use Wallabag\CoreBundle\Tools\Utils;
/**
- * This kind of proxy class take care of getting the content from an url
- * and update the entry with what it found.
+ * This kind of proxy class takes care of getting the content from an url
+ * and updates the entry with what it found.
*/
class ContentProxy
{
protected $graby;
protected $tagger;
+ protected $ignoreOriginProcessor;
protected $validator;
protected $logger;
protected $mimeGuesser;
protected $eventDispatcher;
protected $storeArticleHeaders;
- public function __construct(Graby $graby, RuleBasedTagger $tagger, ValidatorInterface $validator, LoggerInterface $logger, $fetchingErrorMessage, $storeArticleHeaders = false)
+ public function __construct(Graby $graby, RuleBasedTagger $tagger, RuleBasedIgnoreOriginProcessor $ignoreOriginProcessor, ValidatorInterface $validator, LoggerInterface $logger, $fetchingErrorMessage, $storeArticleHeaders = false)
{
$this->graby = $graby;
$this->tagger = $tagger;
+ $this->ignoreOriginProcessor = $ignoreOriginProcessor;
$this->validator = $validator;
$this->logger = $logger;
$this->mimeGuesser = new MimeTypeExtensionGuesser();
*/
public function updateEntry(Entry $entry, $url, array $content = [], $disableContentUpdate = false)
{
+ $this->graby->toggleImgNoReferrer(true);
if (!empty($content['html'])) {
$content['html'] = $this->graby->cleanupHtml($content['html'], $url);
}
if ((empty($content) || false === $this->validateContent($content)) && false === $disableContentUpdate) {
$fetchedContent = $this->graby->fetchContent($url);
- $fetchedContent['title'] = $this->sanitizeContentTitle($fetchedContent['title'], $fetchedContent['content_type']);
+
+ $fetchedContent['title'] = $this->sanitizeContentTitle(
+ $fetchedContent['title'],
+ isset($fetchedContent['headers']['content-type']) ? $fetchedContent['headers']['content-type'] : ''
+ );
// 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
$entry->setUrl($url);
}
+ $entry->setGivenUrl($url);
+
$this->stockEntry($entry, $content);
}
/**
* 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)
/**
* Use a Symfony validator to ensure the preview picture is a real url.
*
- * @param Entry $entry
* @param string $value URL to validate and save
*/
public function updatePreviewPicture(Entry $entry, $value)
/**
* Update date.
*
- * @param Entry $entry
* @param string $value Date to validate and save
*/
public function updatePublishedAt(Entry $entry, $value)
/**
* Helper to extract and save host from entry url.
- *
- * @param Entry $entry
*/
public function setEntryDomainName(Entry $entry)
{
* Helper to set a default title using:
* - url basename, if applicable
* - hostname.
- *
- * @param Entry $entry
*/
public function setDefaultEntryTitle(Entry $entry)
{
/**
* Try to sanitize the title of the fetched content from wrong character encodings and invalid UTF-8 character.
*
- * @param $title
- * @param $contentType
+ * @param string $title
+ * @param string $contentType
*
* @return string
*/
if (!empty($content['title'])) {
$entry->setTitle($content['title']);
- } elseif (!empty($content['open_graph']['og_title'])) {
- $entry->setTitle($content['open_graph']['og_title']);
}
- $html = $content['html'];
- if (false === $html) {
- $html = $this->fetchingErrorMessage;
+ if (empty($content['html'])) {
+ $content['html'] = $this->fetchingErrorMessage;
- if (!empty($content['open_graph']['og_description'])) {
- $html .= '<p><i>But we found a short description: </i></p>';
- $html .= $content['open_graph']['og_description'];
+ if (!empty($content['description'])) {
+ $content['html'] .= '<p><i>But we found a short description: </i></p>';
+ $content['html'] .= $content['description'];
}
}
- $entry->setContent($html);
- $entry->setReadingTime(Utils::getReadingTime($html));
+ $entry->setContent($content['html']);
+ $entry->setReadingTime(Utils::getReadingTime($content['html']));
if (!empty($content['status'])) {
$entry->setHttpStatus($content['status']);
$entry->setPublishedBy($content['authors']);
}
- if (!empty($content['all_headers']) && $this->storeArticleHeaders) {
- $entry->setHeaders($content['all_headers']);
+ if (!empty($content['headers'])) {
+ $entry->setHeaders($content['headers']);
}
if (!empty($content['date'])) {
$this->updateLanguage($entry, $content['language']);
}
- if (!empty($content['open_graph']['og_image'])) {
- $this->updatePreviewPicture($entry, $content['open_graph']['og_image']);
+ $previewPictureUrl = '';
+ if (!empty($content['image'])) {
+ $previewPictureUrl = $content['image'];
}
// if content is an image, define it as a preview too
- if (!empty($content['content_type']) && \in_array($this->mimeGuesser->guess($content['content_type']), ['jpeg', 'jpg', 'gif', 'png'], true)) {
- $this->updatePreviewPicture($entry, $content['url']);
+ if (!empty($content['headers']['content-type']) && \in_array($this->mimeGuesser->guess($content['headers']['content-type']), ['jpeg', 'jpg', 'gif', 'png'], true)) {
+ $previewPictureUrl = $content['url'];
+ } elseif (empty($previewPictureUrl)) {
+ $this->logger->debug('Extracting images from content to provide a default preview picture');
+ $imagesUrls = DownloadImages::extractImagesUrlsFromHtml($content['html']);
+ $this->logger->debug(\count($imagesUrls) . ' pictures found');
+
+ if (!empty($imagesUrls)) {
+ $previewPictureUrl = $imagesUrls[0];
+ }
+ }
+
+ if (!empty($content['headers']['content-type'])) {
+ $entry->setMimetype($content['headers']['content-type']);
}
- if (!empty($content['content_type'])) {
- $entry->setMimetype($content['content_type']);
+ if (!empty($previewPictureUrl)) {
+ $this->updatePreviewPicture($entry, $previewPictureUrl);
}
try {
* Update the origin_url field when a redirection occurs
* This field is set if it is empty and new url does not match ignore list.
*
- * @param Entry $entry
* @param string $url
*/
private function updateOriginUrl(Entry $entry, $url)
{
- if (!empty($url) && $entry->getUrl() !== $url) {
- $parsed_entry_url = parse_url($entry->getUrl());
- $parsed_content_url = parse_url($url);
-
- $diff_ec = array_diff_assoc($parsed_entry_url, $parsed_content_url);
- $diff_ce = array_diff_assoc($parsed_content_url, $parsed_entry_url);
-
- $diff = array_merge($diff_ec, $diff_ce);
- $diff_keys = array_keys($diff);
- sort($diff_keys);
-
- if ($this->ignoreUrl($entry->getUrl())) {
- $entry->setUrl($url);
- } else {
- switch ($diff_keys) {
- case ['path']:
- if (($parsed_entry_url['path'] . '/' === $parsed_content_url['path']) // diff is trailing slash, we only replace the url of the entry
- || ($url === urldecode($entry->getUrl()))) { // we update entry url if new url is a decoded version of it, see EntryRepository#findByUrlAndUserId
- $entry->setUrl($url);
- }
- break;
- case ['scheme']:
- $entry->setUrl($url);
- break;
- case ['fragment']:
- case ['query']:
- case ['fragment', 'query']:
- // noop
- break;
- default:
- if (empty($entry->getOriginUrl())) {
- $entry->setOriginUrl($entry->getUrl());
- }
- $entry->setUrl($url);
- break;
- }
- }
+ if (empty($url) || $entry->getUrl() === $url) {
+ return false;
}
- }
- /**
- * Check entry url against an ignore list to replace with content url.
- *
- * XXX: move the ignore list in the database to let users handle it
- *
- * @param string $url url to test
- *
- * @return bool true if url matches ignore list otherwise false
- */
- private function ignoreUrl($url)
- {
- $ignored_hosts = ['feedproxy.google.com', 'feeds.reuters.com'];
- $ignored_patterns = ['https?://www\.lemonde\.fr/tiny.*'];
-
- $parsed_url = parse_url($url);
-
- $filtered = array_filter($ignored_hosts, function ($var) use ($parsed_url) {
- return $var === $parsed_url['host'];
- });
+ $parsed_entry_url = parse_url($entry->getUrl());
+ $parsed_content_url = parse_url($url);
+
+ /**
+ * The following part computes the list of part changes between two
+ * parse_url arrays.
+ *
+ * As array_diff_assoc only computes changes to go from the left array
+ * to the right one, we make two differents arrays to have both
+ * directions. We merge these two arrays and sort keys before passing
+ * the result to the switch.
+ *
+ * The resulting array gives us all changing parts between the two
+ * urls: scheme, host, path, query and/or fragment.
+ */
+ $diff_ec = array_diff_assoc($parsed_entry_url, $parsed_content_url);
+ $diff_ce = array_diff_assoc($parsed_content_url, $parsed_entry_url);
+
+ $diff = array_merge($diff_ec, $diff_ce);
+ $diff_keys = array_keys($diff);
+ sort($diff_keys);
+
+ if ($this->ignoreOriginProcessor->process($entry)) {
+ $entry->setUrl($url);
- if ([] !== $filtered) {
- return true;
+ return false;
}
- $filtered = array_filter($ignored_patterns, function ($var) use ($url) {
- return preg_match("`$var`i", $url);
- });
-
- if ([] !== $filtered) {
- return true;
+ /**
+ * This switch case lets us apply different behaviors according to
+ * changing parts of urls.
+ *
+ * As $diff_keys is an array, we provide arrays as cases. ['path'] means
+ * 'only the path is different between the two urls' whereas
+ * ['fragment', 'query'] means 'only fragment and query string parts are
+ * different between the two urls'.
+ *
+ * Note that values in $diff_keys are sorted.
+ */
+ switch ($diff_keys) {
+ case ['path']:
+ if (($parsed_entry_url['path'] . '/' === $parsed_content_url['path']) // diff is trailing slash, we only replace the url of the entry
+ || ($url === urldecode($entry->getUrl()))) { // we update entry url if new url is a decoded version of it, see EntryRepository#findByUrlAndUserId
+ $entry->setUrl($url);
+ }
+ break;
+ case ['scheme']:
+ $entry->setUrl($url);
+ break;
+ case ['fragment']:
+ // noop
+ break;
+ default:
+ if (empty($entry->getOriginUrl())) {
+ $entry->setOriginUrl($entry->getUrl());
+ }
+ $entry->setUrl($url);
+ break;
}
-
- return false;
}
/**
* Validate that the given content has at least a title, an html and a url.
*
- * @param array $content
- *
* @return bool true if valid otherwise false
*/
private function validateContent(array $content)