From 4dface66707688ea440a6a7569795a85631d1ff0 Mon Sep 17 00:00:00 2001 From: Thomas Citharel Date: Sun, 26 Jun 2016 10:27:31 +0200 Subject: first draft (from v1) --- src/Wallabag/CoreBundle/Helper/ContentProxy.php | 168 ++++++++++++++++++++++++ 1 file changed, 168 insertions(+) (limited to 'src/Wallabag/CoreBundle/Helper/ContentProxy.php') diff --git a/src/Wallabag/CoreBundle/Helper/ContentProxy.php b/src/Wallabag/CoreBundle/Helper/ContentProxy.php index 8019df42..ddeffa77 100644 --- a/src/Wallabag/CoreBundle/Helper/ContentProxy.php +++ b/src/Wallabag/CoreBundle/Helper/ContentProxy.php @@ -151,4 +151,172 @@ 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); + } + } } -- cgit v1.2.3 From 419214d7221e0821ef2b73eb2b3db816ed0cf173 Mon Sep 17 00:00:00 2001 From: Thomas Citharel Date: Tue, 28 Jun 2016 19:07:55 +0200 Subject: Download pictures successfully Needs to rewrite them properly (get base url) --- src/Wallabag/CoreBundle/Helper/ContentProxy.php | 178 ++---------------------- 1 file changed, 9 insertions(+), 169 deletions(-) (limited to 'src/Wallabag/CoreBundle/Helper/ContentProxy.php') 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); - } - } } -- cgit v1.2.3 From 156bf62758080153668a65db611c4241d0fc8a00 Mon Sep 17 00:00:00 2001 From: Jeremy Benoist Date: Sat, 22 Oct 2016 09:22:30 +0200 Subject: CS --- src/Wallabag/CoreBundle/Helper/ContentProxy.php | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'src/Wallabag/CoreBundle/Helper/ContentProxy.php') diff --git a/src/Wallabag/CoreBundle/Helper/ContentProxy.php b/src/Wallabag/CoreBundle/Helper/ContentProxy.php index bbad705f..8ed11205 100644 --- a/src/Wallabag/CoreBundle/Helper/ContentProxy.php +++ b/src/Wallabag/CoreBundle/Helper/ContentProxy.php @@ -76,7 +76,7 @@ class ContentProxy } if (true) { - $this->logger->log('debug','Starting to download images'); + $this->logger->log('debug', 'Starting to download images'); $downloadImages = new DownloadImages($html, $url, $this->logger); $html = $downloadImages->process(); } -- cgit v1.2.3 From 45fd7e09d75995bd0b9a731ffd70054b7ae6ee1f Mon Sep 17 00:00:00 2001 From: Jeremy Benoist Date: Sun, 30 Oct 2016 09:58:53 +0100 Subject: Cleanup --- src/Wallabag/CoreBundle/Helper/ContentProxy.php | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'src/Wallabag/CoreBundle/Helper/ContentProxy.php') diff --git a/src/Wallabag/CoreBundle/Helper/ContentProxy.php b/src/Wallabag/CoreBundle/Helper/ContentProxy.php index 8ed11205..219b90d3 100644 --- a/src/Wallabag/CoreBundle/Helper/ContentProxy.php +++ b/src/Wallabag/CoreBundle/Helper/ContentProxy.php @@ -3,7 +3,7 @@ namespace Wallabag\CoreBundle\Helper; use Graby\Graby; -use Psr\Log\LoggerInterface as Logger; +use Psr\Log\LoggerInterface; use Wallabag\CoreBundle\Entity\Entry; use Wallabag\CoreBundle\Entity\Tag; use Wallabag\CoreBundle\Tools\Utils; @@ -20,7 +20,7 @@ class ContentProxy protected $logger; protected $tagRepository; - public function __construct(Graby $graby, RuleBasedTagger $tagger, TagRepository $tagRepository, Logger $logger) + public function __construct(Graby $graby, RuleBasedTagger $tagger, TagRepository $tagRepository, LoggerInterface $logger) { $this->graby = $graby; $this->tagger = $tagger; -- cgit v1.2.3 From 7f55941856549a3f5f45c42fdc171d66ff7ee297 Mon Sep 17 00:00:00 2001 From: Jeremy Benoist Date: Sun, 30 Oct 2016 10:48:29 +0100 Subject: Use doctrine event to download images --- src/Wallabag/CoreBundle/Helper/ContentProxy.php | 6 ------ 1 file changed, 6 deletions(-) (limited to 'src/Wallabag/CoreBundle/Helper/ContentProxy.php') diff --git a/src/Wallabag/CoreBundle/Helper/ContentProxy.php b/src/Wallabag/CoreBundle/Helper/ContentProxy.php index 219b90d3..d90d3dc8 100644 --- a/src/Wallabag/CoreBundle/Helper/ContentProxy.php +++ b/src/Wallabag/CoreBundle/Helper/ContentProxy.php @@ -75,12 +75,6 @@ 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'])) { -- cgit v1.2.3 From 48656e0eaac006a80f21e9aec8900747fe76283a Mon Sep 17 00:00:00 2001 From: Jeremy Benoist Date: Sun, 30 Oct 2016 11:27:09 +0100 Subject: Fixing tests --- src/Wallabag/CoreBundle/Helper/ContentProxy.php | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'src/Wallabag/CoreBundle/Helper/ContentProxy.php') diff --git a/src/Wallabag/CoreBundle/Helper/ContentProxy.php b/src/Wallabag/CoreBundle/Helper/ContentProxy.php index d90d3dc8..1986ab33 100644 --- a/src/Wallabag/CoreBundle/Helper/ContentProxy.php +++ b/src/Wallabag/CoreBundle/Helper/ContentProxy.php @@ -65,6 +65,7 @@ class ContentProxy $entry->setUrl($content['url'] ?: $url); $entry->setTitle($title); + $entry->setContent($html); $entry->setLanguage($content['language']); $entry->setMimetype($content['content_type']); @@ -75,8 +76,6 @@ class ContentProxy $entry->setDomainName($domainName); } - $entry->setContent($html); - if (isset($content['open_graph']['og_image'])) { $entry->setPreviewPicture($content['open_graph']['og_image']); } -- cgit v1.2.3