3 namespace Wallabag\CoreBundle\Helper
;
6 use Psr\Log\LoggerInterface
;
7 use Symfony\Component\HttpFoundation\File\MimeType\MimeTypeExtensionGuesser
;
8 use Symfony\Component\Validator\Constraints\Locale
as LocaleConstraint
;
9 use Symfony\Component\Validator\Constraints\Url
as UrlConstraint
;
10 use Symfony\Component\Validator\Validator\ValidatorInterface
;
11 use Wallabag\CoreBundle\Entity\Entry
;
12 use Wallabag\CoreBundle\Tools\Utils
;
15 * This kind of proxy class takes care of getting the content from an url
16 * and updates the entry with what it found.
22 protected $ignoreOriginProcessor;
25 protected $mimeGuesser;
26 protected $fetchingErrorMessage;
27 protected $eventDispatcher;
28 protected $storeArticleHeaders;
30 public function __construct(Graby
$graby, RuleBasedTagger
$tagger, RuleBasedIgnoreOriginProcessor
$ignoreOriginProcessor, ValidatorInterface
$validator, LoggerInterface
$logger, $fetchingErrorMessage, $storeArticleHeaders = false)
32 $this->graby
= $graby;
33 $this->tagger
= $tagger;
34 $this->ignoreOriginProcessor
= $ignoreOriginProcessor;
35 $this->validator
= $validator;
36 $this->logger
= $logger;
37 $this->mimeGuesser
= new MimeTypeExtensionGuesser();
38 $this->fetchingErrorMessage
= $fetchingErrorMessage;
39 $this->storeArticleHeaders
= $storeArticleHeaders;
43 * Update entry using either fetched or provided content.
45 * @param Entry $entry Entry to update
46 * @param string $url Url of the content
47 * @param array $content Array with content provided for import with AT LEAST keys title, html, url to skip the fetchContent from the url
48 * @param bool $disableContentUpdate Whether to skip trying to fetch content using Graby
50 public function updateEntry(Entry
$entry, $url, array $content = [], $disableContentUpdate = false)
52 $this->graby
->toggleImgNoReferrer(true);
53 if (!empty($content['html'])) {
54 $content['html'] = $this->graby
->cleanupHtml($content['html'], $url);
57 if ((empty($content) || false === $this->validateContent($content)) && false === $disableContentUpdate) {
58 $fetchedContent = $this->graby
->fetchContent($url);
60 $fetchedContent['title'] = $this->sanitizeContentTitle(
61 $fetchedContent['title'],
62 isset($fetchedContent['headers']['content-type']) ? $fetchedContent['headers']['content-type'] : ''
65 // when content is imported, we have information in $content
66 // in case fetching content goes bad, we'll keep the imported information instead of overriding them
67 if (empty($content) || $fetchedContent['html'] !== $this->fetchingErrorMessage
) {
68 $content = $fetchedContent;
72 // be sure to keep the url in case of error
73 // so we'll be able to refetch it in the future
74 $content['url'] = !empty($content['url']) ? $content['url'] : $url;
76 // In one case (at least in tests), url is empty here
77 // so we set it using $url provided in the updateEntry call.
78 // Not sure what are the other possible cases where this property is empty
79 if (empty($entry->getUrl()) && !empty($url)) {
83 $entry->setGivenUrl($url);
85 $this->stockEntry($entry, $content);
89 * Use a Symfony validator to ensure the language is well formatted.
91 * @param string $value Language to validate and save
93 public function updateLanguage(Entry
$entry, $value)
95 // some lang are defined as fr-FR, es-ES.
96 // replacing - by _ might increase language support
97 $value = str_replace('-', '_', $value);
99 $errors = $this->validator
->validate(
101 (new LocaleConstraint())
104 if (0 === \
count($errors)) {
105 $entry->setLanguage($value);
110 $this->logger
->warning('Language validation failed. ' . (string) $errors);
114 * Use a Symfony validator to ensure the preview picture is a real url.
116 * @param string $value URL to validate and save
118 public function updatePreviewPicture(Entry
$entry, $value)
120 $errors = $this->validator
->validate(
122 (new UrlConstraint())
125 if (0 === \
count($errors)) {
126 $entry->setPreviewPicture($value);
131 $this->logger
->warning('PreviewPicture validation failed. ' . (string) $errors);
137 * @param string $value Date to validate and save
139 public function updatePublishedAt(Entry
$entry, $value)
143 // is it a timestamp?
144 if (false !== filter_var($date, FILTER_VALIDATE_INT
)) {
149 // is it already a DateTime?
150 // (it's inside the try/catch in case of fail to be parse time string)
151 if (!$date instanceof \DateTime
) {
152 $date = new \
DateTime($date);
155 $entry->setPublishedAt($date);
156 } catch (\Exception
$e) {
157 $this->logger
->warning('Error while defining date', ['e' => $e, 'url' => $entry->getUrl(), 'date' => $value]);
162 * Helper to extract and save host from entry url.
164 public function setEntryDomainName(Entry
$entry)
166 $domainName = parse_url($entry->getUrl(), PHP_URL_HOST
);
167 if (false !== $domainName) {
168 $entry->setDomainName($domainName);
173 * Helper to set a default title using:
174 * - url basename, if applicable
177 public function setDefaultEntryTitle(Entry
$entry)
179 $url = parse_url($entry->getUrl());
180 $path = pathinfo($url['path'], PATHINFO_BASENAME
);
183 $path = $url['host'];
186 $entry->setTitle($path);
190 * Try to sanitize the title of the fetched content from wrong character encodings and invalid UTF-8 character.
192 * @param string $title
193 * @param string $contentType
197 private function sanitizeContentTitle($title, $contentType)
199 if ('application/pdf' === $contentType) {
200 $title = $this->convertPdfEncodingToUTF8($title);
203 return $this->sanitizeUTF8Text($title);
207 * If the title from the fetched content comes from a PDF, then its very possible that the character encoding is not
208 * UTF-8. This methods tries to identify the character encoding and translate the title to UTF-8.
212 * @return string (maybe contains invalid UTF-8 character)
214 private function convertPdfEncodingToUTF8($title)
216 // first try UTF-8 because its easier to detect its present/absence
217 foreach (['UTF-8', 'UTF-16BE', 'WINDOWS-1252'] as $encoding) {
218 if (mb_check_encoding($title, $encoding)) {
219 return mb_convert_encoding($title, 'UTF-8', $encoding);
227 * Remove invalid UTF-8 characters from the given string.
229 * @param string $rawText
233 private function sanitizeUTF8Text($rawText)
235 if (mb_check_encoding($rawText, 'UTF-8')) {
239 return iconv('UTF-8', 'UTF-8//IGNORE', $rawText);
243 * Stock entry with fetched or imported content.
244 * Will fall back to OpenGraph data if available.
246 * @param Entry $entry Entry to stock
247 * @param array $content Array with at least title, url & html
249 private function stockEntry(Entry
$entry, array $content)
251 $this->updateOriginUrl($entry, $content['url']);
253 $this->setEntryDomainName($entry);
255 if (!empty($content['title'])) {
256 $entry->setTitle($content['title']);
259 if (empty($content['html'])) {
260 $content['html'] = $this->fetchingErrorMessage
;
262 if (!empty($content['description'])) {
263 $content['html'] .= '<p><i>But we found a short description: </i></p>';
264 $content['html'] .= $content['description'];
268 $entry->setContent($content['html']);
269 $entry->setReadingTime(Utils
::getReadingTime($content['html']));
271 if (!empty($content['status'])) {
272 $entry->setHttpStatus($content['status']);
275 if (!empty($content['authors']) && \
is_array($content['authors'])) {
276 $entry->setPublishedBy($content['authors']);
279 if (!empty($content['headers'])) {
280 $entry->setHeaders($content['headers']);
283 if (!empty($content['date'])) {
284 $this->updatePublishedAt($entry, $content['date']);
287 if (!empty($content['language'])) {
288 $this->updateLanguage($entry, $content['language']);
291 $previewPictureUrl = '';
292 if (!empty($content['image'])) {
293 $previewPictureUrl = $content['image'];
296 // if content is an image, define it as a preview too
297 if (!empty($content['headers']['content-type']) && \
in_array($this->mimeGuesser
->guess($content['headers']['content-type']), ['jpeg', 'jpg', 'gif', 'png'], true)) {
298 $previewPictureUrl = $content['url'];
299 } elseif (empty($previewPictureUrl)) {
300 $this->logger
->debug('Extracting images from content to provide a default preview picture');
301 $imagesUrls = DownloadImages
::extractImagesUrlsFromHtml($content['html']);
302 $this->logger
->debug(\
count($imagesUrls) . ' pictures found');
304 if (!empty($imagesUrls)) {
305 $previewPictureUrl = $imagesUrls[0];
309 if (!empty($content['headers']['content-type'])) {
310 $entry->setMimetype($content['headers']['content-type']);
313 if (!empty($previewPictureUrl)) {
314 $this->updatePreviewPicture($entry, $previewPictureUrl);
318 $this->tagger
->tag($entry);
319 } catch (\Exception
$e) {
320 $this->logger
->error('Error while trying to automatically tag an entry.', [
321 'entry_url' => $content['url'],
322 'error_msg' => $e->getMessage(),
328 * Update the origin_url field when a redirection occurs
329 * This field is set if it is empty and new url does not match ignore list.
333 private function updateOriginUrl(Entry
$entry, $url)
335 if (empty($url) || $entry->getUrl() === $url) {
339 $parsed_entry_url = parse_url($entry->getUrl());
340 $parsed_content_url = parse_url($url);
343 * The following part computes the list of part changes between two
346 * As array_diff_assoc only computes changes to go from the left array
347 * to the right one, we make two differents arrays to have both
348 * directions. We merge these two arrays and sort keys before passing
349 * the result to the switch.
351 * The resulting array gives us all changing parts between the two
352 * urls: scheme, host, path, query and/or fragment.
354 $diff_ec = array_diff_assoc($parsed_entry_url, $parsed_content_url);
355 $diff_ce = array_diff_assoc($parsed_content_url, $parsed_entry_url);
357 $diff = array_merge($diff_ec, $diff_ce);
358 $diff_keys = array_keys($diff);
361 if ($this->ignoreOriginProcessor
->process($entry)) {
362 $entry->setUrl($url);
368 * This switch case lets us apply different behaviors according to
369 * changing parts of urls.
371 * As $diff_keys is an array, we provide arrays as cases. ['path'] means
372 * 'only the path is different between the two urls' whereas
373 * ['fragment', 'query'] means 'only fragment and query string parts are
374 * different between the two urls'.
376 * Note that values in $diff_keys are sorted.
378 switch ($diff_keys) {
380 if (($parsed_entry_url['path'] . '/' === $parsed_content_url['path']) // diff is trailing slash, we only replace the url of the entry
381 || ($url === urldecode($entry->getUrl()))) { // we update entry url if new url is a decoded version of it, see EntryRepository#findByUrlAndUserId
382 $entry->setUrl($url);
386 $entry->setUrl($url);
392 if (empty($entry->getOriginUrl())) {
393 $entry->setOriginUrl($entry->getUrl());
395 $entry->setUrl($url);
401 * Validate that the given content has at least a title, an html and a url.
403 * @return bool true if valid otherwise false
405 private function validateContent(array $content)
407 return !empty($content['title']) && !empty($content['html']) && !empty($content['url']);