dumphtml.inc

来自「php 开发的内容管理系统」· INC 代码 · 共 651 行 · 第 1/2 页

<?php
/**
 * @package MediaWiki
 * @subpackage Maintenance
 */

define( 'REPORTING_INTERVAL', 10 );

require_once( 'includes/ImagePage.php' );
require_once( 'includes/CategoryPage.php' );
require_once( 'includes/RawPage.php' );

class DumpHTML {
	# Destination directory
	var $dest;

	# Show interlanguage links?
	var $interwiki = true;

	# Depth of HTML directory tree
	var $depth = 3;

	# Directory that commons images are copied into
	var $sharedStaticPath;

	# Relative path to image directory
	var $imageRel = 'upload';

	# Copy commons images instead of symlinking
	var $forceCopy = false;

	# Make links assuming the script path is in the same directory as
	# the destination
	var $alternateScriptPath = false;

	# Original values of various globals
	var $oldArticlePath = false, $oldCopyrightIcon = false;

	# Has setupGlobals been called?
	var $setupDone = false;

	# List of raw pages used in the current article
	var $rawPages;
	
	# Skin to use
	var $skin = 'dumphtml';

	function DumpHTML( $settings ) {
		foreach ( $settings as $var => $value ) {
			$this->$var = $value;
		}
	}

	/**
	 * Write a set of articles specified by start and end page_id
	 * Skip categories and images, they will be done separately
	 */
	function doArticles( $start, $end = false ) {
		$fname = 'DumpHTML::doArticles';

		$this->setupGlobals();

		if ( $end === false ) {
			$dbr =& wfGetDB( DB_SLAVE );
			$end = $dbr->selectField( 'page', 'max(page_id)', false, $fname );
		}

		$mainPageObj = Title::newMainPage();
		$mainPage = $mainPageObj->getPrefixedDBkey();


		for ($id = $start; $id <= $end; $id++) {
			wfWaitForSlaves( 20 );
			if ( !($id % REPORTING_INTERVAL) ) {
				print "Processing ID: $id\r";
			}
			if ( !($id % (REPORTING_INTERVAL*10) ) ) {
				print "\n";
			}
			$title = Title::newFromID( $id );
			if ( $title ) {
				$ns = $title->getNamespace() ;
				if ( $ns != NS_CATEGORY && $title->getPrefixedDBkey() != $mainPage ) {
					$this->doArticle( $title );
				}
			}
		}
		print "\n";
	}

	function doSpecials() {
		$this->doMainPage();

		$this->setupGlobals();
		print "Special:Categories...";
		$this->doArticle( Title::makeTitle( NS_SPECIAL, 'Categories' ) );
		print "\n";
	}

	/** Write the main page as index.html */
	function doMainPage() {

		print "Making index.html  ";

		// Set up globals with no ../../.. in the link URLs
		$this->setupGlobals( 0 );

		$title = Title::newMainPage();
		$text = $this->getArticleHTML( $title );
		$file = fopen( "{$this->dest}/index.html", "w" );
		if ( !$file ) {
			print "\nCan't open index.html for writing\n";
			return false;
		}
		fwrite( $file, $text );
		fclose( $file );
		print "\n";
	}

	function doImageDescriptions() {
		global $wgSharedUploadDirectory;

		$fname = 'DumpHTML::doImageDescriptions';

		$this->setupGlobals();

		/**
		 * Dump image description pages that don't have an associated article, but do
		 * have a local image
		 */
		$dbr =& wfGetDB( DB_SLAVE );
		extract( $dbr->tableNames( 'image', 'page' ) );
		$res = $dbr->select( 'image', array( 'img_name' ), false, $fname );

		$i = 0;
		print "Writing image description pages for local images\n";
		$num = $dbr->numRows( $res );
		while ( $row = $dbr->fetchObject( $res ) ) {
			wfWaitForSlaves( 10 );
			if ( !( ++$i % REPORTING_INTERVAL ) ) {
				print "Done $i of $num\r";
			}
			$title = Title::makeTitle( NS_IMAGE, $row->img_name );
			if ( $title->getArticleID() ) {
				// Already done by dumpHTML
				continue;
			}
			$this->doArticle( $title );
		}
		print "\n";

		/**
		 * Dump images which only have a real description page on commons
		 */
		print "Writing description pages for commons images\n";
		$i = 0;
		for ( $hash = 0; $hash < 256; $hash++ ) {
			$dir = sprintf( "%01x/%02x", intval( $hash / 16 ), $hash );
			$paths = array_merge( glob( "{$this->sharedStaticPath}/$dir/*" ),
				glob( "{$this->sharedStaticPath}/thumb/$dir/*" ) );

			foreach ( $paths as $path ) {
				$file = basename( $path );
				if ( !(++$i % REPORTING_INTERVAL ) ) {
					print "$i\r";
				}

				$title = Title::makeTitle( NS_IMAGE, $file );
				$this->doArticle( $title );
			}
		}
		print "\n";
	}

	function doCategories() {
		$fname = 'DumpHTML::doCategories';
		$this->setupGlobals();

		$dbr =& wfGetDB( DB_SLAVE );
		print "Selecting categories...";
		$sql = 'SELECT DISTINCT cl_to FROM ' . $dbr->tableName( 'categorylinks' );
		$res = $dbr->query( $sql, $fname );

		print "\nWriting " . $dbr->numRows( $res ).  " category pages\n";
		$i = 0;
		while ( $row = $dbr->fetchObject( $res ) ) {
			wfWaitForSlaves( 10 );
			if ( !(++$i % REPORTING_INTERVAL ) ) {
				print "$i\r";
			}
			$title = Title::makeTitle( NS_CATEGORY, $row->cl_to );
			$this->doArticle( $title );
		}
		print "\n";
	}

	function doRedirects() {
		print "Doing redirects...\n";
		$fname = 'DumpHTML::doRedirects';
		$this->setupGlobals();
		$dbr =& wfGetDB( DB_SLAVE );

		$res = $dbr->select( 'page', array( 'page_namespace', 'page_title' ),
			array( 'page_is_redirect' => 1 ), $fname );
		$num = $dbr->numRows( $res );
		print "$num redirects to do...\n";
		$i = 0;
		while ( $row = $dbr->fetchObject( $res ) ) {
			$title = Title::makeTitle( $row->page_namespace, $row->page_title );
                        if ( !(++$i % (REPORTING_INTERVAL*10) ) ) {
                                print "Done $i of $num\n";
                        }
			$this->doArticle( $title );
		}
	}

	/** Write an article specified by title */
	function doArticle( $title ) {
		global $wgTitle, $wgSharedUploadPath, $wgSharedUploadDirectory;
		global $wgUploadDirectory;

		$this->rawPages = array();
		$text = $this->getArticleHTML( $title );

		if ( $text === false ) {
			return;
		}

		# Parse the XHTML to find the images
		$images = $this->findImages( $text );
		$this->copyImages( $images );

		# Write to file
		$this->writeArticle( $title, $text );

		# Do raw pages
		wfMkdirParents( "{$this->dest}/raw", 0755 );
		foreach( $this->rawPages as $record ) {
			list( $file, $title, $params ) = $record;

			$path = "{$this->dest}/raw/$file";
			if ( !file_exists( $path ) ) {
				$article = new Article( $title );
				$request = new FauxRequest( $params );
				$rp = new RawPage( $article, $request );
				$text = $rp->getRawText();

				print "Writing $file\n";
				$file = fopen( $path, 'w' );
				if ( !$file ) {
					print("Can't open file $fullName for writing\n");
					continue;
				}
				fwrite( $file, $text );
				fclose( $file );
			}
		}
	}

	/** Write the given text to the file identified by the given title object */
	function writeArticle( &$title, $text ) {
		$filename = $this->getHashedFilename( $title );
		$fullName = "{$this->dest}/$filename";
		$fullDir = dirname( $fullName );

		wfMkdirParents( $fullDir, 0755 );

		$file = fopen( $fullName, 'w' );
		if ( !$file ) {
			print("Can't open file $fullName for writing\n");
			return;
		}

		fwrite( $file, $text );
		fclose( $file );
	}

	/** Set up globals required for parsing */
	function setupGlobals( $currentDepth = NULL ) {
		global $wgUser, $wgTitle, $wgStylePath, $wgArticlePath, $wgMathPath;
		global $wgUploadPath, $wgLogo, $wgMaxCredits, $wgSharedUploadPath;
		global $wgHideInterlanguageLinks, $wgUploadDirectory, $wgThumbnailScriptPath;
		global $wgSharedThumbnailScriptPath, $wgEnableParserCache, $wgHooks, $wgServer;
		global $wgRightsUrl, $wgRightsText, $wgCopyrightIcon;

		static $oldLogo = NULL;

		if ( !$this->setupDone ) {
			$wgHooks['GetLocalURL'][] =& $this;
			$wgHooks['GetFullURL'][] =& $this;
			$this->oldArticlePath = $wgServer . $wgArticlePath;
		}

		if ( is_null( $currentDepth ) ) {
			$currentDepth = $this->depth;
		}

		if ( $this->alternateScriptPath ) {
			if ( $currentDepth == 0 ) {
				$wgScriptPath = '.';
			} else {
				$wgScriptPath = '..' . str_repeat( '/..', $currentDepth - 1 );
			}
		} else {
			$wgScriptPath = '..' . str_repeat( '/..', $currentDepth );
		}

		$wgArticlePath = str_repeat( '../', $currentDepth ) . '$1';

		# Logo image
		# Allow for repeated setup
		if ( !is_null( $oldLogo ) ) {
			$wgLogo = $oldLogo;
		} else {
			$oldLogo = $wgLogo;
		}

		if ( strpos( $wgLogo, $wgUploadPath ) === 0 ) {
			# If it's in the upload directory, rewrite it to the new upload directory
			$wgLogo = "$wgScriptPath/{$this->imageRel}/" . substr( $wgLogo, strlen( $wgUploadPath ) + 1 );
		} elseif ( $wgLogo{0} == '/' ) {
			# This is basically heuristic
			# Rewrite an absolute logo path to one relative to the the script path
			$wgLogo = $wgScriptPath . $wgLogo;
		}