X-Git-Url: https://git.immae.eu/?a=blobdiff_plain;ds=sidebyside;f=src%2FWallabag%2FCoreBundle%2FHelper%2FContentProxy.php;h=7e93249ddd8dd8d708ea3a868065b0556a487c96;hb=HEAD;hp=fab052685ccef748c6aba14591e05fc7866780cb;hpb=f80f16dfc858ec90da76daacd405b0cfdaa32f74;p=github%2Fwallabag%2Fwallabag.git
diff --git a/src/Wallabag/CoreBundle/Helper/ContentProxy.php b/src/Wallabag/CoreBundle/Helper/ContentProxy.php
index fab05268..7e93249d 100644
--- a/src/Wallabag/CoreBundle/Helper/ContentProxy.php
+++ b/src/Wallabag/CoreBundle/Helper/ContentProxy.php
@@ -12,13 +12,14 @@ use Wallabag\CoreBundle\Entity\Entry;
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;
@@ -26,10 +27,11 @@ class ContentProxy
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();
@@ -47,13 +49,18 @@ class ContentProxy
*/
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
@@ -66,55 +73,21 @@ class ContentProxy
// so we'll be able to refetch it in the future
$content['url'] = !empty($content['url']) ? $content['url'] : $url;
- $this->stockEntry($entry, $content);
- }
-
- /**
- * Try to sanitize the title of the fetched content from wrong character encodings and invalid UTF-8 character.
- * @param $title
- * @param $contentType
- * @return string
- */
- private function sanitizeContentTitle($title, $contentType) {
- if ('application/pdf' === $contentType) {
- $convertedTitle = $this->convertPdfEncodingToUTF8($title);
- return $this->sanitizeUTF8Text($convertedTitle);
+ // 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);
}
- return $this->sanitizeUTF8Text($title);
- }
- /**
- * If the title from the fetched content comes from a PDF, then its very possible that the character encoding is not
- * UTF-8. This methods tries to identify the character encoding and translate the title to UTF-8.
- * @param $title
- * @return string (maybe contains invalid UTF-8 character)
- */
- private function convertPdfEncodingToUTF8($title) {
- // first try UTF-16 (then UTF-8) because its easier to detect its present/absence
- foreach (array('UTF-16BE', 'UTF-16LE', 'UTF-8', 'WINDOWS-1252') as $encoding) {
- if (mb_check_encoding($title, $encoding)) {
- return mb_convert_encoding($title, 'UTF-8', $encoding);
- }
- }
- return $title;
- }
+ $entry->setGivenUrl($url);
- /**
- * Remove invalid UTF-8 characters from the given string.
- * @param String $rawText
- * @return string
- */
- private function sanitizeUTF8Text($rawText) {
- if (mb_check_encoding($rawText, 'UTF-8')) {
- return $rawText;
- }
- return iconv("UTF-8", "UTF-8//IGNORE", $rawText);
+ $this->stockEntry($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)
@@ -140,7 +113,6 @@ class ContentProxy
/**
* 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)
@@ -162,7 +134,6 @@ class ContentProxy
/**
* Update date.
*
- * @param Entry $entry
* @param string $value Date to validate and save
*/
public function updatePublishedAt(Entry $entry, $value)
@@ -189,8 +160,6 @@ class ContentProxy
/**
* Helper to extract and save host from entry url.
- *
- * @param Entry $entry
*/
public function setEntryDomainName(Entry $entry)
{
@@ -204,8 +173,6 @@ class ContentProxy
* Helper to set a default title using:
* - url basename, if applicable
* - hostname.
- *
- * @param Entry $entry
*/
public function setDefaultEntryTitle(Entry $entry)
{
@@ -219,6 +186,59 @@ class ContentProxy
$entry->setTitle($path);
}
+ /**
+ * Try to sanitize the title of the fetched content from wrong character encodings and invalid UTF-8 character.
+ *
+ * @param string $title
+ * @param string $contentType
+ *
+ * @return string
+ */
+ private function sanitizeContentTitle($title, $contentType)
+ {
+ if ('application/pdf' === $contentType) {
+ $title = $this->convertPdfEncodingToUTF8($title);
+ }
+
+ return $this->sanitizeUTF8Text($title);
+ }
+
+ /**
+ * If the title from the fetched content comes from a PDF, then its very possible that the character encoding is not
+ * UTF-8. This methods tries to identify the character encoding and translate the title to UTF-8.
+ *
+ * @param $title
+ *
+ * @return string (maybe contains invalid UTF-8 character)
+ */
+ private function convertPdfEncodingToUTF8($title)
+ {
+ // first try UTF-8 because its easier to detect its present/absence
+ foreach (['UTF-8', 'UTF-16BE', 'WINDOWS-1252'] as $encoding) {
+ if (mb_check_encoding($title, $encoding)) {
+ return mb_convert_encoding($title, 'UTF-8', $encoding);
+ }
+ }
+
+ return $title;
+ }
+
+ /**
+ * Remove invalid UTF-8 characters from the given string.
+ *
+ * @param string $rawText
+ *
+ * @return string
+ */
+ private function sanitizeUTF8Text($rawText)
+ {
+ if (mb_check_encoding($rawText, 'UTF-8')) {
+ return $rawText;
+ }
+
+ return iconv('UTF-8', 'UTF-8//IGNORE', $rawText);
+ }
+
/**
* Stock entry with fetched or imported content.
* Will fall back to OpenGraph data if available.
@@ -228,28 +248,25 @@ class ContentProxy
*/
private function stockEntry(Entry $entry, array $content)
{
- $entry->setUrl($content['url']);
+ $this->updateOriginUrl($entry, $content['url']);
$this->setEntryDomainName($entry);
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 .= 'But we found a short description:
';
- $html .= $content['open_graph']['og_description'];
+ if (!empty($content['description'])) {
+ $content['html'] .= 'But we found a short description:
';
+ $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']);
@@ -259,8 +276,8 @@ class ContentProxy
$entry->setPublishedBy($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'])) {
@@ -271,17 +288,30 @@ class ContentProxy
$this->updateLanguage($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['content_type'])) {
- $entry->setMimetype($content['content_type']);
+ if (!empty($content['headers']['content-type'])) {
+ $entry->setMimetype($content['headers']['content-type']);
+ }
+
+ if (!empty($previewPictureUrl)) {
+ $this->updatePreviewPicture($entry, $previewPictureUrl);
}
try {
@@ -295,9 +325,80 @@ class ContentProxy
}
/**
- * Validate that the given content has at least a title, an html and a url.
+ * 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 array $content
+ * @param string $url
+ */
+ private function updateOriginUrl(Entry $entry, $url)
+ {
+ if (empty($url) || $entry->getUrl() === $url) {
+ return false;
+ }
+
+ $parsed_entry_url = parse_url($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);
+
+ return false;
+ }
+
+ /**
+ * 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;
+ }
+ }
+
+ /**
+ * Validate that the given content has at least a title, an html and a url.
*
* @return bool true if valid otherwise false
*/