aboutsummaryrefslogtreecommitdiffhomepage
path: root/inc/3rdparty/libraries/readability
diff options
context:
space:
mode:
authorNicolas LÅ“uillet <nicolas@loeuillet.org>2014-05-29 18:54:06 +0200
committerNicolas LÅ“uillet <nicolas@loeuillet.org>2014-05-29 18:54:06 +0200
commita9f5e572dde4f986a498d2fbe92a38a1b22f9595 (patch)
tree80b5bfc9836ae92cc4929a4d72ae0b2730e568bc /inc/3rdparty/libraries/readability
parent96834a47b09985e1c82b82857fc108f20e8b8f2b (diff)
parent8038b38802769031e050c753fc0a388a2276629e (diff)
downloadwallabag-a9f5e572dde4f986a498d2fbe92a38a1b22f9595.tar.gz
wallabag-a9f5e572dde4f986a498d2fbe92a38a1b22f9595.tar.zst
wallabag-a9f5e572dde4f986a498d2fbe92a38a1b22f9595.zip
Merge pull request #712 from wallabag/dev1.7.0
1.7, call me "Premium version"
Diffstat (limited to 'inc/3rdparty/libraries/readability')
-rw-r--r--inc/3rdparty/libraries/readability/Readability.php2274
1 files changed, 1137 insertions, 1137 deletions
diff --git a/inc/3rdparty/libraries/readability/Readability.php b/inc/3rdparty/libraries/readability/Readability.php
index 2e8991cc..d0f09d74 100644
--- a/inc/3rdparty/libraries/readability/Readability.php
+++ b/inc/3rdparty/libraries/readability/Readability.php
@@ -1,1138 +1,1138 @@
1<?php 1<?php
2/** 2/**
3* Arc90's Readability ported to PHP for FiveFilters.org 3* Arc90's Readability ported to PHP for FiveFilters.org
4* Based on readability.js version 1.7.1 (without multi-page support) 4* Based on readability.js version 1.7.1 (without multi-page support)
5* Updated to allow HTML5 parsing with html5lib 5* Updated to allow HTML5 parsing with html5lib
6* Updated with lightClean mode to preserve more images and youtube/vimeo/viddler embeds 6* Updated with lightClean mode to preserve more images and youtube/vimeo/viddler embeds
7* ------------------------------------------------------ 7* ------------------------------------------------------
8* Original URL: http://lab.arc90.com/experiments/readability/js/readability.js 8* Original URL: http://lab.arc90.com/experiments/readability/js/readability.js
9* Arc90's project URL: http://lab.arc90.com/experiments/readability/ 9* Arc90's project URL: http://lab.arc90.com/experiments/readability/
10* JS Source: http://code.google.com/p/arc90labs-readability 10* JS Source: http://code.google.com/p/arc90labs-readability
11* Ported by: Keyvan Minoukadeh, http://www.keyvan.net 11* Ported by: Keyvan Minoukadeh, http://www.keyvan.net
12* More information: http://fivefilters.org/content-only/ 12* More information: http://fivefilters.org/content-only/
13* License: Apache License, Version 2.0 13* License: Apache License, Version 2.0
14* Requires: PHP5 14* Requires: PHP5
15* Date: 2012-09-19 15* Date: 2012-09-19
16* 16*
17* Differences between the PHP port and the original 17* Differences between the PHP port and the original
18* ------------------------------------------------------ 18* ------------------------------------------------------
19* Arc90's Readability is designed to run in the browser. It works on the DOM 19* Arc90's Readability is designed to run in the browser. It works on the DOM
20* tree (the parsed HTML) after the page's CSS styles have been applied and 20* tree (the parsed HTML) after the page's CSS styles have been applied and
21* Javascript code executed. This PHP port does not run inside a browser. 21* Javascript code executed. This PHP port does not run inside a browser.
22* We use PHP's ability to parse HTML to build our DOM tree, but we cannot 22* We use PHP's ability to parse HTML to build our DOM tree, but we cannot
23* rely on CSS or Javascript support. As such, the results will not always 23* rely on CSS or Javascript support. As such, the results will not always
24* match Arc90's Readability. (For example, if a web page contains CSS style 24* match Arc90's Readability. (For example, if a web page contains CSS style
25* rules or Javascript code which hide certain HTML elements from display, 25* rules or Javascript code which hide certain HTML elements from display,
26* Arc90's Readability will dismiss those from consideration but our PHP port, 26* Arc90's Readability will dismiss those from consideration but our PHP port,
27* unable to understand CSS or Javascript, will not know any better.) 27* unable to understand CSS or Javascript, will not know any better.)
28* 28*
29* Another significant difference is that the aim of Arc90's Readability is 29* Another significant difference is that the aim of Arc90's Readability is
30* to re-present the main content block of a given web page so users can 30* to re-present the main content block of a given web page so users can
31* read it more easily in their browsers. Correct identification, clean up, 31* read it more easily in their browsers. Correct identification, clean up,
32* and separation of the content block is only a part of this process. 32* and separation of the content block is only a part of this process.
33* This PHP port is only concerned with this part, it does not include code 33* This PHP port is only concerned with this part, it does not include code
34* that relates to presentation in the browser - Arc90 already do 34* that relates to presentation in the browser - Arc90 already do
35* that extremely well, and for PDF output there's FiveFilters.org's 35* that extremely well, and for PDF output there's FiveFilters.org's
36* PDF Newspaper: http://fivefilters.org/pdf-newspaper/. 36* PDF Newspaper: http://fivefilters.org/pdf-newspaper/.
37* 37*
38* Finally, this class contains methods that might be useful for developers 38* Finally, this class contains methods that might be useful for developers
39* working on HTML document fragments. So without deviating too much from 39* working on HTML document fragments. So without deviating too much from
40* the original code (which I don't want to do because it makes debugging 40* the original code (which I don't want to do because it makes debugging
41* and updating more difficult), I've tried to make it a little more 41* and updating more difficult), I've tried to make it a little more
42* developer friendly. You should be able to use the methods here on 42* developer friendly. You should be able to use the methods here on
43* existing DOMElement objects without passing an entire HTML document to 43* existing DOMElement objects without passing an entire HTML document to
44* be parsed. 44* be parsed.
45*/ 45*/
46 46
47// This class allows us to do JavaScript like assignements to innerHTML 47// This class allows us to do JavaScript like assignements to innerHTML
48require_once(dirname(__FILE__).'/JSLikeHTMLElement.php'); 48require_once(dirname(__FILE__).'/JSLikeHTMLElement.php');
49 49
50// Alternative usage (for testing only!) 50// Alternative usage (for testing only!)
51// uncomment the lines below and call Readability.php in your browser 51// uncomment the lines below and call Readability.php in your browser
52// passing it the URL of the page you'd like content from, e.g.: 52// passing it the URL of the page you'd like content from, e.g.:
53// Readability.php?url=http://medialens.org/alerts/09/090615_the_guardian_climate.php 53// Readability.php?url=http://medialens.org/alerts/09/090615_the_guardian_climate.php
54 54
55/* 55/*
56if (!isset($_GET['url']) || $_GET['url'] == '') { 56if (!isset($_GET['url']) || $_GET['url'] == '') {
57 die('Please pass a URL to the script. E.g. Readability.php?url=bla.com/story.html'); 57 die('Please pass a URL to the script. E.g. Readability.php?url=bla.com/story.html');
58} 58}
59$url = $_GET['url']; 59$url = $_GET['url'];
60if (!preg_match('!^https?://!i', $url)) $url = 'http://'.$url; 60if (!preg_match('!^https?://!i', $url)) $url = 'http://'.$url;
61$html = file_get_contents($url); 61$html = file_get_contents($url);
62$r = new Readability($html, $url); 62$r = new Readability($html, $url);
63$r->init(); 63$r->init();
64echo $r->articleContent->innerHTML; 64echo $r->articleContent->innerHTML;
65*/ 65*/
66 66
67class Readability 67class Readability
68{ 68{
69 public $version = '1.7.1-without-multi-page'; 69 public $version = '1.7.1-without-multi-page';
70 public $convertLinksToFootnotes = false; 70 public $convertLinksToFootnotes = false;
71 public $revertForcedParagraphElements = true; 71 public $revertForcedParagraphElements = true;
72 public $articleTitle; 72 public $articleTitle;
73 public $articleContent; 73 public $articleContent;
74 public $dom; 74 public $dom;
75 public $url = null; // optional - URL where HTML was retrieved 75 public $url = null; // optional - URL where HTML was retrieved
76 public $debug = false; 76 public $debug = false;
77 public $lightClean = true; // preserves more content (experimental) added 2012-09-19 77 public $lightClean = true; // preserves more content (experimental) added 2012-09-19
78 protected $body = null; // 78 protected $body = null; //
79 protected $bodyCache = null; // Cache the body HTML in case we need to re-use it later 79 protected $bodyCache = null; // Cache the body HTML in case we need to re-use it later
80 protected $flags = 7; // 1 | 2 | 4; // Start with all flags set. 80 protected $flags = 7; // 1 | 2 | 4; // Start with all flags set.
81 protected $success = false; // indicates whether we were able to extract or not 81 protected $success = false; // indicates whether we were able to extract or not
82 82
83 /** 83 /**
84 * All of the regular expressions in use within readability. 84 * All of the regular expressions in use within readability.
85 * Defined up here so we don't instantiate them repeatedly in loops. 85 * Defined up here so we don't instantiate them repeatedly in loops.
86 **/ 86 **/
87 public $regexps = array( 87 public $regexps = array(
88 'unlikelyCandidates' => '/combx|comment|community|disqus|extra|foot|header|menu|remark|rss|shoutbox|sidebar|sponsor|ad-break|agegate|pagination|pager|popup/i', 88 'unlikelyCandidates' => '/combx|comment|community|disqus|extra|foot|header|menu|remark|rss|shoutbox|sidebar|sponsor|ad-break|agegate|pagination|pager|popup/i',
89 'okMaybeItsACandidate' => '/and|article|body|column|main|shadow/i', 89 'okMaybeItsACandidate' => '/and|article|body|column|main|shadow/i',
90 'positive' => '/article|body|content|entry|hentry|main|page|attachment|pagination|post|text|blog|story/i', 90 'positive' => '/article|body|content|entry|hentry|main|page|attachment|pagination|post|text|blog|story/i',
91 'negative' => '/combx|comment|com-|contact|foot|footer|_nav|footnote|masthead|media|meta|outbrain|promo|related|scroll|shoutbox|sidebar|sponsor|shopping|tags|tool|widget/i', 91 'negative' => '/combx|comment|com-|contact|foot|footer|_nav|footnote|masthead|media|meta|outbrain|promo|related|scroll|shoutbox|sidebar|sponsor|shopping|tags|tool|widget/i',
92 'divToPElements' => '/<(a|blockquote|dl|div|img|ol|p|pre|table|ul)/i', 92 'divToPElements' => '/<(a|blockquote|dl|div|img|ol|p|pre|table|ul)/i',
93 'replaceBrs' => '/(<br[^>]*>[ \n\r\t]*){2,}/i', 93 'replaceBrs' => '/(<br[^>]*>[ \n\r\t]*){2,}/i',
94 'replaceFonts' => '/<(\/?)font[^>]*>/i', 94 'replaceFonts' => '/<(\/?)font[^>]*>/i',
95 // 'trimRe' => '/^\s+|\s+$/g', // PHP has trim() 95 // 'trimRe' => '/^\s+|\s+$/g', // PHP has trim()
96 'normalize' => '/\s{2,}/', 96 'normalize' => '/\s{2,}/',
97 'killBreaks' => '/(<br\s*\/?>(\s|&nbsp;?)*){1,}/', 97 'killBreaks' => '/(<br\s*\/?>(\s|&nbsp;?)*){1,}/',
98 'video' => '!//(player\.|www\.)?(youtube|vimeo|viddler)\.com!i', 98 'video' => '!//(player\.|www\.)?(youtube|vimeo|viddler)\.com!i',
99 'skipFootnoteLink' => '/^\s*(\[?[a-z0-9]{1,2}\]?|^|edit|citation needed)\s*$/i' 99 'skipFootnoteLink' => '/^\s*(\[?[a-z0-9]{1,2}\]?|^|edit|citation needed)\s*$/i'
100 ); 100 );
101 101
102 /* constants */ 102 /* constants */
103 const FLAG_STRIP_UNLIKELYS = 1; 103 const FLAG_STRIP_UNLIKELYS = 1;
104 const FLAG_WEIGHT_CLASSES = 2; 104 const FLAG_WEIGHT_CLASSES = 2;
105 const FLAG_CLEAN_CONDITIONALLY = 4; 105 const FLAG_CLEAN_CONDITIONALLY = 4;
106 106
107 /** 107 /**
108 * Create instance of Readability 108 * Create instance of Readability
109 * @param string UTF-8 encoded string 109 * @param string UTF-8 encoded string
110 * @param string (optional) URL associated with HTML (used for footnotes) 110 * @param string (optional) URL associated with HTML (used for footnotes)
111 * @param string which parser to use for turning raw HTML into a DOMDocument (either 'libxml' or 'html5lib') 111 * @param string which parser to use for turning raw HTML into a DOMDocument (either 'libxml' or 'html5lib')
112 */ 112 */
113 function __construct($html, $url=null, $parser='libxml') 113 function __construct($html, $url=null, $parser='libxml')
114 { 114 {
115 $this->url = $url; 115 $this->url = $url;
116 /* Turn all double br's into p's */ 116 /* Turn all double br's into p's */
117 $html = preg_replace($this->regexps['replaceBrs'], '</p><p>', $html); 117 $html = preg_replace($this->regexps['replaceBrs'], '</p><p>', $html);
118 $html = preg_replace($this->regexps['replaceFonts'], '<$1span>', $html); 118 $html = preg_replace($this->regexps['replaceFonts'], '<$1span>', $html);
119 $html = mb_convert_encoding($html, 'HTML-ENTITIES', "UTF-8"); 119 $html = mb_convert_encoding($html, 'HTML-ENTITIES', "UTF-8");
120 if (trim($html) == '') $html = '<html></html>'; 120 if (trim($html) == '') $html = '<html></html>';
121 if ($parser=='html5lib' && ($this->dom = HTML5_Parser::parse($html))) { 121 if ($parser=='html5lib' && ($this->dom = HTML5_Parser::parse($html))) {
122 // all good 122 // all good
123 } else { 123 } else {
124 $this->dom = new DOMDocument(); 124 $this->dom = new DOMDocument();
125 $this->dom->preserveWhiteSpace = false; 125 $this->dom->preserveWhiteSpace = false;
126 @$this->dom->loadHTML($html); 126 @$this->dom->loadHTML($html);
127 } 127 }
128 $this->dom->registerNodeClass('DOMElement', 'JSLikeHTMLElement'); 128 $this->dom->registerNodeClass('DOMElement', 'JSLikeHTMLElement');
129 } 129 }
130 130
131 /** 131 /**
132 * Get article title element 132 * Get article title element
133 * @return DOMElement 133 * @return DOMElement
134 */ 134 */
135 public function getTitle() { 135 public function getTitle() {
136 return $this->articleTitle; 136 return $this->articleTitle;
137 } 137 }
138 138
139 /** 139 /**
140 * Get article content element 140 * Get article content element
141 * @return DOMElement 141 * @return DOMElement
142 */ 142 */
143 public function getContent() { 143 public function getContent() {
144 return $this->articleContent; 144 return $this->articleContent;
145 } 145 }
146 146
147 /** 147 /**
148 * Runs readability. 148 * Runs readability.
149 * 149 *
150 * Workflow: 150 * Workflow:
151 * 1. Prep the document by removing script tags, css, etc. 151 * 1. Prep the document by removing script tags, css, etc.
152 * 2. Build readability's DOM tree. 152 * 2. Build readability's DOM tree.
153 * 3. Grab the article content from the current dom tree. 153 * 3. Grab the article content from the current dom tree.
154 * 4. Replace the current DOM tree with the new one. 154 * 4. Replace the current DOM tree with the new one.
155 * 5. Read peacefully. 155 * 5. Read peacefully.
156 * 156 *
157 * @return boolean true if we found content, false otherwise 157 * @return boolean true if we found content, false otherwise
158 **/ 158 **/
159 public function init() 159 public function init()
160 { 160 {
161 if (!isset($this->dom->documentElement)) return false; 161 if (!isset($this->dom->documentElement)) return false;
162 $this->removeScripts($this->dom); 162 $this->removeScripts($this->dom);
163 //die($this->getInnerHTML($this->dom->documentElement)); 163 //die($this->getInnerHTML($this->dom->documentElement));
164 164
165 // Assume successful outcome 165 // Assume successful outcome
166 $this->success = true; 166 $this->success = true;
167 167
168 $bodyElems = $this->dom->getElementsByTagName('body'); 168 $bodyElems = $this->dom->getElementsByTagName('body');
169 if ($bodyElems->length > 0) { 169 if ($bodyElems->length > 0) {
170 if ($this->bodyCache == null) { 170 if ($this->bodyCache == null) {
171 $this->bodyCache = $bodyElems->item(0)->innerHTML; 171 $this->bodyCache = $bodyElems->item(0)->innerHTML;
172 } 172 }
173 if ($this->body == null) { 173 if ($this->body == null) {
174 $this->body = $bodyElems->item(0); 174 $this->body = $bodyElems->item(0);
175 } 175 }
176 } 176 }
177 177
178 $this->prepDocument(); 178 $this->prepDocument();
179 179
180 //die($this->dom->documentElement->parentNode->nodeType); 180 //die($this->dom->documentElement->parentNode->nodeType);
181 //$this->setInnerHTML($this->dom->documentElement, $this->getInnerHTML($this->dom->documentElement)); 181 //$this->setInnerHTML($this->dom->documentElement, $this->getInnerHTML($this->dom->documentElement));
182 //die($this->getInnerHTML($this->dom->documentElement)); 182 //die($this->getInnerHTML($this->dom->documentElement));
183 183
184 /* Build readability's DOM tree */ 184 /* Build readability's DOM tree */
185 $overlay = $this->dom->createElement('div'); 185 $overlay = $this->dom->createElement('div');
186 $innerDiv = $this->dom->createElement('div'); 186 $innerDiv = $this->dom->createElement('div');
187 $articleTitle = $this->getArticleTitle(); 187 $articleTitle = $this->getArticleTitle();
188 $articleContent = $this->grabArticle(); 188 $articleContent = $this->grabArticle();
189 189
190 if (!$articleContent) { 190 if (!$articleContent) {
191 $this->success = false; 191 $this->success = false;
192 $articleContent = $this->dom->createElement('div'); 192 $articleContent = $this->dom->createElement('div');
193 $articleContent->setAttribute('id', 'readability-content'); 193 $articleContent->setAttribute('id', 'readability-content');
194 $articleContent->innerHTML = '<p>Sorry, Readability was unable to parse this page for content.</p>'; 194 $articleContent->innerHTML = '<p>Sorry, Readability was unable to parse this page for content.</p>';
195 } 195 }
196 196
197 $overlay->setAttribute('id', 'readOverlay'); 197 $overlay->setAttribute('id', 'readOverlay');
198 $innerDiv->setAttribute('id', 'readInner'); 198 $innerDiv->setAttribute('id', 'readInner');
199 199
200 /* Glue the structure of our document together. */ 200 /* Glue the structure of our document together. */
201 $innerDiv->appendChild($articleTitle); 201 $innerDiv->appendChild($articleTitle);
202 $innerDiv->appendChild($articleContent); 202 $innerDiv->appendChild($articleContent);
203 $overlay->appendChild($innerDiv); 203 $overlay->appendChild($innerDiv);
204 204
205 /* Clear the old HTML, insert the new content. */ 205 /* Clear the old HTML, insert the new content. */
206 $this->body->innerHTML = ''; 206 $this->body->innerHTML = '';
207 $this->body->appendChild($overlay); 207 $this->body->appendChild($overlay);
208 //document.body.insertBefore(overlay, document.body.firstChild); 208 //document.body.insertBefore(overlay, document.body.firstChild);
209 $this->body->removeAttribute('style'); 209 $this->body->removeAttribute('style');
210 210
211 $this->postProcessContent($articleContent); 211 $this->postProcessContent($articleContent);
212 212
213 // Set title and content instance variables 213 // Set title and content instance variables
214 $this->articleTitle = $articleTitle; 214 $this->articleTitle = $articleTitle;
215 $this->articleContent = $articleContent; 215 $this->articleContent = $articleContent;
216 216
217 return $this->success; 217 return $this->success;
218 } 218 }
219 219
220 /** 220 /**
221 * Debug 221 * Debug
222 */ 222 */
223 protected function dbg($msg) { 223 protected function dbg($msg) {
224 if ($this->debug) echo '* ',$msg, "\n"; 224 if ($this->debug) echo '* ',$msg, "\n";
225 } 225 }
226 226
227 /** 227 /**
228 * Run any post-process modifications to article content as necessary. 228 * Run any post-process modifications to article content as necessary.
229 * 229 *
230 * @param DOMElement 230 * @param DOMElement
231 * @return void 231 * @return void
232 */ 232 */
233 public function postProcessContent($articleContent) { 233 public function postProcessContent($articleContent) {
234 if ($this->convertLinksToFootnotes && !preg_match('/wikipedia\.org/', @$this->url)) { 234 if ($this->convertLinksToFootnotes && !preg_match('/wikipedia\.org/', @$this->url)) {
235 $this->addFootnotes($articleContent); 235 $this->addFootnotes($articleContent);
236 } 236 }
237 } 237 }
238 238
239 /** 239 /**
240 * Get the article title as an H1. 240 * Get the article title as an H1.
241 * 241 *
242 * @return DOMElement 242 * @return DOMElement
243 */ 243 */
244 protected function getArticleTitle() { 244 protected function getArticleTitle() {
245 $curTitle = ''; 245 $curTitle = '';
246 $origTitle = ''; 246 $origTitle = '';
247 247
248 try { 248 try {
249 $curTitle = $origTitle = $this->getInnerText($this->dom->getElementsByTagName('title')->item(0)); 249 $curTitle = $origTitle = $this->getInnerText($this->dom->getElementsByTagName('title')->item(0));
250 } catch(Exception $e) {} 250 } catch(Exception $e) {}
251 251
252 if (preg_match('/ [\|\-] /', $curTitle)) 252 if (preg_match('/ [\|\-] /', $curTitle))
253 { 253 {
254 $curTitle = preg_replace('/(.*)[\|\-] .*/i', '$1', $origTitle); 254 $curTitle = preg_replace('/(.*)[\|\-] .*/i', '$1', $origTitle);
255 255
256 if (count(explode(' ', $curTitle)) < 3) { 256 if (count(explode(' ', $curTitle)) < 3) {
257 $curTitle = preg_replace('/[^\|\-]*[\|\-](.*)/i', '$1', $origTitle); 257 $curTitle = preg_replace('/[^\|\-]*[\|\-](.*)/i', '$1', $origTitle);
258 } 258 }
259 } 259 }
260 else if (strpos($curTitle, ': ') !== false) 260 else if (strpos($curTitle, ': ') !== false)
261 { 261 {
262 $curTitle = preg_replace('/.*:(.*)/i', '$1', $origTitle); 262 $curTitle = preg_replace('/.*:(.*)/i', '$1', $origTitle);
263 263
264 if (count(explode(' ', $curTitle)) < 3) { 264 if (count(explode(' ', $curTitle)) < 3) {
265 $curTitle = preg_replace('/[^:]*[:](.*)/i','$1', $origTitle); 265 $curTitle = preg_replace('/[^:]*[:](.*)/i','$1', $origTitle);
266 } 266 }
267 } 267 }
268 else if(strlen($curTitle) > 150 || strlen($curTitle) < 15) 268 else if(strlen($curTitle) > 150 || strlen($curTitle) < 15)
269 { 269 {
270 $hOnes = $this->dom->getElementsByTagName('h1'); 270 $hOnes = $this->dom->getElementsByTagName('h1');
271 if($hOnes->length == 1) 271 if($hOnes->length == 1)
272 { 272 {
273 $curTitle = $this->getInnerText($hOnes->item(0)); 273 $curTitle = $this->getInnerText($hOnes->item(0));
274 } 274 }
275 } 275 }
276 276
277 $curTitle = trim($curTitle); 277 $curTitle = trim($curTitle);
278 278
279 if (count(explode(' ', $curTitle)) <= 4) { 279 if (count(explode(' ', $curTitle)) <= 4) {
280 $curTitle = $origTitle; 280 $curTitle = $origTitle;
281 } 281 }
282 282
283 $articleTitle = $this->dom->createElement('h1'); 283 $articleTitle = $this->dom->createElement('h1');
284 $articleTitle->innerHTML = $curTitle; 284 $articleTitle->innerHTML = $curTitle;
285 285
286 return $articleTitle; 286 return $articleTitle;
287 } 287 }
288 288
289 /** 289 /**
290 * Prepare the HTML document for readability to scrape it. 290 * Prepare the HTML document for readability to scrape it.
291 * This includes things like stripping javascript, CSS, and handling terrible markup. 291 * This includes things like stripping javascript, CSS, and handling terrible markup.
292 * 292 *
293 * @return void 293 * @return void
294 **/ 294 **/
295 protected function prepDocument() { 295 protected function prepDocument() {
296 /** 296 /**
297 * In some cases a body element can't be found (if the HTML is totally hosed for example) 297 * In some cases a body element can't be found (if the HTML is totally hosed for example)
298 * so we create a new body node and append it to the document. 298 * so we create a new body node and append it to the document.
299 */ 299 */
300 if ($this->body == null) 300 if ($this->body == null)
301 { 301 {
302 $this->body = $this->dom->createElement('body'); 302 $this->body = $this->dom->createElement('body');
303 $this->dom->documentElement->appendChild($this->body); 303 $this->dom->documentElement->appendChild($this->body);
304 } 304 }
305 $this->body->setAttribute('id', 'readabilityBody'); 305 $this->body->setAttribute('id', 'readabilityBody');
306 306
307 /* Remove all style tags in head */ 307 /* Remove all style tags in head */
308 $styleTags = $this->dom->getElementsByTagName('style'); 308 $styleTags = $this->dom->getElementsByTagName('style');
309 for ($i = $styleTags->length-1; $i >= 0; $i--) 309 for ($i = $styleTags->length-1; $i >= 0; $i--)
310 { 310 {
311 $styleTags->item($i)->parentNode->removeChild($styleTags->item($i)); 311 $styleTags->item($i)->parentNode->removeChild($styleTags->item($i));
312 } 312 }
313 313
314 /* Turn all double br's into p's */ 314 /* Turn all double br's into p's */
315 /* Note, this is pretty costly as far as processing goes. Maybe optimize later. */ 315 /* Note, this is pretty costly as far as processing goes. Maybe optimize later. */
316 //document.body.innerHTML = document.body.innerHTML.replace(readability.regexps.replaceBrs, '</p><p>').replace(readability.regexps.replaceFonts, '<$1span>'); 316 //document.body.innerHTML = document.body.innerHTML.replace(readability.regexps.replaceBrs, '</p><p>').replace(readability.regexps.replaceFonts, '<$1span>');
317 // We do this in the constructor for PHP as that's when we have raw HTML - before parsing it into a DOM tree. 317 // We do this in the constructor for PHP as that's when we have raw HTML - before parsing it into a DOM tree.
318 // Manipulating innerHTML as it's done in JS is not possible in PHP. 318 // Manipulating innerHTML as it's done in JS is not possible in PHP.
319 } 319 }
320 320
321 /** 321 /**
322 * For easier reading, convert this document to have footnotes at the bottom rather than inline links. 322 * For easier reading, convert this document to have footnotes at the bottom rather than inline links.
323 * @see http://www.roughtype.com/archives/2010/05/experiments_in.php 323 * @see http://www.roughtype.com/archives/2010/05/experiments_in.php
324 * 324 *
325 * @return void 325 * @return void
326 **/ 326 **/
327 public function addFootnotes($articleContent) { 327 public function addFootnotes($articleContent) {
328 $footnotesWrapper = $this->dom->createElement('div'); 328 $footnotesWrapper = $this->dom->createElement('div');
329 $footnotesWrapper->setAttribute('id', 'readability-footnotes'); 329 $footnotesWrapper->setAttribute('id', 'readability-footnotes');
330 $footnotesWrapper->innerHTML = '<h3>References</h3>'; 330 $footnotesWrapper->innerHTML = '<h3>References</h3>';
331 331
332 $articleFootnotes = $this->dom->createElement('ol'); 332 $articleFootnotes = $this->dom->createElement('ol');
333 $articleFootnotes->setAttribute('id', 'readability-footnotes-list'); 333 $articleFootnotes->setAttribute('id', 'readability-footnotes-list');
334 $footnotesWrapper->appendChild($articleFootnotes); 334 $footnotesWrapper->appendChild($articleFootnotes);
335 335
336 $articleLinks = $articleContent->getElementsByTagName('a'); 336 $articleLinks = $articleContent->getElementsByTagName('a');
337 337
338 $linkCount = 0; 338 $linkCount = 0;
339 for ($i = 0; $i < $articleLinks->length; $i++) 339 for ($i = 0; $i < $articleLinks->length; $i++)
340 { 340 {
341 $articleLink = $articleLinks->item($i); 341 $articleLink = $articleLinks->item($i);
342 $footnoteLink = $articleLink->cloneNode(true); 342 $footnoteLink = $articleLink->cloneNode(true);
343 $refLink = $this->dom->createElement('a'); 343 $refLink = $this->dom->createElement('a');
344 $footnote = $this->dom->createElement('li'); 344 $footnote = $this->dom->createElement('li');
345 $linkDomain = @parse_url($footnoteLink->getAttribute('href'), PHP_URL_HOST); 345 $linkDomain = @parse_url($footnoteLink->getAttribute('href'), PHP_URL_HOST);
346 if (!$linkDomain && isset($this->url)) $linkDomain = @parse_url($this->url, PHP_URL_HOST); 346 if (!$linkDomain && isset($this->url)) $linkDomain = @parse_url($this->url, PHP_URL_HOST);
347 //linkDomain = footnoteLink.host ? footnoteLink.host : document.location.host, 347 //linkDomain = footnoteLink.host ? footnoteLink.host : document.location.host,
348 $linkText = $this->getInnerText($articleLink); 348 $linkText = $this->getInnerText($articleLink);
349 349
350 if ((strpos($articleLink->getAttribute('class'), 'readability-DoNotFootnote') !== false) || preg_match($this->regexps['skipFootnoteLink'], $linkText)) { 350 if ((strpos($articleLink->getAttribute('class'), 'readability-DoNotFootnote') !== false) || preg_match($this->regexps['skipFootnoteLink'], $linkText)) {
351 continue; 351 continue;
352 } 352 }
353 353
354 $linkCount++; 354 $linkCount++;
355 355
356 /** Add a superscript reference after the article link */ 356 /** Add a superscript reference after the article link */
357 $refLink->setAttribute('href', '#readabilityFootnoteLink-' . $linkCount); 357 $refLink->setAttribute('href', '#readabilityFootnoteLink-' . $linkCount);
358 $refLink->innerHTML = '<small><sup>[' . $linkCount . ']</sup></small>'; 358 $refLink->innerHTML = '<small><sup>[' . $linkCount . ']</sup></small>';
359 $refLink->setAttribute('class', 'readability-DoNotFootnote'); 359 $refLink->setAttribute('class', 'readability-DoNotFootnote');
360 $refLink->setAttribute('style', 'color: inherit;'); 360 $refLink->setAttribute('style', 'color: inherit;');
361 361
362 //TODO: does this work or should we use DOMNode.isSameNode()? 362 //TODO: does this work or should we use DOMNode.isSameNode()?
363 if ($articleLink->parentNode->lastChild == $articleLink) { 363 if ($articleLink->parentNode->lastChild == $articleLink) {
364 $articleLink->parentNode->appendChild($refLink); 364 $articleLink->parentNode->appendChild($refLink);
365 } else { 365 } else {
366 $articleLink->parentNode->insertBefore($refLink, $articleLink->nextSibling); 366 $articleLink->parentNode->insertBefore($refLink, $articleLink->nextSibling);
367 } 367 }
368 368
369 $articleLink->setAttribute('style', 'color: inherit; text-decoration: none;'); 369 $articleLink->setAttribute('style', 'color: inherit; text-decoration: none;');
370 $articleLink->setAttribute('name', 'readabilityLink-' . $linkCount); 370 $articleLink->setAttribute('name', 'readabilityLink-' . $linkCount);
371 371
372 $footnote->innerHTML = '<small><sup><a href="#readabilityLink-' . $linkCount . '" title="Jump to Link in Article">^</a></sup></small> '; 372 $footnote->innerHTML = '<small><sup><a href="#readabilityLink-' . $linkCount . '" title="Jump to Link in Article">^</a></sup></small> ';
373 373
374 $footnoteLink->innerHTML = ($footnoteLink->getAttribute('title') != '' ? $footnoteLink->getAttribute('title') : $linkText); 374 $footnoteLink->innerHTML = ($footnoteLink->getAttribute('title') != '' ? $footnoteLink->getAttribute('title') : $linkText);
375 $footnoteLink->setAttribute('name', 'readabilityFootnoteLink-' . $linkCount); 375 $footnoteLink->setAttribute('name', 'readabilityFootnoteLink-' . $linkCount);
376 376
377 $footnote->appendChild($footnoteLink); 377 $footnote->appendChild($footnoteLink);
378 if ($linkDomain) $footnote->innerHTML = $footnote->innerHTML . '<small> (' . $linkDomain . ')</small>'; 378 if ($linkDomain) $footnote->innerHTML = $footnote->innerHTML . '<small> (' . $linkDomain . ')</small>';
379 379
380 $articleFootnotes->appendChild($footnote); 380 $articleFootnotes->appendChild($footnote);
381 } 381 }
382 382
383 if ($linkCount > 0) { 383 if ($linkCount > 0) {
384 $articleContent->appendChild($footnotesWrapper); 384 $articleContent->appendChild($footnotesWrapper);
385 } 385 }
386 } 386 }
387 387
388 /** 388 /**
389 * Reverts P elements with class 'readability-styled' 389 * Reverts P elements with class 'readability-styled'
390 * to text nodes - which is what they were before. 390 * to text nodes - which is what they were before.
391 * 391 *
392 * @param DOMElement 392 * @param DOMElement
393 * @return void 393 * @return void
394 */ 394 */
395 function revertReadabilityStyledElements($articleContent) { 395 function revertReadabilityStyledElements($articleContent) {
396 $xpath = new DOMXPath($articleContent->ownerDocument); 396 $xpath = new DOMXPath($articleContent->ownerDocument);
397 $elems = $xpath->query('.//p[@class="readability-styled"]', $articleContent); 397 $elems = $xpath->query('.//p[@class="readability-styled"]', $articleContent);
398 //$elems = $articleContent->getElementsByTagName('p'); 398 //$elems = $articleContent->getElementsByTagName('p');
399 for ($i = $elems->length-1; $i >= 0; $i--) { 399 for ($i = $elems->length-1; $i >= 0; $i--) {
400 $e = $elems->item($i); 400 $e = $elems->item($i);
401 $e->parentNode->replaceChild($articleContent->ownerDocument->createTextNode($e->textContent), $e); 401 $e->parentNode->replaceChild($articleContent->ownerDocument->createTextNode($e->textContent), $e);
402 //if ($e->hasAttribute('class') && $e->getAttribute('class') == 'readability-styled') { 402 //if ($e->hasAttribute('class') && $e->getAttribute('class') == 'readability-styled') {
403 // $e->parentNode->replaceChild($this->dom->createTextNode($e->textContent), $e); 403 // $e->parentNode->replaceChild($this->dom->createTextNode($e->textContent), $e);
404 //} 404 //}
405 } 405 }
406 } 406 }
407 407
408 /** 408 /**
409 * Prepare the article node for display. Clean out any inline styles, 409 * Prepare the article node for display. Clean out any inline styles,
410 * iframes, forms, strip extraneous <p> tags, etc. 410 * iframes, forms, strip extraneous <p> tags, etc.
411 * 411 *
412 * @param DOMElement 412 * @param DOMElement
413 * @return void 413 * @return void
414 */ 414 */
415 function prepArticle($articleContent) { 415 function prepArticle($articleContent) {
416 $this->cleanStyles($articleContent); 416 $this->cleanStyles($articleContent);
417 $this->killBreaks($articleContent); 417 $this->killBreaks($articleContent);
418 if ($this->revertForcedParagraphElements) { 418 if ($this->revertForcedParagraphElements) {
419 $this->revertReadabilityStyledElements($articleContent); 419 $this->revertReadabilityStyledElements($articleContent);
420 } 420 }
421 421
422 /* Clean out junk from the article content */ 422 /* Clean out junk from the article content */
423 $this->cleanConditionally($articleContent, 'form'); 423 $this->cleanConditionally($articleContent, 'form');
424 $this->clean($articleContent, 'object'); 424 $this->clean($articleContent, 'object');
425 $this->clean($articleContent, 'h1'); 425 $this->clean($articleContent, 'h1');
426 426
427 /** 427 /**
428 * If there is only one h2, they are probably using it 428 * If there is only one h2, they are probably using it
429 * as a header and not a subheader, so remove it since we already have a header. 429 * as a header and not a subheader, so remove it since we already have a header.
430 ***/ 430 ***/
431 if (!$this->lightClean && ($articleContent->getElementsByTagName('h2')->length == 1)) { 431 if (!$this->lightClean && ($articleContent->getElementsByTagName('h2')->length == 1)) {
432 $this->clean($articleContent, 'h2'); 432 $this->clean($articleContent, 'h2');
433 } 433 }
434 $this->clean($articleContent, 'iframe'); 434 $this->clean($articleContent, 'iframe');
435 435
436 $this->cleanHeaders($articleContent); 436 $this->cleanHeaders($articleContent);
437 437
438 /* Do these last as the previous stuff may have removed junk that will affect these */ 438 /* Do these last as the previous stuff may have removed junk that will affect these */
439 $this->cleanConditionally($articleContent, 'table'); 439 $this->cleanConditionally($articleContent, 'table');
440 $this->cleanConditionally($articleContent, 'ul'); 440 $this->cleanConditionally($articleContent, 'ul');
441 $this->cleanConditionally($articleContent, 'div'); 441 $this->cleanConditionally($articleContent, 'div');
442 442
443 /* Remove extra paragraphs */ 443 /* Remove extra paragraphs */
444 $articleParagraphs = $articleContent->getElementsByTagName('p'); 444 $articleParagraphs = $articleContent->getElementsByTagName('p');
445 for ($i = $articleParagraphs->length-1; $i >= 0; $i--) 445 for ($i = $articleParagraphs->length-1; $i >= 0; $i--)
446 { 446 {
447 $imgCount = $articleParagraphs->item($i)->getElementsByTagName('img')->length; 447 $imgCount = $articleParagraphs->item($i)->getElementsByTagName('img')->length;
448 $embedCount = $articleParagraphs->item($i)->getElementsByTagName('embed')->length; 448 $embedCount = $articleParagraphs->item($i)->getElementsByTagName('embed')->length;
449 $objectCount = $articleParagraphs->item($i)->getElementsByTagName('object')->length; 449 $objectCount = $articleParagraphs->item($i)->getElementsByTagName('object')->length;
450 $iframeCount = $articleParagraphs->item($i)->getElementsByTagName('iframe')->length; 450 $iframeCount = $articleParagraphs->item($i)->getElementsByTagName('iframe')->length;
451 451
452 if ($imgCount === 0 && $embedCount === 0 && $objectCount === 0 && $iframeCount === 0 && $this->getInnerText($articleParagraphs->item($i), false) == '') 452 if ($imgCount === 0 && $embedCount === 0 && $objectCount === 0 && $iframeCount === 0 && $this->getInnerText($articleParagraphs->item($i), false) == '')
453 { 453 {
454 $articleParagraphs->item($i)->parentNode->removeChild($articleParagraphs->item($i)); 454 $articleParagraphs->item($i)->parentNode->removeChild($articleParagraphs->item($i));
455 } 455 }
456 } 456 }
457 457
458 try { 458 try {
459 $articleContent->innerHTML = preg_replace('/<br[^>]*>\s*<p/i', '<p', $articleContent->innerHTML); 459 $articleContent->innerHTML = preg_replace('/<br[^>]*>\s*<p/i', '<p', $articleContent->innerHTML);
460 //articleContent.innerHTML = articleContent.innerHTML.replace(/<br[^>]*>\s*<p/gi, '<p'); 460 //articleContent.innerHTML = articleContent.innerHTML.replace(/<br[^>]*>\s*<p/gi, '<p');
461 } 461 }
462 catch (Exception $e) { 462 catch (Exception $e) {
463 $this->dbg("Cleaning innerHTML of breaks failed. This is an IE strict-block-elements bug. Ignoring.: " . $e); 463 $this->dbg("Cleaning innerHTML of breaks failed. This is an IE strict-block-elements bug. Ignoring.: " . $e);
464 } 464 }
465 } 465 }
466 466
467 /** 467 /**
468 * Initialize a node with the readability object. Also checks the 468 * Initialize a node with the readability object. Also checks the
469 * className/id for special names to add to its score. 469 * className/id for special names to add to its score.
470 * 470 *
471 * @param Element 471 * @param Element
472 * @return void 472 * @return void
473 **/ 473 **/
474 protected function initializeNode($node) { 474 protected function initializeNode($node) {
475 $readability = $this->dom->createAttribute('readability'); 475 $readability = $this->dom->createAttribute('readability');
476 $readability->value = 0; // this is our contentScore 476 $readability->value = 0; // this is our contentScore
477 $node->setAttributeNode($readability); 477 $node->setAttributeNode($readability);
478 478
479 switch (strtoupper($node->tagName)) { // unsure if strtoupper is needed, but using it just in case 479 switch (strtoupper($node->tagName)) { // unsure if strtoupper is needed, but using it just in case
480 case 'DIV': 480 case 'DIV':
481 $readability->value += 5; 481 $readability->value += 5;
482 break; 482 break;
483 483
484 case 'PRE': 484 case 'PRE':
485 case 'TD': 485 case 'TD':
486 case 'BLOCKQUOTE': 486 case 'BLOCKQUOTE':
487 $readability->value += 3; 487 $readability->value += 3;
488 break; 488 break;
489 489
490 case 'ADDRESS': 490 case 'ADDRESS':
491 case 'OL': 491 case 'OL':
492 case 'UL': 492 case 'UL':
493 case 'DL': 493 case 'DL':
494 case 'DD': 494 case 'DD':
495 case 'DT': 495 case 'DT':
496 case 'LI': 496 case 'LI':
497 case 'FORM': 497 case 'FORM':
498 $readability->value -= 3; 498 $readability->value -= 3;
499 break; 499 break;
500 500
501 case 'H1': 501 case 'H1':
502 case 'H2': 502 case 'H2':
503 case 'H3': 503 case 'H3':
504 case 'H4': 504 case 'H4':
505 case 'H5': 505 case 'H5':
506 case 'H6': 506 case 'H6':
507 case 'TH': 507 case 'TH':
508 $readability->value -= 5; 508 $readability->value -= 5;
509 break; 509 break;
510 } 510 }
511 $readability->value += $this->getClassWeight($node); 511 $readability->value += $this->getClassWeight($node);
512 } 512 }
513 513
514 /*** 514 /***
515 * grabArticle - Using a variety of metrics (content score, classname, element types), find the content that is 515 * grabArticle - Using a variety of metrics (content score, classname, element types), find the content that is
516 * most likely to be the stuff a user wants to read. Then return it wrapped up in a div. 516 * most likely to be the stuff a user wants to read. Then return it wrapped up in a div.
517 * 517 *
518 * @return DOMElement 518 * @return DOMElement
519 **/ 519 **/
520 protected function grabArticle($page=null) { 520 protected function grabArticle($page=null) {
521 $stripUnlikelyCandidates = $this->flagIsActive(self::FLAG_STRIP_UNLIKELYS); 521 $stripUnlikelyCandidates = $this->flagIsActive(self::FLAG_STRIP_UNLIKELYS);
522 if (!$page) $page = $this->dom; 522 if (!$page) $page = $this->dom;
523 $allElements = $page->getElementsByTagName('*'); 523 $allElements = $page->getElementsByTagName('*');
524 /** 524 /**
525 * First, node prepping. Trash nodes that look cruddy (like ones with the class name "comment", etc), and turn divs 525 * First, node prepping. Trash nodes that look cruddy (like ones with the class name "comment", etc), and turn divs
526 * into P tags where they have been used inappropriately (as in, where they contain no other block level elements.) 526 * into P tags where they have been used inappropriately (as in, where they contain no other block level elements.)
527 * 527 *
528 * Note: Assignment from index for performance. See http://www.peachpit.com/articles/article.aspx?p=31567&seqNum=5 528 * Note: Assignment from index for performance. See http://www.peachpit.com/articles/article.aspx?p=31567&seqNum=5
529 * TODO: Shouldn't this be a reverse traversal? 529 * TODO: Shouldn't this be a reverse traversal?
530 **/ 530 **/
531 $node = null; 531 $node = null;
532 $nodesToScore = array(); 532 $nodesToScore = array();
533 for ($nodeIndex = 0; ($node = $allElements->item($nodeIndex)); $nodeIndex++) { 533 for ($nodeIndex = 0; ($node = $allElements->item($nodeIndex)); $nodeIndex++) {
534 //for ($nodeIndex=$targetList->length-1; $nodeIndex >= 0; $nodeIndex--) { 534 //for ($nodeIndex=$targetList->length-1; $nodeIndex >= 0; $nodeIndex--) {
535 //$node = $targetList->item($nodeIndex); 535 //$node = $targetList->item($nodeIndex);
536 $tagName = strtoupper($node->tagName); 536 $tagName = strtoupper($node->tagName);
537 /* Remove unlikely candidates */ 537 /* Remove unlikely candidates */
538 if ($stripUnlikelyCandidates) { 538 if ($stripUnlikelyCandidates) {
539 $unlikelyMatchString = $node->getAttribute('class') . $node->getAttribute('id'); 539 $unlikelyMatchString = $node->getAttribute('class') . $node->getAttribute('id');
540 if ( 540 if (
541 preg_match($this->regexps['unlikelyCandidates'], $unlikelyMatchString) && 541 preg_match($this->regexps['unlikelyCandidates'], $unlikelyMatchString) &&
542 !preg_match($this->regexps['okMaybeItsACandidate'], $unlikelyMatchString) && 542 !preg_match($this->regexps['okMaybeItsACandidate'], $unlikelyMatchString) &&
543 $tagName != 'BODY' 543 $tagName != 'BODY'
544 ) 544 )
545 { 545 {
546 $this->dbg('Removing unlikely candidate - ' . $unlikelyMatchString); 546 $this->dbg('Removing unlikely candidate - ' . $unlikelyMatchString);
547 //$nodesToRemove[] = $node; 547 //$nodesToRemove[] = $node;
548 $node->parentNode->removeChild($node); 548 $node->parentNode->removeChild($node);
549 $nodeIndex--; 549 $nodeIndex--;
550 continue; 550 continue;
551 } 551 }
552 } 552 }
553 553
554 if ($tagName == 'P' || $tagName == 'TD' || $tagName == 'PRE') { 554 if ($tagName == 'P' || $tagName == 'TD' || $tagName == 'PRE') {
555 $nodesToScore[] = $node; 555 $nodesToScore[] = $node;
556 } 556 }
557 557
558 /* Turn all divs that don't have children block level elements into p's */ 558 /* Turn all divs that don't have children block level elements into p's */
559 if ($tagName == 'DIV') { 559 if ($tagName == 'DIV') {
560 if (!preg_match($this->regexps['divToPElements'], $node->innerHTML)) { 560 if (!preg_match($this->regexps['divToPElements'], $node->innerHTML)) {
561 //$this->dbg('Altering div to p'); 561 //$this->dbg('Altering div to p');
562 $newNode = $this->dom->createElement('p'); 562 $newNode = $this->dom->createElement('p');
563 try { 563 try {
564 $newNode->innerHTML = $node->innerHTML; 564 $newNode->innerHTML = $node->innerHTML;
565 //$nodesToReplace[] = array('new'=>$newNode, 'old'=>$node); 565 //$nodesToReplace[] = array('new'=>$newNode, 'old'=>$node);
566 $node->parentNode->replaceChild($newNode, $node); 566 $node->parentNode->replaceChild($newNode, $node);
567 $nodeIndex--; 567 $nodeIndex--;
568 $nodesToScore[] = $node; // or $newNode? 568 $nodesToScore[] = $node; // or $newNode?
569 } 569 }
570 catch(Exception $e) { 570 catch(Exception $e) {
571 $this->dbg('Could not alter div to p, reverting back to div.: ' . $e); 571 $this->dbg('Could not alter div to p, reverting back to div.: ' . $e);
572 } 572 }
573 } 573 }
574 else 574 else
575 { 575 {
576 /* EXPERIMENTAL */ 576 /* EXPERIMENTAL */
577 // TODO: change these p elements back to text nodes after processing 577 // TODO: change these p elements back to text nodes after processing
578 for ($i = 0, $il = $node->childNodes->length; $i < $il; $i++) { 578 for ($i = 0, $il = $node->childNodes->length; $i < $il; $i++) {
579 $childNode = $node->childNodes->item($i); 579 $childNode = $node->childNodes->item($i);
580 if ($childNode->nodeType == 3) { // XML_TEXT_NODE 580 if ($childNode->nodeType == 3) { // XML_TEXT_NODE
581 //$this->dbg('replacing text node with a p tag with the same content.'); 581 //$this->dbg('replacing text node with a p tag with the same content.');
582 $p = $this->dom->createElement('p'); 582 $p = $this->dom->createElement('p');
583 $p->innerHTML = $childNode->nodeValue; 583 $p->innerHTML = $childNode->nodeValue;
584 $p->setAttribute('style', 'display: inline;'); 584 $p->setAttribute('style', 'display: inline;');
585 $p->setAttribute('class', 'readability-styled'); 585 $p->setAttribute('class', 'readability-styled');
586 $childNode->parentNode->replaceChild($p, $childNode); 586 $childNode->parentNode->replaceChild($p, $childNode);
587 } 587 }
588 } 588 }
589 } 589 }
590 } 590 }
591 } 591 }
592 592
593 /** 593 /**
594 * Loop through all paragraphs, and assign a score to them based on how content-y they look. 594 * Loop through all paragraphs, and assign a score to them based on how content-y they look.
595 * Then add their score to their parent node. 595 * Then add their score to their parent node.
596 * 596 *
597 * A score is determined by things like number of commas, class names, etc. Maybe eventually link density. 597 * A score is determined by things like number of commas, class names, etc. Maybe eventually link density.
598 **/ 598 **/
599 $candidates = array(); 599 $candidates = array();
600 for ($pt=0; $pt < count($nodesToScore); $pt++) { 600 for ($pt=0; $pt < count($nodesToScore); $pt++) {
601 $parentNode = $nodesToScore[$pt]->parentNode; 601 $parentNode = $nodesToScore[$pt]->parentNode;
602 // $grandParentNode = $parentNode ? $parentNode->parentNode : null; 602 // $grandParentNode = $parentNode ? $parentNode->parentNode : null;
603 $grandParentNode = !$parentNode ? null : (($parentNode->parentNode instanceof DOMElement) ? $parentNode->parentNode : null); 603 $grandParentNode = !$parentNode ? null : (($parentNode->parentNode instanceof DOMElement) ? $parentNode->parentNode : null);
604 $innerText = $this->getInnerText($nodesToScore[$pt]); 604 $innerText = $this->getInnerText($nodesToScore[$pt]);
605 605
606 if (!$parentNode || !isset($parentNode->tagName)) { 606 if (!$parentNode || !isset($parentNode->tagName)) {
607 continue; 607 continue;
608 } 608 }
609 609
610 /* If this paragraph is less than 25 characters, don't even count it. */ 610 /* If this paragraph is less than 25 characters, don't even count it. */
611 if(strlen($innerText) < 25) { 611 if(strlen($innerText) < 25) {
612 continue; 612 continue;
613 } 613 }
614 614
615 /* Initialize readability data for the parent. */ 615 /* Initialize readability data for the parent. */
616 if (!$parentNode->hasAttribute('readability')) 616 if (!$parentNode->hasAttribute('readability'))
617 { 617 {
618 $this->initializeNode($parentNode); 618 $this->initializeNode($parentNode);
619 $candidates[] = $parentNode; 619 $candidates[] = $parentNode;
620 } 620 }
621 621
622 /* Initialize readability data for the grandparent. */ 622 /* Initialize readability data for the grandparent. */
623 if ($grandParentNode && !$grandParentNode->hasAttribute('readability') && isset($grandParentNode->tagName)) 623 if ($grandParentNode && !$grandParentNode->hasAttribute('readability') && isset($grandParentNode->tagName))
624 { 624 {
625 $this->initializeNode($grandParentNode); 625 $this->initializeNode($grandParentNode);
626 $candidates[] = $grandParentNode; 626 $candidates[] = $grandParentNode;
627 } 627 }
628 628
629 $contentScore = 0; 629 $contentScore = 0;
630 630
631 /* Add a point for the paragraph itself as a base. */ 631 /* Add a point for the paragraph itself as a base. */
632 $contentScore++; 632 $contentScore++;
633 633
634 /* Add points for any commas within this paragraph */ 634 /* Add points for any commas within this paragraph */
635 $contentScore += count(explode(',', $innerText)); 635 $contentScore += count(explode(',', $innerText));
636 636
637 /* For every 100 characters in this paragraph, add another point. Up to 3 points. */ 637 /* For every 100 characters in this paragraph, add another point. Up to 3 points. */
638 $contentScore += min(floor(strlen($innerText) / 100), 3); 638 $contentScore += min(floor(strlen($innerText) / 100), 3);
639 639
640 /* Add the score to the parent. The grandparent gets half. */ 640 /* Add the score to the parent. The grandparent gets half. */
641 $parentNode->getAttributeNode('readability')->value += $contentScore; 641 $parentNode->getAttributeNode('readability')->value += $contentScore;
642 642
643 if ($grandParentNode) { 643 if ($grandParentNode) {
644 $grandParentNode->getAttributeNode('readability')->value += $contentScore/2; 644 $grandParentNode->getAttributeNode('readability')->value += $contentScore/2;
645 } 645 }
646 } 646 }
647 647
648 /** 648 /**
649 * After we've calculated scores, loop through all of the possible candidate nodes we found 649 * After we've calculated scores, loop through all of the possible candidate nodes we found
650 * and find the one with the highest score. 650 * and find the one with the highest score.
651 **/ 651 **/
652 $topCandidate = null; 652 $topCandidate = null;
653 for ($c=0, $cl=count($candidates); $c < $cl; $c++) 653 for ($c=0, $cl=count($candidates); $c < $cl; $c++)
654 { 654 {
655 /** 655 /**
656 * Scale the final candidates score based on link density. Good content should have a 656 * Scale the final candidates score based on link density. Good content should have a
657 * relatively small link density (5% or less) and be mostly unaffected by this operation. 657 * relatively small link density (5% or less) and be mostly unaffected by this operation.
658 **/ 658 **/
659 $readability = $candidates[$c]->getAttributeNode('readability'); 659 $readability = $candidates[$c]->getAttributeNode('readability');
660 $readability->value = $readability->value * (1-$this->getLinkDensity($candidates[$c])); 660 $readability->value = $readability->value * (1-$this->getLinkDensity($candidates[$c]));
661 661
662 $this->dbg('Candidate: ' . $candidates[$c]->tagName . ' (' . $candidates[$c]->getAttribute('class') . ':' . $candidates[$c]->getAttribute('id') . ') with score ' . $readability->value); 662 $this->dbg('Candidate: ' . $candidates[$c]->tagName . ' (' . $candidates[$c]->getAttribute('class') . ':' . $candidates[$c]->getAttribute('id') . ') with score ' . $readability->value);
663 663
664 if (!$topCandidate || $readability->value > (int)$topCandidate->getAttribute('readability')) { 664 if (!$topCandidate || $readability->value > (int)$topCandidate->getAttribute('readability')) {
665 $topCandidate = $candidates[$c]; 665 $topCandidate = $candidates[$c];
666 } 666 }
667 } 667 }
668 668
669 /** 669 /**
670 * If we still have no top candidate, just use the body as a last resort. 670 * If we still have no top candidate, just use the body as a last resort.
671 * We also have to copy the body node so it is something we can modify. 671 * We also have to copy the body node so it is something we can modify.
672 **/ 672 **/
673 if ($topCandidate === null || strtoupper($topCandidate->tagName) == 'BODY') 673 if ($topCandidate === null || strtoupper($topCandidate->tagName) == 'BODY')
674 { 674 {
675 $topCandidate = $this->dom->createElement('div'); 675 $topCandidate = $this->dom->createElement('div');
676 if ($page instanceof DOMDocument) { 676 if ($page instanceof DOMDocument) {
677 if (!isset($page->documentElement)) { 677 if (!isset($page->documentElement)) {
678 // we don't have a body either? what a mess! :) 678 // we don't have a body either? what a mess! :)
679 } else { 679 } else {
680 $topCandidate->innerHTML = $page->documentElement->innerHTML; 680 $topCandidate->innerHTML = $page->documentElement->innerHTML;
681 $page->documentElement->innerHTML = ''; 681 $page->documentElement->innerHTML = '';
682 $page->documentElement->appendChild($topCandidate); 682 $page->documentElement->appendChild($topCandidate);
683 } 683 }
684 } else { 684 } else {
685 $topCandidate->innerHTML = $page->innerHTML; 685 $topCandidate->innerHTML = $page->innerHTML;
686 $page->innerHTML = ''; 686 $page->innerHTML = '';
687 $page->appendChild($topCandidate); 687 $page->appendChild($topCandidate);
688 } 688 }
689 $this->initializeNode($topCandidate); 689 $this->initializeNode($topCandidate);
690 } 690 }
691 691
692 /** 692 /**
693 * Now that we have the top candidate, look through its siblings for content that might also be related. 693 * Now that we have the top candidate, look through its siblings for content that might also be related.
694 * Things like preambles, content split by ads that we removed, etc. 694 * Things like preambles, content split by ads that we removed, etc.
695 **/ 695 **/
696 $articleContent = $this->dom->createElement('div'); 696 $articleContent = $this->dom->createElement('div');
697 $articleContent->setAttribute('id', 'readability-content'); 697 $articleContent->setAttribute('id', 'readability-content');
698 $siblingScoreThreshold = max(10, ((int)$topCandidate->getAttribute('readability')) * 0.2); 698 $siblingScoreThreshold = max(10, ((int)$topCandidate->getAttribute('readability')) * 0.2);
699 $siblingNodes = $topCandidate->parentNode->childNodes; 699 $siblingNodes = $topCandidate->parentNode->childNodes;
700 if (!isset($siblingNodes)) { 700 if (!isset($siblingNodes)) {
701 $siblingNodes = new stdClass; 701 $siblingNodes = new stdClass;
702 $siblingNodes->length = 0; 702 $siblingNodes->length = 0;
703 } 703 }
704 704
705 for ($s=0, $sl=$siblingNodes->length; $s < $sl; $s++) 705 for ($s=0, $sl=$siblingNodes->length; $s < $sl; $s++)
706 { 706 {
707 $siblingNode = $siblingNodes->item($s); 707 $siblingNode = $siblingNodes->item($s);
708 $append = false; 708 $append = false;
709 709
710 $this->dbg('Looking at sibling node: ' . $siblingNode->nodeName . (($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->hasAttribute('readability')) ? (' with score ' . $siblingNode->getAttribute('readability')) : '')); 710 $this->dbg('Looking at sibling node: ' . $siblingNode->nodeName . (($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->hasAttribute('readability')) ? (' with score ' . $siblingNode->getAttribute('readability')) : ''));
711 711
712 //dbg('Sibling has score ' . ($siblingNode->readability ? siblingNode.readability.contentScore : 'Unknown')); 712 //dbg('Sibling has score ' . ($siblingNode->readability ? siblingNode.readability.contentScore : 'Unknown'));
713 713
714 if ($siblingNode === $topCandidate) 714 if ($siblingNode === $topCandidate)
715 // or if ($siblingNode->isSameNode($topCandidate)) 715 // or if ($siblingNode->isSameNode($topCandidate))
716 { 716 {
717 $append = true; 717 $append = true;
718 } 718 }
719 719
720 $contentBonus = 0; 720 $contentBonus = 0;
721 /* Give a bonus if sibling nodes and top candidates have the example same classname */ 721 /* Give a bonus if sibling nodes and top candidates have the example same classname */
722 if ($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->getAttribute('class') == $topCandidate->getAttribute('class') && $topCandidate->getAttribute('class') != '') { 722 if ($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->getAttribute('class') == $topCandidate->getAttribute('class') && $topCandidate->getAttribute('class') != '') {
723 $contentBonus += ((int)$topCandidate->getAttribute('readability')) * 0.2; 723 $contentBonus += ((int)$topCandidate->getAttribute('readability')) * 0.2;
724 } 724 }
725 725
726 if ($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->hasAttribute('readability') && (((int)$siblingNode->getAttribute('readability')) + $contentBonus) >= $siblingScoreThreshold) 726 if ($siblingNode->nodeType === XML_ELEMENT_NODE && $siblingNode->hasAttribute('readability') && (((int)$siblingNode->getAttribute('readability')) + $contentBonus) >= $siblingScoreThreshold)
727 { 727 {
728 $append = true; 728 $append = true;
729 } 729 }
730 730
731 if (strtoupper($siblingNode->nodeName) == 'P') { 731 if (strtoupper($siblingNode->nodeName) == 'P') {
732 $linkDensity = $this->getLinkDensity($siblingNode); 732 $linkDensity = $this->getLinkDensity($siblingNode);
733 $nodeContent = $this->getInnerText($siblingNode); 733 $nodeContent = $this->getInnerText($siblingNode);
734 $nodeLength = strlen($nodeContent); 734 $nodeLength = strlen($nodeContent);
735 735
736 if ($nodeLength > 80 && $linkDensity < 0.25) 736 if ($nodeLength > 80 && $linkDensity < 0.25)
737 { 737 {
738 $append = true; 738 $append = true;
739 } 739 }
740 else if ($nodeLength < 80 && $linkDensity === 0 && preg_match('/\.( |$)/', $nodeContent)) 740 else if ($nodeLength < 80 && $linkDensity === 0 && preg_match('/\.( |$)/', $nodeContent))
741 { 741 {
742 $append = true; 742 $append = true;
743 } 743 }
744 } 744 }
745 745
746 if ($append) 746 if ($append)
747 { 747 {
748 $this->dbg('Appending node: ' . $siblingNode->nodeName); 748 $this->dbg('Appending node: ' . $siblingNode->nodeName);
749 749
750 $nodeToAppend = null; 750 $nodeToAppend = null;
751 $sibNodeName = strtoupper($siblingNode->nodeName); 751 $sibNodeName = strtoupper($siblingNode->nodeName);
752 if ($sibNodeName != 'DIV' && $sibNodeName != 'P') { 752 if ($sibNodeName != 'DIV' && $sibNodeName != 'P') {
753 /* 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. */ 753 /* 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. */
754 754
755 $this->dbg('Altering siblingNode of ' . $sibNodeName . ' to div.'); 755 $this->dbg('Altering siblingNode of ' . $sibNodeName . ' to div.');
756 $nodeToAppend = $this->dom->createElement('div'); 756 $nodeToAppend = $this->dom->createElement('div');
757 try { 757 try {
758 $nodeToAppend->setAttribute('id', $siblingNode->getAttribute('id')); 758 $nodeToAppend->setAttribute('id', $siblingNode->getAttribute('id'));
759 $nodeToAppend->innerHTML = $siblingNode->innerHTML; 759 $nodeToAppend->innerHTML = $siblingNode->innerHTML;
760 } 760 }
761 catch(Exception $e) 761 catch(Exception $e)
762 { 762 {
763 $this->dbg('Could not alter siblingNode to div, reverting back to original.'); 763 $this->dbg('Could not alter siblingNode to div, reverting back to original.');
764 $nodeToAppend = $siblingNode; 764 $nodeToAppend = $siblingNode;
765 $s--; 765 $s--;
766 $sl--; 766 $sl--;
767 } 767 }
768 } else { 768 } else {
769 $nodeToAppend = $siblingNode; 769 $nodeToAppend = $siblingNode;
770 $s--; 770 $s--;
771 $sl--; 771 $sl--;
772 } 772 }
773 773
774 /* To ensure a node does not interfere with readability styles, remove its classnames */ 774 /* To ensure a node does not interfere with readability styles, remove its classnames */
775 $nodeToAppend->removeAttribute('class'); 775 $nodeToAppend->removeAttribute('class');
776 776
777 /* Append sibling and subtract from our list because it removes the node when you append to another node */ 777 /* Append sibling and subtract from our list because it removes the node when you append to another node */
778 $articleContent->appendChild($nodeToAppend); 778 $articleContent->appendChild($nodeToAppend);
779 } 779 }
780 } 780 }
781 781
782 /** 782 /**
783 * So we have all of the content that we need. Now we clean it up for presentation. 783 * So we have all of the content that we need. Now we clean it up for presentation.
784 **/ 784 **/
785 $this->prepArticle($articleContent); 785 $this->prepArticle($articleContent);
786 786
787 /** 787 /**
788 * Now that we've gone through the full algorithm, check to see if we got any meaningful content. 788 * Now that we've gone through the full algorithm, check to see if we got any meaningful content.
789 * If we didn't, we may need to re-run grabArticle with different flags set. This gives us a higher 789 * If we didn't, we may need to re-run grabArticle with different flags set. This gives us a higher
790 * likelihood of finding the content, and the sieve approach gives us a higher likelihood of 790 * likelihood of finding the content, and the sieve approach gives us a higher likelihood of
791 * finding the -right- content. 791 * finding the -right- content.
792 **/ 792 **/
793 if (strlen($this->getInnerText($articleContent, false)) < 250) 793 if (strlen($this->getInnerText($articleContent, false)) < 250)
794 { 794 {
795 // TODO: find out why element disappears sometimes, e.g. for this URL http://www.businessinsider.com/6-hedge-fund-etfs-for-average-investors-2011-7 795 // 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
796 // in the meantime, we check and create an empty element if it's not there. 796 // in the meantime, we check and create an empty element if it's not there.
797 if (!isset($this->body->childNodes)) $this->body = $this->dom->createElement('body'); 797 if (!isset($this->body->childNodes)) $this->body = $this->dom->createElement('body');
798 $this->body->innerHTML = $this->bodyCache; 798 $this->body->innerHTML = $this->bodyCache;
799 799
800 if ($this->flagIsActive(self::FLAG_STRIP_UNLIKELYS)) { 800 if ($this->flagIsActive(self::FLAG_STRIP_UNLIKELYS)) {
801 $this->removeFlag(self::FLAG_STRIP_UNLIKELYS); 801 $this->removeFlag(self::FLAG_STRIP_UNLIKELYS);
802 return $this->grabArticle($this->body); 802 return $this->grabArticle($this->body);
803 } 803 }
804 else if ($this->flagIsActive(self::FLAG_WEIGHT_CLASSES)) { 804 else if ($this->flagIsActive(self::FLAG_WEIGHT_CLASSES)) {
805 $this->removeFlag(self::FLAG_WEIGHT_CLASSES); 805 $this->removeFlag(self::FLAG_WEIGHT_CLASSES);
806 return $this->grabArticle($this->body); 806 return $this->grabArticle($this->body);
807 } 807 }
808 else if ($this->flagIsActive(self::FLAG_CLEAN_CONDITIONALLY)) { 808 else if ($this->flagIsActive(self::FLAG_CLEAN_CONDITIONALLY)) {
809 $this->removeFlag(self::FLAG_CLEAN_CONDITIONALLY); 809 $this->removeFlag(self::FLAG_CLEAN_CONDITIONALLY);
810 return $this->grabArticle($this->body); 810 return $this->grabArticle($this->body);
811 } 811 }
812 else { 812 else {
813 return false; 813 return false;
814 } 814 }
815 } 815 }
816 return $articleContent; 816 return $articleContent;
817 } 817 }
818 818
819 /** 819 /**
820 * Remove script tags from document 820 * Remove script tags from document
821 * 821 *
822 * @param DOMElement 822 * @param DOMElement
823 * @return void 823 * @return void
824 */ 824 */
825 public function removeScripts($doc) { 825 public function removeScripts($doc) {
826 $scripts = $doc->getElementsByTagName('script'); 826 $scripts = $doc->getElementsByTagName('script');
827 for($i = $scripts->length-1; $i >= 0; $i--) 827 for($i = $scripts->length-1; $i >= 0; $i--)
828 { 828 {
829 $scripts->item($i)->parentNode->removeChild($scripts->item($i)); 829 $scripts->item($i)->parentNode->removeChild($scripts->item($i));
830 } 830 }
831 } 831 }
832 832
833 /** 833 /**
834 * Get the inner text of a node. 834 * Get the inner text of a node.
835 * This also strips out any excess whitespace to be found. 835 * This also strips out any excess whitespace to be found.
836 * 836 *
837 * @param DOMElement $ 837 * @param DOMElement $
838 * @param boolean $normalizeSpaces (default: true) 838 * @param boolean $normalizeSpaces (default: true)
839 * @return string 839 * @return string
840 **/ 840 **/
841 public function getInnerText($e, $normalizeSpaces=true) { 841 public function getInnerText($e, $normalizeSpaces=true) {
842 $textContent = ''; 842 $textContent = '';
843 843
844 if (!isset($e->textContent) || $e->textContent == '') { 844 if (!isset($e->textContent) || $e->textContent == '') {
845 return ''; 845 return '';
846 } 846 }
847 847
848 $textContent = trim($e->textContent); 848 $textContent = trim($e->textContent);
849 849
850 if ($normalizeSpaces) { 850 if ($normalizeSpaces) {
851 return preg_replace($this->regexps['normalize'], ' ', $textContent); 851 return preg_replace($this->regexps['normalize'], ' ', $textContent);
852 } else { 852 } else {
853 return $textContent; 853 return $textContent;
854 } 854 }
855 } 855 }
856 856
857 /** 857 /**
858 * Get the number of times a string $s appears in the node $e. 858 * Get the number of times a string $s appears in the node $e.
859 * 859 *
860 * @param DOMElement $e 860 * @param DOMElement $e
861 * @param string - what to count. Default is "," 861 * @param string - what to count. Default is ","
862 * @return number (integer) 862 * @return number (integer)
863 **/ 863 **/
864 public function getCharCount($e, $s=',') { 864 public function getCharCount($e, $s=',') {
865 return substr_count($this->getInnerText($e), $s); 865 return substr_count($this->getInnerText($e), $s);
866 } 866 }
867 867
868 /** 868 /**
869 * Remove the style attribute on every $e and under. 869 * Remove the style attribute on every $e and under.
870 * 870 *
871 * @param DOMElement $e 871 * @param DOMElement $e
872 * @return void 872 * @return void
873 */ 873 */
874 public function cleanStyles($e) { 874 public function cleanStyles($e) {
875 if (!is_object($e)) return; 875 if (!is_object($e)) return;
876 $elems = $e->getElementsByTagName('*'); 876 $elems = $e->getElementsByTagName('*');
877 foreach ($elems as $elem) { 877 foreach ($elems as $elem) {
878 $elem->removeAttribute('style'); 878 $elem->removeAttribute('style');
879 } 879 }
880 } 880 }
881 881
882 /** 882 /**
883 * Get the density of links as a percentage of the content 883 * Get the density of links as a percentage of the content
884 * This is the amount of text that is inside a link divided by the total text in the node. 884 * This is the amount of text that is inside a link divided by the total text in the node.
885 * 885 *
886 * @param DOMElement $e 886 * @param DOMElement $e
887 * @return number (float) 887 * @return number (float)
888 */ 888 */
889 public function getLinkDensity($e) { 889 public function getLinkDensity($e) {
890 $links = $e->getElementsByTagName('a'); 890 $links = $e->getElementsByTagName('a');
891 $textLength = strlen($this->getInnerText($e)); 891 $textLength = strlen($this->getInnerText($e));
892 $linkLength = 0; 892 $linkLength = 0;
893 for ($i=0, $il=$links->length; $i < $il; $i++) 893 for ($i=0, $il=$links->length; $i < $il; $i++)
894 { 894 {
895 $linkLength += strlen($this->getInnerText($links->item($i))); 895 $linkLength += strlen($this->getInnerText($links->item($i)));
896 } 896 }
897 if ($textLength > 0) { 897 if ($textLength > 0) {
898 return $linkLength / $textLength; 898 return $linkLength / $textLength;
899 } else { 899 } else {
900 return 0; 900 return 0;
901 } 901 }
902 } 902 }
903 903
904 /** 904 /**
905 * Get an elements class/id weight. Uses regular expressions to tell if this 905 * Get an elements class/id weight. Uses regular expressions to tell if this
906 * element looks good or bad. 906 * element looks good or bad.
907 * 907 *
908 * @param DOMElement $e 908 * @param DOMElement $e
909 * @return number (Integer) 909 * @return number (Integer)
910 */ 910 */
911 public function getClassWeight($e) { 911 public function getClassWeight($e) {
912 if(!$this->flagIsActive(self::FLAG_WEIGHT_CLASSES)) { 912 if(!$this->flagIsActive(self::FLAG_WEIGHT_CLASSES)) {
913 return 0; 913 return 0;
914 } 914 }
915 915
916 $weight = 0; 916 $weight = 0;
917 917
918 /* Look for a special classname */ 918 /* Look for a special classname */
919 if ($e->hasAttribute('class') && $e->getAttribute('class') != '') 919 if ($e->hasAttribute('class') && $e->getAttribute('class') != '')
920 { 920 {
921 if (preg_match($this->regexps['negative'], $e->getAttribute('class'))) { 921 if (preg_match($this->regexps['negative'], $e->getAttribute('class'))) {
922 $weight -= 25; 922 $weight -= 25;
923 } 923 }
924 if (preg_match($this->regexps['positive'], $e->getAttribute('class'))) { 924 if (preg_match($this->regexps['positive'], $e->getAttribute('class'))) {
925 $weight += 25; 925 $weight += 25;
926 } 926 }
927 } 927 }
928 928
929 /* Look for a special ID */ 929 /* Look for a special ID */
930 if ($e->hasAttribute('id') && $e->getAttribute('id') != '') 930 if ($e->hasAttribute('id') && $e->getAttribute('id') != '')
931 { 931 {
932 if (preg_match($this->regexps['negative'], $e->getAttribute('id'))) { 932 if (preg_match($this->regexps['negative'], $e->getAttribute('id'))) {
933 $weight -= 25; 933 $weight -= 25;
934 } 934 }
935 if (preg_match($this->regexps['positive'], $e->getAttribute('id'))) { 935 if (preg_match($this->regexps['positive'], $e->getAttribute('id'))) {
936 $weight += 25; 936 $weight += 25;
937 } 937 }
938 } 938 }
939 return $weight; 939 return $weight;
940 } 940 }
941 941
942 /** 942 /**
943 * Remove extraneous break tags from a node. 943 * Remove extraneous break tags from a node.
944 * 944 *
945 * @param DOMElement $node 945 * @param DOMElement $node
946 * @return void 946 * @return void
947 */ 947 */
948 public function killBreaks($node) { 948 public function killBreaks($node) {
949 $html = $node->innerHTML; 949 $html = $node->innerHTML;
950 $html = preg_replace($this->regexps['killBreaks'], '<br />', $html); 950 $html = preg_replace($this->regexps['killBreaks'], '<br />', $html);
951 $node->innerHTML = $html; 951 $node->innerHTML = $html;
952 } 952 }
953 953
954 /** 954 /**
955 * Clean a node of all elements of type "tag". 955 * Clean a node of all elements of type "tag".
956 * (Unless it's a youtube/vimeo video. People love movies.) 956 * (Unless it's a youtube/vimeo video. People love movies.)
957 * 957 *
958 * Updated 2012-09-18 to preserve youtube/vimeo iframes 958 * Updated 2012-09-18 to preserve youtube/vimeo iframes
959 * 959 *
960 * @param DOMElement $e 960 * @param DOMElement $e
961 * @param string $tag 961 * @param string $tag
962 * @return void 962 * @return void
963 */ 963 */
964 public function clean($e, $tag) { 964 public function clean($e, $tag) {
965 $targetList = $e->getElementsByTagName($tag); 965 $targetList = $e->getElementsByTagName($tag);
966 $isEmbed = ($tag == 'iframe' || $tag == 'object' || $tag == 'embed'); 966 $isEmbed = ($tag == 'iframe' || $tag == 'object' || $tag == 'embed');
967 967
968 for ($y=$targetList->length-1; $y >= 0; $y--) { 968 for ($y=$targetList->length-1; $y >= 0; $y--) {
969 /* Allow youtube and vimeo videos through as people usually want to see those. */ 969 /* Allow youtube and vimeo videos through as people usually want to see those. */
970 if ($isEmbed) { 970 if ($isEmbed) {
971 $attributeValues = ''; 971 $attributeValues = '';
972 for ($i=0, $il=$targetList->item($y)->attributes->length; $i < $il; $i++) { 972 for ($i=0, $il=$targetList->item($y)->attributes->length; $i < $il; $i++) {
973 $attributeValues .= $targetList->item($y)->attributes->item($i)->value . '|'; // DOMAttr? (TODO: test) 973 $attributeValues .= $targetList->item($y)->attributes->item($i)->value . '|'; // DOMAttr? (TODO: test)
974 } 974 }
975 975
976 /* First, check the elements attributes to see if any of them contain youtube or vimeo */ 976 /* First, check the elements attributes to see if any of them contain youtube or vimeo */
977 if (preg_match($this->regexps['video'], $attributeValues)) { 977 if (preg_match($this->regexps['video'], $attributeValues)) {
978 continue; 978 continue;
979 } 979 }
980 980
981 /* Then check the elements inside this element for the same. */ 981 /* Then check the elements inside this element for the same. */
982 if (preg_match($this->regexps['video'], $targetList->item($y)->innerHTML)) { 982 if (preg_match($this->regexps['video'], $targetList->item($y)->innerHTML)) {
983 continue; 983 continue;
984 } 984 }
985 } 985 }
986 $targetList->item($y)->parentNode->removeChild($targetList->item($y)); 986 $targetList->item($y)->parentNode->removeChild($targetList->item($y));
987 } 987 }
988 } 988 }
989 989
990 /** 990 /**
991 * Clean an element of all tags of type "tag" if they look fishy. 991 * Clean an element of all tags of type "tag" if they look fishy.
992 * "Fishy" is an algorithm based on content length, classnames, 992 * "Fishy" is an algorithm based on content length, classnames,
993 * link density, number of images & embeds, etc. 993 * link density, number of images & embeds, etc.
994 * 994 *
995 * @param DOMElement $e 995 * @param DOMElement $e
996 * @param string $tag 996 * @param string $tag
997 * @return void 997 * @return void
998 */ 998 */
999 public function cleanConditionally($e, $tag) { 999 public function cleanConditionally($e, $tag) {
1000 if (!$this->flagIsActive(self::FLAG_CLEAN_CONDITIONALLY)) { 1000 if (!$this->flagIsActive(self::FLAG_CLEAN_CONDITIONALLY)) {
1001 return; 1001 return;
1002 } 1002 }
1003 1003
1004 $tagsList = $e->getElementsByTagName($tag); 1004 $tagsList = $e->getElementsByTagName($tag);
1005 $curTagsLength = $tagsList->length; 1005 $curTagsLength = $tagsList->length;
1006 1006
1007 /** 1007 /**
1008 * Gather counts for other typical elements embedded within. 1008 * Gather counts for other typical elements embedded within.
1009 * Traverse backwards so we can remove nodes at the same time without effecting the traversal. 1009 * Traverse backwards so we can remove nodes at the same time without effecting the traversal.
1010 * 1010 *
1011 * TODO: Consider taking into account original contentScore here. 1011 * TODO: Consider taking into account original contentScore here.
1012 */ 1012 */
1013 for ($i=$curTagsLength-1; $i >= 0; $i--) { 1013 for ($i=$curTagsLength-1; $i >= 0; $i--) {
1014 $weight = $this->getClassWeight($tagsList->item($i)); 1014 $weight = $this->getClassWeight($tagsList->item($i));
1015 $contentScore = ($tagsList->item($i)->hasAttribute('readability')) ? (int)$tagsList->item($i)->getAttribute('readability') : 0; 1015 $contentScore = ($tagsList->item($i)->hasAttribute('readability')) ? (int)$tagsList->item($i)->getAttribute('readability') : 0;
1016 1016
1017 $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')) : '')); 1017 $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')) : ''));
1018 1018
1019 if ($weight + $contentScore < 0) { 1019 if ($weight + $contentScore < 0) {
1020 $tagsList->item($i)->parentNode->removeChild($tagsList->item($i)); 1020 $tagsList->item($i)->parentNode->removeChild($tagsList->item($i));
1021 } 1021 }
1022 else if ( $this->getCharCount($tagsList->item($i), ',') < 10) { 1022 else if ( $this->getCharCount($tagsList->item($i), ',') < 10) {
1023 /** 1023 /**
1024 * If there are not very many commas, and the number of 1024 * If there are not very many commas, and the number of
1025 * non-paragraph elements is more than paragraphs or other ominous signs, remove the element. 1025 * non-paragraph elements is more than paragraphs or other ominous signs, remove the element.
1026 **/ 1026 **/
1027 $p = $tagsList->item($i)->getElementsByTagName('p')->length; 1027 $p = $tagsList->item($i)->getElementsByTagName('p')->length;
1028 $img = $tagsList->item($i)->getElementsByTagName('img')->length; 1028 $img = $tagsList->item($i)->getElementsByTagName('img')->length;
1029 $li = $tagsList->item($i)->getElementsByTagName('li')->length-100; 1029 $li = $tagsList->item($i)->getElementsByTagName('li')->length-100;
1030 $input = $tagsList->item($i)->getElementsByTagName('input')->length; 1030 $input = $tagsList->item($i)->getElementsByTagName('input')->length;
1031 $a = $tagsList->item($i)->getElementsByTagName('a')->length; 1031 $a = $tagsList->item($i)->getElementsByTagName('a')->length;
1032 1032
1033 $embedCount = 0; 1033 $embedCount = 0;
1034 $embeds = $tagsList->item($i)->getElementsByTagName('embed'); 1034 $embeds = $tagsList->item($i)->getElementsByTagName('embed');
1035 for ($ei=0, $il=$embeds->length; $ei < $il; $ei++) { 1035 for ($ei=0, $il=$embeds->length; $ei < $il; $ei++) {
1036 if (preg_match($this->regexps['video'], $embeds->item($ei)->getAttribute('src'))) { 1036 if (preg_match($this->regexps['video'], $embeds->item($ei)->getAttribute('src'))) {
1037 $embedCount++; 1037 $embedCount++;
1038 } 1038 }
1039 } 1039 }
1040 $embeds = $tagsList->item($i)->getElementsByTagName('iframe'); 1040 $embeds = $tagsList->item($i)->getElementsByTagName('iframe');
1041 for ($ei=0, $il=$embeds->length; $ei < $il; $ei++) { 1041 for ($ei=0, $il=$embeds->length; $ei < $il; $ei++) {
1042 if (preg_match($this->regexps['video'], $embeds->item($ei)->getAttribute('src'))) { 1042 if (preg_match($this->regexps['video'], $embeds->item($ei)->getAttribute('src'))) {
1043 $embedCount++; 1043 $embedCount++;
1044 } 1044 }
1045 } 1045 }
1046 1046
1047 $linkDensity = $this->getLinkDensity($tagsList->item($i)); 1047 $linkDensity = $this->getLinkDensity($tagsList->item($i));
1048 $contentLength = strlen($this->getInnerText($tagsList->item($i))); 1048 $contentLength = strlen($this->getInnerText($tagsList->item($i)));
1049 $toRemove = false; 1049 $toRemove = false;
1050 1050
1051 if ($this->lightClean) { 1051 if ($this->lightClean) {
1052 $this->dbg('Light clean...'); 1052 $this->dbg('Light clean...');
1053 if ( ($img > $p) && ($img > 4) ) { 1053 if ( ($img > $p) && ($img > 4) ) {
1054 $this->dbg(' more than 4 images and more image elements than paragraph elements'); 1054 $this->dbg(' more than 4 images and more image elements than paragraph elements');
1055 $toRemove = true; 1055 $toRemove = true;
1056 } else if ($li > $p && $tag != 'ul' && $tag != 'ol') { 1056 } else if ($li > $p && $tag != 'ul' && $tag != 'ol') {
1057 $this->dbg(' too many <li> elements, and parent is not <ul> or <ol>'); 1057 $this->dbg(' too many <li> elements, and parent is not <ul> or <ol>');
1058 $toRemove = true; 1058 $toRemove = true;
1059 } else if ( $input > floor($p/3) ) { 1059 } else if ( $input > floor($p/3) ) {
1060 $this->dbg(' too many <input> elements'); 1060 $this->dbg(' too many <input> elements');
1061 $toRemove = true; 1061 $toRemove = true;
1062 } else if ($contentLength < 25 && ($embedCount === 0 && ($img === 0 || $img > 2))) { 1062 } else if ($contentLength < 10 && ($embedCount === 0 && ($img === 0 || $img > 2))) {
1063 $this->dbg(' content length less than 25 chars, 0 embeds and either 0 images or more than 2 images'); 1063 $this->dbg(' content length less than 10 chars, 0 embeds and either 0 images or more than 2 images');
1064 $toRemove = true; 1064 $toRemove = true;
1065 } else if($weight < 25 && $linkDensity > 0.2) { 1065 } else if($weight < 25 && $linkDensity > 0.2) {
1066 $this->dbg(' weight smaller than 25 and link density above 0.2'); 1066 $this->dbg(' weight smaller than 25 and link density above 0.2');
1067 $toRemove = true; 1067 $toRemove = true;
1068 } else if($a > 2 && ($weight >= 25 && $linkDensity > 0.5)) { 1068 } else if($a > 2 && ($weight >= 25 && $linkDensity > 0.5)) {
1069 $this->dbg(' more than 2 links and weight above 25 but link density greater than 0.5'); 1069 $this->dbg(' more than 2 links and weight above 25 but link density greater than 0.5');
1070 $toRemove = true; 1070 $toRemove = true;
1071 } else if($embedCount > 3) { 1071 } else if($embedCount > 3) {
1072 $this->dbg(' more than 3 embeds'); 1072 $this->dbg(' more than 3 embeds');
1073 $toRemove = true; 1073 $toRemove = true;
1074 } 1074 }
1075 } else { 1075 } else {
1076 $this->dbg('Standard clean...'); 1076 $this->dbg('Standard clean...');
1077 if ( $img > $p ) { 1077 if ( $img > $p ) {
1078 $this->dbg(' more image elements than paragraph elements'); 1078 $this->dbg(' more image elements than paragraph elements');
1079 $toRemove = true; 1079 $toRemove = true;
1080 } else if ($li > $p && $tag != 'ul' && $tag != 'ol') { 1080 } else if ($li > $p && $tag != 'ul' && $tag != 'ol') {
1081 $this->dbg(' too many <li> elements, and parent is not <ul> or <ol>'); 1081 $this->dbg(' too many <li> elements, and parent is not <ul> or <ol>');
1082 $toRemove = true; 1082 $toRemove = true;
1083 } else if ( $input > floor($p/3) ) { 1083 } else if ( $input > floor($p/3) ) {
1084 $this->dbg(' too many <input> elements'); 1084 $this->dbg(' too many <input> elements');
1085 $toRemove = true; 1085 $toRemove = true;
1086 } else if ($contentLength < 25 && ($img === 0 || $img > 2) ) { 1086 } else if ($contentLength < 25 && ($img === 0 || $img > 2) ) {
1087 $this->dbg(' content length less than 25 chars and 0 images, or more than 2 images'); 1087 $this->dbg(' content length less than 25 chars and 0 images, or more than 2 images');
1088 $toRemove = true; 1088 $toRemove = true;
1089 } else if($weight < 25 && $linkDensity > 0.2) { 1089 } else if($weight < 25 && $linkDensity > 0.2) {
1090 $this->dbg(' weight smaller than 25 and link density above 0.2'); 1090 $this->dbg(' weight smaller than 25 and link density above 0.2');
1091 $toRemove = true; 1091 $toRemove = true;
1092 } else if($weight >= 25 && $linkDensity > 0.5) { 1092 } else if($weight >= 25 && $linkDensity > 0.5) {
1093 $this->dbg(' weight above 25 but link density greater than 0.5'); 1093 $this->dbg(' weight above 25 but link density greater than 0.5');
1094 $toRemove = true; 1094 $toRemove = true;
1095 } else if(($embedCount == 1 && $contentLength < 75) || $embedCount > 1) { 1095 } else if(($embedCount == 1 && $contentLength < 75) || $embedCount > 1) {
1096 $this->dbg(' 1 embed and content length smaller than 75 chars, or more than one embed'); 1096 $this->dbg(' 1 embed and content length smaller than 75 chars, or more than one embed');
1097 $toRemove = true; 1097 $toRemove = true;
1098 } 1098 }
1099 } 1099 }
1100 1100
1101 if ($toRemove) { 1101 if ($toRemove) {
1102 //$this->dbg('Removing: '.$tagsList->item($i)->innerHTML); 1102 //$this->dbg('Removing: '.$tagsList->item($i)->innerHTML);
1103 $tagsList->item($i)->parentNode->removeChild($tagsList->item($i)); 1103 $tagsList->item($i)->parentNode->removeChild($tagsList->item($i));
1104 } 1104 }
1105 } 1105 }
1106 } 1106 }
1107 } 1107 }
1108 1108
1109 /** 1109 /**
1110 * Clean out spurious headers from an Element. Checks things like classnames and link density. 1110 * Clean out spurious headers from an Element. Checks things like classnames and link density.
1111 * 1111 *
1112 * @param DOMElement $e 1112 * @param DOMElement $e
1113 * @return void 1113 * @return void
1114 */ 1114 */
1115 public function cleanHeaders($e) { 1115 public function cleanHeaders($e) {
1116 for ($headerIndex = 1; $headerIndex < 3; $headerIndex++) { 1116 for ($headerIndex = 1; $headerIndex < 3; $headerIndex++) {
1117 $headers = $e->getElementsByTagName('h' . $headerIndex); 1117 $headers = $e->getElementsByTagName('h' . $headerIndex);
1118 for ($i=$headers->length-1; $i >=0; $i--) { 1118 for ($i=$headers->length-1; $i >=0; $i--) {
1119 if ($this->getClassWeight($headers->item($i)) < 0 || $this->getLinkDensity($headers->item($i)) > 0.33) { 1119 if ($this->getClassWeight($headers->item($i)) < 0 || $this->getLinkDensity($headers->item($i)) > 0.33) {
1120 $headers->item($i)->parentNode->removeChild($headers->item($i)); 1120 $headers->item($i)->parentNode->removeChild($headers->item($i));
1121 } 1121 }
1122 } 1122 }
1123 } 1123 }
1124 } 1124 }
1125 1125
1126 public function flagIsActive($flag) { 1126 public function flagIsActive($flag) {
1127 return ($this->flags & $flag) > 0; 1127 return ($this->flags & $flag) > 0;
1128 } 1128 }
1129 1129
1130 public function addFlag($flag) { 1130 public function addFlag($flag) {
1131 $this->flags = $this->flags | $flag; 1131 $this->flags = $this->flags | $flag;
1132 } 1132 }
1133 1133
1134 public function removeFlag($flag) { 1134 public function removeFlag($flag) {
1135 $this->flags = $this->flags & ~$flag; 1135 $this->flags = $this->flags & ~$flag;
1136 } 1136 }
1137} 1137}
1138?> \ No newline at end of file 1138?> \ No newline at end of file