A type-safe tRPC API layer has been discussed many times in the community, but as versions iterate, many conclusions need updating. This article revisits the topic based on the latest version.

Getting Started

Building on this foundation, we can further optimize:

import { ApolloClient, InMemoryCache, createHttpLink } from '@apollo/client'
import { setContext } from '@apollo/client/link/context'

const httpLink = createHttpLink({ uri: '/graphql' })
const authLink = setContext((_, { headers }) => {
  const token = localStorage.getItem('token')
  return { headers: { ...headers, authorization: `Bearer ${token}` } }
})

const client = new ApolloClient({
  link: authLink.concat(httpLink),
  cache: new InMemoryCache({
    typePolicies: {
      Query: {
        fields: {
          users: { keyArgs: ['filter'], merge: (e, i) => ({ ...i, edges: [...(e?.edges||[]), ...i.edges] }) }
        }
      }
    }
  })
})

This pattern is very practical in large projects and can significantly reduce maintenance costs.

Source Code Analysis

Usage in real projects is a bit more involved:

import { ApolloClient, InMemoryCache, createHttpLink } from '@apollo/client'
import { setContext } from '@apollo/client/link/context'

const httpLink = createHttpLink({ uri: '/graphql' })
const authLink = setContext((_, { headers }) => {
  const token = localStorage.getItem('token')
  return { headers: { ...headers, authorization: `Bearer ${token}` } }
})

const client = new ApolloClient({
  link: authLink.concat(httpLink),
  cache: new InMemoryCache({
    typePolicies: {
      Query: {
        fields: {
          users: { keyArgs: ['filter'], merge: (e, i) => ({ ...i, edges: [...(e?.edges||[]), ...i.edges] }) }
        }
      }
    }
  })
})

Through this approach, both the testability and scalability of the code are improved.

Real-World Applications

Here is a complete example:

import { ApolloClient, InMemoryCache, createHttpLink } from '@apollo/client'
import { setContext } from '@apollo/client/link/context'

const httpLink = createHttpLink({ uri: '/graphql' })
const authLink = setContext((_, { headers }) => {
  const token = localStorage.getItem('token')
  return { headers: { ...headers, authorization: `Bearer ${token}` } }
})

const client = new ApolloClient({
  link: authLink.concat(httpLink),
  cache: new InMemoryCache({
    typePolicies: {
      Query: {
        fields: {
          users: { keyArgs: ['filter'], merge: (e, i) => ({ ...i, edges: [...(e?.edges||[]), ...i.edges] }) }
        }
      }
    }
  })
})

Pay attention to edge case handling, which is critical in production.

Optimization Tips

The key lies in understanding the core logic:

import { ApolloClient, InMemoryCache, createHttpLink } from '@apollo/client'
import { setContext } from '@apollo/client/link/context'

const httpLink = createHttpLink({ uri: '/graphql' })
const authLink = setContext((_, { headers }) => {
  const token = localStorage.getItem('token')
  return { headers: { ...headers, authorization: `Bearer ${token}` } }
})

const client = new ApolloClient({
  link: authLink.concat(httpLink),
  cache: new InMemoryCache({
    typePolicies: {
      Query: {
        fields: {
          users: { keyArgs: ['filter'], merge: (e, i) => ({ ...i, edges: [...(e?.edges||[]), ...i.edges] }) }
        }
      }
    }
  })
})

Performance optimization should be tailored to specific scenarios; not all cases require over-optimization.

Pitfall Guide

We can improve it in the following ways:

import { ApolloClient, InMemoryCache, createHttpLink } from '@apollo/client'
import { setContext } from '@apollo/client/link/context'

const httpLink = createHttpLink({ uri: '/graphql' })
const authLink = setContext((_, { headers }) => {
  const token = localStorage.getItem('token')
  return { headers: { ...headers, authorization: `Bearer ${token}` } }
})

const client = new ApolloClient({
  link: authLink.concat(httpLink),
  cache: new InMemoryCache({
    typePolicies: {
      Query: {
        fields: {
          users: { keyArgs: ['filter'], merge: (e, i) => ({ ...i, edges: [...(e?.edges||[]), ...i.edges] }) }
        }
      }
    }
  })
})

This approach has been running stably in production for over six months and has been practically validated.

Summary

• Code examples are for reference only and need to be adjusted according to your business scenario
• A type-safe tRPC API layer isn't a silver bullet; pick it based on your project's scale and tech stack
• Understanding underlying principles is more important than memorizing APIs
• Always verify compatibility before using in production