]> git.immae.eu Git - github/wallabag/wallabag.git/blob - src/Wallabag/CoreBundle/Helper/ContentProxy.php
Merge pull request #2608 from wallabag/change-dev-version
[github/wallabag/wallabag.git] / src / Wallabag / CoreBundle / Helper / ContentProxy.php
1 <?php
2
3 namespace Wallabag\CoreBundle\Helper;
4
5 use Graby\Graby;
6 use Psr\Log\LoggerInterface as Logger;
7 use Wallabag\CoreBundle\Entity\Entry;
8 use Wallabag\CoreBundle\Entity\Tag;
9 use Wallabag\CoreBundle\Tools\Utils;
10 use Wallabag\CoreBundle\Repository\TagRepository;
11 use Symfony\Component\HttpFoundation\File\MimeType\MimeTypeExtensionGuesser;
12
13 /**
14 * This kind of proxy class take care of getting the content from an url
15 * and update the entry with what it found.
16 */
17 class ContentProxy
18 {
19 protected $graby;
20 protected $tagger;
21 protected $logger;
22 protected $tagRepository;
23 protected $mimeGuesser;
24
25 public function __construct(Graby $graby, RuleBasedTagger $tagger, TagRepository $tagRepository, Logger $logger)
26 {
27 $this->graby = $graby;
28 $this->tagger = $tagger;
29 $this->logger = $logger;
30 $this->tagRepository = $tagRepository;
31 $this->mimeGuesser = new MimeTypeExtensionGuesser();
32 }
33
34 /**
35 * Fetch content using graby and hydrate given entry with results information.
36 * In case we couldn't find content, we'll try to use Open Graph data.
37 *
38 * We can also force the content, in case of an import from the v1 for example, so the function won't
39 * fetch the content from the website but rather use information given with the $content parameter.
40 *
41 * @param Entry $entry Entry to update
42 * @param string $url Url to grab content for
43 * @param array $content An array with AT LEAST keys title, html, url, language & content_type to skip the fetchContent from the url
44 *
45 * @return Entry
46 */
47 public function updateEntry(Entry $entry, $url, array $content = [])
48 {
49 // do we have to fetch the content or the provided one is ok?
50 if (empty($content) || false === $this->validateContent($content)) {
51 $content = $this->graby->fetchContent($url);
52 }
53
54 $title = $content['title'];
55 if (!$title && isset($content['open_graph']['og_title'])) {
56 $title = $content['open_graph']['og_title'];
57 }
58
59 $html = $content['html'];
60 if (false === $html) {
61 $html = '<p>Unable to retrieve readable content.</p>';
62
63 if (isset($content['open_graph']['og_description'])) {
64 $html .= '<p><i>But we found a short description: </i></p>';
65 $html .= $content['open_graph']['og_description'];
66 }
67 }
68
69 $entry->setUrl($content['url'] ?: $url);
70 $entry->setTitle($title);
71 $entry->setContent($html);
72 $entry->setLanguage($content['language']);
73 $entry->setMimetype($content['content_type']);
74 $entry->setReadingTime(Utils::getReadingTime($html));
75
76 $domainName = parse_url($entry->getUrl(), PHP_URL_HOST);
77 if (false !== $domainName) {
78 $entry->setDomainName($domainName);
79 }
80
81 if (isset($content['open_graph']['og_image'])) {
82 $entry->setPreviewPicture($content['open_graph']['og_image']);
83 }
84
85 // if content is an image define as a preview too
86 if (in_array($this->mimeGuesser->guess($content['content_type']), ['jpeg', 'jpg', 'gif', 'png'], true)) {
87 $entry->setPreviewPicture($content['url']);
88 }
89
90 try {
91 $this->tagger->tag($entry);
92 } catch (\Exception $e) {
93 $this->logger->error('Error while trying to automatically tag an entry.', [
94 'entry_url' => $url,
95 'error_msg' => $e->getMessage(),
96 ]);
97 }
98
99 return $entry;
100 }
101
102 /**
103 * Assign some tags to an entry.
104 *
105 * @param Entry $entry
106 * @param array|string $tags An array of tag or a string coma separated of tag
107 * @param array $entitiesReady Entities from the EntityManager which are persisted but not yet flushed
108 * It is mostly to fix duplicate tag on import @see http://stackoverflow.com/a/7879164/569101
109 */
110 public function assignTagsToEntry(Entry $entry, $tags, array $entitiesReady = [])
111 {
112 if (!is_array($tags)) {
113 $tags = explode(',', $tags);
114 }
115
116 // keeps only Tag entity from the "not yet flushed entities"
117 $tagsNotYetFlushed = [];
118 foreach ($entitiesReady as $entity) {
119 if ($entity instanceof Tag) {
120 $tagsNotYetFlushed[$entity->getLabel()] = $entity;
121 }
122 }
123
124 foreach ($tags as $label) {
125 $label = trim($label);
126
127 // avoid empty tag
128 if (0 === strlen($label)) {
129 continue;
130 }
131
132 if (isset($tagsNotYetFlushed[$label])) {
133 $tagEntity = $tagsNotYetFlushed[$label];
134 } else {
135 $tagEntity = $this->tagRepository->findOneByLabel($label);
136
137 if (is_null($tagEntity)) {
138 $tagEntity = new Tag();
139 $tagEntity->setLabel($label);
140 }
141 }
142
143 // only add the tag on the entry if the relation doesn't exist
144 if (false === $entry->getTags()->contains($tagEntity)) {
145 $entry->addTag($tagEntity);
146 }
147 }
148 }
149
150 /**
151 * Validate that the given content as enough value to be used
152 * instead of fetch the content from the url.
153 *
154 * @param array $content
155 *
156 * @return bool true if valid otherwise false
157 */
158 private function validateContent(array $content)
159 {
160 return isset($content['title']) && isset($content['html']) && isset($content['url']) && isset($content['language']) && isset($content['content_type']);
161 }
162 }