Building a GraphQL API with Nitric

What we'll be doing

GraphQL APIs rely on only one HTTP endpoint, which means that you want it to be reliable, scalable, and performant. By using serverless compute such as Lambda, the GraphQL endpoint can be auto-scaling, whilst maintaining performance and reliability.

We'll be using Nitric to create a GraphQL API, that can be deployed to a cloud of your choice, gaining the benefits of serverless compute.

Create the GraphQL Schema
Write Resolvers
Create handler for GraphQL requests
Run locally for testing
Deploy to a cloud of your choice

Video

Here's a video of this guide built with Node.js:

Serverless GraphQL on any Cloud

Prerequisites

uv - for Python dependency management
The Nitric CLI
(optional) Your choice of an AWS, GCP or Azure account

Getting started

We'll start by creating a new project for our API.

nitric new my-profile-api py-starter

Next, open the project in your editor of choice.

cd my-profile-api

Make sure all dependencies are resolved using uv:

uv sync

The scaffolded project should have the following structure:

+--.venv/
+--services/
|  +-- api.py
+--.env
+--.gitignore
+--nitric.yaml
+--.pythonversion
+--pythonproject.toml
+--uv.lock
+--README.md

Build the GraphQL Schema

GraphQL requests are typesafe, and so they require a schema to be defined to validate queries.

Let's first add the Ariadne library

uv add ariadne

We'll then import our dependencies, and write out the schema.

from ariadne import MutationType, QueryType, gql, make_executable_schema, graphql
from uuid import uuid4

type_defs = gql("""
  type Profile {
    pid: String!
    name: String!
    age: Int!
    home: String!
  }

  type Message {
    msg: String!
  }

  input ProfileInput {
    name: String!
    age: Int!
    home: String!
  }

  type Query {
    getProfile(pid: String!): Profile
  }

  type Mutation {
    createProfile(profile: ProfileInput!): Profile
    updateProfile(pid: String!, profile: ProfileInput!): Profile
  }
""")

Define a Key Value store

Lets define a key value store resource for the resolvers get/set data from.

from nitric.resources import api, kv
from nitric.application import Nitric

profiles = kv('profiles').allow('get','set')

Create Resolvers

We'll need to map our resolvers to mutations or queries using Ariadne's QueryType or MutationType.

query = QueryType()
mutation = MutationType()

We can then use the key value store within these services. Each resolver will receive a parent and info arguments, as well as any query or mutation's arguments as keyword arguments.

An example of this is converting the GraphQL query function into Python:

updateProfile(pid: String!, profile: ProfileInput!): Profile

@mutation.field("updateProfile")
async def update_profiles(obj, info, pid, profile):
  pass

Create a profile

@mutation.field("createProfile")
async def resolve_create_profile(obj, info, profile):
  pid = str(uuid4())

  p = { 'pid': pid, 'name': profile['name'], 'age': profile['age'], 'home': profile['home'] }
  await profiles.set(pid, p)

  return p

Update a profile

@mutation.field("updateProfile")
async def update_profiles(obj, info, pid, profile):
  profile = await profiles.get(pid)
  if profile is None:
    return { 'msg': f'Profile with id {pid} not found.'}

  p = { 'pid': pid, 'name': profile['name'], 'age': profile['age'], 'home': profile['home'] }
  await profiles.doc(pid).set(p)

  return p

Get a profile by its ID

@query.field("getProfile")
async def resolve_get_profile(obj, info, pid):
  profile = await profiles.get(pid)
  return { 'pid': pid, 'name': profile['name'], 'age': profile['age'], 'home': profile['home'] }

GraphQL Handler

We'll define an API to define our GraphQL handler. This API will only have one endpoint, which will handle all the requests.

First load the schema with our queries and mutations.

from nitric.resources import kv, api
from nitric.application import Nitric

graph_api = api("public")
schema = make_executable_schema(type_defs, [query, mutation])

Nitric.run()

Then add the API handler.

@graph_api.post("/")
async def profile_handler(ctx: HttpContext) -> None:
    query = ctx.req.json

    success, result = await graphql(
        schema,
        query
    )

    ctx.res.status = 200 if success else 400
    ctx.res.body = result

Nitric.run()

Run it!

Now that you have an API defined with a handler for the GraphQL requests, it's time to test it out locally.

nitric start

Once it starts, the service will be able to receive requests via the API port.

GraphQL Queries

We can use cURL, postman or any other HTTP Client to test our application, however it's better if the client has GraphQL support.

Get all Profiles using cURL

curl --location -X POST \
  'http://localhost:4001' \
  --header 'Content-Type: application/json' \
  --data-raw '{"query":"query { getProfiles { pid name age home }}","variables":{}}'

{
  "data": {
    "getProfiles": [
      {
        "pid": "3f70ca58-25ed-4e88-8a45-eea1fbbb45d8",
        "name": "Tony Stark",
        "age": 53,
        "home": "Manhattan, New York City"
      },
      {
        "pid": "9c53bd95-199c-4151-a2a6-0da3ae24c29d",
        "name": "Peter Parker",
        "age": 22,
        "home": "Queens, New York City"
      },
      {
        "pid": "9ff191b0-0fbe-4e49-b944-85e79b5caa21",
        "name": "Steve Rogers",
        "age": 105,
        "home": "New York City"
      }
    ]
  }
}

Get a single profile

curl --location -X POST \
  'http://localhost:4001' \
  --header 'Content-Type: application/json' \
  --data-raw '{"query":"query { getProfile(pid: \"3f70ca58-25ed-4e88-8a45-eea1fbbb45d8\") { pid name age home }}","variables":{}}'

{
  "data": {
    "getProfile": {
      "pid": "3f70ca58-25ed-4e88-8a45-eea1fbbb45d8",
      "name": "Tony Stark",
      "age": 53,
      "home": "Manhattan, New York City"
    }
  }
}

Create a profile

curl --location -X POST \
  'http://localhost:4001' \
  --header 'Content-Type: application/json' \
  --data-raw '{"query":"mutation { createProfile(profile: { name: \"Tony Stark\", age: 53, home: \"Manhattan, New York City\" }){ pid name age home }}","variables":{}}'

{
  "data": {
    "getProfile": {
      "pid": "3f70ca58-25ed-4e88-8a45-eea1fbbb45d8",
      "name": "Tony Stark",
      "age": 53,
      "home": "Manhattan, New York City"
    }
  }
}

Update a profile

curl --location -X POST \
  'http://localhost:4001' \
  --header 'Content-Type: application/json' \
  --data-raw '{"query":"mutation { updateProfile(pid: \"3f70ca58-25ed-4e88-8a45-eea1fbbb45d8\",profile: { name: \"Peter Parker\", age: 22, home: \"Queens, New York City\" }){ pid name age home }}","variables":{}}'

{
  "data": {
    "getProfile": {
      "pid": "3f70ca58-25ed-4e88-8a45-eea1fbbb45d8",
      "name": "Peter Parker",
      "age": 22,
      "home": "Queens, New York City"
    }
  }
}

Deploy to the cloud

At this point, you can deploy what you've built to any of the supported cloud providers. In this example we'll deploy to AWS. Start by setting up your credentials and configuration for the nitric/aws provider.

Next, we'll need to create a stack file (deployment target). A stack is a deployed instance of an application. You might want separate stacks for each environment, such as stacks for dev, test, and prod. For now, let's start by creating a file for the dev stack.

The stack new command below will create a stack named dev that uses the aws provider.

nitric stack new dev aws

Edit the stack file nitric.dev.yaml and set your preferred AWS region, for example us-east-1.

provider: nitric/aws@latest
region: us-east-1

You are responsible for staying within the limits of the free tier or any costs associated with deployment.

Let's try deploying the stack with the up command:

nitric up

When the deployment is complete, go to the relevant cloud console and you'll be able to see and interact with your application.

To tear down your application from the cloud, use the down command:

nitric down