Merge pull request #53 from j0k3r/openload.co

Add openload.co to media detection

phpstan.neon

@@ -5,4 +5,4 @@ includes:
 parameters:
     # https://github.com/phpstan/phpstan/issues/694#issuecomment-350724288
     autoload_files:
-        - vendor/bin/.phpunit/phpunit-6.5/vendor/autoload.php
+        - vendor/bin/.phpunit/phpunit-7.4/vendor/autoload.php

phpunit.xml.dist

@@ -7,7 +7,6 @@
     convertWarningsToExceptions="true"
     processIsolation="false"
     stopOnFailure="false"
-    syntaxCheck="false"
     bootstrap="vendor/autoload.php"
     >
 

src/Readability.php

@@ -61,7 +61,7 @@ class Readability implements LoggerAwareInterface
     // constants
     const SCORE_CHARS_IN_PARAGRAPH = 100;
     const SCORE_WORDS_IN_PARAGRAPH = 20;
-    const GRANDPARENT_SCORE_DIVISOR = 2.2;
+    const GRANDPARENT_SCORE_DIVISOR = 2;
     const MIN_PARAGRAPH_LENGTH = 20;
     const MIN_COMMAS_IN_PARAGRAPH = 6;
     const MIN_ARTICLE_LENGTH = 200;
@@ -92,10 +92,10 @@ class Readability implements LoggerAwareInterface
         'unlikelyCandidates' => '/display\s*:\s*none|ignore|\binfos?\b|annoy|clock|date|time|author|intro|hidd?e|about|archive|\bprint|bookmark|tags|tag-list|share|search|social|robot|published|combx|comment|mast(?:head)|subscri|community|category|disqus|extra|head|head(?:er|note)|floor|foot(?:er|note)|menu|tool\b|function|nav|remark|rss|shoutbox|widget|meta|banner|sponsor|adsense|inner-?ad|ad-|sponsor|\badv\b|\bads\b|agr?egate?|pager|sidebar|popup|tweet|twitter/i',
         'okMaybeItsACandidate' => '/article\b|contain|\bcontent|column|general|detail|shadow|lightbox|blog|body|entry|main|page|footnote/i',
         'positive' => '/read|full|article|body|\bcontent|contain|entry|main|markdown|media|page|attach|pagination|post|text|blog|story/i',
-        'negative' => '/bottom|stat|info|discuss|e[\-]?mail|comment|reply|log.{2}(n|ed)|sign|single|combx|com-|contact|_nav|link|media|\bout|promo|\bad-|related|scroll|shoutbox|sidebar|sponsor|shopping|teaser|recommend/i',
+        'negative' => '/bottom|stat|info|discuss|e[\-]?mail|comment|reply|log.{2}(n|ed)|sign|single|combx|com-|contact|_nav|link|media|promo|\bad-|related|scroll|shoutbox|sidebar|sponsor|shopping|teaser|recommend/i',
         'divToPElements' => '/<(?:blockquote|header|section|code|div|article|footer|aside|img|p|pre|dl|ol|ul)/mi',
         'killBreaks' => '/(<br\s*\/?>([ \r\n\s]|&nbsp;?)*)+/',
-        'media' => '!//(?:[^\.\?/]+\.)?(?:youtu(?:be)?|giphy|soundcloud|dailymotion|vimeo|pornhub|xvideos|twitvid|rutube|viddler)\.(?:com|be|org|net)/!i',
+        'media' => '!//(?:[^\.\?/]+\.)?(?:youtu(?:be)?|giphy|soundcloud|dailymotion|vimeo|pornhub|xvideos|twitvid|rutube|openload\.co|viddler)\.(?:com|be|org|net)/!i',
         'skipFootnoteLink' => '/^\s*(\[?[a-z0-9]{1,2}\]?|^|edit|citation needed)\s*$/i',
     ];
     public $tidy_config = [
@@ -1115,7 +1115,7 @@ class Readability implements LoggerAwareInterface
             // Add the score to the parent. The grandparent gets half.
             $parentNode->getAttributeNode('readability')->value += $contentScore;
             if ($grandParentNode) {
-                $grandParentNode->getAttributeNode('readability')->value += $contentScore / self::GRANDPARENT_SCORE_DIVISOR;
+                $grandParentNode->getAttributeNode('readability')->value += round($contentScore / self::GRANDPARENT_SCORE_DIVISOR);
             }
         }
 

tests/ReadabilityTest.php

@@ -193,7 +193,7 @@ class ReadabilityTest extends \PHPUnit\Framework\TestCase
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertContains('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
         $this->assertNotContains('<aside>', $readability->getContent()->getInnerHtml());
-        $this->assertContains('<footer readability="4"/>', $readability->getContent()->getInnerHtml());
+        $this->assertContains('<footer readability="5"/>', $readability->getContent()->getInnerHtml());
     }
 
     public function testWithClasses()