]>
Commit | Line | Data |
---|---|---|
558d9aab JB |
1 | <?php |
2 | ||
3 | namespace Wallabag\CoreBundle\Helper; | |
4 | ||
5 | use Graby\Graby; | |
45fd7e09 | 6 | use Psr\Log\LoggerInterface; |
8d7b4f0e | 7 | use Symfony\Component\HttpFoundation\File\MimeType\MimeTypeExtensionGuesser; |
42f3bb2c | 8 | use Symfony\Component\Validator\Constraints\Locale as LocaleConstraint; |
0d349ea6 JB |
9 | use Symfony\Component\Validator\Constraints\Url as UrlConstraint; |
10 | use Symfony\Component\Validator\Validator\ValidatorInterface; | |
f808b016 JB |
11 | use Wallabag\CoreBundle\Entity\Entry; |
12 | use Wallabag\CoreBundle\Tools\Utils; | |
558d9aab JB |
13 | |
14 | /** | |
15 | * This kind of proxy class take care of getting the content from an url | |
f1e29e69 | 16 | * and update the entry with what it found. |
558d9aab JB |
17 | */ |
18 | class ContentProxy | |
19 | { | |
20 | protected $graby; | |
c3510620 | 21 | protected $tagger; |
be54dfe4 | 22 | protected $validator; |
1c9cd2a7 | 23 | protected $logger; |
8d7b4f0e | 24 | protected $mimeGuesser; |
29dca432 | 25 | protected $fetchingErrorMessage; |
6bc6fb1f | 26 | protected $eventDispatcher; |
8a219854 | 27 | protected $storeArticleHeaders; |
558d9aab | 28 | |
709e21a3 | 29 | public function __construct(Graby $graby, RuleBasedTagger $tagger, ValidatorInterface $validator, LoggerInterface $logger, $fetchingErrorMessage, $storeArticleHeaders = false) |
558d9aab | 30 | { |
347fa6be | 31 | $this->graby = $graby; |
c3510620 | 32 | $this->tagger = $tagger; |
0d349ea6 | 33 | $this->validator = $validator; |
1c9cd2a7 | 34 | $this->logger = $logger; |
8d7b4f0e | 35 | $this->mimeGuesser = new MimeTypeExtensionGuesser(); |
29dca432 | 36 | $this->fetchingErrorMessage = $fetchingErrorMessage; |
8a219854 | 37 | $this->storeArticleHeaders = $storeArticleHeaders; |
558d9aab JB |
38 | } |
39 | ||
40 | /** | |
6acadf8e | 41 | * Update entry using either fetched or provided content. |
4d0ec0e7 | 42 | * |
6acadf8e JB |
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 | |
558d9aab | 47 | */ |
6acadf8e | 48 | public function updateEntry(Entry $entry, $url, array $content = [], $disableContentUpdate = false) |
558d9aab | 49 | { |
2dbb5b23 | 50 | $this->graby->toggleImgNoReferrer(true); |
6acadf8e JB |
51 | if (!empty($content['html'])) { |
52 | $content['html'] = $this->graby->cleanupHtml($content['html'], $url); | |
d5c2cc54 | 53 | } |
e668a812 | 54 | |
6acadf8e | 55 | if ((empty($content) || false === $this->validateContent($content)) && false === $disableContentUpdate) { |
ec970721 | 56 | $fetchedContent = $this->graby->fetchContent($url); |
f80f16df | 57 | $fetchedContent['title'] = $this->sanitizeContentTitle($fetchedContent['title'], $fetchedContent['content_type']); |
106bdbcd JB |
58 | |
59 | // when content is imported, we have information in $content | |
60 | // in case fetching content goes bad, we'll keep the imported information instead of overriding them | |
6acadf8e | 61 | if (empty($content) || $fetchedContent['html'] !== $this->fetchingErrorMessage) { |
29dca432 JC |
62 | $content = $fetchedContent; |
63 | } | |
4d0ec0e7 | 64 | } |
558d9aab | 65 | |
6acadf8e JB |
66 | // be sure to keep the url in case of error |
67 | // so we'll be able to refetch it in the future | |
68 | $content['url'] = !empty($content['url']) ? $content['url'] : $url; | |
69 | ||
4a81360e KD |
70 | // In one case (at least in tests), url is empty here |
71 | // so we set it using $url provided in the updateEntry call. | |
781864b9 KD |
72 | // Not sure what are the other possible cases where this property is empty |
73 | if (empty($entry->getUrl()) && !empty($url)) { | |
4a81360e KD |
74 | $entry->setUrl($url); |
75 | } | |
76 | ||
d0e9b3d6 JC |
77 | $this->stockEntry($entry, $content); |
78 | } | |
79 | ||
c18a2476 JB |
80 | /** |
81 | * Use a Symfony validator to ensure the language is well formatted. | |
82 | * | |
83 | * @param Entry $entry | |
84 | * @param string $value Language to validate and save | |
85 | */ | |
86 | public function updateLanguage(Entry $entry, $value) | |
87 | { | |
88 | // some lang are defined as fr-FR, es-ES. | |
89 | // replacing - by _ might increase language support | |
90 | $value = str_replace('-', '_', $value); | |
91 | ||
92 | $errors = $this->validator->validate( | |
93 | $value, | |
94 | (new LocaleConstraint()) | |
95 | ); | |
96 | ||
2a1ceb67 | 97 | if (0 === \count($errors)) { |
c18a2476 JB |
98 | $entry->setLanguage($value); |
99 | ||
100 | return; | |
101 | } | |
102 | ||
103 | $this->logger->warning('Language validation failed. ' . (string) $errors); | |
104 | } | |
105 | ||
106 | /** | |
107 | * Use a Symfony validator to ensure the preview picture is a real url. | |
108 | * | |
109 | * @param Entry $entry | |
110 | * @param string $value URL to validate and save | |
111 | */ | |
112 | public function updatePreviewPicture(Entry $entry, $value) | |
113 | { | |
114 | $errors = $this->validator->validate( | |
115 | $value, | |
116 | (new UrlConstraint()) | |
117 | ); | |
118 | ||
2a1ceb67 | 119 | if (0 === \count($errors)) { |
c18a2476 JB |
120 | $entry->setPreviewPicture($value); |
121 | ||
122 | return; | |
123 | } | |
124 | ||
125 | $this->logger->warning('PreviewPicture validation failed. ' . (string) $errors); | |
126 | } | |
127 | ||
128 | /** | |
129 | * Update date. | |
130 | * | |
131 | * @param Entry $entry | |
132 | * @param string $value Date to validate and save | |
133 | */ | |
134 | public function updatePublishedAt(Entry $entry, $value) | |
135 | { | |
ff9f89fd | 136 | $date = $value; |
c18a2476 JB |
137 | |
138 | // is it a timestamp? | |
3ef055ce | 139 | if (false !== filter_var($date, FILTER_VALIDATE_INT)) { |
ff9f89fd | 140 | $date = '@' . $date; |
c18a2476 JB |
141 | } |
142 | ||
143 | try { | |
ff9f89fd JB |
144 | // is it already a DateTime? |
145 | // (it's inside the try/catch in case of fail to be parse time string) | |
146 | if (!$date instanceof \DateTime) { | |
147 | $date = new \DateTime($date); | |
148 | } | |
149 | ||
150 | $entry->setPublishedAt($date); | |
c18a2476 JB |
151 | } catch (\Exception $e) { |
152 | $this->logger->warning('Error while defining date', ['e' => $e, 'url' => $entry->getUrl(), 'date' => $value]); | |
153 | } | |
154 | } | |
155 | ||
af29e1bf KD |
156 | /** |
157 | * Helper to extract and save host from entry url. | |
158 | * | |
159 | * @param Entry $entry | |
160 | */ | |
161 | public function setEntryDomainName(Entry $entry) | |
162 | { | |
163 | $domainName = parse_url($entry->getUrl(), PHP_URL_HOST); | |
164 | if (false !== $domainName) { | |
165 | $entry->setDomainName($domainName); | |
166 | } | |
167 | } | |
168 | ||
169 | /** | |
170 | * Helper to set a default title using: | |
171 | * - url basename, if applicable | |
172 | * - hostname. | |
173 | * | |
174 | * @param Entry $entry | |
175 | */ | |
176 | public function setDefaultEntryTitle(Entry $entry) | |
177 | { | |
178 | $url = parse_url($entry->getUrl()); | |
179 | $path = pathinfo($url['path'], PATHINFO_BASENAME); | |
180 | ||
181 | if (empty($path)) { | |
182 | $path = $url['host']; | |
183 | } | |
184 | ||
185 | $entry->setTitle($path); | |
186 | } | |
187 | ||
83f1c327 T |
188 | /** |
189 | * Try to sanitize the title of the fetched content from wrong character encodings and invalid UTF-8 character. | |
190 | * | |
191 | * @param $title | |
192 | * @param $contentType | |
193 | * | |
194 | * @return string | |
195 | */ | |
196 | private function sanitizeContentTitle($title, $contentType) | |
197 | { | |
198 | if ('application/pdf' === $contentType) { | |
199 | $title = $this->convertPdfEncodingToUTF8($title); | |
200 | } | |
201 | ||
202 | return $this->sanitizeUTF8Text($title); | |
203 | } | |
204 | ||
205 | /** | |
206 | * If the title from the fetched content comes from a PDF, then its very possible that the character encoding is not | |
207 | * UTF-8. This methods tries to identify the character encoding and translate the title to UTF-8. | |
208 | * | |
209 | * @param $title | |
210 | * | |
211 | * @return string (maybe contains invalid UTF-8 character) | |
212 | */ | |
213 | private function convertPdfEncodingToUTF8($title) | |
214 | { | |
215 | // first try UTF-8 because its easier to detect its present/absence | |
216 | foreach (['UTF-8', 'UTF-16BE', 'WINDOWS-1252'] as $encoding) { | |
217 | if (mb_check_encoding($title, $encoding)) { | |
218 | return mb_convert_encoding($title, 'UTF-8', $encoding); | |
219 | } | |
220 | } | |
221 | ||
222 | return $title; | |
223 | } | |
224 | ||
225 | /** | |
226 | * Remove invalid UTF-8 characters from the given string. | |
227 | * | |
228 | * @param string $rawText | |
229 | * | |
230 | * @return string | |
231 | */ | |
232 | private function sanitizeUTF8Text($rawText) | |
233 | { | |
234 | if (mb_check_encoding($rawText, 'UTF-8')) { | |
235 | return $rawText; | |
236 | } | |
237 | ||
238 | return iconv('UTF-8', 'UTF-8//IGNORE', $rawText); | |
239 | } | |
240 | ||
d0e9b3d6 JC |
241 | /** |
242 | * Stock entry with fetched or imported content. | |
243 | * Will fall back to OpenGraph data if available. | |
244 | * | |
d5c2cc54 | 245 | * @param Entry $entry Entry to stock |
ec970721 | 246 | * @param array $content Array with at least title, url & html |
d0e9b3d6 JC |
247 | */ |
248 | private function stockEntry(Entry $entry, array $content) | |
249 | { | |
e07fadea | 250 | $this->updateOriginUrl($entry, $content['url']); |
a05b6115 | 251 | |
af29e1bf | 252 | $this->setEntryDomainName($entry); |
a05b6115 JB |
253 | |
254 | if (!empty($content['title'])) { | |
255 | $entry->setTitle($content['title']); | |
256 | } elseif (!empty($content['open_graph']['og_title'])) { | |
257 | $entry->setTitle($content['open_graph']['og_title']); | |
558d9aab JB |
258 | } |
259 | ||
8ca858ee JB |
260 | if (empty($content['html'])) { |
261 | $content['html'] = $this->fetchingErrorMessage; | |
558d9aab | 262 | |
e668a812 | 263 | if (!empty($content['open_graph']['og_description'])) { |
8ca858ee JB |
264 | $content['html'] .= '<p><i>But we found a short description: </i></p>'; |
265 | $content['html'] .= $content['open_graph']['og_description']; | |
558d9aab JB |
266 | } |
267 | } | |
268 | ||
8ca858ee JB |
269 | $entry->setContent($content['html']); |
270 | $entry->setReadingTime(Utils::getReadingTime($content['html'])); | |
f0378b4d | 271 | |
a05b6115 JB |
272 | if (!empty($content['status'])) { |
273 | $entry->setHttpStatus($content['status']); | |
5e9009ce NL |
274 | } |
275 | ||
2a1ceb67 | 276 | if (!empty($content['authors']) && \is_array($content['authors'])) { |
7b0b3622 NL |
277 | $entry->setPublishedBy($content['authors']); |
278 | } | |
279 | ||
8a219854 | 280 | if (!empty($content['all_headers']) && $this->storeArticleHeaders) { |
dda6a6ad NL |
281 | $entry->setHeaders($content['all_headers']); |
282 | } | |
283 | ||
a05b6115 JB |
284 | if (!empty($content['date'])) { |
285 | $this->updatePublishedAt($entry, $content['date']); | |
286 | } | |
0d349ea6 | 287 | |
a05b6115 JB |
288 | if (!empty($content['language'])) { |
289 | $this->updateLanguage($entry, $content['language']); | |
290 | } | |
291 | ||
292 | if (!empty($content['open_graph']['og_image'])) { | |
293 | $this->updatePreviewPicture($entry, $content['open_graph']['og_image']); | |
294 | } | |
0d349ea6 | 295 | |
be54dfe4 | 296 | // if content is an image, define it as a preview too |
2a1ceb67 | 297 | if (!empty($content['content_type']) && \in_array($this->mimeGuesser->guess($content['content_type']), ['jpeg', 'jpg', 'gif', 'png'], true)) { |
d0ec2ddd | 298 | $this->updatePreviewPicture($entry, $content['url']); |
0d349ea6 JB |
299 | } |
300 | ||
a05b6115 JB |
301 | if (!empty($content['content_type'])) { |
302 | $entry->setMimetype($content['content_type']); | |
4d0ec0e7 | 303 | } |
558d9aab | 304 | |
1c9cd2a7 KG |
305 | try { |
306 | $this->tagger->tag($entry); | |
307 | } catch (\Exception $e) { | |
4094ea47 | 308 | $this->logger->error('Error while trying to automatically tag an entry.', [ |
d0e9b3d6 | 309 | 'entry_url' => $content['url'], |
1c9cd2a7 | 310 | 'error_msg' => $e->getMessage(), |
4094ea47 | 311 | ]); |
1c9cd2a7 | 312 | } |
558d9aab | 313 | } |
c2656f96 | 314 | |
e07fadea KD |
315 | /** |
316 | * Update the origin_url field when a redirection occurs | |
317 | * This field is set if it is empty and new url does not match ignore list. | |
318 | * | |
319 | * @param Entry $entry | |
320 | * @param string $url | |
321 | */ | |
322 | private function updateOriginUrl(Entry $entry, $url) | |
323 | { | |
5ba5e22a KD |
324 | if (empty($url) || $entry->getUrl() === $url) { |
325 | return false; | |
326 | } | |
327 | ||
328 | $parsed_entry_url = parse_url($entry->getUrl()); | |
329 | $parsed_content_url = parse_url($url); | |
e07fadea | 330 | |
44e63667 KD |
331 | /** |
332 | * The following part computes the list of part changes between two | |
333 | * parse_url arrays. | |
334 | * | |
335 | * As array_diff_assoc only computes changes to go from the left array | |
336 | * to the right one, we make two differents arrays to have both | |
337 | * directions. We merge these two arrays and sort keys before passing | |
338 | * the result to the switch. | |
339 | * | |
340 | * The resulting array gives us all changing parts between the two | |
341 | * urls: scheme, host, path, query and/or fragment. | |
342 | */ | |
5ba5e22a KD |
343 | $diff_ec = array_diff_assoc($parsed_entry_url, $parsed_content_url); |
344 | $diff_ce = array_diff_assoc($parsed_content_url, $parsed_entry_url); | |
e07fadea | 345 | |
5ba5e22a KD |
346 | $diff = array_merge($diff_ec, $diff_ce); |
347 | $diff_keys = array_keys($diff); | |
348 | sort($diff_keys); | |
349 | ||
350 | if ($this->ignoreUrl($entry->getUrl())) { | |
351 | $entry->setUrl($url); | |
1b220426 | 352 | |
5ba5e22a KD |
353 | return false; |
354 | } | |
e07fadea | 355 | |
44e63667 KD |
356 | /** |
357 | * This switch case lets us apply different behaviors according to | |
358 | * changing parts of urls. | |
359 | * | |
360 | * As $diff_keys is an array, we provide arrays as cases. ['path'] means | |
361 | * 'only the path is different between the two urls' whereas | |
362 | * ['fragment', 'query'] means 'only fragment and query string parts are | |
363 | * different between the two urls'. | |
1b220426 | 364 | * |
44e63667 KD |
365 | * Note that values in $diff_keys are sorted. |
366 | */ | |
5ba5e22a KD |
367 | switch ($diff_keys) { |
368 | case ['path']: | |
369 | if (($parsed_entry_url['path'] . '/' === $parsed_content_url['path']) // diff is trailing slash, we only replace the url of the entry | |
370 | || ($url === urldecode($entry->getUrl()))) { // we update entry url if new url is a decoded version of it, see EntryRepository#findByUrlAndUserId | |
371 | $entry->setUrl($url); | |
372 | } | |
373 | break; | |
374 | case ['scheme']: | |
b49c87ac | 375 | $entry->setUrl($url); |
5ba5e22a KD |
376 | break; |
377 | case ['fragment']: | |
5ba5e22a KD |
378 | // noop |
379 | break; | |
380 | default: | |
381 | if (empty($entry->getOriginUrl())) { | |
382 | $entry->setOriginUrl($entry->getUrl()); | |
b49c87ac | 383 | } |
5ba5e22a KD |
384 | $entry->setUrl($url); |
385 | break; | |
e07fadea KD |
386 | } |
387 | } | |
388 | ||
b49c87ac KD |
389 | /** |
390 | * Check entry url against an ignore list to replace with content url. | |
391 | * | |
392 | * XXX: move the ignore list in the database to let users handle it | |
393 | * | |
394 | * @param string $url url to test | |
395 | * | |
396 | * @return bool true if url matches ignore list otherwise false | |
397 | */ | |
398 | private function ignoreUrl($url) | |
399 | { | |
400 | $ignored_hosts = ['feedproxy.google.com', 'feeds.reuters.com']; | |
401 | $ignored_patterns = ['https?://www\.lemonde\.fr/tiny.*']; | |
402 | ||
403 | $parsed_url = parse_url($url); | |
404 | ||
405 | $filtered = array_filter($ignored_hosts, function ($var) use ($parsed_url) { | |
406 | return $var === $parsed_url['host']; | |
407 | }); | |
408 | ||
409 | if ([] !== $filtered) { | |
410 | return true; | |
411 | } | |
412 | ||
413 | $filtered = array_filter($ignored_patterns, function ($var) use ($url) { | |
414 | return preg_match("`$var`i", $url); | |
415 | }); | |
416 | ||
417 | if ([] !== $filtered) { | |
418 | return true; | |
419 | } | |
420 | ||
421 | return false; | |
422 | } | |
423 | ||
4d0ec0e7 | 424 | /** |
d0e9b3d6 | 425 | * Validate that the given content has at least a title, an html and a url. |
4d0ec0e7 JB |
426 | * |
427 | * @param array $content | |
6acadf8e JB |
428 | * |
429 | * @return bool true if valid otherwise false | |
4d0ec0e7 JB |
430 | */ |
431 | private function validateContent(array $content) | |
432 | { | |
6acadf8e | 433 | return !empty($content['title']) && !empty($content['html']) && !empty($content['url']); |
4d0ec0e7 | 434 | } |
558d9aab | 435 | } |