namespace Wallabag\CoreBundle\Helper;
use Graby\Graby;
-use Psr\Log\LoggerInterface as Logger;
+use Psr\Log\LoggerInterface;
+use Symfony\Component\HttpFoundation\File\MimeType\MimeTypeExtensionGuesser;
+use Symfony\Component\Validator\Constraints\Locale as LocaleConstraint;
+use Symfony\Component\Validator\Constraints\Url as UrlConstraint;
+use Symfony\Component\Validator\Validator\ValidatorInterface;
use Wallabag\CoreBundle\Entity\Entry;
-use Wallabag\CoreBundle\Entity\Tag;
use Wallabag\CoreBundle\Tools\Utils;
-use Wallabag\CoreBundle\Repository\TagRepository;
/**
* This kind of proxy class take care of getting the content from an url
{
protected $graby;
protected $tagger;
+ protected $validator;
protected $logger;
- protected $tagRepository;
+ protected $mimeGuesser;
+ protected $fetchingErrorMessage;
+ protected $eventDispatcher;
+ protected $storeArticleHeaders;
- public function __construct(Graby $graby, RuleBasedTagger $tagger, TagRepository $tagRepository, Logger $logger)
+ public function __construct(Graby $graby, RuleBasedTagger $tagger, ValidatorInterface $validator, LoggerInterface $logger, $fetchingErrorMessage, $storeArticleHeaders = false)
{
$this->graby = $graby;
$this->tagger = $tagger;
+ $this->validator = $validator;
$this->logger = $logger;
- $this->tagRepository = $tagRepository;
+ $this->mimeGuesser = new MimeTypeExtensionGuesser();
+ $this->fetchingErrorMessage = $fetchingErrorMessage;
+ $this->storeArticleHeaders = $storeArticleHeaders;
}
/**
- * Fetch content using graby and hydrate given entry with results information.
- * In case we couldn't find content, we'll try to use Open Graph data.
+ * Update entry using either fetched or provided content.
*
- * We can also force the content, in case of an import from the v1 for example, so the function won't
- * fetch the content from the website but rather use information given with the $content parameter.
- *
- * @param Entry $entry Entry to update
- * @param string $url Url to grab content for
- * @param array $content An array with AT LEAST keys title, html, url, language & content_type to skip the fetchContent from the url
- *
- * @return Entry
+ * @param Entry $entry Entry to update
+ * @param string $url Url of the content
+ * @param array $content Array with content provided for import with AT LEAST keys title, html, url to skip the fetchContent from the url
+ * @param bool $disableContentUpdate Whether to skip trying to fetch content using Graby
*/
- public function updateEntry(Entry $entry, $url, array $content = [])
+ public function updateEntry(Entry $entry, $url, array $content = [], $disableContentUpdate = false)
{
- // do we have to fetch the content or the provided one is ok?
- if (empty($content) || false === $this->validateContent($content)) {
- $content = $this->graby->fetchContent($url);
- }
-
- $title = $content['title'];
- if (!$title && isset($content['open_graph']['og_title'])) {
- $title = $content['open_graph']['og_title'];
+ if (!empty($content['html'])) {
+ $content['html'] = $this->graby->cleanupHtml($content['html'], $url);
}
- $html = $content['html'];
- if (false === $html) {
- $html = '<p>Unable to retrieve readable content.</p>';
+ if ((empty($content) || false === $this->validateContent($content)) && false === $disableContentUpdate) {
+ $fetchedContent = $this->graby->fetchContent($url);
- if (isset($content['open_graph']['og_description'])) {
- $html .= '<p><i>But we found a short description: </i></p>';
- $html .= $content['open_graph']['og_description'];
+ // when content is imported, we have information in $content
+ // in case fetching content goes bad, we'll keep the imported information instead of overriding them
+ if (empty($content) || $fetchedContent['html'] !== $this->fetchingErrorMessage) {
+ $content = $fetchedContent;
}
}
- $entry->setUrl($content['url'] ?: $url);
- $entry->setTitle($title);
- $entry->setContent($html);
- $entry->setLanguage($content['language']);
- $entry->setMimetype($content['content_type']);
- $entry->setReadingTime(Utils::getReadingTime($html));
+ // be sure to keep the url in case of error
+ // so we'll be able to refetch it in the future
+ $content['url'] = !empty($content['url']) ? $content['url'] : $url;
- $domainName = parse_url($entry->getUrl(), PHP_URL_HOST);
- if (false !== $domainName) {
- $entry->setDomainName($domainName);
+ // In one case (at least in tests), url is empty here
+ // so we set it using $url provided in the updateEntry call.
+ // Not sure what are the other possible cases where this property is empty
+ if (empty($entry->getUrl()) && !empty($url)) {
+ $entry->setUrl($url);
}
- if (isset($content['open_graph']['og_image'])) {
- $entry->setPreviewPicture($content['open_graph']['og_image']);
- }
+ $this->stockEntry($entry, $content);
+ }
- try {
- $this->tagger->tag($entry);
- } catch (\Exception $e) {
- $this->logger->error('Error while trying to automatically tag an entry.', [
- 'entry_url' => $url,
- 'error_msg' => $e->getMessage(),
- ]);
+ /**
+ * Use a Symfony validator to ensure the language is well formatted.
+ *
+ * @param Entry $entry
+ * @param string $value Language to validate and save
+ */
+ public function updateLanguage(Entry $entry, $value)
+ {
+ // some lang are defined as fr-FR, es-ES.
+ // replacing - by _ might increase language support
+ $value = str_replace('-', '_', $value);
+
+ $errors = $this->validator->validate(
+ $value,
+ (new LocaleConstraint())
+ );
+
+ if (0 === count($errors)) {
+ $entry->setLanguage($value);
+
+ return;
}
- return $entry;
+ $this->logger->warning('Language validation failed. ' . (string) $errors);
}
/**
- * Assign some tags to an entry.
+ * Use a Symfony validator to ensure the preview picture is a real url.
*
- * @param Entry $entry
- * @param array|string $tags An array of tag or a string coma separated of tag
- * @param array $entitiesReady Entities from the EntityManager which are persisted but not yet flushed
- * It is mostly to fix duplicate tag on import @see http://stackoverflow.com/a/7879164/569101
+ * @param Entry $entry
+ * @param string $value URL to validate and save
*/
- public function assignTagsToEntry(Entry $entry, $tags, array $entitiesReady = [])
+ public function updatePreviewPicture(Entry $entry, $value)
{
- if (!is_array($tags)) {
- $tags = explode(',', $tags);
- }
+ $errors = $this->validator->validate(
+ $value,
+ (new UrlConstraint())
+ );
- // keeps only Tag entity from the "not yet flushed entities"
- $tagsNotYetFlushed = [];
- foreach ($entitiesReady as $entity) {
- if ($entity instanceof Tag) {
- $tagsNotYetFlushed[$entity->getLabel()] = $entity;
- }
+ if (0 === count($errors)) {
+ $entry->setPreviewPicture($value);
+
+ return;
}
- foreach ($tags as $label) {
- $label = trim($label);
+ $this->logger->warning('PreviewPicture validation failed. ' . (string) $errors);
+ }
- // avoid empty tag
- if (0 === strlen($label)) {
- continue;
- }
+ /**
+ * Update date.
+ *
+ * @param Entry $entry
+ * @param string $value Date to validate and save
+ */
+ public function updatePublishedAt(Entry $entry, $value)
+ {
+ $date = $value;
- if (isset($tagsNotYetFlushed[$label])) {
- $tagEntity = $tagsNotYetFlushed[$label];
- } else {
- $tagEntity = $this->tagRepository->findOneByLabel($label);
+ // is it a timestamp?
+ if (false !== filter_var($date, FILTER_VALIDATE_INT)) {
+ $date = '@' . $date;
+ }
- if (is_null($tagEntity)) {
- $tagEntity = new Tag();
- $tagEntity->setLabel($label);
- }
+ try {
+ // is it already a DateTime?
+ // (it's inside the try/catch in case of fail to be parse time string)
+ if (!$date instanceof \DateTime) {
+ $date = new \DateTime($date);
}
- // only add the tag on the entry if the relation doesn't exist
- if (false === $entry->getTags()->contains($tagEntity)) {
- $entry->addTag($tagEntity);
- }
+ $entry->setPublishedAt($date);
+ } catch (\Exception $e) {
+ $this->logger->warning('Error while defining date', ['e' => $e, 'url' => $entry->getUrl(), 'date' => $value]);
}
}
/**
- * Validate that the given content as enough value to be used
- * instead of fetch the content from the url.
+ * Helper to extract and save host from entry url.
*
- * @param array $content
- *
- * @return bool true if valid otherwise false
+ * @param Entry $entry
*/
- private function validateContent(array $content)
+ public function setEntryDomainName(Entry $entry)
{
- return isset($content['title']) && isset($content['html']) && isset($content['url']) && isset($content['language']) && isset($content['content_type']);
+ $domainName = parse_url($entry->getUrl(), PHP_URL_HOST);
+ if (false !== $domainName) {
+ $entry->setDomainName($domainName);
+ }
}
/**
- * Changing pictures URL in article content.
+ * Helper to set a default title using:
+ * - url basename, if applicable
+ * - hostname.
+ *
+ * @param Entry $entry
*/
- public static function filterPicture($content, $url, $id)
+ public function setDefaultEntryTitle(Entry $entry)
{
- $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;
- }
+ $url = parse_url($entry->getUrl());
+ $path = pathinfo($url['path'], PATHINFO_BASENAME);
+
+ if (empty($path)) {
+ $path = $url['host'];
}
- return $content;
+ $entry->setTitle($path);
}
/**
- * Get absolute URL.
+ * Stock entry with fetched or imported content.
+ * Will fall back to OpenGraph data if available.
+ *
+ * @param Entry $entry Entry to stock
+ * @param array $content Array with at least title, url & html
*/
- private static function _getAbsoluteLink($relativeLink, $url)
+ private function stockEntry(Entry $entry, array $content)
{
- /* return if already absolute URL */
- if (parse_url($relativeLink, PHP_URL_SCHEME) != '') {
- return $relativeLink;
- }
-
- /* queries and anchors */
- if ($relativeLink[0] == '#' || $relativeLink[0] == '?') {
- return $url.$relativeLink;
+ // When a redirection occurs while fetching an entry
+ // we move the original url in origin_url property if empty
+ // and set the entry url with the final value
+ if (!empty($content['url']) && $entry->getUrl() !== $content['url']) {
+ if (empty($entry->getOriginUrl())) {
+ $entry->setOriginUrl($entry->getUrl());
+ }
+ $entry->setUrl($content['url']);
}
- /* parse base URL and convert to local variables:
- $scheme, $host, $path */
- extract(parse_url($url));
+ $this->setEntryDomainName($entry);
- /* remove non-directory element from path */
- $path = preg_replace('#/[^/]*$#', '', $path);
-
- /* destroy path if relative url points to root */
- if ($relativeLink[0] == '/') {
- $path = '';
+ if (!empty($content['title'])) {
+ $entry->setTitle($content['title']);
+ } elseif (!empty($content['open_graph']['og_title'])) {
+ $entry->setTitle($content['open_graph']['og_title']);
}
- /* dirty absolute URL */
- $abs = $host.$path.'/'.$relativeLink;
+ $html = $content['html'];
+ if (false === $html) {
+ $html = $this->fetchingErrorMessage;
- /* replace '//' or '/./' or '/foo/../' with '/' */
- $re = array('#(/\.?/)#', '#/(?!\.\.)[^/]+/\.\./#');
- for ($n = 1; $n > 0; $abs = preg_replace($re, '/', $abs, -1, $n)) {
+ if (!empty($content['open_graph']['og_description'])) {
+ $html .= '<p><i>But we found a short description: </i></p>';
+ $html .= $content['open_graph']['og_description'];
+ }
}
- /* 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);
+ $entry->setContent($html);
+ $entry->setReadingTime(Utils::getReadingTime($html));
- if (file_exists($fullpath)) {
- unlink($fullpath);
+ if (!empty($content['status'])) {
+ $entry->setHttpStatus($content['status']);
}
- // 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;
+ if (!empty($content['authors']) && is_array($content['authors'])) {
+ $entry->setPublishedBy($content['authors']);
}
- // 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;
+ if (!empty($content['all_headers']) && $this->storeArticleHeaders) {
+ $entry->setHeaders($content['all_headers']);
}
- // regenerate image
- $im = imagecreatefromstring($rawdata);
- if ($im === false) {
- Tools::logm('error while regenerating image '.$fullpath);
-
- return false;
+ if (!empty($content['date'])) {
+ $this->updatePublishedAt($entry, $content['date']);
}
- switch ($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, $fullpath, ceil(REGENERATE_PICTURES_QUALITY / 100 * 9));
- break;
+ if (!empty($content['language'])) {
+ $this->updateLanguage($entry, $content['language']);
}
- imagedestroy($im);
- return $result;
- }
+ if (!empty($content['open_graph']['og_image'])) {
+ $this->updatePreviewPicture($entry, $content['open_graph']['og_image']);
+ }
- /**
- * 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($assets_path, 0715);
+ // if content is an image, define it as a preview too
+ if (!empty($content['content_type']) && in_array($this->mimeGuesser->guess($content['content_type']), ['jpeg', 'jpg', 'gif', 'png'], true)) {
+ $this->updatePreviewPicture($entry, $content['url']);
}
- $article_directory = $assets_path.$id;
- if (!is_dir($article_directory)) {
- mkdir($article_directory, 0715);
+ if (!empty($content['content_type'])) {
+ $entry->setMimetype($content['content_type']);
}
- return $article_directory;
+ try {
+ $this->tagger->tag($entry);
+ } catch (\Exception $e) {
+ $this->logger->error('Error while trying to automatically tag an entry.', [
+ 'entry_url' => $content['url'],
+ 'error_msg' => $e->getMessage(),
+ ]);
+ }
}
/**
- * Remove the directory.
+ * Validate that the given content has at least a title, an html and a url.
*
- * @param $directory
+ * @param array $content
*
- * @return bool
+ * @return bool true if valid otherwise false
*/
- public static function removeDirectory($directory)
+ private function validateContent(array $content)
{
- if (is_dir($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);
- }
+ return !empty($content['title']) && !empty($content['html']) && !empty($content['url']);
}
}