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

Refactor Linker class, add Target class.

Browse Source
This commit is contained in:
Daniel Kraus
2017-08-27 22:18:22 +02:00
parent 5f7c96459d
commit 35b174771e
5 changed files with 328 additions and 200 deletions
Download Patch File Download Diff File Expand all files Collapse all files

245
includes/Linker.php
View File

@ -27,24 +27,6 @@ namespace LinkTitles;
* Performs the actual linking of content to existing pages.
*/
class Linker {
/// A Title object for the page that is being parsed.
private $currentTitle;
/// A Title object for the target page currently being examined.
private $targetTitle;
// The TitleValue object of the target page
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 $targetContent;
/// Holds the page title of the currently processed target page
/// as a string.
private $targetTitleText;
/**
* LinkTitles configuration.
*
@ -52,6 +34,17 @@ class Linker {
*/
public $config;
/**
* The string representation of the title object for the potential target page
* that is currently being processed.
*
* This is an instance variable (rather than a local method variable) so it
* can be accessed in the preg_replace_callback callbacks.
*
* @var String $targetTitleString
*/
private $targetTitleText;
/**
* Constructs a new instance of the Linker class.
*
@ -61,58 +54,50 @@ class Linker {
$this->config = $config;
}
/*
/**
* Core function of the extension, performs the actual parsing of the content.
*
* @param Parser $parser Parser instance for the current page
* This method receives a Title object and the string representation of the
* source page. It does not work on a WikiPage object directly because the
* callbacks in the Extension class do not always get a WikiPage object in the
* first place.
*
* @param \Title &$title Title object for the current page.
* @param String $text String that holds the article content
* @returns String String with links to target pages
* @return String with links to target pages
*/
public function linkContent( \Title &$title, &$text ) {
( $this->config->firstOnly ) ? $limit = 1 : $limit = -1;
$limitReached = false;
$this->currentTitle = $title;
$newText = $text;
$delimiters = Delimiters::default( $this->config );
$targets = Targets::default( $title, $this->config );
// Iterate through the page titles
// Iterate through the target page titles
foreach( $targets->queryResult as $row ) {
$this->newTarget( $row->page_namespace, $row->page_title );
$target = new Target( $row->page_namespace, $row->page_title, $this->config );
// Don't link current page
if ( $this->targetTitle->equals( $this->currentTitle ) ) { continue; }
// split the page content by [[...]] groups
// credits to inhan @ StackOverflow for suggesting preg_split
// see http://stackoverflow.com/questions/10672286
$arr = preg_split( $delimiters->splitter, $newText, -1, PREG_SPLIT_DELIM_CAPTURE );
// Escape certain special characters in the page title to prevent
// regexp compilation errors
$this->targetTitleText = $this->targetTitle->getText();
$quotedTitle = preg_quote( $this->targetTitleText, '/' );
$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 ( $this->config->capitalLinks && ( $quotedTitle[0] != '\\' )) {
$searchTerm = '((?i)' . $quotedTitle[0] . '(?-i)' .
substr($quotedTitle, 1) . ')';
} else {
$searchTerm = '(' . $quotedTitle . ')';
// Don't link current page and don't link if the target page redirects
// to the current page or has the __NOAUTOLINKTARGET__ magic word
// (as required by the actual LinkTitles configuration).
if ( $target->isSameTitle( $title ) || !$target->mayLinkTo( $title ) ) {
continue;
}
$regex = '/(?<![\:\.\@\/\?\&])' . $delimiters->wordStart . $searchTerm . $delimiters->wordEnd . '/S';
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,
// Split the page content by non-linkable sections.
// Credits to inhan @ StackOverflow for suggesting preg_split.
// See http://stackoverflow.com/questions/10672286
$arr = $delimiters->split( $newText );
$count = 0;
// Cache the target title text for the regex callbacks
$this->targetTitleText = $target->getTitleText();
// Even indexes will point to sections of the text that may be linked
for ( $i = 0; $i < count( $arr ); $i += 2 ) {
$arr[$i] = preg_replace_callback( $target->getCaseSensitiveRegex(),
array( $this, 'simpleModeCallback'),
$arr[$i], $limit, $count );
if ( $this->config->firstOnly && ( $count > 0 ) ) {
@ -126,13 +111,14 @@ class Linker {
// pass on the page and add links with aliases where the case does
// not match.
if ( $this->config->smartMode && !$limitReached ) {
$arr = preg_split( $delimiters->splitter, $newText, -1, PREG_SPLIT_DELIM_CAPTURE );
if ( $count > 0 ) {
// Split the text again because it was changed in the first pass.
$arr = $delimiters->split( $newText );
}
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',
$arr[$i] = preg_replace_callback( $target->getCaseInsensitiveRegex(),
array( $this, 'smartModeCallback'),
$arr[$i], $limit, $count );
if ( $this->config->firstOnly && ( $count > 0 )) {
@ -142,138 +128,57 @@ class Linker {
$newText = implode( '', $arr );
} // $wgLinkTitlesSmartMode
}; // foreach $res as $row
return $newText;
}
// Build an anonymous callback function to be used in simple mode.
/**
* Callback for preg_replace_callback in simple mode.
*
* @param array $matches Matches provided by preg_replace_callback
* @return string Target page title with or without link markup
*/
private function simpleModeCallback( array $matches ) {
if ( $this->checkTargetPage() ) {
$this->ltLog( "Linking '$matches[0]' to '" . $this->targetTitle . "'" );
return '[[' . $matches[0] . ']]';
}
else
{
return $matches[0];
}
return '[[' . $matches[0] . ']]';
}
// Callback function for use with preg_replace_callback.
// This essentially performs a case-sensitive comparison of the
// current page title and the occurrence found on the page; if
// the cases do not match, it builds an aliased (piped) link.
// 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.
/**
* Callback function for use with preg_replace_callback.
* This essentially performs a case-sensitive comparison of the
* current page title and the occurrence found on the page; if
* the cases do not match, it builds an aliased (piped) link.
* 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.
*
* @param array $matches Matches provided by preg_replace_callback
* @return string Target page title with or without link markup
*/
private function smartModeCallback( array $matches ) {
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 ( $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 '[[' . $this->targetTitleText . '|' . $matches[0] . ']]';
}
}
else
{
return $matches[0];
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 '[[' . $this->targetTitleText . '|' . $matches[ 0 ] . ']]';
}
} else {
// If $wgCapitalLinks is false, we can use the simple variant
// of the callback function.
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 '[[' . $this->targetTitleText . '|' . $matches[0] . ']]';
}
}
else
{
return $matches[0];
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 '[[' . $this->targetTitleText . '|' . $matches[ 0 ] . ']]';
}
}
}
/// Sets member variables for the current target page.
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.
/// This function serves to be used in preg_replace_callback callback
/// functions, in order to load the target page content from the
/// database only when needed.
/// @note It is absolutely necessary that the newTarget()
/// function is called for every new page.
private function getTargetContent() {
if ( ! isset( $targetContent ) ) {
$this->targetContent = \WikiPage::factory( $this->targetTitle )->getContent();
};
return $this->targetContent;
}
/// Examines the current target page. Returns true if it may be linked;
/// false if not. This depends on the settings
/// $wgLinkTitlesCheckRedirect and $wgLinkTitlesEnableNoTargetMagicWord
/// and whether the target page is a redirect or contains the
/// __NOAUTOLINKTARGET__ magic word.
/// @returns boolean
private function checkTargetPage() {
global $wgLinkTitlesEnableNoTargetMagicWord;
global $wgLinkTitlesCheckRedirect;
// If checking for redirects is enabled and the target page does
// indeed redirect to the current page, return the page title as-is
// (unlinked).
if ( $wgLinkTitlesCheckRedirect ) {
$redirectTitle = $this->getTargetContent()->getUltimateRedirectTarget();
if ( $redirectTitle && $redirectTitle->equals($this->currentTitle) ) {
return false;
}
};
// If the magic word __NOAUTOLINKTARGET__ is enabled and the target
// page does indeed contain this magic word, return the page title
// as-is (unlinked).
if ( $wgLinkTitlesEnableNoTargetMagicWord ) {
if ( $this->getTargetContent()->matchMagicWord(
\MagicWord::get('MAG_LINKTITLES_NOTARGET') ) ) {
return false;
}
};
return true;
}
/// Local Debugging output function which can send output to console as well
public function ltDebugLog($text) {
if ( $this->config->enableDebugConsoleOutput ) {
print $text . "\n";
}
wfDebugLog( 'LinkTitles', $text , 'private' );
}
/// Local Logging output function which can send output to console as well
public function ltLog($text) {
if ( $this->config->enableConsoleOutput) {
print $text . "\n";
}
wfDebugLog( 'LinkTitles', $text , 'private' );
}
}
// vim: ts=2:sw=2:noet:comments^=\:///