Merge pull request #185 from shikorism/feature/metadata_tag

MetadataResolverからタグ情報を保存できるようにする

app/Http/Controllers/Api/CardController.php

@@ -4,6 +4,7 @@ namespace App\Http\Controllers\Api;
 
 use App\Metadata;
 use App\MetadataResolver\MetadataResolver;
+use App\Tag;
 use App\Utilities\Formatter;
 use Illuminate\Http\Request;
 
@@ -41,6 +42,13 @@ class CardController
                 'image' => $resolved->image,
                 'expires_at' => $resolved->expires_at
             ]);
+
+            $tagIds = [];
+            foreach ($resolved->tags as $tagName) {
+                $tag = Tag::firstOrCreate(['name' => $tagName]);
+                $tagIds[] = $tag->id;
+            }
+            $metadata->tags()->sync($tagIds);
         }
 
         $response = response($metadata);

app/Listeners/LinkCollector.php

@@ -5,6 +5,7 @@ namespace App\Listeners;
 use App\Events\LinkDiscovered;
 use App\Metadata;
 use App\MetadataResolver\MetadataResolver;
+use App\Tag;
 use App\Utilities\Formatter;
 use GuzzleHttp\Exception\TransferException;
 use Illuminate\Contracts\Queue\ShouldQueue;
@@ -47,12 +48,19 @@ class LinkCollector
         if ($metadata == null || ($metadata->expires_at !== null && $metadata->expires_at < now())) {
             try {
                 $resolved = $this->metadataResolver->resolve($url);
-                Metadata::updateOrCreate(['url' => $url], [
+                $metadata = Metadata::updateOrCreate(['url' => $url], [
                     'title' => $resolved->title,
                     'description' => $resolved->description,
                     'image' => $resolved->image,
                     'expires_at' => $resolved->expires_at
                 ]);
+
+                $tagIds = [];
+                foreach ($resolved->tags as $tagName) {
+                    $tag = Tag::firstOrCreate(['name' => $tagName]);
+                    $tagIds[] = $tag->id;
+                }
+                $metadata->tags()->sync($tagIds);
             } catch (TransferException $e) {
                 // 何らかの通信エラーによってメタデータの取得に失敗した時、とりあえずエラーログにURLを残す
                 Log::error(self::class . ': メタデータの取得に失敗 URL=' . $url);

app/Metadata.php

@@ -14,4 +14,9 @@ class Metadata extends Model
     protected $visible = ['url', 'title', 'description', 'image', 'expires_at'];
 
     protected $dates = ['created_at', 'updated_at', 'expires_at'];
+
+    public function tags()
+    {
+        return $this->belongsToMany(Tag::class)->withTimestamps();
+    }
 }

app/MetadataResolver/KomifloResolver.php

@@ -34,6 +34,20 @@ class KomifloResolver implements Resolver
                 ($json['content']['parents'][0]['data']['title'] ?? '?');
             $metadata->image = 'https://t.komiflo.com/564_mobile_large_3x/' . $json['content']['named_imgs']['cover']['filename'];
 
+            // 作者情報
+            if (!empty($json['content']['attributes']['artists']['children'])) {
+                foreach ($json['content']['attributes']['artists']['children'] as $artist) {
+                    $metadata->tags[] = preg_replace('/\s/', '_', $artist['data']['name']);
+                }
+            }
+
+            // タグ
+            if (!empty($json['content']['attributes']['tags']['children'])) {
+                foreach ($json['content']['attributes']['tags']['children'] as $tag) {
+                    $metadata->tags[] = preg_replace('/\s/', '_', $tag['data']['name']);
+                }
+            }
+
             return $metadata;
         } else {
             throw new \RuntimeException("{$res->getStatusCode()}: $url");

app/MetadataResolver/Metadata.php

@@ -6,9 +6,21 @@ use Carbon\Carbon;
 
 class Metadata
 {
+    /** @var string タイトル */
     public $title = '';
+
+    /** @var string 概要 */
     public $description = '';
+
+    /** @var string サムネイルのURL */
     public $image = '';
-    /** @var Carbon|null */
+
+    /** @var Carbon|null メタデータの有効期限 */
     public $expires_at = null;
+
+    /**
+     * @var string[] タグ
+     * チェックインタグと同様に保存されるため、スペースや改行文字を含めてはいけません。
+     */
+    public $tags = [];
 }

app/MetadataResolver/PixivResolver.php

@@ -49,6 +49,43 @@ class PixivResolver implements Resolver
         return str_replace('i.pximg.net', 'i.pixiv.cat', $pixivUrl);
     }
 
+    /**
+     * HTMLからタグとして利用可能な情報を抽出する
+     * @param string $html ページ HTML
+     * @return string[] タグ
+     */
+    public function extractTags(string $html): array
+    {
+        $dom = new \DOMDocument();
+        @$dom->loadHTML(mb_convert_encoding($html, 'HTML-ENTITIES', 'UTF-8'));
+        $xpath = new \DOMXPath($dom);
+
+        $nodes = $xpath->query("//meta[@name='keywords']");
+        if ($nodes->length === 0) {
+            return [];
+        }
+
+        $keywords = $nodes->item(0)->getAttribute('content');
+        $tags = [];
+
+        foreach (mb_split(',', $keywords) as $keyword) {
+            $keyword = trim($keyword);
+
+            if (empty($keyword)) {
+                continue;
+            }
+
+            // 一部の固定キーワードは無視
+            if (array_search($keyword, ['R-18', 'イラスト', 'pixiv', 'ピクシブ'], true)) {
+                continue;
+            }
+
+            $tags[] = preg_replace('/\s/', '_', $keyword);
+        }
+
+        return $tags;
+    }
+
     public function resolve(string $url): Metadata
     {
         parse_str(parse_url($url, PHP_URL_QUERY), $params);
@@ -78,6 +115,8 @@ class PixivResolver implements Resolver
 
             $metadata->image = $this->proxize($illustUrl);
 
+            $metadata->tags = $this->extractTags($res->getBody());
+
             return $metadata;
         } else {
             throw new \RuntimeException("{$res->getStatusCode()}: $url");

database/migrations/2019_04_29_111019_create_metadata_tag_table.php

@@ -0,0 +1,33 @@
+<?php
+
+use Illuminate\Database\Migrations\Migration;
+use Illuminate\Database\Schema\Blueprint;
+use Illuminate\Support\Facades\Schema;
+
+class CreateMetadataTagTable extends Migration
+{
+    /**
+     * Run the migrations.
+     *
+     * @return void
+     */
+    public function up()
+    {
+        Schema::create('metadata_tag', function (Blueprint $table) {
+            $table->increments('id');
+            $table->text('metadata_url')->index();
+            $table->integer('tag_id')->index();
+            $table->timestamps();
+        });
+    }
+
+    /**
+     * Reverse the migrations.
+     *
+     * @return void
+     */
+    public function down()
+    {
+        Schema::dropIfExists('metadata_tag');
+    }
+}