From 419214d7221e0821ef2b73eb2b3db816ed0cf173 Mon Sep 17 00:00:00 2001 From: Thomas Citharel Date: Tue, 28 Jun 2016 19:07:55 +0200 Subject: [PATCH] Download pictures successfully Needs to rewrite them properly (get base url) --- composer.json | 3 +- .../CoreBundle/Helper/ContentProxy.php | 178 +----------------- .../CoreBundle/Helper/DownloadImages.php | 147 +++++++++++++++ 3 files changed, 158 insertions(+), 170 deletions(-) create mode 100644 src/Wallabag/CoreBundle/Helper/DownloadImages.php diff --git a/composer.json b/composer.json index ebc0a7dc..e7b30fa1 100644 --- a/composer.json +++ b/composer.json @@ -82,7 +82,8 @@ "white-october/pagerfanta-bundle": "^1.0", "php-amqplib/rabbitmq-bundle": "^1.8", "predis/predis": "^1.0", - "javibravo/simpleue": "^1.0" + "javibravo/simpleue": "^1.0", + "symfony/dom-crawler": "^3.1" }, "require-dev": { "doctrine/doctrine-fixtures-bundle": "~2.2", diff --git a/src/Wallabag/CoreBundle/Helper/ContentProxy.php b/src/Wallabag/CoreBundle/Helper/ContentProxy.php index ddeffa77..bbad705f 100644 --- a/src/Wallabag/CoreBundle/Helper/ContentProxy.php +++ b/src/Wallabag/CoreBundle/Helper/ContentProxy.php @@ -65,7 +65,7 @@ class ContentProxy $entry->setUrl($content['url'] ?: $url); $entry->setTitle($title); - $entry->setContent($html); + $entry->setLanguage($content['language']); $entry->setMimetype($content['content_type']); $entry->setReadingTime(Utils::getReadingTime($html)); @@ -75,6 +75,14 @@ class ContentProxy $entry->setDomainName($domainName); } + if (true) { + $this->logger->log('debug','Starting to download images'); + $downloadImages = new DownloadImages($html, $url, $this->logger); + $html = $downloadImages->process(); + } + + $entry->setContent($html); + if (isset($content['open_graph']['og_image'])) { $entry->setPreviewPicture($content['open_graph']['og_image']); } @@ -151,172 +159,4 @@ class ContentProxy { return isset($content['title']) && isset($content['html']) && isset($content['url']) && isset($content['language']) && isset($content['content_type']); } - - /** - * Changing pictures URL in article content. - */ - public static function filterPicture($content, $url, $id) - { - $matches = array(); - $processing_pictures = array(); // list of processing image to avoid processing the same pictures twice - preg_match_all('#<\s*(img)[^>]+src="([^"]*)"[^>]*>#Si', $content, $matches, PREG_SET_ORDER); - foreach ($matches as $i => $link) { - $link[1] = trim($link[1]); - if (!preg_match('#^(([a-z]+://)|(\#))#', $link[1])) { - $absolute_path = self::_getAbsoluteLink($link[2], $url); - $filename = basename(parse_url($absolute_path, PHP_URL_PATH)); - $directory = self::_createAssetsDirectory($id); - $fullpath = $directory.'/'.$filename; - - if (in_array($absolute_path, $processing_pictures) === true) { - // replace picture's URL only if processing is OK : already processing -> go to next picture - continue; - } - - if (self::_downloadPictures($absolute_path, $fullpath) === true) { - $content = str_replace($matches[$i][2], Tools::getPocheUrl().$fullpath, $content); - } - - $processing_pictures[] = $absolute_path; - } - } - - return $content; - } - - /** - * Get absolute URL. - */ - private static function _getAbsoluteLink($relativeLink, $url) - { - /* return if already absolute URL */ - if (parse_url($relativeLink, PHP_URL_SCHEME) != '') { - return $relativeLink; - } - - /* queries and anchors */ - if ($relativeLink[0] == '#' || $relativeLink[0] == '?') { - return $url.$relativeLink; - } - - /* parse base URL and convert to local variables: - $scheme, $host, $path */ - extract(parse_url($url)); - - /* remove non-directory element from path */ - $path = preg_replace('#/[^/]*$#', '', $path); - - /* destroy path if relative url points to root */ - if ($relativeLink[0] == '/') { - $path = ''; - } - - /* dirty absolute URL */ - $abs = $host.$path.'/'.$relativeLink; - - /* replace '//' or '/./' or '/foo/../' with '/' */ - $re = array('#(/\.?/)#', '#/(?!\.\.)[^/]+/\.\./#'); - for ($n = 1; $n > 0; $abs = preg_replace($re, '/', $abs, -1, $n)) { - } - - /* absolute URL is ready! */ - return $scheme.'://'.$abs; - } - - /** - * Downloading pictures. - * - * @return bool true if the download and processing is OK, false else - */ - private static function _downloadPictures($absolute_path, $fullpath) - { - $rawdata = Tools::getFile($absolute_path); - $fullpath = urldecode($fullpath); - - if (file_exists($fullpath)) { - unlink($fullpath); - } - - // check extension - $file_ext = strrchr($fullpath, '.'); - $whitelist = array('.jpg', '.jpeg', '.gif', '.png'); - if (!(in_array($file_ext, $whitelist))) { - Tools::logm('processed image with not allowed extension. Skipping '.$fullpath); - - return false; - } - - // check headers - $imageinfo = getimagesize($absolute_path); - if ($imageinfo['mime'] != 'image/gif' && $imageinfo['mime'] != 'image/jpeg' && $imageinfo['mime'] != 'image/jpg' && $imageinfo['mime'] != 'image/png') { - Tools::logm('processed image with bad header. Skipping '.$fullpath); - - return false; - } - - // regenerate image - $im = imagecreatefromstring($rawdata); - if ($im === false) { - Tools::logm('error while regenerating image '.$fullpath); - - return false; - } - - switch ($imageinfo['mime']) { - case 'image/gif': - $result = imagegif($im, $fullpath); - break; - case 'image/jpeg': - case 'image/jpg': - $result = imagejpeg($im, $fullpath, REGENERATE_PICTURES_QUALITY); - break; - case 'image/png': - $result = imagepng($im, $fullpath, ceil(REGENERATE_PICTURES_QUALITY / 100 * 9)); - break; - } - imagedestroy($im); - - return $result; - } - - /** - * Create a directory for an article. - * - * @param $id ID of the article - * - * @return string - */ - private static function _createAssetsDirectory($id) - { - $assets_path = ABS_PATH; - if (!is_dir($assets_path)) { - mkdir($assets_path, 0715); - } - - $article_directory = $assets_path.$id; - if (!is_dir($article_directory)) { - mkdir($article_directory, 0715); - } - - return $article_directory; - } - - /** - * Remove the directory. - * - * @param $directory - * - * @return bool - */ - public static function removeDirectory($directory) - { - if (is_dir($directory)) { - $files = array_diff(scandir($directory), array('.', '..')); - foreach ($files as $file) { - (is_dir("$directory/$file")) ? self::removeDirectory("$directory/$file") : unlink("$directory/$file"); - } - - return rmdir($directory); - } - } } diff --git a/src/Wallabag/CoreBundle/Helper/DownloadImages.php b/src/Wallabag/CoreBundle/Helper/DownloadImages.php new file mode 100644 index 00000000..32a9dbb2 --- /dev/null +++ b/src/Wallabag/CoreBundle/Helper/DownloadImages.php @@ -0,0 +1,147 @@ +html = $html; + $this->url = $url; + $this->setFolder(); + $this->logger = $logger; + } + + public function setFolder($folder = "assets/images") { + // if folder doesn't exist, attempt to create one and store the folder name in property $folder + if(!file_exists($folder)) { + mkdir($folder); + } + $this->folder = $folder; + } + + public function process() { + //instantiate the symfony DomCrawler Component + $crawler = new Crawler($this->html); + // create an array of all scrapped image links + $this->logger->log('debug', 'Finding images inside document'); + $result = $crawler + ->filterXpath('//img') + ->extract(array('src')); + + // download and save the image to the folder + foreach ($result as $image) { + $file = file_get_contents($image); + + // Checks + $absolute_path = self::getAbsoluteLink($image, $this->url); + $filename = basename(parse_url($absolute_path, PHP_URL_PATH)); + $fullpath = $this->folder."/".$filename; + self::checks($file, $fullpath, $absolute_path); + $this->html = str_replace($image, $fullpath, $this->html); + } + + return $this->html; + } + + private function checks($rawdata, $fullpath, $absolute_path) { + $fullpath = urldecode($fullpath); + + if (file_exists($fullpath)) { + unlink($fullpath); + } + + // check extension + $this->logger->log('debug','Checking extension'); + + $file_ext = strrchr($fullpath, '.'); + $whitelist = array('.jpg', '.jpeg', '.gif', '.png'); + if (!(in_array($file_ext, $whitelist))) { + $this->logger->log('debug','processed image with not allowed extension. Skipping '.$fullpath); + + return false; + } + + // check headers + $this->logger->log('debug','Checking headers'); + $imageinfo = getimagesize($absolute_path); + if ($imageinfo['mime'] != 'image/gif' && $imageinfo['mime'] != 'image/jpeg' && $imageinfo['mime'] != 'image/jpg' && $imageinfo['mime'] != 'image/png') { + $this->logger->log('debug','processed image with bad header. Skipping '.$fullpath); + + return false; + } + + // regenerate image + $this->logger->log('debug','regenerating image'); + $im = imagecreatefromstring($rawdata); + if ($im === false) { + $this->logger->log('error','error while regenerating image '.$fullpath); + + return false; + } + + switch ($imageinfo['mime']) { + case 'image/gif': + $result = imagegif($im, $fullpath); + $this->logger->log('debug','Re-creating gif'); + break; + case 'image/jpeg': + case 'image/jpg': + $result = imagejpeg($im, $fullpath, REGENERATE_PICTURES_QUALITY); + $this->logger->log('debug','Re-creating jpg'); + break; + case 'image/png': + $this->logger->log('debug','Re-creating png'); + $result = imagepng($im, $fullpath, ceil(REGENERATE_PICTURES_QUALITY / 100 * 9)); + break; + } + imagedestroy($im); + + return $result; + } + + private static function getAbsoluteLink($relativeLink, $url) + { + /* return if already absolute URL */ + if (parse_url($relativeLink, PHP_URL_SCHEME) != '') { + return $relativeLink; + } + + /* queries and anchors */ + if ($relativeLink[0] == '#' || $relativeLink[0] == '?') { + return $url.$relativeLink; + } + + /* parse base URL and convert to local variables: + $scheme, $host, $path */ + extract(parse_url($url)); + + /* remove non-directory element from path */ + $path = preg_replace('#/[^/]*$#', '', $path); + + /* destroy path if relative url points to root */ + if ($relativeLink[0] == '/') { + $path = ''; + } + + /* dirty absolute URL */ + $abs = $host.$path.'/'.$relativeLink; + + /* replace '//' or '/./' or '/foo/../' with '/' */ + $re = array('#(/\.?/)#', '#/(?!\.\.)[^/]+/\.\./#'); + for ($n = 1; $n > 0; $abs = preg_replace($re, '/', $abs, -1, $n)) { + } + + /* absolute URL is ready! */ + return $scheme.'://'.$abs; + } +} -- 2.41.0