent( RevisionRecord::RAW );
62  $titleText = $page->getTitle()->getPrefixedDBkey();
63  if ( !$content ) {
64  $this->error( "Page has no content: $titleText" );
65  continue;
66  }
67  if ( !$content instanceof TextContent ) {
68  $this->error( "Page has a non-text content model: $titleText" );
69  continue;
70  }
71 
72  $text = $content->getText();
73 
74  if ( $titleOnly ) {
75  if ( preg_match( $delimRegex, $text ) ) {
76  if ( $showWiki ) {
77  echo "$wikiId\t$titleText\n";
78  } else {
79  echo "$titleText\n";
80  }
81  }
82  } else {
83  foreach ( StringUtils::explode( "\n", $text ) as $lineNum => $line ) {
84  $lineNum++;
85  if ( preg_match( $delimRegex, $line ) ) {
86  if ( $showWiki ) {
87  echo "$wikiId\t$titleText:$lineNum:$line\n";
88  } else {
89  echo "$titleText:$lineNum:$line\n";
90  }
91  }
92  }
93  }
94  }
95  }
96 
97  public function findPages( $prefixes = null ) {
98  $dbr = $this->getDB( DB_REPLICA );
99  $orConds = [];
100  if ( $prefixes !== null ) {
101  foreach ( $prefixes as $prefix ) {
102  $colonPos = strpos( $prefix, ':' );
103  if ( $colonPos !== false ) {
104  $ns = $this->contLang->getNsIndex( substr( $prefix, 0, $colonPos ) );
105  $prefixDBkey = substr( $prefix, $colonPos + 1 );
106  } else {
107  $ns = NS_MAIN;
108  $prefixDBkey = $prefix;
109  }
110  $prefixCond = [ 'page_namespace' => $ns ];
111  if ( $prefixDBkey !== '' ) {
112  $prefixCond[] = 'page_title ' . $dbr->buildLike( $prefixDBkey, $dbr->anyString() );
113  }
114  $orConds[] = $dbr->makeList( $prefixCond, LIST_AND );
115  }
116  }
117 
118  $conds = $orConds ? $dbr->makeList( $orConds, LIST_OR ) : [];
119  $pageQuery = WikiPage::getQueryInfo();
120 
121  $res = $dbr->newSelectQueryBuilder()
122  ->queryInfo( $pageQuery )
123  ->where( $conds )
124  ->caller( __METHOD__ )
125  ->fetchResultSet();
126  foreach ( $res as $row ) {
127  $title = Title::newFromRow( $row );
128  yield $this->wikiPageFactory->newFromTitle( $title );
129  }
130  }
131 }
132 
133 $maintClass = GrepPages::class;
134 require_once RUN_MAINTENANCE_IF_MAIN;
const NS_MAIN
Definition: Defines.php:64
const LIST_OR
Definition: Defines.php:46
const LIST_AND
Definition: Defines.php:43
Search pages for a given regex.
Definition: grep.php:16
execute()
Do the actual work.
Definition: grep.php:45
findPages( $prefixes=null)
Definition: grep.php:97
__construct()
Default constructor.
Definition: grep.php:23
Abstract maintenance class for quickly writing and churning out maintenance scripts with minimal effo...
Definition: Maintenance.php:66
getDB( $db, $groups=[], $dbDomain=false)
Returns a database to be used by current maintenance script.
error( $err, $die=0)
Throw an error to the user.
addArg( $arg, $description, $required=true, $multi=false)
Add some args that are needed.
hasOption( $name)
Checks to see if a particular option was set.
getArg( $argId=0, $default=null)
Get an argument.
addDescription( $text)
Set the description text.
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.
Service locator for MediaWiki core services.
Service for creating WikiPage objects.
Page revision base class.
Represents a title within MediaWiki.
Definition: Title.php:82
Helper tools for dealing with other locally-hosted wikis.
Definition: WikiMap.php:33
static explode( $separator, $subject)
Workalike for explode() with limited memory usage.
Content object implementation for representing flat text.
Definition: TextContent.php:40
static getQueryInfo()
Return the tables, fields, and join conditions to be selected to create a new page object.
Definition: WikiPage.php:375
$maintClass
Definition: grep.php:133
const DB_REPLICA
Definition: defines.php:26
$content
Definition: router.php:76