pepepper/tissue
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Merge pull request #260 from eai04191/feature/resolver-iwara-improve

Browse Source 
IwaraResolver 拡張など
This commit is contained in:
shibafu
2019-08-31 03:12:17 +09:00
committed by GitHub
parent 4940b7a9ca 5e02a8ab7a
commit 27970e3ac5
6 changed files with 3469 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

63
app/MetadataResolver/IwaraResolver.php
View File

@@ -3,6 +3,7 @@
namespace App\MetadataResolver;
use GuzzleHttp\Client;
use Symfony\Component\DomCrawler\Crawler;
class IwaraResolver implements Resolver
{
@@ -19,46 +20,40 @@ class IwaraResolver implements Resolver
public function resolve(string $url): Metadata
{
$res = $this->client->get($url);
if ($res->getStatusCode() === 200) {
$dom = new \DOMDocument();
@$dom->loadHTML(mb_convert_encoding($res->getBody(), 'HTML-ENTITIES', 'UTF-8'));
$xpath = new \DOMXPath($dom);
$metadata = new Metadata();
$html = (string) $res->getBody();
$crawler = new Crawler($html);
// find title
foreach ($xpath->query('//title') as $node) {
$content = $node->textContent;
if (!empty($content)) {
$metadata->title = $content;
break;
$infoElements = $crawler->filter('#video-player + div, .field-name-field-video-url + div, .field-name-field-images + div');
$title = $infoElements->filter('h1.title')->text();
$author = $infoElements->filter('.username')->text();
$description = $infoElements->filter('.field-type-text-with-summary')->text('');
$tags = $infoElements->filter('a[href^="/videos"], a[href^="/images"]')->extract('_text');
// 役に立たないタグを削除する
$tags = array_values(array_diff($tags, ['Uncategorized', 'Other']));
array_push($tags, $author);
$metadata->title = $title;
$metadata->description = '投稿者: ' . $author . PHP_EOL . $description;
$metadata->tags = $tags;
// iwara video
if ($crawler->filter('#video-player')->count()) {
$metadata->image = 'https:' . $crawler->filter('#video-player')->attr('poster');
}
// youtube
if ($crawler->filter('iframe[src^="//www.youtube.com"]')->count()) {
if (preg_match('~youtube\.com/embed/(\S+)\?~', $crawler->filter('iframe[src^="//www.youtube.com"]')->attr('src'), $matches) === 1) {
$youtubeId = $matches[1];
$metadata->image = 'https://img.youtube.com/vi/' . $youtubeId . '/maxresdefault.jpg';
}
}
// find thumbnail
foreach ($xpath->query('//*[@id="video-player"]') as $node) {
$poster = $node->getAttribute('poster');
if (!empty($poster)) {
if (strpos($poster, '//') === 0) {
$poster = 'https:' . $poster;
}
$metadata->image = $poster;
break;
}
}
if (empty($metadata->image)) {
// YouTube embedded?
foreach ($xpath->query('//div[@class="embedded-video"]//iframe') as $node) {
$src = $node->getAttribute('src');
if (preg_match('~youtube\.com/embed/(\S+)\?~', $src, $matches) !== -1) {
$youtubeId = $matches[1];
$iwaraThumbUrl = 'https://i.iwara.tv/sites/default/files/styles/thumbnail/public/video_embed_field_thumbnails/youtube/' . $youtubeId . '.jpg';
$metadata->image = $iwaraThumbUrl;
break;
}
}
// images
if ($crawler->filter('.field-name-field-images')->count()) {
$metadata->image = 'https:' . $crawler->filter('.field-name-field-images a')->first()->attr('href');
}
return $metadata;

2
app/MetadataResolver/MetadataResolver.php
View File

@@ -14,7 +14,7 @@ class MetadataResolver implements Resolver
'~komiflo\.com(/#!)?/comics/(\\d+)~' => KomifloResolver::class,
'~www\.melonbooks\.co\.jp/detail/detail\.php~' => MelonbooksResolver::class,
'~ec\.toranoana\.(jp|shop)/(tora|joshi)(_[rd]+)?/(ec|digi)/item/~' => ToranoanaResolver::class,
'~iwara\.tv/videos/.*~' => IwaraResolver::class,
'~iwara\.tv/(videos|images)/.*~' => IwaraResolver::class,
'~www\.dlsite\.com/.*/(work|announce)/=/product_id/..\d+(\.html)?~' => DLsiteResolver::class,
'~dlsite\.jp/...tw/..\d+~' => DLsiteResolver::class,
'~www\.pixiv\.net/member_illust\.php\?illust_id=\d+~' => PixivResolver::class,