src/Wallabag/CoreBundle/Helper/ContentProxy.php

   1 <?php
   2
   3 namespace Wallabag\CoreBundle\Helper;
   4
   5 use Graby\Graby;
   6 use Psr\Log\LoggerInterface as Logger;
   7 use Wallabag\CoreBundle\Entity\Entry;
   8 use Wallabag\CoreBundle\Entity\Tag;
   9 use Wallabag\CoreBundle\Tools\Utils;
  10 use Wallabag\CoreBundle\Repository\TagRepository;
  11
  12 /**
  13  * This kind of proxy class take care of getting the content from an url
  14  * and update the entry with what it found.
  15  */
  16 class ContentProxy
  17 {
  18     protected $graby;
  19     protected $tagger;
  20     protected $logger;
  21     protected $tagRepository;
  22
  23     public function __construct(Graby $graby, RuleBasedTagger $tagger, TagRepository $tagRepository, Logger $logger)
  24     {
  25         $this->graby = $graby;
  26         $this->tagger = $tagger;
  27         $this->logger = $logger;
  28         $this->tagRepository = $tagRepository;
  29     }
  30
  31     /**
  32      * Fetch content using graby and hydrate given entry with results information.
  33      * In case we couldn't find content, we'll try to use Open Graph data.
  34      *
  35      * We can also force the content, in case of an import from the v1 for example, so the function won't
  36      * fetch the content from the website but rather use information given with the $content parameter.
  37      *
  38      * @param Entry  $entry   Entry to update
  39      * @param string $url     Url to grab content for
  40      * @param array  $content An array with AT LEAST keys title, html, url, language & content_type to skip the fetchContent from the url
  41      *
  42      * @return Entry
  43      */
  44     public function updateEntry(Entry $entry, $url, array $content = [])
  45     {
  46         // do we have to fetch the content or the provided one is ok?
  47         if (empty($content) || false === $this->validateContent($content)) {
  48             $content = $this->graby->fetchContent($url);
  49         }
  50
  51         $title = $content['title'];
  52         if (!$title && isset($content['open_graph']['og_title'])) {
  53             $title = $content['open_graph']['og_title'];
  54         }
  55
  56         $html = $content['html'];
  57         if (false === $html) {
  58             $html = '<p>Unable to retrieve readable content.</p>';
  59
  60             if (isset($content['open_graph']['og_description'])) {
  61                 $html .= '<p><i>But we found a short description: </i></p>';
  62                 $html .= $content['open_graph']['og_description'];
  63             }
  64         }
  65
  66         $entry->setUrl($content['url'] ?: $url);
  67         $entry->setTitle($title);
  68         $entry->setContent($html);
  69         $entry->setLanguage($content['language']);
  70         $entry->setMimetype($content['content_type']);
  71         $entry->setReadingTime(Utils::getReadingTime($html));
  72
  73         $domainName = parse_url($entry->getUrl(), PHP_URL_HOST);
  74         if (false !== $domainName) {
  75             $entry->setDomainName($domainName);
  76         }
  77
  78         if (isset($content['open_graph']['og_image'])) {
  79             $entry->setPreviewPicture($content['open_graph']['og_image']);
  80         }
  81
  82         try {
  83             $this->tagger->tag($entry);
  84         } catch (\Exception $e) {
  85             $this->logger->error('Error while trying to automatically tag an entry.', [
  86                 'entry_url' => $url,
  87                 'error_msg' => $e->getMessage(),
  88             ]);
  89         }
  90
  91         return $entry;
  92     }
  93
  94     /**
  95      * Assign some tags to an entry.
  96      *
  97      * @param Entry        $entry
  98      * @param array|string $tags          An array of tag or a string coma separated of tag
  99      * @param array        $entitiesReady Entities from the EntityManager which are persisted but not yet flushed
 100      *                                    It is mostly to fix duplicate tag on import @see http://stackoverflow.com/a/7879164/569101
 101      */
 102     public function assignTagsToEntry(Entry $entry, $tags, array $entitiesReady = [])
 103     {
 104         if (!is_array($tags)) {
 105             $tags = explode(',', $tags);
 106         }
 107
 108         // keeps only Tag entity from the "not yet flushed entities"
 109         $tagsNotYetFlushed = [];
 110         foreach ($entitiesReady as $entity) {
 111             if ($entity instanceof Tag) {
 112                 $tagsNotYetFlushed[$entity->getLabel()] = $entity;
 113             }
 114         }
 115
 116         foreach ($tags as $label) {
 117             $label = trim($label);
 118
 119             // avoid empty tag
 120             if (0 === strlen($label)) {
 121                 continue;
 122             }
 123
 124             if (isset($tagsNotYetFlushed[$label])) {
 125                 $tagEntity = $tagsNotYetFlushed[$label];
 126             } else {
 127                 $tagEntity = $this->tagRepository->findOneByLabel($label);
 128
 129                 if (is_null($tagEntity)) {
 130                     $tagEntity = new Tag();
 131                     $tagEntity->setLabel($label);
 132                 }
 133             }
 134
 135             // only add the tag on the entry if the relation doesn't exist
 136             if (false === $entry->getTags()->contains($tagEntity)) {
 137                 $entry->addTag($tagEntity);
 138             }
 139         }
 140     }
 141
 142     /**
 143      * Validate that the given content as enough value to be used
 144      * instead of fetch the content from the url.
 145      *
 146      * @param array $content
 147      *
 148      * @return bool true if valid otherwise false
 149      */
 150     private function validateContent(array $content)
 151     {
 152         return isset($content['title']) && isset($content['html']) && isset($content['url']) && isset($content['language']) && isset($content['content_type']);
 153     }
 154
 155     /**
 156      * Changing pictures URL in article content.
 157      */
 158     public static function filterPicture($content, $url, $id)
 159     {
 160         $matches = array();
 161         $processing_pictures = array(); // list of processing image to avoid processing the same pictures twice
 162         preg_match_all('#<\s*(img)[^>]+src="([^"]*)"[^>]*>#Si', $content, $matches, PREG_SET_ORDER);
 163         foreach ($matches as $i => $link) {
 164             $link[1] = trim($link[1]);
 165             if (!preg_match('#^(([a-z]+://)|(\#))#', $link[1])) {
 166                 $absolute_path = self::_getAbsoluteLink($link[2], $url);
 167                 $filename = basename(parse_url($absolute_path, PHP_URL_PATH));
 168                 $directory = self::_createAssetsDirectory($id);
 169                 $fullpath = $directory.'/'.$filename;
 170
 171                 if (in_array($absolute_path, $processing_pictures) === true) {
 172                     // replace picture's URL only if processing is OK : already processing -> go to next picture
 173                   continue;
 174                 }
 175
 176                 if (self::_downloadPictures($absolute_path, $fullpath) === true) {
 177                     $content = str_replace($matches[$i][2], Tools::getPocheUrl().$fullpath, $content);
 178                 }
 179
 180                 $processing_pictures[] = $absolute_path;
 181             }
 182         }
 183
 184         return $content;
 185     }
 186
 187     /**
 188      * Get absolute URL.
 189      */
 190     private static function _getAbsoluteLink($relativeLink, $url)
 191     {
 192         /* return if already absolute URL */
 193         if (parse_url($relativeLink, PHP_URL_SCHEME) != '') {
 194             return $relativeLink;
 195         }
 196
 197         /* queries and anchors */
 198         if ($relativeLink[0] == '#' || $relativeLink[0] == '?') {
 199             return $url.$relativeLink;
 200         }
 201
 202         /* parse base URL and convert to local variables:
 203            $scheme, $host, $path */
 204         extract(parse_url($url));
 205
 206         /* remove non-directory element from path */
 207         $path = preg_replace('#/[^/]*$#', '', $path);
 208
 209         /* destroy path if relative url points to root */
 210         if ($relativeLink[0] == '/') {
 211             $path = '';
 212         }
 213
 214         /* dirty absolute URL */
 215         $abs = $host.$path.'/'.$relativeLink;
 216
 217         /* replace '//' or '/./' or '/foo/../' with '/' */
 218         $re = array('#(/\.?/)#', '#/(?!\.\.)[^/]+/\.\./#');
 219         for ($n = 1; $n > 0; $abs = preg_replace($re, '/', $abs, -1, $n)) {
 220         }
 221
 222         /* absolute URL is ready! */
 223         return $scheme.'://'.$abs;
 224     }
 225
 226     /**
 227      * Downloading pictures.
 228      *
 229      * @return bool true if the download and processing is OK, false else
 230      */
 231     private static function _downloadPictures($absolute_path, $fullpath)
 232     {
 233         $rawdata = Tools::getFile($absolute_path);
 234         $fullpath = urldecode($fullpath);
 235
 236         if (file_exists($fullpath)) {
 237             unlink($fullpath);
 238         }
 239
 240         // check extension
 241         $file_ext = strrchr($fullpath, '.');
 242         $whitelist = array('.jpg', '.jpeg', '.gif', '.png');
 243         if (!(in_array($file_ext, $whitelist))) {
 244             Tools::logm('processed image with not allowed extension. Skipping '.$fullpath);
 245
 246             return false;
 247         }
 248
 249         // check headers
 250         $imageinfo = getimagesize($absolute_path);
 251         if ($imageinfo['mime'] != 'image/gif' && $imageinfo['mime'] != 'image/jpeg' && $imageinfo['mime'] != 'image/jpg' && $imageinfo['mime'] != 'image/png') {
 252             Tools::logm('processed image with bad header. Skipping '.$fullpath);
 253
 254             return false;
 255         }
 256
 257         // regenerate image
 258         $im = imagecreatefromstring($rawdata);
 259         if ($im === false) {
 260             Tools::logm('error while regenerating image '.$fullpath);
 261
 262             return false;
 263         }
 264
 265         switch ($imageinfo['mime']) {
 266             case 'image/gif':
 267                 $result = imagegif($im, $fullpath);
 268                 break;
 269             case 'image/jpeg':
 270             case 'image/jpg':
 271                 $result = imagejpeg($im, $fullpath, REGENERATE_PICTURES_QUALITY);
 272                 break;
 273             case 'image/png':
 274                 $result = imagepng($im, $fullpath, ceil(REGENERATE_PICTURES_QUALITY / 100 * 9));
 275                 break;
 276         }
 277         imagedestroy($im);
 278
 279         return $result;
 280     }
 281
 282     /**
 283      * Create a directory for an article.
 284      *
 285      * @param $id ID of the article
 286      *
 287      * @return string
 288      */
 289     private static function _createAssetsDirectory($id)
 290     {
 291         $assets_path = ABS_PATH;
 292         if (!is_dir($assets_path)) {
 293             mkdir($assets_path, 0715);
 294         }
 295
 296         $article_directory = $assets_path.$id;
 297         if (!is_dir($article_directory)) {
 298             mkdir($article_directory, 0715);
 299         }
 300
 301         return $article_directory;
 302     }
 303
 304     /**
 305      * Remove the directory.
 306      *
 307      * @param $directory
 308      *
 309      * @return bool
 310      */
 311     public static function removeDirectory($directory)
 312     {
 313         if (is_dir($directory)) {
 314             $files = array_diff(scandir($directory), array('.', '..'));
 315             foreach ($files as $file) {
 316                 (is_dir("$directory/$file")) ? self::removeDirectory("$directory/$file") : unlink("$directory/$file");
 317             }
 318
 319             return rmdir($directory);
 320         }
 321     }
 322 }