feat: queueing bulk follow/unfollow and block/unblock (#10544)

* wrap follow/unfollow and block/unblock as job queue

* create import job to follow in each iteration

* make relationship jobs concurrent

* replace to job queue if called repeatedly

* use addBulk to import

* omit stream when importing

* fix job caller

* use ThinUser instead of User to reduce redis memory consumption

* createImportFollowingToDbJobの呼び出し方を変える, 型補強

* Force ThinUser

* オブジェクト操作のみのメソッド名はgenerate...Data

* Force ThinUser in generateRelationshipJobData

* silent bulk unfollow at admin api endpoint

---------

Co-authored-by: tamaina <tamaina@hotmail.co.jp>

packages/backend/src/queue/processors/DeleteDriveFilesProcessorService.ts

@@ -7,7 +7,7 @@ import type Logger from '@/logger.js';
 import { DriveService } from '@/core/DriveService.js';
 import { QueueLoggerService } from '../QueueLoggerService.js';
 import type Bull from 'bull';
-import type { DbUserJobData } from '../types.js';
+import type { DbJobDataWithUser } from '../types.js';
 import { bindThis } from '@/decorators.js';
 
 @Injectable()
@@ -31,7 +31,7 @@ export class DeleteDriveFilesProcessorService {
 	}
 
 	@bindThis
-	public async process(job: Bull.Job<DbUserJobData>, done: () => void): Promise<void> {
+	public async process(job: Bull.Job<DbJobDataWithUser>, done: () => void): Promise<void> {
 		this.logger.info(`Deleting drive files of ${job.data.user.id} ...`);
 
 		const user = await this.usersRepository.findOneBy({ id: job.data.user.id });