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

Merge pull request #259 from eai04191/feature/resolver-nijie

Browse Source 
NijieResolver 修正など
This commit is contained in:
shibafu 2019-08-31 03:12:34 +09:00 committed by GitHub
commit ac2077af49
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
6 changed files with 409 additions and 130 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
app/MetadataResolver/NijieResolver.php
View File

@ -3,6 +3,7 @@
namespace App\MetadataResolver; namespace App\MetadataResolver;
use GuzzleHttp\Client; use GuzzleHttp\Client;
use Symfony\Component\DomCrawler\Crawler;
class NijieResolver implements Resolver class NijieResolver implements Resolver
{ {
@ -30,23 +31,29 @@ class NijieResolver implements Resolver
$url = preg_replace('~view_popup\.php~', 'view.php', $url); $url = preg_replace('~view_popup\.php~', 'view.php', $url);
} }
$client = $this->client; $res = $this->client->get($url);
$res = $client->get($url);
if ($res->getStatusCode() === 200) { if ($res->getStatusCode() === 200) {
$metadata = $this->ogpResolver->parse($res->getBody()); $html = (string) $res->getBody();
$metadata = $this->ogpResolver->parse($html);
$crawler = new Crawler($html);
$dom = new \DOMDocument(); // DomCrawler内でjson内の日本語がHTMLエンティティに変換されるのでhtml_entity_decode
@$dom->loadHTML(mb_convert_encoding($res->getBody(), 'HTML-ENTITIES', 'UTF-8')); $json = html_entity_decode($crawler->filter('script[type="application/ld+json"]')->first()->text());
$xpath = new \DOMXPath($dom);
$dataNode = $xpath->query('//script[substring(@type, string-length(@type) - 3, 4) = "json"]'); // 改行がそのまま入っていることがあるのでデコード前にエスケープが必要
foreach ($dataNode as $node) { $data = json_decode(preg_replace('/\r?\n/', '\n', $json), true);
// 改行がそのまま入っていることがあるのでデコード前にエスケープが必要
$imageData = json_decode(preg_replace('/\r?\n/', '\n', $node->nodeValue), true); $metadata->title = $data['name'];
if (isset($imageData['thumbnailUrl']) && !ends_with($imageData['thumbnailUrl'], '.gif') && !ends_with($imageData['thumbnailUrl'], '.mp4')) { $metadata->description = '投稿者: ' . $data['author']['name'] . PHP_EOL . $data['description'];
$metadata->image = preg_replace('~nijie\\.info/.*/nijie_picture/~', 'nijie.info/nijie_picture/', $imageData['thumbnailUrl']); if (
break; isset($data['thumbnailUrl']) &&
} !ends_with($data['thumbnailUrl'], '.gif') &&
!ends_with($data['thumbnailUrl'], '.mp4')
) {
// サムネイルからメイン画像に
$metadata->image = str_replace('__rs_l160x160/', '', $data['thumbnailUrl']);
} }
$metadata->tags = $crawler->filter('#view-tag span.tag_name')->extract('_text');
return $metadata; return $metadata;
} else { } else {

103
tests/Unit/MetadataResolver/NijieResolverTest.php
View File

@ -25,10 +25,10 @@ class NijieResolverTest extends TestCase
$this->createResolver(NijieResolver::class, $responseText); $this->createResolver(NijieResolver::class, $responseText);
$metadata = $this->resolver->resolve('https://nijie.info/view.php?id=66384'); $metadata = $this->resolver->resolve('https://nijie.info/view.php?id=66384');
$this->assertEquals('チンポップくんの日常ep.1「チンポップくんと釣り」 | ニジエ運営', $metadata->title); $this->assertSame('チンポップくんの日常ep.1「チンポップくんと釣り」', $metadata->title);
$this->assertEquals("メールマガジン漫画のバックナンバー第一話です！\r\n最新話はメールマガジンより配信中です。", $metadata->description); $this->assertSame('投稿者: ニジエ運営' . PHP_EOL . 'メールマガジン漫画のバックナンバー第一話です！' . PHP_EOL . '最新話はメールマガジンより配信中です。', $metadata->description);
$this->assertRegExp('/pic\d+\.nijie\.info/', $metadata->image); $this->assertSame('https://pic.nijie.net/04/nijie_picture/38_20131130155623.png', $metadata->image);
$this->assertNotRegExp('~/diff/main/~', $metadata->image); $this->assertSame(['ニジエたん', '釣り', 'チンポップ君の日常', '公式漫画'], $metadata->tags);
if ($this->shouldUseMock()) { if ($this->shouldUseMock()) {
$this->assertSame('https://nijie.info/view.php?id=66384', (string) $this->handler->getLastRequest()->getUri()); $this->assertSame('https://nijie.info/view.php?id=66384', (string) $this->handler->getLastRequest()->getUri());
} }
@ -41,10 +41,10 @@ class NijieResolverTest extends TestCase
$this->createResolver(NijieResolver::class, $responseText); $this->createResolver(NijieResolver::class, $responseText);
$metadata = $this->resolver->resolve('https://nijie.info/view.php?id=202707'); $metadata = $this->resolver->resolve('https://nijie.info/view.php?id=202707');
$this->assertEquals('ニジエ壁紙 | ニジエ運営', $metadata->title); $this->assertSame('ニジエ壁紙', $metadata->title);
$this->assertEquals("ニジエのPCとiphone用(4.7inch推奨)の壁紙です。\r\n保存してご自由にお使いくださいませ。", $metadata->description); $this->assertSame('投稿者: ニジエ運営' . PHP_EOL . 'ニジエのPCとiphone用(4.7inch推奨)の壁紙です。' . PHP_EOL . '保存してご自由にお使いくださいませ。', $metadata->description);
$this->assertRegExp('/pic\d+\.nijie\.info/', $metadata->image); $this->assertSame('https://pic.nijie.net/03/nijie_picture/38_20170209185801_0.png', $metadata->image);
$this->assertNotRegExp('~/diff/main/~', $metadata->image); $this->assertSame(['ニジエたん', '壁紙'], $metadata->tags);
if ($this->shouldUseMock()) { if ($this->shouldUseMock()) {
$this->assertSame('https://nijie.info/view.php?id=202707', (string) $this->handler->getLastRequest()->getUri()); $this->assertSame('https://nijie.info/view.php?id=202707', (string) $this->handler->getLastRequest()->getUri());
} }
@ -57,9 +57,10 @@ class NijieResolverTest extends TestCase
$this->createResolver(NijieResolver::class, $responseText); $this->createResolver(NijieResolver::class, $responseText);
$metadata = $this->resolver->resolve('https://nijie.info/view.php?id=9537'); $metadata = $this->resolver->resolve('https://nijie.info/view.php?id=9537');
$this->assertEquals('ニジエがgifに対応したんだってね　奥さん | 黒末アプコ', $metadata->title); $this->assertSame('ニジエがgifに対応したんだってね　奥さん', $metadata->title);
$this->assertEquals('アニメgifとか専門外なのでよくわかりませんでした', $metadata->description); $this->assertSame('投稿者: 黒末アプコ' . PHP_EOL . 'アニメgifとか専門外なのでよくわかりませんでした', $metadata->description);
$this->assertRegExp('~/nijie\.info/pic/logo~', $metadata->image); $this->assertStringStartsWith('https://nijie.info/pic/logo/nijie_logo_og.png', $metadata->image);
$this->assertSame(['おっぱい', '陥没乳首', '眼鏡', 'GIFアニメ', 'ぶるんぶるん', 'アニメgif'], $metadata->tags);
if ($this->shouldUseMock()) { if ($this->shouldUseMock()) {
$this->assertSame('https://nijie.info/view.php?id=9537', (string) $this->handler->getLastRequest()->getUri()); $this->assertSame('https://nijie.info/view.php?id=9537', (string) $this->handler->getLastRequest()->getUri());
} }
@ -72,74 +73,60 @@ class NijieResolverTest extends TestCase
$this->createResolver(NijieResolver::class, $responseText); $this->createResolver(NijieResolver::class, $responseText);
$metadata = $this->resolver->resolve('https://nijie.info/view.php?id=256283'); $metadata = $this->resolver->resolve('https://nijie.info/view.php?id=256283');
$this->assertEquals('てすと | ニジエ運営', $metadata->title); $this->assertSame('てすと', $metadata->title);
$this->assertEquals("H264動画てすと　あとで消します\r\n\r\n今の所、H264コーデックのみ、出力時に音声なしにしないと投稿できません\r\n動画は勝手にループします", $metadata->description); $this->assertSame('投稿者: ニジエ運営' . PHP_EOL . 'H264動画てすと　あとで消します' . PHP_EOL . PHP_EOL . '今の所、H264コーデックのみ、出力時に音声なしにしないと投稿できません' . PHP_EOL . '動画は勝手にループします', $metadata->description);
$this->assertRegExp('~/nijie\.info/pic/logo~', $metadata->image); $this->assertStringStartsWith('https://nijie.info/pic/logo/nijie_logo_og.png', $metadata->image);
$this->assertSame([], $metadata->tags);
if ($this->shouldUseMock()) { if ($this->shouldUseMock()) {
$this->assertSame('https://nijie.info/view.php?id=256283', (string) $this->handler->getLastRequest()->getUri()); $this->assertSame('https://nijie.info/view.php?id=256283', (string) $this->handler->getLastRequest()->getUri());
} }
} }
public function testStandardPictureSp() public function testViewPopup()
{
$responseText = file_get_contents(__DIR__ . '/../../fixture/Nijie/testStandardPictureResponse.html');
$this->createResolver(NijieResolver::class, $responseText);
$metadata = $this->resolver->resolve('https://nijie.info/view_popup.php?id=66384');
$this->assertSame('チンポップくんの日常ep.1「チンポップくんと釣り」', $metadata->title);
$this->assertSame('投稿者: ニジエ運営' . PHP_EOL . 'メールマガジン漫画のバックナンバー第一話です！' . PHP_EOL . '最新話はメールマガジンより配信中です。', $metadata->description);
$this->assertSame('https://pic.nijie.net/04/nijie_picture/38_20131130155623.png', $metadata->image);
$this->assertSame(['ニジエたん', '釣り', 'チンポップ君の日常', '公式漫画'], $metadata->tags);
if ($this->shouldUseMock()) {
$this->assertSame('https://nijie.info/view.php?id=66384', (string) $this->handler->getLastRequest()->getUri());
}
}
public function testSp()
{ {
$responseText = file_get_contents(__DIR__ . '/../../fixture/Nijie/testStandardPictureResponse.html'); $responseText = file_get_contents(__DIR__ . '/../../fixture/Nijie/testStandardPictureResponse.html');
$this->createResolver(NijieResolver::class, $responseText); $this->createResolver(NijieResolver::class, $responseText);
$metadata = $this->resolver->resolve('https://sp.nijie.info/view.php?id=66384'); $metadata = $this->resolver->resolve('https://sp.nijie.info/view.php?id=66384');
$this->assertEquals('チンポップくんの日常ep.1「チンポップくんと釣り」 | ニジエ運営', $metadata->title); $this->assertSame('チンポップくんの日常ep.1「チンポップくんと釣り」', $metadata->title);
$this->assertEquals("メールマガジン漫画のバックナンバー第一話です！\r\n最新話はメールマガジンより配信中です。", $metadata->description); $this->assertSame('投稿者: ニジエ運営' . PHP_EOL . 'メールマガジン漫画のバックナンバー第一話です！' . PHP_EOL . '最新話はメールマガジンより配信中です。', $metadata->description);
$this->assertRegExp('/pic\d+\.nijie\.info/', $metadata->image); $this->assertSame('https://pic.nijie.net/04/nijie_picture/38_20131130155623.png', $metadata->image);
$this->assertNotRegExp('~/diff/main/~', $metadata->image); $this->assertSame(['ニジエたん', '釣り', 'チンポップ君の日常', '公式漫画'], $metadata->tags);
if ($this->shouldUseMock()) { if ($this->shouldUseMock()) {
$this->assertSame('https://nijie.info/view.php?id=66384', (string) $this->handler->getLastRequest()->getUri()); $this->assertSame('https://nijie.info/view.php?id=66384', (string) $this->handler->getLastRequest()->getUri());
} }
} }
public function testMultiplePictureSp() public function testSpViewPopup()
{ {
$responseText = file_get_contents(__DIR__ . '/../../fixture/Nijie/testMultiplePictureResponse.html'); $responseText = file_get_contents(__DIR__ . '/../../fixture/Nijie/testStandardPictureResponse.html');
$this->createResolver(NijieResolver::class, $responseText); $this->createResolver(NijieResolver::class, $responseText);
$metadata = $this->resolver->resolve('https://sp.nijie.info/view.php?id=202707'); $metadata = $this->resolver->resolve('https://sp.nijie.info/view_popup.php?id=66384');
$this->assertEquals('ニジエ壁紙 | ニジエ運営', $metadata->title); $this->assertSame('チンポップくんの日常ep.1「チンポップくんと釣り」', $metadata->title);
$this->assertEquals("ニジエのPCとiphone用(4.7inch推奨)の壁紙です。\r\n保存してご自由にお使いくださいませ。", $metadata->description); $this->assertSame('投稿者: ニジエ運営' . PHP_EOL . 'メールマガジン漫画のバックナンバー第一話です！' . PHP_EOL . '最新話はメールマガジンより配信中です。', $metadata->description);
$this->assertRegExp('/pic\d+\.nijie\.info/', $metadata->image); $this->assertSame('https://pic.nijie.net/04/nijie_picture/38_20131130155623.png', $metadata->image);
$this->assertNotRegExp('~/diff/main/~', $metadata->image); $this->assertSame(['ニジエたん', '釣り', 'チンポップ君の日常', '公式漫画'], $metadata->tags);
if ($this->shouldUseMock()) { if ($this->shouldUseMock()) {
$this->assertSame('https://nijie.info/view.php?id=202707', (string) $this->handler->getLastRequest()->getUri()); $this->assertSame('https://nijie.info/view.php?id=66384', (string) $this->handler->getLastRequest()->getUri());
}
}
public function testAnimationGifSp()
{
$responseText = file_get_contents(__DIR__ . '/../../fixture/Nijie/testAnimationGifResponse.html');
$this->createResolver(NijieResolver::class, $responseText);
$metadata = $this->resolver->resolve('https://nijie.info/view.php?id=9537');
$this->assertEquals('ニジエがgifに対応したんだってね　奥さん | 黒末アプコ', $metadata->title);
$this->assertEquals('アニメgifとか専門外なのでよくわかりませんでした', $metadata->description);
$this->assertRegExp('~/nijie\.info/pic/logo~', $metadata->image);
if ($this->shouldUseMock()) {
$this->assertSame('https://nijie.info/view.php?id=9537', (string) $this->handler->getLastRequest()->getUri());
}
}
public function testMp4MovieSp()
{
$responseText = file_get_contents(__DIR__ . '/../../fixture/Nijie/testMp4MovieResponse.html');
$this->createResolver(NijieResolver::class, $responseText);
$metadata = $this->resolver->resolve('https://sp.nijie.info/view.php?id=256283');
$this->assertEquals('てすと | ニジエ運営', $metadata->title);
$this->assertEquals("H264動画てすと　あとで消します\r\n\r\n今の所、H264コーデックのみ、出力時に音声なしにしないと投稿できません\r\n動画は勝手にループします", $metadata->description);
$this->assertRegExp('~/nijie\.info/pic/logo~', $metadata->image);
if ($this->shouldUseMock()) {
$this->assertSame('https://nijie.info/view.php?id=256283', (string) $this->handler->getLastRequest()->getUri());
} }
} }
} }

89
tests/fixture/Nijie/testAnimationGifResponse.html vendored
View File

File diff suppressed because one or more lines are too long

124
tests/fixture/Nijie/testMp4MovieResponse.html vendored
View File

File diff suppressed because one or more lines are too long

91
tests/fixture/Nijie/testMultiplePictureResponse.html vendored
View File

File diff suppressed because one or more lines are too long

97
tests/fixture/Nijie/testStandardPictureResponse.html vendored
View File

File diff suppressed because one or more lines are too long