service-layer-architecture（服务层架构）

Service Layer Architecture

Clean, performant API layers with proper separation of concerns and parallel data fetching.

When to Use

• Building REST APIs with complex data aggregation
• GraphQL resolvers needing data from multiple sources
• Any API where responses combine data from multiple queries
• Systems needing testable, maintainable code

Three-Layer Architecture

┌─────────────────────────────────────────────────────┐
│  Controllers   │  HTTP handling, validation        │
├─────────────────────────────────────────────────────┤
│  Services      │  Business logic, data enrichment  │
├─────────────────────────────────────────────────────┤
│  Queries       │  Database access, raw data fetch  │
└─────────────────────────────────────────────────────┘

Layer 1: Controllers (HTTP Only)

// controllers/Entity.ts
import { getEntity, getEntities } from "../services/Entity";

const router = new Router();

router.get("/entity/:entityId", async (ctx) => {
  const { entityId } = ctx.params;

  if (!entityId) {
    ctx.status = 400;
    ctx.body = { error: "Invalid entity ID" };
    return;
  }

  const entity = await getEntity(entityId);
  
  if (!entity) {
    ctx.status = 404;
    ctx.body = { error: "Entity not found" };
    return;
  }
  
  ctx.status = 200;
  ctx.body = entity;
});

Layer 2: Services (Business Logic)

// services/Entity.ts
import { queries } from "@common";

export const getEntityData = async (entity: RawEntity): Promise<EnrichedEntity> => {
  // Parallel fetch all related data
  const [metadata, score, activity, location] = await Promise.all([
    queries.getMetadata(),
    queries.getLatestScore(entity.id),
    queries.getActivity(entity.id),
    queries.getLocation(entity.slotId),
  ]);

  // Transform and combine
  return {
    ...entity,
    bonded: entity.bonded / Math.pow(10, metadata.decimals),
    total: score?.total ?? 0,
    location: location?.city,
    activity: {
      activeCount: activity?.active?.length ?? 0,
      inactiveCount: activity?.inactive?.length ?? 0,
    },
  };
};

export const getEntity = async (entityId: string): Promise<EnrichedEntity | null> => {
  const entity = await queries.getEntityById(entityId);
  if (!entity) return null;
  return getEntityData(entity);
};

export const getEntities = async (): Promise<EnrichedEntity[]> => {
  const all = await queries.allEntities();
  const enriched = await Promise.all(all.map(getEntityData));
  return enriched.sort((a, b) => b.total - a.total);
};

Layer 3: Queries (Database Access)

// queries/Entities.ts
import { EntityModel } from "../models";

export const allEntities = async () => {
  return EntityModel.find({}).lean();  // Always use .lean()
};

export const getEntityById = async (id: string) => {
  return EntityModel.findOne({ id }).lean();
};

export const validEntities = async () => {
  return EntityModel.find({ valid: true }).lean();
};

Parallel Data Fetching

// BAD: Sequential (slow)
const metadata = await queries.getMetadata();
const score = await queries.getScore(id);
const location = await queries.getLocation(id);
// Time: sum of all queries

// GOOD: Parallel (fast)
const [metadata, score, location] = await Promise.all([
  queries.getMetadata(),
  queries.getScore(id),
  queries.getLocation(id),
]);
// Time: max of all queries

Layer Responsibilities

Related Skills

• Related: postgres-job-queue — Background job processing
• Related: realtime/websocket-hub-patterns — Real-time updates from services

NEVER Do

• NEVER put database queries in controllers — Violates separation
• NEVER put HTTP concerns in services — Services must be reusable
• NEVER fetch related data sequentially — Use Promise.all
• NEVER skip .lean() on read queries — 5-10x faster
• NEVER expose raw database errors — Transform to user-friendly messages