diocloid/LinkTitles
1
0
Fork 0
mirror of https://github.com/diocloid/LinkTitles.git synced 2025-07-13 09:49:31 +02:00
Code Issues Packages Projects Releases Wiki Activity

Make Linker class instantiable.

Browse Source
This commit is contained in:
Daniel Kraus
2017-08-27 10:42:10 +02:00
parent 5fe5923bf7
commit 5f7c96459d
4 changed files with 112 additions and 89 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
LinkTitles_Maintenance.php
View File

@ -44,7 +44,7 @@ else
}
};
require_once( __DIR__ . "/includes/LinkTitles_Extension.php" );
require_once( __DIR__ . "/includes/Extension.php" );
/// Core class of the maintanance script.
/// @note Note that the execution of maintenance scripts is prohibited for

5
includes/Config.php
View File

@ -133,6 +133,9 @@ class Config {
*/
public $parseHeadings;
public $enableConsoleOutput;
public $enableDebugConsoleOutput;
/**
* Constructs a new Config object.
*
@ -166,6 +169,8 @@ class Config {
$this->wordEndOnly = $wgLinkTitlesWordEndOnly;
$this->skipTemplates = $wgLinkTitlesSkipTemplates;
$this->parseHeadings = $wgLinkTitlesParseHeadings;
$this->enableConsoleOutput = false;
$this->enableDebugConsoleOutput = false;
}
}

52
includes/Extension.php
View File

@ -1,6 +1,6 @@
<?php
/**
* The LinkTitles\Extension class provides entry points for the extension.
* The LinkTitles\Extension class provides event handlers and entry points for the extension.
*
* Copyright 2012-2017 Daniel Kraus <bovender@bovender.de> ('bovender')
*
@ -24,7 +24,7 @@
namespace LinkTitles;
/**
* Provides entry points for the extension.
* Provides event handlers and entry points for the extension.
*/
class Extension {
@ -33,7 +33,7 @@ class Extension {
$isMinor, $isWatch, $section, &$flags, &$status ) {
global $wgLinkTitlesParseOnEdit;
global $wgLinkTitlesNamespaces;
if ( !$wgLinkTitlesParseOnEdit ) return true;
if ( !$wgLinkTitlesParseOnEdit ) return true; // TODO: refactor with following if
if ( !$isMinor ) {
$title = $wikiPage->getTitle();
@ -43,7 +43,9 @@ class Extension {
if ( in_array( $title->getNamespace(), $wgLinkTitlesNamespaces )) {
$text = $content->getContentHandler()->serializeContent( $content );
if ( !\MagicWord::get( 'MAG_LINKTITLES_NOAUTOLINKS' )->match( $text ) ) {
$newText = Linker::linkContent( $title, $text );
$config = new Config();
$linker = new Linker( $config );
$newText = $linker->linkContent( $title, $text );
if ( $newText != $text ) {
$content = $content->getContentHandler()->unserializeContent( $newText );
}
@ -53,39 +55,43 @@ class Extension {
return true;
}
/// Event handler that is hooked to the InternalParseBeforeLinks event.
/// @param Parser $parser Parser that raised the event.
/// @param $text Preprocessed text of the page.
/*
* Event handler that is hooked to the InternalParseBeforeLinks event.
* @param Parser $parser Parser that raised the event.
* @param $text Preprocessed text of the page
*/
public static function onInternalParseBeforeLinks( \Parser &$parser, &$text ) {
global $wgLinkTitlesParseOnRender;
if (!$wgLinkTitlesParseOnRender) return true;
global $wgLinkTitlesNamespaces;
$config = new Config();
if (!$config->parseOnRender) return true;
$title = $parser->getTitle();
// If the page contains the magic word '__NOAUTOLINKS__', do not parse it.
// Only process if page is in one of our namespaces we want to link
if ( !\MagicWord::get( 'MAG_LINKTITLES_NOAUTOLINKS' )->match( $text ) &&
in_array( $title->getNamespace(), $wgLinkTitlesNamespaces ) ) {
$text = Linker::linkContent( $title, $text );
in_array( $title->getNamespace(), $config->nameSpaces ) ) {
$linker = new Linker( $config );
$text = $linker->linkContent( $title, $text );
}
return true;
}
/// Automatically processes a single page, given a $title Title object.
/// This function is called by the SpecialLinkTitles class and the
/// LinkTitlesJob class.
/// @param Title $title Title object.
/// @param RequestContext $context Current request context.
/// If in doubt, call MediaWiki's `RequestContext::getMain()`
/// to obtain such an object.
/// @returns boolean True if the page exists, false if the page does not exist
/*
* Automatically processes a single page, given a $title Title object.
* This function is called by the SpecialLinkTitles class and the
* LinkTitlesJob class.
* @param Title $title Title object.
* @param RequestContext $context Current request context. If in doubt, call MediaWiki's `RequestContext::getMain()` to obtain such an object.
* @returns bool True if the page exists, false if the page does not exist
*/
public static function processPage( \Title $title, \RequestContext $context ) {
self::ltLog('Processing '. $title->getPrefixedText());
$page = \WikiPage::factory($title);
$content = $page->getContent();
if ( $content != null ) {
$text = $content->getContentHandler()->serializeContent($content);
$newText = Linker::linkContent($title, $text);
$config = new Config();
$linker = new Linker( $config );
$newText = $linker->linkContent($title, $text);
if ( $text != $newText ) {
$content = $content->getContentHandler()->unserializeContent( $newText );
$page->doEditContent(
@ -130,7 +136,9 @@ class Extension {
/// by simply returning the text between the tags (if any).
/// See https://www.mediawiki.org/wiki/Manual:Tag_extensions#How_do_I_render_wikitext_in_my_extension.3F
public static function doAutolinksTag( $input, array $args, \Parser $parser, \PPFrame $frame ) {
$withLinks = Linker::linkContent( $parser->getTitle(), $input );
$config = new Config();
$linker = new Linker( $config );
$withLinks = $linker->linkContent( $parser->getTitle(), $input );
$output = $parser->recursiveTagParse( $withLinks, $frame );
return $output;
}

142
includes/Linker.php
View File

@ -24,56 +24,66 @@
namespace LinkTitles;
/**
* Provides entry points for the extension.
* Performs the actual linking of content to existing pages.
*/
class Linker {
/// A Title object for the page that is being parsed.
private static $currentTitle;
private $currentTitle;
/// A Title object for the target page currently being examined.
private static $targetTitle;
private $targetTitle;
// The TitleValue object of the target page
private static $targetTitleValue;
private $targetTitleValue;
/// The content object for the currently processed target page.
/// This variable is necessary to be able to prevent loading the target
/// content twice.
private static $targetContent;
private $targetContent;
/// Holds the page title of the currently processed target page
/// as a string.
private static $targetTitleText;
private $targetTitleText;
public static $ltConsoleOutput;
public static $ltConsoleOutputDebug;
/**
* LinkTitles configuration.
*
* @var Config $config
*/
public $config;
/// Core function of the extension, performs the actual parsing of the content.
/// @param Parser $parser Parser instance for the current page
/// @param $text String that holds the article content
/// @returns string: parsed text with links added if needed
public static function linkContent( $title, &$text ) {
/**
* Constructs a new instance of the Linker class.
*
* @param Config $config LinkTitles configuration object.
*/
public function __construct( Config &$config ) {
$this->config = $config;
}
// Configuration variables need to be defined here as globals.
global $wgLinkTitlesFirstOnly;
global $wgLinkTitlesSmartMode;
global $wgCapitalLinks;
/*
* Core function of the extension, performs the actual parsing of the content.
*
* @param Parser $parser Parser instance for the current page
* @param String $text String that holds the article content
* @returns String String with links to target pages
*/
public function linkContent( \Title &$title, &$text ) {
( $wgLinkTitlesFirstOnly ) ? $limit = 1 : $limit = -1;
( $this->config->firstOnly ) ? $limit = 1 : $limit = -1;
$limitReached = false;
self::$currentTitle = $title;
$this->currentTitle = $title;
$newText = $text;
$config = new Config();
$delimiters = Delimiters::default( $config );
$targets = Targets::default( $title, $config );
$delimiters = Delimiters::default( $this->config );
$targets = Targets::default( $title, $this->config );
// Iterate through the page titles
foreach( $targets->queryResult as $row ) {
self::newTarget( $row->page_namespace, $row->page_title );
$this->newTarget( $row->page_namespace, $row->page_title );
// Don't link current page
if ( self::$targetTitle->equals( self::$currentTitle ) ) { continue; }
if ( $this->targetTitle->equals( $this->currentTitle ) ) { continue; }
// split the page content by [[...]] groups
// credits to inhan @ StackOverflow for suggesting preg_split
@ -82,17 +92,17 @@ class Linker {
// Escape certain special characters in the page title to prevent
// regexp compilation errors
self::$targetTitleText = self::$targetTitle->getText();
$quotedTitle = preg_quote( self::$targetTitleText, '/' );
$this->targetTitleText = $this->targetTitle->getText();
$quotedTitle = preg_quote( $this->targetTitleText, '/' );
self::ltDebugLog( 'TargetTitle='. self::$targetTitleText, 'private' );
self::ltDebugLog( 'TargetTitleQuoted='. $quotedTitle, 'private' );
$this->ltDebugLog( 'TargetTitle='. $this->targetTitleText, 'private' );
$this->ltDebugLog( 'TargetTitleQuoted='. $quotedTitle, 'private' );
// Depending on the global configuration setting $wgCapitalLinks,
// the title has to be searched for either in a strictly case-sensitive
// way, or in a 'fuzzy' way where the first letter of the title may
// be either case.
if ( $config->capitalLinks && ( $quotedTitle[0] != '\\' )) {
if ( $this->config->capitalLinks && ( $quotedTitle[0] != '\\' )) {
$searchTerm = '((?i)' . $quotedTitle[0] . '(?-i)' .
substr($quotedTitle, 1) . ')';
} else {
@ -103,8 +113,9 @@ class Linker {
for ( $i = 0; $i < count( $arr ); $i+=2 ) {
// even indexes will point to text that is not enclosed by brackets
$arr[$i] = preg_replace_callback( $regex,
'LinkTitles\Linker::simpleModeCallback', $arr[$i], $limit, $count );
if ( $config->firstOnly && ( $count > 0 ) ) {
array( $this, 'simpleModeCallback'),
$arr[$i], $limit, $count );
if ( $this->config->firstOnly && ( $count > 0 ) ) {
$limitReached = true;
break;
};
@ -114,16 +125,17 @@ class Linker {
// If smart mode is turned on, the extension will perform a second
// pass on the page and add links with aliases where the case does
// not match.
if ( $config->smartMode && !$limitReached ) {
if ( $this->config->smartMode && !$limitReached ) {
$arr = preg_split( $delimiters->splitter, $newText, -1, PREG_SPLIT_DELIM_CAPTURE );
for ( $i = 0; $i < count( $arr ); $i+=2 ) {
// even indexes will point to text that is not enclosed by brackets
$arr[$i] = preg_replace_callback( '/(?<![\:\.\@\/\?\&])' .
$delimiters->wordStart . '(' . $quotedTitle . ')' .
$delimiters->wordEnd . '/iS', 'LinkTitles\Linker::smartModeCallback',
$delimiters->wordEnd . '/iS',
array( $this, 'smartModeCallback'),
$arr[$i], $limit, $count );
if ( $config->firstOnly && ( $count > 0 )) {
if ( $this->config->firstOnly && ( $count > 0 )) {
break;
};
};
@ -134,9 +146,9 @@ class Linker {
}
// Build an anonymous callback function to be used in simple mode.
private static function simpleModeCallback( array $matches ) {
if ( self::checkTargetPage() ) {
self::ltLog( "Linking '$matches[0]' to '" . self::$targetTitle . "'" );
private function simpleModeCallback( array $matches ) {
if ( $this->checkTargetPage() ) {
$this->ltLog( "Linking '$matches[0]' to '" . $this->targetTitle . "'" );
return '[[' . $matches[0] . ']]';
}
else
@ -152,22 +164,21 @@ class Linker {
// If $wgCapitalLinks is set to true, the case of the first
// letter is ignored by MediaWiki and we don't need to build a
// piped link if only the case of the first letter is different.
private static function smartModeCallback( array $matches ) {
global $wgCapitalLinks;
private function smartModeCallback( array $matches ) {
if ( $wgCapitalLinks ) {
if ( $this->config->capitalLinks ) {
// With $wgCapitalLinks set to true we have a slightly more
// complicated version of the callback than if it were false;
// we need to ignore the first letter of the page titles, as
// it does not matter for linking.
if ( self::checkTargetPage() ) {
self::ltLog( "Linking (smart) '$matches[0]' to '" . self::$targetTitle . "'" );
if ( strcmp(substr(self::$targetTitleText, 1), substr($matches[0], 1)) == 0 ) {
if ( $this->checkTargetPage() ) {
$this->ltLog( "Linking (smart) '$matches[0]' to '" . $this->targetTitle . "'" );
if ( strcmp(substr($this->targetTitleText, 1), substr($matches[0], 1)) == 0 ) {
// Case-sensitive match: no need to bulid piped link.
return '[[' . $matches[0] . ']]';
} else {
// Case-insensitive match: build piped link.
return '[[' . self::$targetTitleText . '|' . $matches[0] . ']]';
return '[[' . $this->targetTitleText . '|' . $matches[0] . ']]';
}
}
else
@ -177,14 +188,14 @@ class Linker {
} else {
// If $wgCapitalLinks is false, we can use the simple variant
// of the callback function.
if ( self::checkTargetPage() ) {
self::ltLog( "Linking (smart) '$matches[0]' to '" . self::$targetTitle . "'" );
if ( strcmp(self::$targetTitleText, $matches[0]) == 0 ) {
if ( $this->checkTargetPage() ) {
$this->ltLog( "Linking (smart) '$matches[0]' to '" . $this->targetTitle . "'" );
if ( strcmp($this->targetTitleText, $matches[0]) == 0 ) {
// Case-sensitive match: no need to bulid piped link.
return '[[' . $matches[0] . ']]';
} else {
// Case-insensitive match: build piped link.
return '[[' . self::$targetTitleText . '|' . $matches[0] . ']]';
return '[[' . $this->targetTitleText . '|' . $matches[0] . ']]';
}
}
else
@ -195,12 +206,12 @@ class Linker {
}
/// Sets member variables for the current target page.
private static function newTarget( $ns, $title ) {
self::$targetTitle = \Title::makeTitleSafe( $ns, $title );
self::ltDebugLog( 'newtarget='. self::$targetTitle->getText(), "private" );
self::$targetTitleValue = self::$targetTitle->getTitleValue();
self::ltDebugLog( 'altTarget='. self::$targetTitleValue->getText(), "private" );
self::$targetContent = null;
private function newTarget( $ns, $title ) {
$this->targetTitle = \Title::makeTitleSafe( $ns, $title );
$this->ltDebugLog( 'newtarget='. $this->targetTitle->getText(), "private" );
$this->targetTitleValue = $this->targetTitle->getTitleValue();
$this->ltDebugLog( 'altTarget='. $this->targetTitleValue->getText(), "private" );
$this->targetContent = null;
}
/// Returns the content of the current target page.
@ -209,12 +220,11 @@ class Linker {
/// database only when needed.
/// @note It is absolutely necessary that the newTarget()
/// function is called for every new page.
private static function getTargetContent() {
private function getTargetContent() {
if ( ! isset( $targetContent ) ) {
self::$targetContent = \WikiPage::factory(
self::$targetTitle)->getContent();
$this->targetContent = \WikiPage::factory( $this->targetTitle )->getContent();
};
return self::$targetContent;
return $this->targetContent;
}
/// Examines the current target page. Returns true if it may be linked;
@ -223,7 +233,7 @@ class Linker {
/// and whether the target page is a redirect or contains the
/// __NOAUTOLINKTARGET__ magic word.
/// @returns boolean
private static function checkTargetPage() {
private function checkTargetPage() {
global $wgLinkTitlesEnableNoTargetMagicWord;
global $wgLinkTitlesCheckRedirect;
@ -231,8 +241,8 @@ class Linker {
// indeed redirect to the current page, return the page title as-is
// (unlinked).
if ( $wgLinkTitlesCheckRedirect ) {
$redirectTitle = self::getTargetContent()->getUltimateRedirectTarget();
if ( $redirectTitle && $redirectTitle->equals(self::$currentTitle) ) {
$redirectTitle = $this->getTargetContent()->getUltimateRedirectTarget();
if ( $redirectTitle && $redirectTitle->equals($this->currentTitle) ) {
return false;
}
};
@ -241,7 +251,7 @@ class Linker {
// page does indeed contain this magic word, return the page title
// as-is (unlinked).
if ( $wgLinkTitlesEnableNoTargetMagicWord ) {
if ( self::getTargetContent()->matchMagicWord(
if ( $this->getTargetContent()->matchMagicWord(
\MagicWord::get('MAG_LINKTITLES_NOTARGET') ) ) {
return false;
}
@ -250,16 +260,16 @@ class Linker {
}
/// Local Debugging output function which can send output to console as well
public static function ltDebugLog($text) {
if ( self::$ltConsoleOutputDebug ) {
public function ltDebugLog($text) {
if ( $this->config->enableDebugConsoleOutput ) {
print $text . "\n";
}
wfDebugLog( 'LinkTitles', $text , 'private' );
}
/// Local Logging output function which can send output to console as well
public static function ltLog($text) {
if (self::$ltConsoleOutput) {
public function ltLog($text) {
if ( $this->config->enableConsoleOutput) {
print $text . "\n";
}
wfDebugLog( 'LinkTitles', $text , 'private' );