Fix #47 - update Readability class (thx to http://code.fivefilters.org/php-readability/)

author: nicosomb <nicolas@loeuillet.org> 2013-04-23 15:57:23 +0200
committer: nicosomb <nicolas@loeuillet.org> 2013-04-23 15:57:23 +0200
commit: 0ace6cab0b934b7b8e4d4de4acca16294b7b12df (patch)
tree: db5a4314ff13882753bc947ea0f9ee2168179f23 /inc
parent: c1c9f252f7f0133903dc4143a0db8ed8104f8733 (diff)
download: wallabag-0ace6cab0b934b7b8e4d4de4acca16294b7b12df.tar.gz
wallabag-0ace6cab0b934b7b8e4d4de4acca16294b7b12df.tar.zst
wallabag-0ace6cab0b934b7b8e4d4de4acca16294b7b12df.zip
1 files changed, 1099 insertions, 1066 deletions
diff --git a/inc/Readability.php b/inc/Readability.php
index 2ce90f6d..d28d28f9 100644
--- a/inc/Readability.php
+++ b/inc/Readability.php
@@ -1,7 +1,9 @@
 <?php
-/** 
+/**
 * Arc90's Readability ported to PHP for FiveFilters.org
 * Based on readability.js version 1.7.1 (without multi-page support)
+* Updated to allow HTML5 parsing with html5lib
+* Updated with lightClean mode to preserve more images and youtube/vimeo/viddler embeds
 * ------------------------------------------------------
 * Original URL: http://lab.arc90.com/experiments/readability/js/readability.js
 * Arc90's project URL: http://lab.arc90.com/experiments/readability/
@@ -10,35 +12,35 @@
 * More information: http://fivefilters.org/content-only/
 * License: Apache License, Version 2.0
 * Requires: PHP5
-* Date: 2011-07-22
+* Date: 2012-09-19
-* 
+*
 * Differences between the PHP port and the original
 * ------------------------------------------------------
-* Arc90's Readability is designed to run in the browser. It works on the DOM 
+* Arc90's Readability is designed to run in the browser. It works on the DOM
-* tree (the parsed HTML) after the page's CSS styles have been applied and 
+* tree (the parsed HTML) after the page's CSS styles have been applied and
-* Javascript code executed. This PHP port does not run inside a browser. 
+* Javascript code executed. This PHP port does not run inside a browser.
-* We use PHP's ability to parse HTML to build our DOM tree, but we cannot 
+* We use PHP's ability to parse HTML to build our DOM tree, but we cannot
-* rely on CSS or Javascript support. As such, the results will not always 
+* rely on CSS or Javascript support. As such, the results will not always
-* match Arc90's Readability. (For example, if a web page contains CSS style 
+* match Arc90's Readability. (For example, if a web page contains CSS style
-* rules or Javascript code which hide certain HTML elements from display, 
+* rules or Javascript code which hide certain HTML elements from display,
-* Arc90's Readability will dismiss those from consideration but our PHP port, 
+* Arc90's Readability will dismiss those from consideration but our PHP port,
 * unable to understand CSS or Javascript, will not know any better.)
-* 
+*
-* Another significant difference is that the aim of Arc90's Readability is 
+* Another significant difference is that the aim of Arc90's Readability is
-* to re-present the main content block of a given web page so users can 
+* to re-present the main content block of a given web page so users can
-* read it more easily in their browsers. Correct identification, clean up, 
+* read it more easily in their browsers. Correct identification, clean up,
-* and separation of the content block is only a part of this process. 
+* and separation of the content block is only a part of this process.
-* This PHP port is only concerned with this part, it does not include code 
+* This PHP port is only concerned with this part, it does not include code
-* that relates to presentation in the browser - Arc90 already do 
+* that relates to presentation in the browser - Arc90 already do
-* that extremely well, and for PDF output there's FiveFilters.org's 
+* that extremely well, and for PDF output there's FiveFilters.org's
 * PDF Newspaper: http://fivefilters.org/pdf-newspaper/.
-* 
+*
-* Finally, this class contains methods that might be useful for developers 
+* Finally, this class contains methods that might be useful for developers
-* working on HTML document fragments. So without deviating too much from 
+* working on HTML document fragments. So without deviating too much from
-* the original code (which I don't want to do because it makes debugging 
+* the original code (which I don't want to do because it makes debugging
-* and updating more difficult), I've tried to make it a little more 
+* and updating more difficult), I've tried to make it a little more
-* developer friendly. You should be able to use the methods here on 
+* developer friendly. You should be able to use the methods here on
-* existing DOMElement objects without passing an entire HTML document to 
+* existing DOMElement objects without passing an entire HTML document to
 * be parsed.
 */
@@ -46,13 +48,13 @@
 require_once(dirname(__FILE__).'/JSLikeHTMLElement.php');
 // Alternative usage (for testing only!)
-// uncomment the lines below and call Readability.php in your browser 
+// uncomment the lines below and call Readability.php in your browser
 // passing it the URL of the page you'd like content from, e.g.:
 // Readability.php?url=http://medialens.org/alerts/09/090615_the_guardian_climate.php
 /*
 if (!isset($_GET['url']) || $_GET['url'] == '') {
-        die('Please pass a URL to the script. E.g. Readability.php?url=bla.com/story.html');
+    die('Please pass a URL to the script. E.g. Readability.php?url=bla.com/story.html');
 }
 $url = $_GET['url'];
 if (!preg_match('!^https?://!i', $url)) $url = 'http://'.$url;
@@ -62,1043 +64,1074 @@ $r->init();
 echo $r->articleContent->innerHTML;
 */
-        
 class Readability
 {
-        /* constants */
+    public $version = '1.7.1-without-multi-page';
-        const FLAG_STRIP_UNLIKELYS = 1;
+    public $convertLinksToFootnotes = false;
-        const FLAG_WEIGHT_CLASSES = 2;
+    public $revertForcedParagraphElements = true;
-        const FLAG_CLEAN_CONDITIONALLY = 4;
+    public $articleTitle;
+    public $articleContent;
-        public $version = '1.7.1-without-multi-page';
+    public $dom;
-        public $convertLinksToFootnotes = false;
+    public $url = null; // optional - URL where HTML was retrieved
-        public $revertForcedParagraphElements = true;
+    public $debug = false;
-        public $articleTitle;
+    public $lightClean = true; // preserves more content (experimental) added 2012-09-19
-        public $articleContent;
+    protected $body = null; //
-        public $dom;
+    protected $bodyCache = null; // Cache the body HTML in case we need to re-use it later
-        public $url = null; // optional - URL where HTML was retrieved
+    protected $flags = 7; // 1 | 2 | 4;   // Start with all flags set.
-        public $debug = false;
+    protected $success = false; // indicates whether we were able to extract or not
-        protected $body = null; // 
-        protected $bodyCache = null; // Cache the body HTML in case we need to re-use it later
+    /**
-        protected $flags =  self::FLAG_CLEAN_CONDITIONALLY; // 1 | 2 | 4;   // Start with all flags set.
+    * All of the regular expressions in use within readability.
-        protected $success = false; // indicates whether we were able to extract or not
+    * Defined up here so we don't instantiate them repeatedly in loops.
-        
+    **/
-        /**
+    public $regexps = array(
-        * All of the regular expressions in use within readability.
+        'unlikelyCandidates' => '/combx|comment|community|disqus|extra|foot|header|menu|remark|rss|shoutbox|sidebar|sponsor|ad-break|agegate|pagination|pager|popup/i',
-        * Defined up here so we don't instantiate them repeatedly in loops.
+        'okMaybeItsACandidate' => '/and|article|body|column|main|shadow/i',
-        **/
+        'positive' => '/article|body|content|entry|hentry|main|page|attachment|pagination|post|text|blog|story/i',
-        public $regexps = array(
+        'negative' => '/combx|comment|com-|contact|foot|footer|_nav|footnote|masthead|media|meta|outbrain|promo|related|scroll|shoutbox|sidebar|sponsor|shopping|tags|tool|widget/i',
-                'unlikelyCandidates' => '/combx|comment|comments|community|disqus|extra|foot|header|menu|remark|rss|shoutbox|sidebar|sponsor|ad-break|agegate|pagination|pager|popup|tweet|twitter/i',
+        'divToPElements' => '/<(a|blockquote|dl|div|img|ol|p|pre|table|ul)/i',
-                'okMaybeItsACandidate' => '/and|article|body|column|main|shadow/i',
+        'replaceBrs' => '/(<br[^>]*>[ \n\r\t]*){2,}/i',
-                'positive' => '/article|body|content|entry|hentry|main|page|pagination|post|text|blog|story|attachment/i',
+        'replaceFonts' => '/<(\/?)font[^>]*>/i',
-                'negative' => '/combx|comment|comments|com-|contact|foot|footer|footnote|masthead|media|meta|outbrain|promo|related|scroll|shoutbox|sidebar|sponsor|shopping|tags|tool|widget/i',
+        // 'trimRe' => '/^\s+|\s+$/g', // PHP has trim()
-                'divToPElements' => '/<(a|blockquote|dl|div|ol|p|pre|table|ul)/i',
+        'normalize' => '/\s{2,}/',
-                'replaceBrs' => '/(<br[^>]*>[ \n\r\t]*){2,}/i',
+        'killBreaks' => '/(<br\s*\/?>(\s|&nbsp;?)*){1,}/',
-                'replaceFonts' => '/<(\/?)font[^>]*>/i',
+        'video' => '!//(player\.|www\.)?(youtube|vimeo|viddler)\.com!i',
-                // 'trimRe' => '/^\s+|\s+$/g', // PHP has trim()
+        'skipFootnoteLink' => '/^\s*(\[?[a-z0-9]{1,2}\]?|^|edit|citation needed)\s*$/i'
-                'normalize' => '/\s{2,}/',
+    );
-                'killBreaks' => '/(<br\s*\/?>(\s|&nbsp;?)*){1,}/',
-                'video' => '/http:\/\/(www\.)?(youtube|vimeo|dailymotion)\.com/i',
+    /* constants */
-                'skipFootnoteLink' => '/^\s*(\[?[a-z0-9]{1,2}\]?|^|edit|citation needed)\s*$/i'
+    const FLAG_STRIP_UNLIKELYS = 1;
-        );      
+    const FLAG_WEIGHT_CLASSES = 2;
-        
+    const FLAG_CLEAN_CONDITIONALLY = 4;
-        /**
-        * Create instance of Readability
+    /**
-        * @param string UTF-8 encoded string
+    * Create instance of Readability
-        * @param string (optional) URL associated with HTML (used for footnotes)
+    * @param string UTF-8 encoded string
-        */
+    * @param string (optional) URL associated with HTML (used for footnotes)
-        function __construct($html, $url=null)
+    * @param string which parser to use for turning raw HTML into a DOMDocument (either 'libxml' or 'html5lib')
-        {
+    */
-                /* Turn all double br's into p's */
+    function __construct($html, $url=null, $parser='libxml')
-                $html = preg_replace($this->regexps['replaceBrs'], '</p><p>', $html);
+    {
-                $html = preg_replace($this->regexps['replaceFonts'], '<$1span>', $html);
+        $this->url = $url;
-                $html = mb_convert_encoding($html, 'HTML-ENTITIES', "UTF-8");
+        /* Turn all double br's into p's */
-                $this->dom = new DOMDocument();
+        $html = preg_replace($this->regexps['replaceBrs'], '</p><p>', $html);
-                $this->dom->preserveWhiteSpace = false;
+        $html = preg_replace($this->regexps['replaceFonts'], '<$1span>', $html);
-                $this->dom->registerNodeClass('DOMElement', 'JSLikeHTMLElement');
+        $html = mb_convert_encoding($html, 'HTML-ENTITIES', "UTF-8");
-                if (trim($html) == '') $html = '<html></html>';
+        if (trim($html) == '') $html = '<html></html>';
-                @$this->dom->loadHTML($html);
+        if ($parser=='html5lib' && ($this->dom = HTML5_Parser::parse($html))) {
-                $this->url = $url;
+            // all good
-        }
+        } else {
+            $this->dom = new DOMDocument();
-        /**
+            $this->dom->preserveWhiteSpace = false;
-        * Get article title element
+            @$this->dom->loadHTML($html);
-        * @return DOMElement
+        }
-        */
+        $this->dom->registerNodeClass('DOMElement', 'JSLikeHTMLElement');
-        public function getTitle() {
+    }
-                return $this->articleTitle;
-        }
+    /**
-        
+    * Get article title element
-        /**
+    * @return DOMElement
-        * Get article content element
+    */
-        * @return DOMElement
+    public function getTitle() {
-        */
+        return $this->articleTitle;
-        public function getContent() {
+    }
-                return $this->articleContent;
-        }       
+    /**
-        
+    * Get article content element
-        /**
+    * @return DOMElement
-        * Runs readability.
+    */
-        * 
+    public function getContent() {
-        * Workflow:
+        return $this->articleContent;
-        *  1. Prep the document by removing script tags, css, etc.
+    }
-        *  2. Build readability's DOM tree.
-        *  3. Grab the article content from the current dom tree.
+    /**
-        *  4. Replace the current DOM tree with the new one.
+    * Runs readability.
-        *  5. Read peacefully.
+    *
-        *
+    * Workflow:
-        * @return boolean true if we found content, false otherwise
+    *  1. Prep the document by removing script tags, css, etc.
-        **/
+    *  2. Build readability's DOM tree.
-        public function init()
+    *  3. Grab the article content from the current dom tree.
-        {
+    *  4. Replace the current DOM tree with the new one.
-                if (!isset($this->dom->documentElement)) return false;
+    *  5. Read peacefully.
-                $this->removeScripts($this->dom);
+    *
-                //die($this->getInnerHTML($this->dom->documentElement));
+    * @return boolean true if we found content, false otherwise
-                
+    **/
-                // Assume successful outcome
+    public function init()
-                $this->success = true;
+    {
+        if (!isset($this->dom->documentElement)) return false;
-                $bodyElems = $this->dom->getElementsByTagName('body');
+        $this->removeScripts($this->dom);
-                if ($bodyElems->length > 0) {
+        //die($this->getInnerHTML($this->dom->documentElement));
-                        if ($this->bodyCache == null) {
-                                $this->bodyCache = $bodyElems->item(0)->innerHTML;
+        // Assume successful outcome
-                        }
+        $this->success = true;
-                        if ($this->body == null) {
-                                $this->body = $bodyElems->item(0);
+        $bodyElems = $this->dom->getElementsByTagName('body');
-                        }
+        if ($bodyElems->length > 0) {
-                }
+            if ($this->bodyCache == null) {
+                $this->bodyCache = $bodyElems->item(0)->innerHTML;
-                $this->prepDocument();
+            }
-                
+            if ($this->body == null) {
-                //die($this->dom->documentElement->parentNode->nodeType);
+                $this->body = $bodyElems->item(0);
-                //$this->setInnerHTML($this->dom->documentElement, $this->getInnerHTML($this->dom->documentElement));
+            }
-                //die($this->getInnerHTML($this->dom->documentElement));
+        }
-                /* Build readability's DOM tree */
+        $this->prepDocument();
-                $overlay        = $this->dom->createElement('div');
-                $innerDiv       = $this->dom->createElement('div');
+        //die($this->dom->documentElement->parentNode->nodeType);
-                $articleTitle   = $this->getArticleTitle();
+        //$this->setInnerHTML($this->dom->documentElement, $this->getInnerHTML($this->dom->documentElement));
-                $articleContent = $this->grabArticle();
+        //die($this->getInnerHTML($this->dom->documentElement));
-                if (!$articleContent) {
+        /* Build readability's DOM tree */
-                        $this->success = false;
+        $overlay        = $this->dom->createElement('div');
-                        $articleContent = $this->dom->createElement('div');
+        $innerDiv       = $this->dom->createElement('div');
-                        $articleContent->setAttribute('id', 'readability-content');
+        $articleTitle   = $this->getArticleTitle();
-                        $articleContent->innerHTML = '<p>Sorry, Readability was unable to parse this page for content.</p>';            
+        $articleContent = $this->grabArticle();
-                        return $this->success;
-                }
+        if (!$articleContent) {
-                
+            $this->success = false;
-                $overlay->setAttribute('id', 'readOverlay');
+            $articleContent = $this->dom->createElement('div');
-                $innerDiv->setAttribute('id', 'readInner');
+            $articleContent->setAttribute('id', 'readability-content');
+            $articleContent->innerHTML = '<p>Sorry, Readability was unable to parse this page for content.</p>';
-                /* Glue the structure of our document together. */
+        }
-                $innerDiv->appendChild($articleTitle);
-                $innerDiv->appendChild($articleContent);
+        $overlay->setAttribute('id', 'readOverlay');
-                $overlay->appendChild($innerDiv);
+        $innerDiv->setAttribute('id', 'readInner');
-                
-                /* Clear the old HTML, insert the new content. */
+        /* Glue the structure of our document together. */
-                $this->body->innerHTML = '';
+        $innerDiv->appendChild($articleTitle);
-                $this->body->appendChild($overlay);
+        $innerDiv->appendChild($articleContent);
-                //document.body.insertBefore(overlay, document.body.firstChild);
+        $overlay->appendChild($innerDiv);
-                $this->body->removeAttribute('style');
+        /* Clear the old HTML, insert the new content. */
-                $this->postProcessContent($articleContent);
+        $this->body->innerHTML = '';
-                
+        $this->body->appendChild($overlay);
-                // Set title and content instance variables
+        //document.body.insertBefore(overlay, document.body.firstChild);
-                $this->articleTitle = $articleTitle;
+        $this->body->removeAttribute('style');
-                $this->articleContent = $articleContent;
-                
+        $this->postProcessContent($articleContent);
-                return $this->success;
-        }
+        // Set title and content instance variables
-        
+        $this->articleTitle = $articleTitle;
-        /**
+        $this->articleContent = $articleContent;
-        * Debug
-        */
+        return $this->success;
-        protected function dbg($msg) {
+    }
-                if ($this->debug) echo '* ',$msg, '<br />', "\n";
-        }
+    /**
-        
+    * Debug
-        /**
+    */
-        * Run any post-process modifications to article content as necessary.
+    protected function dbg($msg) {
-        *
+        if ($this->debug) echo '* ',$msg, "\n";
-        * @param DOMElement
+    }
-        * @return void
-        */
+    /**
-        public function postProcessContent($articleContent) {
+    * Run any post-process modifications to article content as necessary.
-                if ($this->convertLinksToFootnotes && !preg_match('/wikipedia\.org/', @$this->url)) { 
+    *
-                        $this->addFootnotes($articleContent);
+    * @param DOMElement
-                }
+    * @return void
-        }
+    */
-        
+    public function postProcessContent($articleContent) {
-        /**
+        if ($this->convertLinksToFootnotes && !preg_match('/wikipedia\.org/', @$this->url)) {
-        * Get the article title as an H1.
+            $this->addFootnotes($articleContent);
-        *
+        }
-        * @return DOMElement
+    }
-        */
-        protected function getArticleTitle() {
+    /**
-                $curTitle = '';
+    * Get the article title as an H1.
-                $origTitle = '';
+    *
+    * @return DOMElement
-                try {
+    */
-                        $curTitle = $origTitle = $this->getInnerText($this->dom->getElementsByTagName('title')->item(0));
+    protected function getArticleTitle() {
-                } catch(Exception $e) {}
+        $curTitle = '';
-                
+        $origTitle = '';
-                if (preg_match('/ [\|\-] /', $curTitle))
-                {
+        try {
-                        $curTitle = preg_replace('/(.*)[\|\-] .*/i', '$1', $origTitle);
+            $curTitle = $origTitle = $this->getInnerText($this->dom->getElementsByTagName('title')->item(0));
-                        
+        } catch(Exception $e) {}
-                        if (count(explode(' ', $curTitle)) < 3) {
-                                $curTitle = preg_replace('/[^\|\-]*[\|\-](.*)/i', '$1', $origTitle);
+        if (preg_match('/ [\|\-] /', $curTitle))
-                        }
+        {
-                }
+            $curTitle = preg_replace('/(.*)[\|\-] .*/i', '$1', $origTitle);
-                else if (strpos($curTitle, ': ') !== false)
-                {
+            if (count(explode(' ', $curTitle)) < 3) {
-                        $curTitle = preg_replace('/.*:(.*)/i', '$1', $origTitle);
+                $curTitle = preg_replace('/[^\|\-]*[\|\-](.*)/i', '$1', $origTitle);
+            }
-                        if (count(explode(' ', $curTitle)) < 3) {
+        }
-                                $curTitle = preg_replace('/[^:]*[:](.*)/i','$1', $origTitle);
+        else if (strpos($curTitle, ': ') !== false)
-                        }
+        {
-                }
+            $curTitle = preg_replace('/.*:(.*)/i', '$1', $origTitle);
-                else if(strlen($curTitle) > 150 || strlen($curTitle) < 15)
-                {
+            if (count(explode(' ', $curTitle)) < 3) {
-                        $hOnes = $this->dom->getElementsByTagName('h1');
+                $curTitle = preg_replace('/[^:]*[:](.*)/i','$1', $origTitle);
-                        if($hOnes->length == 1)
+            }
-                        {
+        }
-                                $curTitle = $this->getInnerText($hOnes->item(0));
+        else if(strlen($curTitle) > 150 || strlen($curTitle) < 15)
-                        }
+        {
-                }
+            $hOnes = $this->dom->getElementsByTagName('h1');
+            if($hOnes->length == 1)
-                $curTitle = trim($curTitle);
+            {
+                $curTitle = $this->getInnerText($hOnes->item(0));
-                if (count(explode(' ', $curTitle)) <= 4) {
+            }
-                        $curTitle = $origTitle;
+        }
-                }
-                
+        $curTitle = trim($curTitle);
-                $articleTitle = $this->dom->createElement('h1');
-                $articleTitle->innerHTML = $curTitle;
+        if (count(explode(' ', $curTitle)) <= 4) {
-                
+            $curTitle = $origTitle;
-                return $articleTitle;
+        }
-        }
-        
+        $articleTitle = $this->dom->createElement('h1');
-        /**
+        $articleTitle->innerHTML = $curTitle;
-        * Prepare the HTML document for readability to scrape it.
-        * This includes things like stripping javascript, CSS, and handling terrible markup.
+        return $articleTitle;
-        * 
+    }
-        * @return void
-        **/
+    /**
-        protected function prepDocument() {
+    * Prepare the HTML document for readability to scrape it.
-                /**
+    * This includes things like stripping javascript, CSS, and handling terrible markup.
-                * In some cases a body element can't be found (if the HTML is totally hosed for example)
+    *
-                * so we create a new body node and append it to the document.
+    * @return void
-                */
+    **/
-                if ($this->body == null)
+    protected function prepDocument() {
-                {
+        /**
-                        $this->body = $this->dom->createElement('body');
+        * In some cases a body element can't be found (if the HTML is totally hosed for example)
-                        $this->dom->documentElement->appendChild($this->body);
+        * so we create a new body node and append it to the document.
-                }
+        */
-                $this->body->setAttribute('id', 'readabilityBody');
+        if ($this->body == null)
+        {
-                /* Remove all style tags in head */
+            $this->body = $this->dom->createElement('body');
-                $styleTags = $this->dom->getElementsByTagName('style');
+            $this->dom->documentElement->appendChild($this->body);
-                for ($i = $styleTags->length-1; $i >= 0; $i--)
+        }
-                {
+        $this->body->setAttribute('id', 'readabilityBody');
-                        $styleTags->item($i)->parentNode->removeChild($styleTags->item($i));
-                }
+        /* Remove all style tags in head */
+        $styleTags = $this->dom->getElementsByTagName('style');
-                /* Turn all double br's into p's */
+        for ($i = $styleTags->length-1; $i >= 0; $i--)
-                /* Note, this is pretty costly as far as processing goes. Maybe optimize later. */
+        {
-                //document.body.innerHTML = document.body.innerHTML.replace(readability.regexps.replaceBrs, '</p><p>').replace(readability.regexps.replaceFonts, '<$1span>');
+            $styleTags->item($i)->parentNode->removeChild($styleTags->item($i));
-                // We do this in the constructor for PHP as that's when we have raw HTML - before parsing it into a DOM tree.
+        }
-                // Manipulating innerHTML as it's done in JS is not possible in PHP.
-        }
+        /* Turn all double br's into p's */
+        /* Note, this is pretty costly as far as processing goes. Maybe optimize later. */
-        /**
+        //document.body.innerHTML = document.body.innerHTML.replace(readability.regexps.replaceBrs, '</p><p>').replace(readability.regexps.replaceFonts, '<$1span>');
-        * For easier reading, convert this document to have footnotes at the bottom rather than inline links.
+        // We do this in the constructor for PHP as that's when we have raw HTML - before parsing it into a DOM tree.
-        * @see http://www.roughtype.com/archives/2010/05/experiments_in.php
+        // Manipulating innerHTML as it's done in JS is not possible in PHP.
-        *
+    }
-        * @return void
-        **/
+    /**
-        public function addFootnotes($articleContent) {
+    * For easier reading, convert this document to have footnotes at the bottom rather than inline links.
-                $footnotesWrapper = $this->dom->createElement('div');
+    * @see http://www.roughtype.com/archives/2010/05/experiments_in.php
-                $footnotesWrapper->setAttribute('id', 'readability-footnotes');
+    *
-                $footnotesWrapper->innerHTML = '<h3>References</h3>';
+    * @return void
-                
+    **/
-                $articleFootnotes = $this->dom->createElement('ol');
+    public function addFootnotes($articleContent) {
-                $articleFootnotes->setAttribute('id', 'readability-footnotes-list');
+        $footnotesWrapper = $this->dom->createElement('div');
-                $footnotesWrapper->appendChild($articleFootnotes);
+        $footnotesWrapper->setAttribute('id', 'readability-footnotes');
-                
+        $footnotesWrapper->innerHTML = '<h3>References</h3>';
-                $articleLinks = $articleContent->getElementsByTagName('a');
-                
+        $articleFootnotes = $this->dom->createElement('ol');
-                $linkCount = 0;
+        $articleFootnotes->setAttribute('id', 'readability-footnotes-list');
-                for ($i = 0; $i < $articleLinks->length; $i++)
+        $footnotesWrapper->appendChild($articleFootnotes);
-                {
-                        $articleLink  = $articleLinks->item($i);
+        $articleLinks = $articleContent->getElementsByTagName('a');
-                        $footnoteLink = $articleLink->cloneNode(true);
-                        $refLink      = $this->dom->createElement('a');
+        $linkCount = 0;
-                        $footnote     = $this->dom->createElement('li');
+        for ($i = 0; $i < $articleLinks->length; $i++)
-                        $linkDomain   = @parse_url($footnoteLink->getAttribute('href'), PHP_URL_HOST);
+        {
-                        if (!$linkDomain && isset($this->url)) $linkDomain = @parse_url($this->url, PHP_URL_HOST);
+            $articleLink  = $articleLinks->item($i);
-                        //linkDomain   = footnoteLink.host ? footnoteLink.host : document.location.host,
+            $footnoteLink = $articleLink->cloneNode(true);
-                        $linkText     = $this->getInnerText($articleLink);
+            $refLink      = $this->dom->createElement('a');
-                        
+            $footnote     = $this->dom->createElement('li');
-                        if ((strpos($articleLink->getAttribute('class'), 'readability-DoNotFootnote') !== false) || preg_match($this->regexps['skipFootnoteLink'], $linkText)) {
+            $linkDomain   = @parse_url($footnoteLink->getAttribute('href'), PHP_URL_HOST);
-                                continue;
+            if (!$linkDomain && isset($this->url)) $linkDomain = @parse_url($this->url, PHP_URL_HOST);
-                        }
+            //linkDomain   = footnoteLink.host ? footnoteLink.host : document.location.host,
-                        
+            $linkText     = $this->getInnerText($articleLink);
-                        $linkCount++;
+            if ((strpos($articleLink->getAttribute('class'), 'readability-DoNotFootnote') !== false) || preg_match($this->regexps['skipFootnoteLink'], $linkText)) {
-                        /** Add a superscript reference after the article link */
+                continue;
-                        $refLink->setAttribute('href', '#readabilityFootnoteLink-' . $linkCount);
+            }
-                        $refLink->innerHTML = '<small><sup>[' . $linkCount . ']</sup></small>';
-                        $refLink->setAttribute('class', 'readability-DoNotFootnote');
+            $linkCount++;
-                        $refLink->setAttribute('style', 'color: inherit;');
-                        
+            /** Add a superscript reference after the article link */
-                        //TODO: does this work or should we use DOMNode.isSameNode()?
+            $refLink->setAttribute('href', '#readabilityFootnoteLink-' . $linkCount);
-                        if ($articleLink->parentNode->lastChild == $articleLink) {
+            $refLink->innerHTML = '<small><sup>[' . $linkCount . ']</sup></small>';
-                                $articleLink->parentNode->appendChild($refLink);
+            $refLink->setAttribute('class', 'readability-DoNotFootnote');
-                        } else {
+            $refLink->setAttribute('style', 'color: inherit;');
-                                $articleLink->parentNode->insertBefore($refLink, $articleLink->nextSibling);
-                        }
+            //TODO: does this work or should we use DOMNode.isSameNode()?
+            if ($articleLink->parentNode->lastChild == $articleLink) {
-                        $articleLink->setAttribute('style', 'color: inherit; text-decoration: none;');
+                $articleLink->parentNode->appendChild($refLink);
-                        $articleLink->setAttribute('name', 'readabilityLink-' . $linkCount);
+            } else {
+                $articleLink->parentNode->insertBefore($refLink, $articleLink->nextSibling);
-                        $footnote->innerHTML = '<small><sup><a href="#readabilityLink-' . $linkCount . '" title="Jump to Link in Article">^</a></sup></small> ';
+            }
-                        $footnoteLink->innerHTML = ($footnoteLink->getAttribute('title') != '' ? $footnoteLink->getAttribute('title') : $linkText);
+            $articleLink->setAttribute('style', 'color: inherit; text-decoration: none;');
-                        $footnoteLink->setAttribute('name', 'readabilityFootnoteLink-' . $linkCount);
+            $articleLink->setAttribute('name', 'readabilityLink-' . $linkCount);
-                        
-                        $footnote->appendChild($footnoteLink);
+            $footnote->innerHTML = '<small><sup><a href="#readabilityLink-' . $linkCount . '" title="Jump to Link in Article">^</a></sup></small> ';
-                        if ($linkDomain) $footnote->innerHTML = $footnote->innerHTML . '<small> (' . $linkDomain . ')</small>';
-                        
+            $footnoteLink->innerHTML = ($footnoteLink->getAttribute('title') != '' ? $footnoteLink->getAttribute('title') : $linkText);
-                        $articleFootnotes->appendChild($footnote);
+            $footnoteLink->setAttribute('name', 'readabilityFootnoteLink-' . $linkCount);
-                }
+            $footnote->appendChild($footnoteLink);
-                if ($linkCount > 0) {
+            if ($linkDomain) $footnote->innerHTML = $footnote->innerHTML . '<small> (' . $linkDomain . ')</small>';
-                        $articleContent->appendChild($footnotesWrapper);           
-                }
+            $articleFootnotes->appendChild($footnote);
-        }
+        }
-        /**
+        if ($linkCount > 0) {
-        * Reverts P elements with class 'readability-styled'
+            $articleContent->appendChild($footnotesWrapper);
-        * to text nodes - which is what they were before.
+        }
-        *
+    }
-        * @param DOMElement
-        * @return void
+    /**
-        */
+    * Reverts P elements with class 'readability-styled'
-        function revertReadabilityStyledElements($articleContent) {
+    * to text nodes - which is what they were before.
-                $xpath = new DOMXPath($articleContent->ownerDocument);
+    *
-                $elems = $xpath->query('.//p[@class="readability-styled"]', $articleContent);
+    * @param DOMElement
-                //$elems = $articleContent->getElementsByTagName('p');
+    * @return void
-                for ($i = $elems->length-1; $i >= 0; $i--) {
+    */
-                        $e = $elems->item($i);
+    function revertReadabilityStyledElements($articleContent) {
-                        $e->parentNode->replaceChild($articleContent->ownerDocument->createTextNode($e->textContent), $e);
+        $xpath = new DOMXPath($articleContent->ownerDocument);
-                        //if ($e->hasAttribute('class') && $e->getAttribute('class') == 'readability-styled') {
+        $elems = $xpath->query('.//p[@class="readability-styled"]', $articleContent);
-                        //      $e->parentNode->replaceChild($this->dom->createTextNode($e->textContent), $e);
+        //$elems = $articleContent->getElementsByTagName('p');
-                        //}
+        for ($i = $elems->length-1; $i >= 0; $i--) {
-                }
+            $e = $elems->item($i);
-        }
+            $e->parentNode->replaceChild($articleContent->ownerDocument->createTextNode($e->textContent), $e);
-        
+            //if ($e->hasAttribute('class') && $e->getAttribute('class') == 'readability-styled') {
-        /**
+            //  $e->parentNode->replaceChild($this->dom->createTextNode($e->textContent), $e);
-        * Prepare the article node for display. Clean out any inline styles,
+            //}
-        * iframes, forms, strip extraneous <p> tags, etc.
+        }
-        *
+    }
-        * @param DOMElement
-        * @return void
+    /**
-        */
+    * Prepare the article node for display. Clean out any inline styles,
-        function prepArticle($articleContent) {
+    * iframes, forms, strip extraneous <p> tags, etc.
-                $this->cleanStyles($articleContent);
+    *
-                $this->killBreaks($articleContent);
+    * @param DOMElement
-                if ($this->revertForcedParagraphElements) {
+    * @return void
-                        $this->revertReadabilityStyledElements($articleContent);
+    */
-                }
+    function prepArticle($articleContent) {
+        $this->cleanStyles($articleContent);
-                /* Clean out junk from the article content */
+        $this->killBreaks($articleContent);
-                $this->cleanConditionally($articleContent, 'form');
+        if ($this->revertForcedParagraphElements) {
-                $this->clean($articleContent, 'object');
+            $this->revertReadabilityStyledElements($articleContent);
-                $this->clean($articleContent, 'h1');
+        }
-                /**
+        /* Clean out junk from the article content */
-                * If there is only one h2, they are probably using it
+        $this->cleanConditionally($articleContent, 'form');
-                * as a header and not a subheader, so remove it since we already have a header.
+        $this->clean($articleContent, 'object');
-                ***/
+        $this->clean($articleContent, 'h1');
-                if ($articleContent->getElementsByTagName('h2')->length == 1) {
-                        $this->clean($articleContent, 'h2'); 
+        /**
-                }
+        * If there is only one h2, they are probably using it
-                $this->clean($articleContent, 'iframe');
+        * as a header and not a subheader, so remove it since we already have a header.
+        ***/
-                $this->cleanHeaders($articleContent);
+        if (!$this->lightClean && ($articleContent->getElementsByTagName('h2')->length == 1)) {
+            $this->clean($articleContent, 'h2');
-                /* Do these last as the previous stuff may have removed junk that will affect these */
+        }
-                $this->cleanConditionally($articleContent, 'table');
+        $this->clean($articleContent, 'iframe');
-                $this->cleanConditionally($articleContent, 'ul');
-                $this->cleanConditionally($articleContent, 'div');
+        $this->cleanHeaders($articleContent);
-                /* Remove extra paragraphs */
+        /* Do these last as the previous stuff may have removed junk that will affect these */
-                $articleParagraphs = $articleContent->getElementsByTagName('p');
+        $this->cleanConditionally($articleContent, 'table');
-                for ($i = $articleParagraphs->length-1; $i >= 0; $i--)
+        $this->cleanConditionally($articleContent, 'ul');
-                {
+        $this->cleanConditionally($articleContent, 'div');
-                        $imgCount    = $articleParagraphs->item($i)->getElementsByTagName('img')->length;
-                        $embedCount  = $articleParagraphs->item($i)->getElementsByTagName('embed')->length;
+        /* Remove extra paragraphs */
-                        $objectCount = $articleParagraphs->item($i)->getElementsByTagName('object')->length;
+        $articleParagraphs = $articleContent->getElementsByTagName('p');
-                        
+        for ($i = $articleParagraphs->length-1; $i >= 0; $i--)
-                        if ($imgCount === 0 && $embedCount === 0 && $objectCount === 0 && $this->getInnerText($articleParagraphs->item($i), false) == '')
+        {
-                        {
+            $imgCount    = $articleParagraphs->item($i)->getElementsByTagName('img')->length;
-                                $articleParagraphs->item($i)->parentNode->removeChild($articleParagraphs->item($i));
+            $embedCount  = $articleParagraphs->item($i)->getElementsByTagName('embed')->length;
-                        }
+            $objectCount = $articleParagraphs->item($i)->getElementsByTagName('object')->length;
-                }
+            $iframeCount = $articleParagraphs->item($i)->getElementsByTagName('iframe')->length;
-                try {
+            if ($imgCount === 0 && $embedCount === 0 && $objectCount === 0 && $iframeCount === 0 && $this->getInnerText($articleParagraphs->item($i), false) == '')
-                        $articleContent->innerHTML = preg_replace('/<br[^>]*>\s*<p/i', '<p', $articleContent->innerHTML);
+            {
-                        //articleContent.innerHTML = articleContent.innerHTML.replace(/<br[^>]*>\s*<p/gi, '<p');      
+                $articleParagraphs->item($i)->parentNode->removeChild($articleParagraphs->item($i));
-                }
+            }
-                catch (Exception $e) {
+        }
-                        $this->dbg("Cleaning innerHTML of breaks failed. This is an IE strict-block-elements bug. Ignoring.: " . $e);
-                }
+        try {
-        }
+            $articleContent->innerHTML = preg_replace('/<br[^>]*>\s*<p/i', '<p', $articleContent->innerHTML);
-        
+            //articleContent.innerHTML = articleContent.innerHTML.replace(/<br[^>]*>\s*<p/gi, '<p');
-        /**
+        }
-        * Initialize a node with the readability object. Also checks the
+        catch (Exception $e) {
-        * className/id for special names to add to its score.
+            $this->dbg("Cleaning innerHTML of breaks failed. This is an IE strict-block-elements bug. Ignoring.: " . $e);
-        *
+        }
-        * @param Element
+    }
-        * @return void
-        **/
+    /**
-        protected function initializeNode($node) {
+    * Initialize a node with the readability object. Also checks the
-                $readability = $this->dom->createAttribute('readability');
+    * className/id for special names to add to its score.
-                $readability->value = 0; // this is our contentScore
+    *
-                $node->setAttributeNode($readability);                   
+    * @param Element
+    * @return void
-                switch (strtoupper($node->tagName)) { // unsure if strtoupper is needed, but using it just in case
+    **/
-                        case 'DIV':
+    protected function initializeNode($node) {
-                                $readability->value += 5;
+        $readability = $this->dom->createAttribute('readability');
-                                break;
+        $readability->value = 0; // this is our contentScore
+        $node->setAttributeNode($readability);
-                        case 'PRE':
-                        case 'TD':
+        switch (strtoupper($node->tagName)) { // unsure if strtoupper is needed, but using it just in case
-                        case 'BLOCKQUOTE':
+            case 'DIV':
-                                $readability->value += 3;
+                $readability->value += 5;
-                                break;
+                break;
-                                
-                        case 'ADDRESS':
+            case 'PRE':
-                        case 'OL':
+            case 'TD':
-                        case 'UL':
+            case 'BLOCKQUOTE':
-                        case 'DL':
+                $readability->value += 3;
-                        case 'DD':
+                break;
-                        case 'DT':
-                        case 'LI':
+            case 'ADDRESS':
-                        case 'FORM':
+            case 'OL':
-                                $readability->value -= 3;
+            case 'UL':
-                                break;
+            case 'DL':
+            case 'DD':
-                        case 'H1':
+            case 'DT':
-                        case 'H2':
+            case 'LI':
-                        case 'H3':
+            case 'FORM':
-                        case 'H4':
+                $readability->value -= 3;
-                        case 'H5':
+                break;
-                        case 'H6':
-                        case 'TH':
+            case 'H1':
-                                $readability->value -= 5;
+            case 'H2':
-                                break;
+            case 'H3':
-                }
+            case 'H4':
-                $readability->value += $this->getClassWeight($node);
+            case 'H5':
-        }
+            case 'H6':
-        
+            case 'TH':
-        /***
+                $readability->value -= 5;
-        * grabArticle - Using a variety of metrics (content score, classname, element types), find the content that is
+                break;
-        *               most likely to be the stuff a user wants to read. Then return it wrapped up in a div.
+        }
-        *
+        $readability->value += $this->getClassWeight($node);
-        * @return DOMElement
+    }
-        **/
-        protected function grabArticle($page=null) {
+    /***
-                $stripUnlikelyCandidates = $this->flagIsActive(self::FLAG_STRIP_UNLIKELYS);
+    * grabArticle - Using a variety of metrics (content score, classname, element types), find the content that is
-                if (!$page) $page = $this->dom;
+    *               most likely to be the stuff a user wants to read. Then return it wrapped up in a div.
-                $allElements = $page->getElementsByTagName('*');
+    *
-                /**
+    * @return DOMElement
-                * First, node prepping. Trash nodes that look cruddy (like ones with the class name "comment", etc), and turn divs
+    **/
-                * into P tags where they have been used inappropriately (as in, where they contain no other block level elements.)
+    protected function grabArticle($page=null) {
-                *
+        $stripUnlikelyCandidates = $this->flagIsActive(self::FLAG_STRIP_UNLIKELYS);
-                * Note: Assignment from index for performance. See http://www.peachpit.com/articles/article.aspx?p=31567&seqNum=5
+        if (!$page) $page = $this->dom;
-                * TODO: Shouldn't this be a reverse traversal?
+        $allElements = $page->getElementsByTagName('*');
-                **/
+        /**
-                $node = null;
+        * First, node prepping. Trash nodes that look cruddy (like ones with the class name "comment", etc), and turn divs
-                $nodesToScore = array();
+        * into P tags where they have been used inappropriately (as in, where they contain no other block level elements.)
-                for ($nodeIndex = 0; ($node = $allElements->item($nodeIndex)); $nodeIndex++) {
+        *
-                //for ($nodeIndex=$targetList->length-1; $nodeIndex >= 0; $nodeIndex--) {
+        * Note: Assignment from index for performance. See http://www.peachpit.com/articles/article.aspx?p=31567&seqNum=5
-                        //$node = $targetList->item($nodeIndex);
+        * TODO: Shouldn't this be a reverse traversal?
-                        $tagName = strtoupper($node->tagName);
+        **/
-                        /* Remove unlikely candidates */
+        $node = null;
-                        if ($stripUnlikelyCandidates) {
+        $nodesToScore = array();
-                                $unlikelyMatchString = $node->getAttribute('class') . $node->getAttribute('id');
+        for ($nodeIndex = 0; ($node = $allElements->item($nodeIndex)); $nodeIndex++) {
-                                if (
+        //for ($nodeIndex=$targetList->length-1; $nodeIndex >= 0; $nodeIndex--) {
-                                        preg_match($this->regexps['unlikelyCandidates'], $unlikelyMatchString) &&
+            //$node = $targetList->item($nodeIndex);
-                                        !preg_match($this->regexps['okMaybeItsACandidate'], $unlikelyMatchString) &&
+            $tagName = strtoupper($node->tagName);
-                                        $tagName != 'BODY'
+            /* Remove unlikely candidates */
-                                )
+            if ($stripUnlikelyCandidates) {
-                                {
+                $unlikelyMatchString = $node->getAttribute('class') . $node->getAttribute('id');
-                                        $this->dbg('Removing unlikely candidate - ' . $unlikelyMatchString);
+                if (
-                                        //$nodesToRemove[] = $node;
+                    preg_match($this->regexps['unlikelyCandidates'], $unlikelyMatchString) &&
-                                        $node->parentNode->removeChild($node);
+                    !preg_match($this->regexps['okMaybeItsACandidate'], $unlikelyMatchString) &&
-                                        $nodeIndex--;
+                    $tagName != 'BODY'
-                                        continue;
+                )
-                                }               
+                {
-                        }
+                    $this->dbg('Removing unlikely candidate - ' . $unlikelyMatchString);
+                    //$nodesToRemove[] = $node;
-                        if ($tagName == 'P' || $tagName == 'TD' || $tagName == 'PRE') {
+                    $node->parentNode->removeChild($node);
-                                $nodesToScore[] = $node;
+                    $nodeIndex--;
-                        }
+                    continue;
+                }
-                        /* Turn all divs that don't have children block level elements into p's */
+            }
-                        if ($tagName == 'DIV') {
-                                if (!preg_match($this->regexps['divToPElements'], $node->innerHTML)) {
+            if ($tagName == 'P' || $tagName == 'TD' || $tagName == 'PRE') {
-                                        //$this->dbg('Altering div to p');
+                $nodesToScore[] = $node;
-                                        $newNode = $this->dom->createElement('p');
+            }
-                                        try {
-                                                $newNode->innerHTML = $node->innerHTML;
+            /* Turn all divs that don't have children block level elements into p's */
-                                                //$nodesToReplace[] = array('new'=>$newNode, 'old'=>$node);
+            if ($tagName == 'DIV') {
-                                                $node->parentNode->replaceChild($newNode, $node);
+                if (!preg_match($this->regexps['divToPElements'], $node->innerHTML)) {
-                                                $nodeIndex--;
+                    //$this->dbg('Altering div to p');
-                                                $nodesToScore[] = $node; // or $newNode?
+                    $newNode = $this->dom->createElement('p');
-                                        }
+                    try {
-                                        catch(Exception $e) {
+                        $newNode->innerHTML = $node->innerHTML;
-                                                $this->dbg('Could not alter div to p, reverting back to div.: ' . $e);
+                        //$nodesToReplace[] = array('new'=>$newNode, 'old'=>$node);
-                                        }
+                        $node->parentNode->replaceChild($newNode, $node);
-                                }
+                        $nodeIndex--;
-                                else
+                        $nodesToScore[] = $node; // or $newNode?
-                                {
+                    }
-                                        // EXPERIMENTAL
+                    catch(Exception $e) {
-                                        // TODO: change these p elements back to text nodes after processing
+                        $this->dbg('Could not alter div to p, reverting back to div.: ' . $e);
-                                        for ($i = 0, $il = $node->childNodes->length; $i < $il; $i++) {
+                    }
-                                                $childNode = $node->childNodes->item($i);
+                }
-                                                if ($childNode->nodeType == 3) { // XML_TEXT_NODE
+                else
-                                                        //$this->dbg('replacing text node with a p tag with the same content.');
+                {
-                                                        $p = $this->dom->createElement('p');
+                    /* EXPERIMENTAL */
-                                                        $p->innerHTML = $childNode->nodeValue;
+                    // TODO: change these p elements back to text nodes after processing
-                                                        $p->setAttribute('style', 'display: inline;');
+                    for ($i = 0, $il = $node->childNodes->length; $i < $il; $i++) {
-                                                        $p->setAttribute('class', 'readability-styled');
+                        $childNode = $node->childNodes->item($i);
-                                                        $childNode->parentNode->replaceChild($p, $childNode);
+                        if ($childNode->nodeType == 3) { // XML_TEXT_NODE
-                                                }
+                            //$this->dbg('replacing text node with a p tag with the same content.');
-                                        }
+                            $p = $this->dom->createElement('p');
-                                }
+                            $p->innerHTML = $childNode->nodeValue;
-                        }
+                            $p->setAttribute('style', 'display: inline;');
-                }
+                            $p->setAttribute('class', 'readability-styled');
-                
+                            $childNode->parentNode->replaceChild($p, $childNode);
-                /**
+                        }
-                * Loop through all paragraphs, and assign a score to them based on how content-y they look.
+                    }
-                * Then add their score to their parent node.
+                }
-                *
+            }
-                * A score is determined by things like number of commas, class names, etc. Maybe eventually link density.
+        }
-                **/
-                $candidates = array();
+        /**
-                for ($pt=0; $pt < count($nodesToScore); $pt++) {
+        * Loop through all paragraphs, and assign a score to them based on how content-y they look.
-                        $parentNode      = $nodesToScore[$pt]->parentNode;
+        * Then add their score to their parent node.
-                        // $grandParentNode = $parentNode ? $parentNode->parentNode : null;
+        *
-                        $grandParentNode = !$parentNode ? null : (($parentNode->parentNode instanceof DOMElement) ? $parentNode->parentNode : null);
+        * A score is determined by things like number of commas, class names, etc. Maybe eventually link density.
-                        $innerText       = $this->getInnerText($nodesToScore[$pt]);
+        **/
+        $candidates = array();
-                        if (!$parentNode || !isset($parentNode->tagName)) {
+        for ($pt=0; $pt < count($nodesToScore); $pt++) {
-                                continue;
+            $parentNode      = $nodesToScore[$pt]->parentNode;
-                        }
+            // $grandParentNode = $parentNode ? $parentNode->parentNode : null;
+            $grandParentNode = !$parentNode ? null : (($parentNode->parentNode instanceof DOMElement) ? $parentNode->parentNode : null);
-                        /* If this paragraph is less than 25 characters, don't even count it. */
+            $innerText       = $this->getInnerText($nodesToScore[$pt]);
-                        if(strlen($innerText) < 25) {
-                                continue;
+            if (!$parentNode || !isset($parentNode->tagName)) {
-                        }
+                continue;
+            }
-                        /* Initialize readability data for the parent. */
-                        if (!$parentNode->hasAttribute('readability')) 
+            /* If this paragraph is less than 25 characters, don't even count it. */
-                        {
+            if(strlen($innerText) < 25) {
-                                $this->initializeNode($parentNode);
+                continue;
-                                $candidates[] = $parentNode;
+            }
-                        }
+            /* Initialize readability data for the parent. */
-                        /* Initialize readability data for the grandparent. */
+            if (!$parentNode->hasAttribute('readability'))
-                        if ($grandParentNode && !$grandParentNode->hasAttribute('readability') && isset($grandParentNode->tagName))
+            {
-                        {
+                $this->initializeNode($parentNode);
-                                $this->initializeNode($grandParentNode);
+                $candidates[] = $parentNode;
-                                $candidates[] = $grandParentNode;
+            }
-                        }
+            /* Initialize readability data for the grandparent. */
-                        $contentScore = 0;
+            if ($grandParentNode && !$grandParentNode->hasAttribute('readability') && isset($grandParentNode->tagName))
+            {
-                        /* Add a point for the paragraph itself as a base. */
+                $this->initializeNode($grandParentNode);
-                        $contentScore++;
+                $candidates[] = $grandParentNode;
+            }
-                        /* Add points for any commas within this paragraph */
-                        $contentScore += count(explode(',', $innerText));
+            $contentScore = 0;
-                        
-                        /* For every 100 characters in this paragraph, add another point. Up to 3 points. */
+            /* Add a point for the paragraph itself as a base. */
-                        $contentScore += min(floor(strlen($innerText) / 100), 3);
+            $contentScore++;
-                        
-                        /* Add the score to the parent. The grandparent gets half. */
+            /* Add points for any commas within this paragraph */
-                        $parentNode->getAttributeNode('readability')->value += $contentScore;
+            $contentScore += count(explode(',', $innerText));
-                        if ($grandParentNode) {
+            /* For every 100 characters in this paragraph, add another point. Up to 3 points. */
-                                $grandParentNode->getAttributeNode('readability')->value += $contentScore/2;             
+            $contentScore += min(floor(strlen($innerText) / 100), 3);
-                        }
-                }
+            /* Add the score to the parent. The grandparent gets half. */
+            $parentNode->getAttributeNode('readability')->value += $contentScore;
-                /**
-                * After we've calculated scores, loop through all of the possible candidate nodes we found
+            if ($grandParentNode) {
-                * and find the one with the highest score.
+                $grandParentNode->getAttributeNode('readability')->value += $contentScore/2;
-                **/
+            }
-                $topCandidate = null;
+        }
-                for ($c=0, $cl=count($candidates); $c < $cl; $c++)
-                {
+        /**
-                        /**
+        * After we've calculated scores, loop through all of the possible candidate nodes we found
-                        * Scale the final candidates score based on link density. Good content should have a
+        * and find the one with the highest score.
-                        * relatively small link density (5% or less) and be mostly unaffected by this operation.
+        **/
-                        **/
+        $topCandidate = null;
-                        $readability = $candidates[$c]->getAttributeNode('readability');
+        for ($c=0, $cl=count($candidates); $c < $cl; $c++)
-                        $readability->value = $readability->value * (1-$this->getLinkDensity($candidates[$c]));
+        {
+            /**
-                        $this->dbg('Candidate: ' . $candidates[$c]->tagName . ' (' . $candidates[$c]->getAttribute('class') . ':' . $candidates[$c]->getAttribute('id') . ') with score ' . $readability->value);
+            * Scale the final candidates score based on link density. Good content should have a
+            * relatively small link density (5% or less) and be mostly unaffected by this operation.
-                        if (!$topCandidate || $readability->value > (int)$topCandidate->getAttribute('readability')) {
+            **/
-                                $topCandidate = $candidates[$c];
+            $readability = $candidates[$c]->getAttributeNode('readability');
-                        }
+            $readability->value = $readability->value * (1-$this->getLinkDensity($candidates[$c]));
-                }
+            $this->dbg('Candidate: ' . $candidates[$c]->tagName . ' (' . $candidates[$c]->getAttribute('class') . ':' . $candidates[$c]->getAttribute('id') . ') with score ' . $readability->value);
-                /**
-                * If we still have no top candidate, just use the body as a last resort.
+            if (!$topCandidate || $readability->value > (int)$topCandidate->getAttribute('readability')) {
-                * We also have to copy the body node so it is something we can modify.
+                $topCandidate = $candidates[$c];
-                **/
+            }
-                if ($topCandidate === null || strtoupper($topCandidate->tagName) == 'BODY')
+        }
-                {
-                        $topCandidate = $this->dom->createElement('div');
+        /**
-                        if ($page instanceof DOMDocument) {
+        * If we still have no top candidate, just use the body as a last resort.
-                                if (!isset($page->documentElement)) {
+        * We also have to copy the body node so it is something we can modify.
-                                        // we don't have a body either? what a mess! :)
+        **/
-                                } else {
+        if ($topCandidate === null || strtoupper($topCandidate->tagName) == 'BODY')
-                                        $topCandidate->innerHTML = $page->documentElement->innerHTML;
+        {
-                                        $page->documentElement->innerHTML = '';
+            $topCandidate = $this->dom->createElement('div');
-                                        $page->documentElement->appendChild($topCandidate);
+            if ($page instanceof DOMDocument) {
-                                }
+                if (!isset($page->documentElement)) {
-                        } else {
+                    // we don't have a body either? what a mess! :)
-                                $topCandidate->innerHTML = $page->innerHTML;
+                } else {
-                                $page->innerHTML = '';
+                    $topCandidate->innerHTML = $page->documentElement->innerHTML;
-                                $page->appendChild($topCandidate);
+                    $page->documentElement->innerHTML = '';
-                        }
+                    $page->documentElement->appendChild($topCandidate);
-                        $this->initializeNode($topCandidate);
+                }
-                }
+            } else {
+                $topCandidate->innerHTML = $page->innerHTML;
-                /**
+                $page->innerHTML = '';
-                * Now that we have the top candidate, look through its siblings for content that might also be related.
+                $page->appendChild($topCandidate);
-                * Things like preambles, content split by ads that we removed, etc.
+            }
-                **/
+            $this->initializeNode($topCandidate);
-                $articleContent        = $this->dom->createElement('div');
+        }
-                $articleContent->setAttribute('id', 'readability-content');
-                $siblingScoreThreshold = max(10, ((int)$topCandidate->getAttribute('readability')) * 0.2);
+        /**
-                $siblingNodes          = $topCandidate->parentNode->childNodes;
+        * Now that we have the top candidate, look through its siblings for content that might also be related.
-                if (!isset($siblingNodes)) {
+        * Things like preambles, content split by ads that we removed, etc.
-                        $siblingNodes = new stdClass;
+        **/
-                        $siblingNodes->length = 0;
+        $articleContent        = $this->dom->createElement('div');
-                }
+        $articleContent->setAttribute('id', 'readability-content');
+        $siblingScoreThreshold = max(10, ((int)$topCandidate->getAttribute('readability')) * 0.2);
-                for ($s=0, $sl=$siblingNodes->length; $s < $sl; $s++)
+        $siblingNodes          = $topCandidate->parentNode->childNodes;
-                {
+        if (!isset($siblingNodes)) {
-                        $siblingNode = $siblingNodes->item($s);
+            $siblingNodes = new stdClass;
-                        $append      = false;
+            $siblingNodes->length = 0;
+        }
-                        $this->dbg('Looking at sibling node: ' . $siblingNode->nodeName . (($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->hasAttribute('readability')) ? (' with score ' . $siblingNode->getAttribute('readability')) : ''));
+        for ($s=0, $sl=$siblingNodes->length; $s < $sl; $s++)
-                        //dbg('Sibling has score ' . ($siblingNode->readability ? siblingNode.readability.contentScore : 'Unknown'));
+        {
+            $siblingNode = $siblingNodes->item($s);
-                        if ($siblingNode === $topCandidate)
+            $append      = false;
-                        // or if ($siblingNode->isSameNode($topCandidate))
-                        {
+            $this->dbg('Looking at sibling node: ' . $siblingNode->nodeName . (($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->hasAttribute('readability')) ? (' with score ' . $siblingNode->getAttribute('readability')) : ''));
-                                $append = true;
-                        }
+            //dbg('Sibling has score ' . ($siblingNode->readability ? siblingNode.readability.contentScore : 'Unknown'));
-                        $contentBonus = 0;
+            if ($siblingNode === $topCandidate)
-                        /* Give a bonus if sibling nodes and top candidates have the example same classname */
+            // or if ($siblingNode->isSameNode($topCandidate))
-                        if ($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->getAttribute('class') == $topCandidate->getAttribute('class') && $topCandidate->getAttribute('class') != '') {
+            {
-                                $contentBonus += ((int)$topCandidate->getAttribute('readability')) * 0.2;
+                $append = true;
-                        }
+            }
-                        if ($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->hasAttribute('readability') && (((int)$siblingNode->getAttribute('readability')) + $contentBonus) >= $siblingScoreThreshold)
+            $contentBonus = 0;
-                        {
+            /* Give a bonus if sibling nodes and top candidates have the example same classname */
-                                $append = true;
+            if ($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->getAttribute('class') == $topCandidate->getAttribute('class') && $topCandidate->getAttribute('class') != '') {
-                        }
+                $contentBonus += ((int)$topCandidate->getAttribute('readability')) * 0.2;
-                        
+            }
-                        if (strtoupper($siblingNode->nodeName) == 'P') {
-                                $linkDensity = $this->getLinkDensity($siblingNode);
+            if ($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->hasAttribute('readability') && (((int)$siblingNode->getAttribute('readability')) + $contentBonus) >= $siblingScoreThreshold)
-                                $nodeContent = $this->getInnerText($siblingNode);
+            {
-                                $nodeLength  = strlen($nodeContent);
+                $append = true;
-                                
+            }
-                                if ($nodeLength > 80 && $linkDensity < 0.25)
-                                {
+            if (strtoupper($siblingNode->nodeName) == 'P') {
-                                        $append = true;
+                $linkDensity = $this->getLinkDensity($siblingNode);
-                                }
+                $nodeContent = $this->getInnerText($siblingNode);
-                                else if ($nodeLength < 80 && $linkDensity === 0 && preg_match('/\.( |$)/', $nodeContent))
+                $nodeLength  = strlen($nodeContent);
-                                {
-                                        $append = true;
+                if ($nodeLength > 80 && $linkDensity < 0.25)
-                                }
+                {
-                        }
+                    $append = true;
+                }
-                        /* Look for a special classname */
+                else if ($nodeLength < 80 && $linkDensity === 0 && preg_match('/\.( |$)/', $nodeContent))
-                        if ($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->hasAttribute('class') && $siblingNode->getAttribute('class') != '')
+                {
-                        {
+                    $append = true;
-                                if (preg_match($this->regexps['okMaybeItsACandidate'], $siblingNode->getAttribute('class'))) {
+                }
-                                        $append = true;
+            }
-                                }
-                        }
+            if ($append)
+            {
-                        /* Look for a special classname */
+                $this->dbg('Appending node: ' . $siblingNode->nodeName);
-                        if ($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->hasAttribute('id') && $siblingNode->getAttribute('id') != '')
-                        {
+                $nodeToAppend = null;
-                                if (preg_match($this->regexps['okMaybeItsACandidate'], $siblingNode->getAttribute('id'))) {
+                $sibNodeName = strtoupper($siblingNode->nodeName);
-                                        $append = true;
+                if ($sibNodeName != 'DIV' && $sibNodeName != 'P') {
-                                }
+                    /* We have a node that isn't a common block level element, like a form or td tag. Turn it into a div so it doesn't get filtered out later by accident. */
-                        }
+                    $this->dbg('Altering siblingNode of ' . $sibNodeName . ' to div.');
+                    $nodeToAppend = $this->dom->createElement('div');
-                        if ($append)
+                    try {
-                        {
+                        $nodeToAppend->setAttribute('id', $siblingNode->getAttribute('id'));
-                                $this->dbg('Appending node: ' . $siblingNode->nodeName);
+                        $nodeToAppend->innerHTML = $siblingNode->innerHTML;
+                    }
-                                $nodeToAppend = null;
+                    catch(Exception $e)
-                                $sibNodeName = strtoupper($siblingNode->nodeName);
+                    {
-                                if ($sibNodeName != 'DIV' && $sibNodeName != 'P') {
+                        $this->dbg('Could not alter siblingNode to div, reverting back to original.');
-                                        /* We have a node that isn't a common block level element, like a form or td tag. Turn it into a div so it doesn't get filtered out later by accident. */
+                        $nodeToAppend = $siblingNode;
-                                        
+                        $s--;
-                                        $this->dbg('Altering siblingNode of ' . $sibNodeName . ' to div.');
+                        $sl--;
-                                        $nodeToAppend = $this->dom->createElement('div');
+                    }
-                                        try {
+                } else {
-                                                $nodeToAppend->setAttribute('id', $siblingNode->getAttribute('id'));
+                    $nodeToAppend = $siblingNode;
-                                                $nodeToAppend->innerHTML = $siblingNode->innerHTML;
+                    $s--;
-                                        }
+                    $sl--;
-                                        catch(Exception $e)
+                }
-                                        {
-                                                $this->dbg('Could not alter siblingNode to div, reverting back to original.');
+                /* To ensure a node does not interfere with readability styles, remove its classnames */
-                                                $nodeToAppend = $siblingNode;
+                $nodeToAppend->removeAttribute('class');
-                                                $s--;
-                                                $sl--;
+                /* Append sibling and subtract from our list because it removes the node when you append to another node */
-                                        }
+                $articleContent->appendChild($nodeToAppend);
-                                } else {
+            }
-                                        $nodeToAppend = $siblingNode;
+        }
-                                        $s--;
-                                        $sl--;
+        /**
-                                }
+        * So we have all of the content that we need. Now we clean it up for presentation.
-                                
+        **/
-                                /* To ensure a node does not interfere with readability styles, remove its classnames */
+        $this->prepArticle($articleContent);
-                                $nodeToAppend->removeAttribute('class');
+        /**
-                                /* Append sibling and subtract from our list because it removes the node when you append to another node */
+        * Now that we've gone through the full algorithm, check to see if we got any meaningful content.
-                                $articleContent->appendChild($nodeToAppend);
+        * If we didn't, we may need to re-run grabArticle with different flags set. This gives us a higher
-                        }
+        * likelihood of finding the content, and the sieve approach gives us a higher likelihood of
-                }
+        * finding the -right- content.
+        **/
-                /**
+        if (strlen($this->getInnerText($articleContent, false)) < 250)
-                * So we have all of the content that we need. Now we clean it up for presentation.
+        {
-                **/
+            // TODO: find out why element disappears sometimes, e.g. for this URL http://www.businessinsider.com/6-hedge-fund-etfs-for-average-investors-2011-7
-                $this->prepArticle($articleContent);
+            // in the meantime, we check and create an empty element if it's not there.
+            if (!isset($this->body->childNodes)) $this->body = $this->dom->createElement('body');
-                /**
+            $this->body->innerHTML = $this->bodyCache;
-                * Now that we've gone through the full algorithm, check to see if we got any meaningful content.
-                * If we didn't, we may need to re-run grabArticle with different flags set. This gives us a higher
+            if ($this->flagIsActive(self::FLAG_STRIP_UNLIKELYS)) {
-                * likelihood of finding the content, and the sieve approach gives us a higher likelihood of
+                $this->removeFlag(self::FLAG_STRIP_UNLIKELYS);
-                * finding the -right- content.
+                return $this->grabArticle($this->body);
-                **/
+            }
-                if (strlen($this->getInnerText($articleContent, false)) < 250)
+            else if ($this->flagIsActive(self::FLAG_WEIGHT_CLASSES)) {
-                {
+                $this->removeFlag(self::FLAG_WEIGHT_CLASSES);
-                        // TODO: find out why element disappears sometimes, e.g. for this URL http://www.businessinsider.com/6-hedge-fund-etfs-for-average-investors-2011-7
+                return $this->grabArticle($this->body);
-                        // in the meantime, we check and create an empty element if it's not there.
+            }
-                        if (!isset($this->body->childNodes)) $this->body = $this->dom->createElement('body');
+            else if ($this->flagIsActive(self::FLAG_CLEAN_CONDITIONALLY)) {
-                        $this->body->innerHTML = $this->bodyCache;
+                $this->removeFlag(self::FLAG_CLEAN_CONDITIONALLY);
-                        
+                return $this->grabArticle($this->body);
-                        if ($this->flagIsActive(self::FLAG_STRIP_UNLIKELYS)) {
+            }
-                                $this->removeFlag(self::FLAG_STRIP_UNLIKELYS);
+            else {
-                                return $this->grabArticle($this->body);
+                return false;
-                        }
+            }
-                        else if ($this->flagIsActive(self::FLAG_WEIGHT_CLASSES)) {
+        }
-                                $this->removeFlag(self::FLAG_WEIGHT_CLASSES);
+        return $articleContent;
-                                return $this->grabArticle($this->body);              
+    }
-                        }
-                        else if ($this->flagIsActive(self::FLAG_CLEAN_CONDITIONALLY)) {
+    /**
-                                $this->removeFlag(self::FLAG_CLEAN_CONDITIONALLY);
+    * Remove script tags from document
-                                return $this->grabArticle($this->body);
+    *
-                        }
+    * @param DOMElement
-                        else {
+    * @return void
-                                return false;
+    */
-                        }
+    public function removeScripts($doc) {
-                }
+        $scripts = $doc->getElementsByTagName('script');
-                return $articleContent;
+        for($i = $scripts->length-1; $i >= 0; $i--)
-        }
+        {
-        
+            $scripts->item($i)->parentNode->removeChild($scripts->item($i));
-        /**
+        }
-        * Remove script tags from document
+    }
-        *
-        * @param DOMElement
+    /**
-        * @return void
+    * Get the inner text of a node.
-        */
+    * This also strips out any excess whitespace to be found.
-        public function removeScripts($doc) {
+    *
-                $scripts = $doc->getElementsByTagName('script');
+    * @param DOMElement $
-                for($i = $scripts->length-1; $i >= 0; $i--)
+    * @param boolean $normalizeSpaces (default: true)
-                {
+    * @return string
-                        $scripts->item($i)->parentNode->removeChild($scripts->item($i));
+    **/
-                }
+    public function getInnerText($e, $normalizeSpaces=true) {
-        }
+        $textContent = '';
-        
-        /**
+        if (!isset($e->textContent) || $e->textContent == '') {
-        * Get the inner text of a node.
+            return '';
-        * This also strips out any excess whitespace to be found.
+        }
-        *
-        * @param DOMElement $
+        $textContent = trim($e->textContent);
-        * @param boolean $normalizeSpaces (default: true)
-        * @return string
+        if ($normalizeSpaces) {
-        **/
+            return preg_replace($this->regexps['normalize'], ' ', $textContent);
-        public function getInnerText($e, $normalizeSpaces=true) {
+        } else {
-                $textContent = '';
+            return $textContent;
+        }
-                if (!isset($e->textContent) || $e->textContent == '') {
+    }
-                        return '';
-                }
+    /**
+    * Get the number of times a string $s appears in the node $e.
-                $textContent = trim($e->textContent);
+    *
+    * @param DOMElement $e
-                if ($normalizeSpaces) {
+    * @param string - what to count. Default is ","
-                        return preg_replace($this->regexps['normalize'], ' ', $textContent);
+    * @return number (integer)
-                } else {
+    **/
-                        return $textContent;
+    public function getCharCount($e, $s=',') {
-                }
+        return substr_count($this->getInnerText($e), $s);
-        }
+    }
-        /**
+    /**
-        * Get the number of times a string $s appears in the node $e.
+    * Remove the style attribute on every $e and under.
-        *
+    *
-        * @param DOMElement $e
+    * @param DOMElement $e
-        * @param string - what to count. Default is ","
+    * @return void
-        * @return number (integer)
+    */
-        **/
+    public function cleanStyles($e) {
-        public function getCharCount($e, $s=',') {
+        if (!is_object($e)) return;
-                return substr_count($this->getInnerText($e), $s);
+        $elems = $e->getElementsByTagName('*');
-        }
+        foreach ($elems as $elem) {
+            $elem->removeAttribute('style');
-        /**
+        }
-        * Remove the style attribute on every $e and under.
+    }
-        *
-        * @param DOMElement $e
+    /**
-        * @return void
+    * Get the density of links as a percentage of the content
-        */
+    * This is the amount of text that is inside a link divided by the total text in the node.
-        public function cleanStyles($e) {
+    *
-                if (!is_object($e)) return;
+    * @param DOMElement $e
-                $elems = $e->getElementsByTagName('*');
+    * @return number (float)
-                foreach ($elems as $elem) {
+    */
-                        $elem->removeAttribute('style');
+    public function getLinkDensity($e) {
-                }
+        $links      = $e->getElementsByTagName('a');
-        }
+        $textLength = strlen($this->getInnerText($e));
-        
+        $linkLength = 0;
-        /**
+        for ($i=0, $il=$links->length; $i < $il; $i++)
-        * Get the density of links as a percentage of the content
+        {
-        * This is the amount of text that is inside a link divided by the total text in the node.
+            $linkLength += strlen($this->getInnerText($links->item($i)));
-        * 
+        }
-        * @param DOMElement $e
+        if ($textLength > 0) {
-        * @return number (float)
+            return $linkLength / $textLength;
-        */
+        } else {
-        public function getLinkDensity($e) {
+            return 0;
-                $links      = $e->getElementsByTagName('a');
+        }
-                $textLength = strlen($this->getInnerText($e));
+    }
-                $linkLength = 0;
-                for ($i=0, $il=$links->length; $i < $il; $i++)
+    /**
-                {
+    * Get an elements class/id weight. Uses regular expressions to tell if this
-                        $linkLength += strlen($this->getInnerText($links->item($i)));
+    * element looks good or bad.
-                }
+    *
-                if ($textLength > 0) {
+    * @param DOMElement $e
-                        return $linkLength / $textLength;
+    * @return number (Integer)
-                } else {
+    */
-                        return 0;
+    public function getClassWeight($e) {
-                }
+        if(!$this->flagIsActive(self::FLAG_WEIGHT_CLASSES)) {
-        }
+            return 0;
-        
+        }
-        /**
-        * Get an elements class/id weight. Uses regular expressions to tell if this 
+        $weight = 0;
-        * element looks good or bad.
-        *
+        /* Look for a special classname */
-        * @param DOMElement $e
+        if ($e->hasAttribute('class') && $e->getAttribute('class') != '')
-        * @return number (Integer)
+        {
-        */
+            if (preg_match($this->regexps['negative'], $e->getAttribute('class'))) {
-        public function getClassWeight($e) {
+                $weight -= 25;
-                if(!$this->flagIsActive(self::FLAG_WEIGHT_CLASSES)) {
+            }
-                        return 0;
+            if (preg_match($this->regexps['positive'], $e->getAttribute('class'))) {
-                }
+                $weight += 25;
+            }
-                $weight = 0;
+        }
-                /* Look for a special classname */
+        /* Look for a special ID */
-                if ($e->hasAttribute('class') && $e->getAttribute('class') != '')
+        if ($e->hasAttribute('id') && $e->getAttribute('id') != '')
-                {
+        {
-                        if (preg_match($this->regexps['negative'], $e->getAttribute('class'))) {
+            if (preg_match($this->regexps['negative'], $e->getAttribute('id'))) {
-                                $weight -= 25;
+                $weight -= 25;
-                        }
+            }
-                        if (preg_match($this->regexps['positive'], $e->getAttribute('class'))) {
+            if (preg_match($this->regexps['positive'], $e->getAttribute('id'))) {
-                                $weight += 25;
+                $weight += 25;
-                        }
+            }
-                }
+        }
+        return $weight;
-                /* Look for a special ID */
+    }
-                if ($e->hasAttribute('id') && $e->getAttribute('id') != '')
-                {
+    /**
-                        if (preg_match($this->regexps['negative'], $e->getAttribute('id'))) {
+    * Remove extraneous break tags from a node.
-                                $weight -= 25;
+    *
-                        }
+    * @param DOMElement $node
-                        if (preg_match($this->regexps['positive'], $e->getAttribute('id'))) {
+    * @return void
-                                $weight += 25;
+    */
-                        }
+    public function killBreaks($node) {
-                }
+        $html = $node->innerHTML;
-                return $weight;
+        $html = preg_replace($this->regexps['killBreaks'], '<br />', $html);
-        }
+        $node->innerHTML = $html;
+    }
-        /**
-        * Remove extraneous break tags from a node.
+    /**
-        *
+    * Clean a node of all elements of type "tag".
-        * @param DOMElement $node
+    * (Unless it's a youtube/vimeo video. People love movies.)
-        * @return void
+    *
-        */
+    * Updated 2012-09-18 to preserve youtube/vimeo iframes
-        public function killBreaks($node) {
+    *
-                $html = $node->innerHTML;
+    * @param DOMElement $e
-                $html = preg_replace($this->regexps['killBreaks'], '<br />', $html);
+    * @param string $tag
-                $node->innerHTML = $html;
+    * @return void
-        }
+    */
+    public function clean($e, $tag) {
-        /**
+        $targetList = $e->getElementsByTagName($tag);
-        * Clean a node of all elements of type "tag".
+        $isEmbed = ($tag == 'iframe' || $tag == 'object' || $tag == 'embed');
-        * (Unless it's a youtube/vimeo video. People love movies.)
-        *
+        for ($y=$targetList->length-1; $y >= 0; $y--) {
-        * @param DOMElement $e
+            /* Allow youtube and vimeo videos through as people usually want to see those. */
-        * @param string $tag
+            if ($isEmbed) {
-        * @return void
+                $attributeValues = '';
-        */
+                for ($i=0, $il=$targetList->item($y)->attributes->length; $i < $il; $i++) {
-        public function clean($e, $tag) {
+                    $attributeValues .= $targetList->item($y)->attributes->item($i)->value . '|'; // DOMAttr? (TODO: test)
-                $targetList = $e->getElementsByTagName($tag);
+                }
-                $isEmbed = ($tag == 'object' || $tag == 'embed');
-                
+                /* First, check the elements attributes to see if any of them contain youtube or vimeo */
-                for ($y=$targetList->length-1; $y >= 0; $y--) {
+                if (preg_match($this->regexps['video'], $attributeValues)) {
-                        /* Allow youtube and vimeo videos through as people usually want to see those. */
+                    continue;
-                        if ($isEmbed) {
+                }
-                                $attributeValues = '';
-                                for ($i=0, $il=$targetList->item($y)->attributes->length; $i < $il; $i++) {
+                /* Then check the elements inside this element for the same. */
-                                        $attributeValues .= $targetList->item($y)->attributes->item($i)->value . '|'; // DOMAttr? (TODO: test)
+                if (preg_match($this->regexps['video'], $targetList->item($y)->innerHTML)) {
-                                }
+                    continue;
-                                
+                }
-                                /* First, check the elements attributes to see if any of them contain youtube or vimeo */
+            }
-                                if (preg_match($this->regexps['video'], $attributeValues)) {
+            $targetList->item($y)->parentNode->removeChild($targetList->item($y));
-                                        continue;
+        }
-                                }
+    }
-                                /* Then check the elements inside this element for the same. */
+    /**
-                                if (preg_match($this->regexps['video'], $targetList->item($y)->innerHTML)) {
+    * Clean an element of all tags of type "tag" if they look fishy.
-                                        continue;
+    * "Fishy" is an algorithm based on content length, classnames,
-                                }
+    * link density, number of images & embeds, etc.
-                        }
+    *
-                        $targetList->item($y)->parentNode->removeChild($targetList->item($y));
+    * @param DOMElement $e
-                }
+    * @param string $tag
-        }
+    * @return void
-        
+    */
-        /**
+    public function cleanConditionally($e, $tag) {
-        * Clean an element of all tags of type "tag" if they look fishy.
+        if (!$this->flagIsActive(self::FLAG_CLEAN_CONDITIONALLY)) {
-        * "Fishy" is an algorithm based on content length, classnames, 
+            return;
-        * link density, number of images & embeds, etc.
+        }
-        *
-        * @param DOMElement $e
+        $tagsList = $e->getElementsByTagName($tag);
-        * @param string $tag
+        $curTagsLength = $tagsList->length;
-        * @return void
-        */
+        /**
-        public function cleanConditionally($e, $tag) {
+        * Gather counts for other typical elements embedded within.
-                if (!$this->flagIsActive(self::FLAG_CLEAN_CONDITIONALLY)) {
+        * Traverse backwards so we can remove nodes at the same time without effecting the traversal.
-                        return;
+        *
-                }
+        * TODO: Consider taking into account original contentScore here.
+        */
-                $tagsList = $e->getElementsByTagName($tag);
+        for ($i=$curTagsLength-1; $i >= 0; $i--) {
-                $curTagsLength = $tagsList->length;
+            $weight = $this->getClassWeight($tagsList->item($i));
+            $contentScore = ($tagsList->item($i)->hasAttribute('readability')) ? (int)$tagsList->item($i)->getAttribute('readability') : 0;
-                /**
-                * Gather counts for other typical elements embedded within.
+            $this->dbg('Cleaning Conditionally ' . $tagsList->item($i)->tagName . ' (' . $tagsList->item($i)->getAttribute('class') . ':' . $tagsList->item($i)->getAttribute('id') . ')' . (($tagsList->item($i)->hasAttribute('readability')) ? (' with score ' . $tagsList->item($i)->getAttribute('readability')) : ''));
-                * Traverse backwards so we can remove nodes at the same time without effecting the traversal.
-                *
+            if ($weight + $contentScore < 0) {
-                * TODO: Consider taking into account original contentScore here.
+                $tagsList->item($i)->parentNode->removeChild($tagsList->item($i));
-                */
+            }
-                for ($i=$curTagsLength-1; $i >= 0; $i--) {
+            else if ( $this->getCharCount($tagsList->item($i), ',') < 10) {
-                        $weight = $this->getClassWeight($tagsList->item($i));
+                /**
-                        $contentScore = ($tagsList->item($i)->hasAttribute('readability')) ? (int)$tagsList->item($i)->getAttribute('readability') : 0;
+                * If there are not very many commas, and the number of
-                        
+                * non-paragraph elements is more than paragraphs or other ominous signs, remove the element.
-                        $this->dbg('Cleaning Conditionally ' . $tagsList->item($i)->tagName . ' (' . $tagsList->item($i)->getAttribute('class') . ':' . $tagsList->item($i)->getAttribute('id') . ')' . (($tagsList->item($i)->hasAttribute('readability')) ? (' with score ' . $tagsList->item($i)->getAttribute('readability')) : ''));
+                **/
+                $p      = $tagsList->item($i)->getElementsByTagName('p')->length;
-                        if ($weight + $contentScore < 0) {
+                $img    = $tagsList->item($i)->getElementsByTagName('img')->length;
-                                $tagsList->item($i)->parentNode->removeChild($tagsList->item($i));
+                $li     = $tagsList->item($i)->getElementsByTagName('li')->length-100;
-                        }
+                $input  = $tagsList->item($i)->getElementsByTagName('input')->length;
-                        else if ( $this->getCharCount($tagsList->item($i), ',') < 10) {
+                $a      = $tagsList->item($i)->getElementsByTagName('a')->length;
-                                /**
-                                * If there are not very many commas, and the number of
+                $embedCount = 0;
-                                * non-paragraph elements is more than paragraphs or other ominous signs, remove the element.
+                $embeds = $tagsList->item($i)->getElementsByTagName('embed');
-                                **/
+                for ($ei=0, $il=$embeds->length; $ei < $il; $ei++) {
-                                $p      = $tagsList->item($i)->getElementsByTagName('p')->length;
+                    if (preg_match($this->regexps['video'], $embeds->item($ei)->getAttribute('src'))) {
-                                $img    = $tagsList->item($i)->getElementsByTagName('img')->length;
+                        $embedCount++;
-                                $li     = $tagsList->item($i)->getElementsByTagName('li')->length-100;
+                    }
-                                $input  = $tagsList->item($i)->getElementsByTagName('input')->length;
+                }
+                $embeds = $tagsList->item($i)->getElementsByTagName('iframe');
-                                $embedCount = 0;
+                for ($ei=0, $il=$embeds->length; $ei < $il; $ei++) {
-                                $embeds = $tagsList->item($i)->getElementsByTagName('embed');
+                    if (preg_match($this->regexps['video'], $embeds->item($ei)->getAttribute('src'))) {
-                                for ($ei=0, $il=$embeds->length; $ei < $il; $ei++) {
+                        $embedCount++;
-                                        if (preg_match($this->regexps['video'], $embeds->item($ei)->getAttribute('src'))) {
+                    }
-                                        $embedCount++; 
+                }
-                                        }
-                                }
+                $linkDensity   = $this->getLinkDensity($tagsList->item($i));
+                $contentLength = strlen($this->getInnerText($tagsList->item($i)));
-                                $linkDensity   = $this->getLinkDensity($tagsList->item($i));
+                $toRemove      = false;
-                                $contentLength = strlen($this->getInnerText($tagsList->item($i)));
-                                $toRemove      = false;
+                if ($this->lightClean) {
+                    $this->dbg('Light clean...');
-                                if ( $img > $p ) {
+                    if ( ($img > $p) && ($img > 4) ) {
-                                        $toRemove = true;
+                        $this->dbg(' more than 4 images and more image elements than paragraph elements');
-                                } else if ($li > $p && $tag != 'ul' && $tag != 'ol') {
+                        $toRemove = true;
-                                        $toRemove = true;
+                    } else if ($li > $p && $tag != 'ul' && $tag != 'ol') {
-                                } else if ( $input > floor($p/3) ) {
+                        $this->dbg(' too many <li> elements, and parent is not <ul> or <ol>');
-                                        $toRemove = true; 
+                        $toRemove = true;
-                                } else if ($contentLength < 25 && ($img === 0 || $img > 2) ) {
+                    } else if ( $input > floor($p/3) ) {
-                                        $toRemove = true;
+                        $this->dbg(' too many <input> elements');
-                                } else if($weight < 25 && $linkDensity > 0.2) {
+                        $toRemove = true;
-                                        $toRemove = true;
+                    } else if ($contentLength < 25 && ($embedCount === 0 && ($img === 0 || $img > 2))) {
-                                } else if($weight >= 25 && $linkDensity > 0.5) {
+                        $this->dbg(' content length less than 25 chars, 0 embeds and either 0 images or more than 2 images');
-                                        $toRemove = true;
+                        $toRemove = true;
-                                } else if(($embedCount == 1 && $contentLength < 75) || $embedCount > 1) {
+                    } else if($weight < 25 && $linkDensity > 0.2) {
-                                        $toRemove = true;
+                        $this->dbg(' weight smaller than 25 and link density above 0.2');
-                                }
+                        $toRemove = true;
+                    } else if($a > 2 && ($weight >= 25 && $linkDensity > 0.5)) {
-                                if ($toRemove) {
+                        $this->dbg(' more than 2 links and weight above 25 but link density greater than 0.5');
-                                        $tagsList->item($i)->parentNode->removeChild($tagsList->item($i));
+                        $toRemove = true;
-                                }
+                    } else if($embedCount > 3) {
-                        }
+                        $this->dbg(' more than 3 embeds');
-                }
+                        $toRemove = true;
-        }
+                    }
+                } else {
-        /**
+                    $this->dbg('Standard clean...');
-        * Clean out spurious headers from an Element. Checks things like classnames and link density.
+                    if ( $img > $p ) {
-        *
+                        $this->dbg(' more image elements than paragraph elements');
-        * @param DOMElement $e
+                        $toRemove = true;
-        * @return void
+                    } else if ($li > $p && $tag != 'ul' && $tag != 'ol') {
-        */
+                        $this->dbg(' too many <li> elements, and parent is not <ul> or <ol>');
-        public function cleanHeaders($e) {
+                        $toRemove = true;
-                for ($headerIndex = 1; $headerIndex < 3; $headerIndex++) {
+                    } else if ( $input > floor($p/3) ) {
-                        $headers = $e->getElementsByTagName('h' . $headerIndex);
+                        $this->dbg(' too many <input> elements');
-                        for ($i=$headers->length-1; $i >=0; $i--) {
+                        $toRemove = true;
-                                if ($this->getClassWeight($headers->item($i)) < 0 || $this->getLinkDensity($headers->item($i)) > 0.33) {
+                    } else if ($contentLength < 25 && ($img === 0 || $img > 2) ) {
-                                        $headers->item($i)->parentNode->removeChild($headers->item($i));
+                        $this->dbg(' content length less than 25 chars and 0 images, or more than 2 images');
-                                }
+                        $toRemove = true;
-                        }
+                    } else if($weight < 25 && $linkDensity > 0.2) {
-                }
+                        $this->dbg(' weight smaller than 25 and link density above 0.2');
-        }
+                        $toRemove = true;
+                    } else if($weight >= 25 && $linkDensity > 0.5) {
-        public function flagIsActive($flag) {
+                        $this->dbg(' weight above 25 but link density greater than 0.5');
-                return ($this->flags & $flag) > 0;
+                        $toRemove = true;
-        }
+                    } else if(($embedCount == 1 && $contentLength < 75) || $embedCount > 1) {
-        
+                        $this->dbg(' 1 embed and content length smaller than 75 chars, or more than one embed');
-        public function addFlag($flag) {
+                        $toRemove = true;
-                $this->flags = $this->flags | $flag;
+                    }
-        }
+                }
-        
-        public function removeFlag($flag) {
+                if ($toRemove) {
-                $this->flags = $this->flags & ~$flag;
+                    //$this->dbg('Removing: '.$tagsList->item($i)->innerHTML);
-        }
+                    $tagsList->item($i)->parentNode->removeChild($tagsList->item($i));
-}
+                }
-?>
-\ No newline at end of file
+            }
+        }
+    }
+    /**
+    * Clean out spurious headers from an Element. Checks things like classnames and link density.
+    *
+    * @param DOMElement $e
+    * @return void
+    */
+    public function cleanHeaders($e) {
+        for ($headerIndex = 1; $headerIndex < 3; $headerIndex++) {
+            $headers = $e->getElementsByTagName('h' . $headerIndex);
+            for ($i=$headers->length-1; $i >=0; $i--) {
+                if ($this->getClassWeight($headers->item($i)) < 0 || $this->getLinkDensity($headers->item($i)) > 0.33) {
+                    $headers->item($i)->parentNode->removeChild($headers->item($i));
+                }
+            }
+        }
+    }
+    public function flagIsActive($flag) {
+        return ($this->flags & $flag) > 0;
+    }
+    public function addFlag($flag) {
+        $this->flags = $this->flags | $flag;
+    }
+    public function removeFlag($flag) {
+        $this->flags = $this->flags & ~$flag;
+    }
+}
+\ No newline at end of file
author	nicosomb <nicolas@loeuillet.org>	2013-04-23 15:57:23 +0200
committer	nicosomb <nicolas@loeuillet.org>	2013-04-23 15:57:23 +0200
commit	0ace6cab0b934b7b8e4d4de4acca16294b7b12df (patch)
tree	db5a4314ff13882753bc947ea0f9ee2168179f23 /inc
parent	c1c9f252f7f0133903dc4143a0db8ed8104f8733 (diff)
download	wallabag-0ace6cab0b934b7b8e4d4de4acca16294b7b12df.tar.gz wallabag-0ace6cab0b934b7b8e4d4de4acca16294b7b12df.tar.zst wallabag-0ace6cab0b934b7b8e4d4de4acca16294b7b12df.zip