enhance(backend): summalyの結果をキャッシュするように (#17367)

* enhance(backend): summalyの結果をキャッシュするように

* fix

* fix: unify memorykvcache

* fix

* refactor

* fix

* Update Changelog

CHANGELOG.md

@@ -40,6 +40,7 @@
 - Enhance: Sentry バックエンドの自動計装を `sentryForBackend.disabledIntegrations` で個別に無効化できるように
 - Enhance: Node.js 22.23.0以降、24.17.0以降、26.4.0以降をサポートするように
 - Enhance: Docker Image の Node.js を 26.4.0 に、Debian を trixie (v13) に更新
+- Enhance: URLプレビューの結果を内部でキャッシュするように
 - Fix: `/stats` API のレスポンス型が正しくない問題を修正
 - Fix: ハッシュタグに関連するデータを更新する際のエラーハンドリングを修正
 - Fix: Sentry 使用環境下にて、Misskey が発行した SQL クエリが span に含まれない問題を修正

packages/backend/src/misc/cache.ts

@@ -204,14 +204,13 @@ export class RedisSingleCache<T> {
 	}
 }
 
-// TODO: メモリ節約のためあまり参照されないキーを定期的に削除できるようにする？
-
 export class MemoryKVCache<T> {
 	private readonly cache = new Map<string, { date: number; value: T; }>();
 	private readonly gcIntervalHandle = setInterval(() => this.gc(), 1000 * 60 * 3); // 3m
 
 	constructor(
 		private readonly lifetime: number,
+		private readonly limit: number = Infinity,
 	) {}
 
 	@bindThis
@@ -220,6 +219,25 @@ export class MemoryKVCache<T> {
 	 * @deprecated これを直接呼び出すべきではない。InternalEventなどで変更を全てのプロセス/マシンに通知するべき
 	 */
 	public set(key: string, value: T): void {
+		if (this.limit <= 0) {
+			throw new Error('Limit must be greater than 0');
+		}
+
+		if (this.limit !== Infinity) {
+			this.gc();
+
+			// 挿入順を更新して LRU を保つため、同一キーは一度削除する
+			this.cache.delete(key);
+
+			while (this.cache.size >= this.limit) {
+				const oldestKey = this.cache.keys().next().value;
+				if (oldestKey == null) {
+					throw new Error('Cache is empty but size exceeds the limit');
+				}
+				this.cache.delete(oldestKey);
+			}
+		}
+
 		this.cache.set(key, {
 			date: Date.now(),
 			value,
@@ -234,6 +252,11 @@ export class MemoryKVCache<T> {
 			this.cache.delete(key);
 			return undefined;
 		}
+		if (this.limit !== Infinity) {
+			// access 順を更新して LRU を保つ
+			this.cache.delete(key);
+			this.cache.set(key, cached);
+		}
 		return cached.value;
 	}
 

packages/backend/src/server/web/UrlPreviewService.ts

@@ -3,13 +3,14 @@
  * SPDX-License-Identifier: AGPL-3.0-only
  */
 
-import { Inject, Injectable } from '@nestjs/common';
+import { Inject, Injectable, OnApplicationShutdown } from '@nestjs/common';
 import type { SummalyResult } from '@misskey-dev/summaly';
 import { DI } from '@/di-symbols.js';
 import type { Config } from '@/config.js';
 import { HttpRequestService } from '@/core/HttpRequestService.js';
 import type Logger from '@/logger.js';
 import { query } from '@/misc/prelude/url.js';
+import { MemoryKVCache } from '@/misc/cache.js';
 import { LoggerService } from '@/core/LoggerService.js';
 import { UtilityService } from '@/core/UtilityService.js';
 import { bindThis } from '@/decorators.js';
@@ -18,8 +19,9 @@ import { MiMeta } from '@/models/Meta.js';
 import type { FastifyRequest, FastifyReply } from 'fastify';
 
 @Injectable()
-export class UrlPreviewService {
+export class UrlPreviewService implements OnApplicationShutdown {
 	private logger: Logger;
+	private summaryCache: MemoryKVCache<SummalyResult>;
 	private readonly summalyDefaultUserAgent: string;
 
 	constructor(
@@ -34,6 +36,7 @@ export class UrlPreviewService {
 		private loggerService: LoggerService,
 	) {
 		this.logger = this.loggerService.getLogger('url-preview');
+		this.summaryCache = new MemoryKVCache<SummalyResult>(1000 * 60 * 60, 100); // 1h, 100件
 		this.summalyDefaultUserAgent = `SummalyBot/${_SUMMALY_VERSION_} (${this.config.url}; +https://github.com/misskey-dev/summaly/blob/master/README.md)`;
 	}
 
@@ -80,20 +83,32 @@ export class UrlPreviewService {
 			: `Getting preview of ${url}@${lang} ...`);
 
 		try {
-			const summary = this.meta.urlPreviewSummaryProxyUrl
-				? await this.fetchSummaryFromProxy(url, this.meta, lang)
-				: await this.fetchSummary(url, this.meta, lang);
+			const fetcher = async () => {
+				const result = await (
+					this.meta.urlPreviewSummaryProxyUrl
+					? this.fetchSummaryFromProxy(url, lang)
+					: this.fetchSummary(url, lang)
+				);
+
+				if (!result.url.startsWith('http://') && !result.url.startsWith('https://')) {
+					return undefined;
+				}
+
+				if (result.player.url && !result.player.url.startsWith('http://') && !result.player.url.startsWith('https://')) {
+					return undefined;
+				}
+
+				return result;
+			};
+
+			const summary = await this.summaryCache.fetchMaybe(`${url}@${lang ?? '_DEFAULT_'}`, fetcher);
+
+			if (summary == null) {
+				throw new Error('Invalid summary');
+			}
 
 			this.logger.succ(`Got preview of ${url}: ${summary.title}`);
 
-			if (!(summary.url.startsWith('http://') || summary.url.startsWith('https://'))) {
-				throw new Error('unsupported schema included');
-			}
-
-			if (summary.player.url && !(summary.player.url.startsWith('http://') || summary.player.url.startsWith('https://'))) {
-				throw new Error('unsupported schema included');
-			}
-
 			summary.icon = this.wrap(summary.icon);
 			summary.thumbnail = this.wrap(summary.thumbnail);
 
@@ -120,7 +135,8 @@ export class UrlPreviewService {
 		}
 	}
 
-	private async fetchSummary(url: string, meta: MiMeta, lang?: string): Promise<SummalyResult> {
+	@bindThis
+	private async fetchSummary(url: string, lang?: string): Promise<SummalyResult> {
 		const { summaly } = await import('@misskey-dev/summaly');
 
 		return summaly(url, {
@@ -130,25 +146,36 @@ export class UrlPreviewService {
 				http: this.httpRequestService.httpAgent,
 				https: this.httpRequestService.httpsAgent,
 			},
-			userAgent: meta.urlPreviewUserAgent ?? this.summalyDefaultUserAgent,
-			operationTimeout: meta.urlPreviewTimeout,
-			contentLengthLimit: meta.urlPreviewMaximumContentLength,
-			contentLengthRequired: meta.urlPreviewRequireContentLength,
+			userAgent: this.meta.urlPreviewUserAgent ?? this.summalyDefaultUserAgent,
+			operationTimeout: this.meta.urlPreviewTimeout,
+			contentLengthLimit: this.meta.urlPreviewMaximumContentLength,
+			contentLengthRequired: this.meta.urlPreviewRequireContentLength,
 		});
 	}
 
-	private fetchSummaryFromProxy(url: string, meta: MiMeta, lang?: string): Promise<SummalyResult> {
-		const proxy = meta.urlPreviewSummaryProxyUrl!;
+	@bindThis
+	private fetchSummaryFromProxy(url: string, lang?: string): Promise<SummalyResult> {
+		const proxy = this.meta.urlPreviewSummaryProxyUrl!;
 		const queryStr = query({
 			url: url,
 			lang: lang ?? 'ja-JP',
 			followRedirects: this.meta.urlPreviewAllowRedirect,
-			userAgent: meta.urlPreviewUserAgent ?? this.summalyDefaultUserAgent,
-			operationTimeout: meta.urlPreviewTimeout,
-			contentLengthLimit: meta.urlPreviewMaximumContentLength,
-			contentLengthRequired: meta.urlPreviewRequireContentLength,
+			userAgent: this.meta.urlPreviewUserAgent ?? this.summalyDefaultUserAgent,
+			operationTimeout: this.meta.urlPreviewTimeout,
+			contentLengthLimit: this.meta.urlPreviewMaximumContentLength,
+			contentLengthRequired: this.meta.urlPreviewRequireContentLength,
 		});
 
 		return this.httpRequestService.getJson<SummalyResult>(`${proxy}?${queryStr}`, 'application/json, */*', undefined, true);
 	}
+
+	@bindThis
+	public dispose(): void {
+		this.summaryCache.dispose();
+	}
+
+	@bindThis
+	public onApplicationShutdown(): void {
+		this.dispose();
+	}
 }

packages/backend/test/unit/misc/cache.ts

@@ -38,6 +38,29 @@ describe('misc:MemoryKVCache', () => {
 		cache.dispose();
 	});
 
+	test('keeps current behavior when limit is omitted', () => {
+		const cache = new MemoryKVCache<number>(1000 * 60);
+		cache.set('a', 1);
+		cache.set('b', 2);
+		cache.set('c', 3);
+		expect(cache.get('a')).toBe(1);
+		expect(cache.get('b')).toBe(2);
+		expect(cache.get('c')).toBe(3);
+		cache.dispose();
+	});
+
+	test('evicts least recently used entry when limit is reached', () => {
+		const cache = new MemoryKVCache<number>(1000 * 60, 2);
+		cache.set('a', 1);
+		cache.set('b', 2);
+		expect(cache.get('a')).toBe(1);
+		cache.set('c', 3);
+		expect(cache.get('a')).toBe(1);
+		expect(cache.get('b')).toBeUndefined();
+		expect(cache.get('c')).toBe(3);
+		cache.dispose();
+	});
+
 	describe('gc()', () => {
 		test('removes expired entries', () => {
 			const cache = new MemoryKVCache<string>(1000);