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.
24 protected $mimeGuesser;
25 protected $fetchingErrorMessage;
26 protected $eventDispatcher;
27 protected $storeArticleHeaders;
29 public function __construct(Graby
$graby, RuleBasedTagger
$tagger, ValidatorInterface
$validator, LoggerInterface
$logger, $fetchingErrorMessage, $storeArticleHeaders = false)
31 $this->graby
= $graby;
32 $this->tagger
= $tagger;
33 $this->validator
= $validator;
34 $this->logger
= $logger;
35 $this->mimeGuesser
= new MimeTypeExtensionGuesser();
36 $this->fetchingErrorMessage
= $fetchingErrorMessage;
37 $this->storeArticleHeaders
= $storeArticleHeaders;
41 * Update entry using either fetched or provided content.
43 * @param Entry $entry Entry to update
44 * @param string $url Url of the content
45 * @param array $content Array with content provided for import with AT LEAST keys title, html, url to skip the fetchContent from the url
46 * @param bool $disableContentUpdate Whether to skip trying to fetch content using Graby
48 public function updateEntry(Entry
$entry, $url, array $content = [], $disableContentUpdate = false)
50 $this->graby
->toggleImgNoReferrer(true);
51 if (!empty($content['html'])) {
52 $content['html'] = $this->graby
->cleanupHtml($content['html'], $url);
55 if ((empty($content) || false === $this->validateContent($content)) && false === $disableContentUpdate) {
56 $fetchedContent = $this->graby
->fetchContent($url);
58 $fetchedContent['title'] = $this->sanitizeContentTitle(
59 $fetchedContent['title'],
60 isset($fetchedContent['headers']['content-type']) ? $fetchedContent['headers']['content-type'] : ''
63 // when content is imported, we have information in $content
64 // in case fetching content goes bad, we'll keep the imported information instead of overriding them
65 if (empty($content) || $fetchedContent['html'] !== $this->fetchingErrorMessage
) {
66 $content = $fetchedContent;
70 // be sure to keep the url in case of error
71 // so we'll be able to refetch it in the future
72 $content['url'] = !empty($content['url']) ? $content['url'] : $url;
74 // In one case (at least in tests), url is empty here
75 // so we set it using $url provided in the updateEntry call.
76 // Not sure what are the other possible cases where this property is empty
77 if (empty($entry->getUrl()) && !empty($url)) {
79 $entry->setGivenUrl($url);
82 $this->stockEntry($entry, $content);
86 * Use a Symfony validator to ensure the language is well formatted.
89 * @param string $value Language to validate and save
91 public function updateLanguage(Entry
$entry, $value)
93 // some lang are defined as fr-FR, es-ES.
94 // replacing - by _ might increase language support
95 $value = str_replace('-', '_', $value);
97 $errors = $this->validator
->validate(
99 (new LocaleConstraint())
102 if (0 === \
count($errors)) {
103 $entry->setLanguage($value);
108 $this->logger
->warning('Language validation failed. ' . (string) $errors);
112 * Use a Symfony validator to ensure the preview picture is a real url.
114 * @param Entry $entry
115 * @param string $value URL to validate and save
117 public function updatePreviewPicture(Entry
$entry, $value)
119 $errors = $this->validator
->validate(
121 (new UrlConstraint())
124 if (0 === \
count($errors)) {
125 $entry->setPreviewPicture($value);
130 $this->logger
->warning('PreviewPicture validation failed. ' . (string) $errors);
136 * @param Entry $entry
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 * @param Entry $entry
166 public function setEntryDomainName(Entry
$entry)
168 $domainName = parse_url($entry->getUrl(), PHP_URL_HOST
);
169 if (false !== $domainName) {
170 $entry->setDomainName($domainName);
175 * Helper to set a default title using:
176 * - url basename, if applicable
179 * @param Entry $entry
181 public function setDefaultEntryTitle(Entry
$entry)
183 $url = parse_url($entry->getUrl());
184 $path = pathinfo($url['path'], PATHINFO_BASENAME
);
187 $path = $url['host'];
190 $entry->setTitle($path);
194 * Try to sanitize the title of the fetched content from wrong character encodings and invalid UTF-8 character.
196 * @param string $title
197 * @param string $contentType
201 private function sanitizeContentTitle($title, $contentType)
203 if ('application/pdf' === $contentType) {
204 $title = $this->convertPdfEncodingToUTF8($title);
207 return $this->sanitizeUTF8Text($title);
211 * If the title from the fetched content comes from a PDF, then its very possible that the character encoding is not
212 * UTF-8. This methods tries to identify the character encoding and translate the title to UTF-8.
216 * @return string (maybe contains invalid UTF-8 character)
218 private function convertPdfEncodingToUTF8($title)
220 // first try UTF-8 because its easier to detect its present/absence
221 foreach (['UTF-8', 'UTF-16BE', 'WINDOWS-1252'] as $encoding) {
222 if (mb_check_encoding($title, $encoding)) {
223 return mb_convert_encoding($title, 'UTF-8', $encoding);
231 * Remove invalid UTF-8 characters from the given string.
233 * @param string $rawText
237 private function sanitizeUTF8Text($rawText)
239 if (mb_check_encoding($rawText, 'UTF-8')) {
243 return iconv('UTF-8', 'UTF-8//IGNORE', $rawText);
247 * Stock entry with fetched or imported content.
248 * Will fall back to OpenGraph data if available.
250 * @param Entry $entry Entry to stock
251 * @param array $content Array with at least title, url & html
253 private function stockEntry(Entry
$entry, array $content)
255 $this->updateOriginUrl($entry, $content['url']);
257 $this->setEntryDomainName($entry);
259 if (!empty($content['title'])) {
260 $entry->setTitle($content['title']);
263 if (empty($content['html'])) {
264 $content['html'] = $this->fetchingErrorMessage
;
266 if (!empty($content['description'])) {
267 $content['html'] .= '<p><i>But we found a short description: </i></p>';
268 $content['html'] .= $content['description'];
272 $entry->setContent($content['html']);
273 $entry->setReadingTime(Utils
::getReadingTime($content['html']));
275 if (!empty($content['status'])) {
276 $entry->setHttpStatus($content['status']);
279 if (!empty($content['authors']) && \
is_array($content['authors'])) {
280 $entry->setPublishedBy($content['authors']);
283 if (!empty($content['headers'])) {
284 $entry->setHeaders($content['headers']);
287 if (!empty($content['date'])) {
288 $this->updatePublishedAt($entry, $content['date']);
291 if (!empty($content['language'])) {
292 $this->updateLanguage($entry, $content['language']);
295 $previewPictureUrl = '';
296 if (!empty($content['image'])) {
297 $previewPictureUrl = $content['image'];
300 // if content is an image, define it as a preview too
301 if (!empty($content['headers']['content-type']) && \
in_array($this->mimeGuesser
->guess($content['headers']['content-type']), ['jpeg', 'jpg', 'gif', 'png'], true)) {
302 $previewPictureUrl = $content['url'];
303 } elseif (empty($previewPictureUrl)) {
304 $this->logger
->debug('Extracting images from content to provide a default preview picture');
305 $imagesUrls = DownloadImages
::extractImagesUrlsFromHtml($content['html']);
306 $this->logger
->debug(\
count($imagesUrls) . ' pictures found');
308 if (!empty($imagesUrls)) {
309 $previewPictureUrl = $imagesUrls[0];
313 if (!empty($content['headers']['content-type'])) {
314 $entry->setMimetype($content['headers']['content-type']);
317 if (!empty($previewPictureUrl)) {
318 $this->updatePreviewPicture($entry, $previewPictureUrl);
322 $this->tagger
->tag($entry);
323 } catch (\Exception
$e) {
324 $this->logger
->error('Error while trying to automatically tag an entry.', [
325 'entry_url' => $content['url'],
326 'error_msg' => $e->getMessage(),
332 * Update the origin_url field when a redirection occurs
333 * This field is set if it is empty and new url does not match ignore list.
335 * @param Entry $entry
338 private function updateOriginUrl(Entry
$entry, $url)
340 if (empty($url) || $entry->getUrl() === $url) {
344 $parsed_entry_url = parse_url($entry->getUrl());
345 $parsed_content_url = parse_url($url);
348 * The following part computes the list of part changes between two
351 * As array_diff_assoc only computes changes to go from the left array
352 * to the right one, we make two differents arrays to have both
353 * directions. We merge these two arrays and sort keys before passing
354 * the result to the switch.
356 * The resulting array gives us all changing parts between the two
357 * urls: scheme, host, path, query and/or fragment.
359 $diff_ec = array_diff_assoc($parsed_entry_url, $parsed_content_url);
360 $diff_ce = array_diff_assoc($parsed_content_url, $parsed_entry_url);
362 $diff = array_merge($diff_ec, $diff_ce);
363 $diff_keys = array_keys($diff);
366 if ($this->ignoreUrl($entry->getUrl())) {
367 $entry->setUrl($url);
373 * This switch case lets us apply different behaviors according to
374 * changing parts of urls.
376 * As $diff_keys is an array, we provide arrays as cases. ['path'] means
377 * 'only the path is different between the two urls' whereas
378 * ['fragment', 'query'] means 'only fragment and query string parts are
379 * different between the two urls'.
381 * Note that values in $diff_keys are sorted.
383 switch ($diff_keys) {
385 if (($parsed_entry_url['path'] . '/' === $parsed_content_url['path']) // diff is trailing slash, we only replace the url of the entry
386 || ($url === urldecode($entry->getUrl()))) { // we update entry url if new url is a decoded version of it, see EntryRepository#findByUrlAndUserId
387 $entry->setUrl($url);
391 $entry->setUrl($url);
397 if (empty($entry->getOriginUrl())) {
398 $entry->setOriginUrl($entry->getUrl());
400 $entry->setUrl($url);
406 * Check entry url against an ignore list to replace with content url.
408 * XXX: move the ignore list in the database to let users handle it
410 * @param string $url url to test
412 * @return bool true if url matches ignore list otherwise false
414 private function ignoreUrl($url)
416 $ignored_hosts = ['feedproxy.google.com', 'feeds.reuters.com'];
417 $ignored_patterns = ['https?://www\.lemonde\.fr/tiny.*'];
419 $parsed_url = parse_url($url);
421 $filtered = array_filter($ignored_hosts, function ($var) use ($parsed_url) {
422 return $var === $parsed_url['host'];
425 if ([] !== $filtered) {
429 $filtered = array_filter($ignored_patterns, function ($var) use ($url) {
430 return preg_match("`$var`i", $url);
433 if ([] !== $filtered) {
441 * Validate that the given content has at least a title, an html and a url.
443 * @param array $content
445 * @return bool true if valid otherwise false
447 private function validateContent(array $content)
449 return !empty($content['title']) && !empty($content['html']) && !empty($content['url']);