Merge pull request #54 from andreskrey/development

Prepare for release

.coveralls.yml

@@ -0,0 +1,3 @@
+coverage_clover: test/clover.xml
+json_path: test/coveralls-upload.json
+service_name: travis-ci

.travis.yml

@@ -1,11 +1,19 @@
 language: php
 
-install: composer install
+install:
+ - composer install
 
 php:
   - "5.6"
   - "7.0"
   - "7.1"
   - "7.2"
 
+script:
+  - ./vendor/bin/phpunit --coverage-clover ./test/clover.xml
+
+after_script:
+  - composer require php-coveralls/php-coveralls:^2.0
+  - php ./vendor/php-coveralls/php-coveralls/bin/php-coveralls -v
+
 sudo: false

CHANGELOG.md

@@ -3,9 +3,17 @@ All notable changes to this project will be documented in this file.
 
 ## Unreleased
 
+- Merged PR#49 (Missing object when calling `->getContent()`)
+- Imported all changes from Readability.js as of 2 March 2018 ([8525c6a](https://github.com/mozilla/readability/commit/8525c6af36d3badbe27c4672a6f2dd99ddb4097f)):
+    - Check for `<base>` elements before converting URLs to absolute.
+    - Clean `<link>` tags on `prepArticle()`
+    - Attempt to return at least some text if all the algorithm runs fail (Check PR [#423](https://github.com/mozilla/readability/pull/423) on JS version)
+    - Add new test cases for the previous changes
+    - And all other changes reflected [in this diff](https://github.com/mozilla/readability/compare/c3ff1a2d2c94c1db257b2c9aa88a4b8fbeb221c5...8525c6af36d3badbe27c4672a6f2dd99ddb4097f)
+
 ## [v1.1.1](https://github.com/andreskrey/readability.php/releases/tag/v1.1.1)
 
-- Switched from assertEquals to assertSame on unit testing to avoid weak comparisons.
+- Switched from assertEquals to assertSame on unit testing to avoid weak comparisons. 
 - Added a safe check to avoid sending the DOMDocument as a node when scanning for node ancestors.
 - Fix issue #45: Small mistake in documentation
 - Fix issue #46: Added `data-src` as a image source path

README.md

@@ -1,20 +1,20 @@
 # Readability.php
-[![Latest Stable Version](https://poser.pugx.org/andreskrey/readability.php/v/stable)](https://packagist.org/packages/andreskrey/readability.php) [![StyleCI](https://styleci.io/repos/71042668/shield?branch=master)](https://styleci.io/repos/71042668) [![Build Status](https://travis-ci.org/andreskrey/readability.php.svg?branch=master)](https://travis-ci.org/andreskrey/readability.php) [![Total Downloads](https://poser.pugx.org/andreskrey/readability.php/downloads)](https://packagist.org/packages/andreskrey/readability.php) [![Monthly Downloads](https://poser.pugx.org/andreskrey/readability.php/d/monthly)](https://packagist.org/packages/andreskrey/readability.php)
+[![Latest Stable Version](https://poser.pugx.org/andreskrey/readability.php/v/stable)](https://packagist.org/packages/andreskrey/readability.php) [![Build Status](https://travis-ci.org/andreskrey/readability.php.svg?branch=master)](https://travis-ci.org/andreskrey/readability.php) [![Coverage Status](https://coveralls.io/repos/github/andreskrey/readability.php/badge.svg?branch=master)](https://coveralls.io/github/andreskrey/readability.php/?branch=master) [![StyleCI](https://styleci.io/repos/71042668/shield?branch=master)](https://styleci.io/repos/71042668) [![Total Downloads](https://poser.pugx.org/andreskrey/readability.php/downloads)](https://packagist.org/packages/andreskrey/readability.php) [![Monthly Downloads](https://poser.pugx.org/andreskrey/readability.php/d/monthly)](https://packagist.org/packages/andreskrey/readability.php)
 
 PHP port of *Mozilla's* **[Readability.js](https://github.com/mozilla/readability)**. Parses html text (usually news and other articles) and returns **title**, **author**, **main image** and **text content** without nav bars, ads, footers, or anything that isn't the main body of the text. Analyzes each node, gives them a score, and determines what's relevant and what can be discarded.
 
 ![Screenshot](https://raw.githubusercontent.com/andreskrey/readability.php/assets/screenshot.png)
 
 The project aim is to be a 1 to 1 port of Mozilla's version and to follow closely all changes introduced there, but there are some major differences on the structure. Most of the code is a 1:1 copy –even the comments were imported– but some functions and structures were adapted to suit better the PHP language.
 
+**Lead Developer**: Andres Rey
+
 ## Requirements
 
 PHP 5.6+, ext-dom, ext-xml, and ext-mbstring. To install all this dependencies (in the rare case your system does not have them already), you could try something like this in *nix like environments:
 
 `$ sudo apt-get install php7.1-xml php7.1-mbstring`
 
-**Lead Developer**: Andres Rey
-
 ## How to use it
 
 First you have to require the library using composer:
@@ -152,7 +152,7 @@ Self closing tags like `<br />` get automatically expanded to `<br></br`. No way
 
 ## Dependencies
 
-Readability.php uses the [PSR Log](https://github.com/php-fig/log) interface to define the allowed type of loggers.
+Readability.php uses the [PSR Log](https://github.com/php-fig/log) interface to define the allowed type of loggers. [Monolog](https://github.com/Seldaek/monolog) is only required on development installations. (`--dev` option during `composer install`).
 
 ## To-do
 
@@ -165,7 +165,7 @@ Readability parses all the text with DOMDocument, scans the text nodes and gives
 
 ## Code porting
 
-Up to date with readability.js as of [16 Oct 2017](https://github.com/mozilla/readability/commit/c3ff1a2d2c94c1db257b2c9aa88a4b8fbeb221c5).
+Up to date with readability.js as of [2 Mar 2018](https://github.com/mozilla/readability/commit/8525c6af36d3badbe27c4672a6f2dd99ddb4097f).
 
 ## License
 

composer.json

@@ -25,7 +25,8 @@
         "psr/log": "^1.0"
     },
     "require-dev": {
-        "phpunit/phpunit": "^5.7"
+        "phpunit/phpunit": "^5.7",
+        "monolog/monolog": "^1.23"
     },
     "suggest": {
         "monolog/monolog": "Allow logging debug information"

src/Configuration.php

@@ -114,6 +114,18 @@ public function getLogger()
         }
     }
 
+    /**
+     * @param LoggerInterface $logger
+     *
+     * @return Configuration
+     */
+    public function setLogger(LoggerInterface $logger)
+    {
+        $this->logger = $logger;
+
+        return $this;
+    }
+
     /**
      * @return int
      */

src/Nodes/DOM/DOMDocument.php

@@ -20,10 +20,11 @@ public function __construct($version, $encoding)
         $this->registerNodeClass('DOMDocumentFragment', DOMDocumentFragment::class);
         $this->registerNodeClass('DOMDocumentType', DOMDocumentType::class);
         $this->registerNodeClass('DOMElement', DOMElement::class);
+        $this->registerNodeClass('DOMEntity', DOMEntity::class);
+        $this->registerNodeClass('DOMEntityReference', DOMEntityReference::class);
         $this->registerNodeClass('DOMNode', DOMNode::class);
         $this->registerNodeClass('DOMNotation', DOMNotation::class);
         $this->registerNodeClass('DOMProcessingInstruction', DOMProcessingInstruction::class);
         $this->registerNodeClass('DOMText', DOMText::class);
-        $this->registerNodeClass('DOMEntityReference', DOMEntityReference::class);
     }
 }

src/Nodes/DOM/DOMEntity.php

@@ -0,0 +1,10 @@
+<?php
+
+namespace andreskrey\Readability\Nodes\DOM;
+
+use andreskrey\Readability\Nodes\NodeTrait;
+
+class DOMEntity extends \DOMEntity
+{
+    use NodeTrait;
+}

src/Nodes/NodeTrait.php

@@ -7,6 +7,9 @@
 use andreskrey\Readability\Nodes\DOM\DOMNode;
 use andreskrey\Readability\Nodes\DOM\DOMText;
 
+/**
+ * @method \DOMNode removeAttribute($name)
+ */
 trait NodeTrait
 {
     /**

src/Readability.php

@@ -77,6 +77,13 @@ class Readability
      */
     private $logger;
 
+    /**
+     * Collection of attempted text extractions.
+     *
+     * @var array
+     */
+    private $attempts = [];
+
     /**
      * @var array
      */
@@ -155,54 +162,76 @@ public function parse($html)
              * finding the -right- content.
              */
 
-            $length = 0;
-            foreach ($result->getElementsByTagName('p') as $p) {
-                $length += mb_strlen($p->textContent);
-            }
+            $length = mb_strlen(preg_replace(NodeUtility::$regexps['onlyWhitespace'], '', $result->textContent));
 
             $this->logger->info(sprintf('[Parsing] Article parsed. Amount of words: %s. Current threshold is: %s', $length, $this->configuration->getWordThreshold()));
 
-            if ($result && mb_strlen(preg_replace('/\s/', '', $result->textContent)) < $this->configuration->getWordThreshold()) {
+            $parseSuccessful = true;
+
+            if ($result && $length < $this->configuration->getWordThreshold()) {
                 $this->dom = $this->loadHTML($html);
                 $root = $this->dom->getElementsByTagName('body')->item(0);
+                $parseSuccessful = false;
 
                 if ($this->configuration->getStripUnlikelyCandidates()) {
                     $this->logger->debug('[Parsing] Threshold not met, trying again setting StripUnlikelyCandidates as false');
                     $this->configuration->setStripUnlikelyCandidates(false);
+                    $this->attempts[] = ['articleContent' => $result, 'textLength' => $length];
                 } elseif ($this->configuration->getWeightClasses()) {
                     $this->logger->debug('[Parsing] Threshold not met, trying again setting WeightClasses as false');
                     $this->configuration->setWeightClasses(false);
+                    $this->attempts[] = ['articleContent' => $result, 'textLength' => $length];
                 } elseif ($this->configuration->getCleanConditionally()) {
                     $this->logger->debug('[Parsing] Threshold not met, trying again setting CleanConditionally as false');
                     $this->configuration->setCleanConditionally(false);
+                    $this->attempts[] = ['articleContent' => $result, 'textLength' => $length];
                 } else {
-                    $this->logger->emergency('[Parsing] Could not parse text, giving up :(');
+                    $this->logger->debug('[Parsing] Threshold not met, searching across attempts for some content.');
+                    $this->attempts[] = ['articleContent' => $result, 'textLength' => $length];
+
+                    // No luck after removing flags, just return the longest text we found during the different loops
+                    usort($this->attempts, function ($a, $b) {
+                        return $a['textLength'] < $b['textLength'];
+                    });
+
+                    // But first check if we actually have something
+                    if (!$this->attempts[0]['textLength']) {
+                        $this->logger->emergency('[Parsing] Could not parse text, giving up :(');
 
-                    throw new ParseException('Could not parse text.');
+                        throw new ParseException('Could not parse text.');
+                    }
+
+                    $this->logger->debug('[Parsing] Threshold not met, but found some content in previous attempts.');
+
+                    $result = $this->attempts[0]['articleContent'];
+                    $parseSuccessful = true;
+                    break;
                 }
             } else {
                 break;
             }
         }
 
-        $result = $this->postProcessContent($result);
-
-        // If we haven't found an excerpt in the article's metadata, use the article's
-        // first paragraph as the excerpt. This can be used for displaying a preview of
-        // the article's content.
-        if (!$this->getExcerpt()) {
-            $this->logger->debug('[Parsing] No excerpt text found on metadata, extracting first p node and using it as excerpt.');
-            $paragraphs = $result->getElementsByTagName('p');
-            if ($paragraphs->length > 0) {
-                $this->setExcerpt(trim($paragraphs->item(0)->textContent));
+        if ($parseSuccessful) {
+            $result = $this->postProcessContent($result);
+
+            // If we haven't found an excerpt in the article's metadata, use the article's
+            // first paragraph as the excerpt. This can be used for displaying a preview of
+            // the article's content.
+            if (!$this->getExcerpt()) {
+                $this->logger->debug('[Parsing] No excerpt text found on metadata, extracting first p node and using it as excerpt.');
+                $paragraphs = $result->getElementsByTagName('p');
+                if ($paragraphs->length > 0) {
+                    $this->setExcerpt(trim($paragraphs->item(0)->textContent));
+                }
             }
-        }
 
-        $this->setContent($result);
+            $this->setContent($result);
 
-        $this->logger->info('*** Parse successful :)');
+            $this->logger->info('*** Parse successful :)');
 
-        return true;
+            return true;
+        }
     }
 
     /**
@@ -468,6 +497,10 @@ private function getArticleTitle()
                 if (count(preg_split('/\s+/', $curTitle)) < 3) {
                     $curTitle = substr($originalTitle, strpos($originalTitle, ':') + 1);
                     $this->logger->info(sprintf('[Metadata] Title too short, using the first part of the title instead: \'%s\'', $curTitle));
+                } elseif (count(preg_split('/\s+/', substr($curTitle, 0, strpos($curTitle, ':')))) > 5) {
+                    // But if we have too many words before the colon there's something weird
+                    // with the titles and the H tags so let's just use the original title instead
+                    $curTitle = $originalTitle;
                 }
             }
         } elseif (mb_strlen($curTitle) > 150 || mb_strlen($curTitle) < 15) {
@@ -549,7 +582,19 @@ private function toAbsoluteURI($uri)
      */
     public function getPathInfo($url)
     {
-        $pathBase = parse_url($url, PHP_URL_SCHEME) . '://' . parse_url($url, PHP_URL_HOST) . dirname(parse_url($url, PHP_URL_PATH)) . '/';
+        // Check for base URLs
+        if ($this->dom->baseURI !== null) {
+            if (substr($this->dom->baseURI, 0, 1) === '/') {
+                // URLs starting with '/' override completely the URL defined in the link
+                $pathBase = parse_url($url, PHP_URL_SCHEME) . '://' . parse_url($url, PHP_URL_HOST) . $this->dom->baseURI;
+            } else {
+                // Otherwise just prepend the base to the actual path
+                $pathBase = parse_url($url, PHP_URL_SCHEME) . '://' . parse_url($url, PHP_URL_HOST) . dirname(parse_url($url, PHP_URL_PATH)) . '/' . rtrim($this->dom->baseURI, '/') . '/';
+            }
+        } else {
+            $pathBase = parse_url($url, PHP_URL_SCHEME) . '://' . parse_url($url, PHP_URL_HOST) . dirname(parse_url($url, PHP_URL_PATH)) . '/';
+        }
+
         $scheme = parse_url($pathBase, PHP_URL_SCHEME);
         $prePath = $scheme . '://' . parse_url($pathBase, PHP_URL_HOST);
 
@@ -1129,6 +1174,7 @@ public function prepArticle(DOMDocument $article)
         $this->_clean($article, 'embed');
         $this->_clean($article, 'h1');
         $this->_clean($article, 'footer');
+        $this->_clean($article, 'link');
 
         // Clean out elements have "share" in their id/class combinations from final top candidates,
         // which means we don't remove the top candidates even they have "share".
@@ -1479,6 +1525,28 @@ public function _cleanHeaders(DOMDocument $article)
         }
     }
 
+    /**
+     * Removes the class="" attribute from every element in the given
+     * subtree.
+     *
+     * Readability.js has a special filter to avoid cleaning the classes that the algorithm adds. We don't add classes
+     * here so no need to filter those.
+     *
+     * @param DOMDocument|DOMNode $node
+     *
+     * @return void
+     **/
+    public function _cleanClasses($node)
+    {
+        if ($node->getAttribute('class') !== '') {
+            $node->removeAttribute('class');
+        }
+
+        for ($node = $node->firstChild; $node !== null; $node = $node->nextSibling) {
+            $this->_cleanClasses($node);
+        }
+    }
+
     /**
      * @param DOMDocument $article
      *
@@ -1532,6 +1600,8 @@ public function postProcessContent(DOMDocument $article)
             }
         }
 
+        $this->_cleanClasses($article);
+
         return $article;
     }
 
@@ -1564,7 +1634,7 @@ protected function setTitle($title)
      */
     public function getContent()
     {
-        return $this->content->C14N();
+        return ($this->content instanceof DOMDocument) ? $this->content->C14N() : null;
     }
 
     /**