You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

7.3 KiB

FastAPI Integration

Introduction

Good news: Redis OM was specifically designed to integrate with FastAPI!

This section includes a complete example showing how to integrate Redis OM with FastAPI.

Concepts

Every Redis OM Model is also a Pydantic model

Every Redis OM model is also a Pydantic model, so you can define a model and then use the model class anywhere that FastAPI expects a Pydantic model.

This means a couple of things:

  1. A Redis OM model can be used for request body validation
  2. Redis OM models show up in the auto-generated API documentation

Cache vs. Data

Redis works well as either a durable data store or a cache, but the optimal Redis configuration is often different between these two use cases.

You almost always want to use a Redis instance tuned for caching when you're caching and a separate Redis instance tuned for data durability for storing application state.

This example shows how to manage these two uses of Redis within the same application. The app uses a FastAPI caching framework and dedicated caching instance of Redis for caching, and a separate Redis instance tuned for durability for Redis OM models.

Example app code

Let's look at an example FastAPI app that uses Redis OM.

NOTE: This example code requires dependencies to run. To install the dependencies, first clone the redis-om-fastapi repository from GitHub. Then follow the installation steps later in this document or in that repository's README.md file.

import datetime
from typing import Optional

import redis

from fastapi import FastAPI, HTTPException
from starlette.requests import Request
from starlette.responses import Response

from fastapi_cache import FastAPICache
from fastapi_cache.backends.redis import RedisBackend
from fastapi_cache.decorator import cache

from pydantic import EmailStr

from redis_om import HashModel, NotFoundError
from redis_om import get_redis_connection

# This Redis instance is tuned for durability.
REDIS_DATA_URL = "redis://localhost:6380"

# This Redis instance is tuned for cache performance.
REDIS_CACHE_URL = "redis://localhost:6381"


class Customer(HashModel):
    first_name: str
    last_name: str
    email: EmailStr
    join_date: datetime.date
    age: int
    bio: Optional[str]


app = FastAPI()


@app.post("/customer")
async def save_customer(customer: Customer):
    # We can save the model to Redis by calling `save()`:
    return customer.save()


@app.get("/customers")
async def list_customers(request: Request, response: Response):
    # To retrieve this customer with its primary key, we use `Customer.get()`:
    return {"customers": Customer.all_pks()}


@app.get("/customer/{pk}")
@cache(expire=10)
async def get_customer(pk: str, request: Request, response: Response):
    # To retrieve this customer with its primary key, we use `Customer.get()`:
    try:
        return Customer.get(pk)
    except NotFoundError:
        raise HTTPException(status_code=404, detail="Customer not found")


@app.on_event("startup")
async def startup():
    r = redis.asyncio.from_url(REDIS_CACHE_URL, encoding="utf8",
                          decode_responses=True)
    FastAPICache.init(RedisBackend(r), prefix="fastapi-cache")

    # You can set the Redis OM URL using the REDIS_OM_URL environment
    # variable, or by manually creating the connection using your model's
    # Meta object.
    Customer.Meta.database = get_redis_connection(url=REDIS_DATA_URL,
                                                  decode_responses=True)

Testing the app

You should install the app's dependencies first. This app uses Poetry, so you'll want to make sure you have Poetry installed first:

$ pip install poetry

Then install the dependencies:

$ poetry install

Next, start the server:

$ poetry run uvicorn --reload main:app

Then, in another shell, create a customer:

    $ curl -X POST  "http://localhost:8000/customer" -H 'Content-Type: application/json' -d '{"first_name":"Andrew","last_name":"Brookins","email":"a@example.com","age":"38","join_date":"2020
-01-02"}'
    {"pk":"01FM2G8EP38AVMH7PMTAJ123TA","first_name":"Andrew","last_name":"Brookins","email":"a@example.com","join_date":"2020-01-02","age":38,"bio":""}

Get a copy of the value for "pk," which is the model's primary key, and make another request to get that customer:

$ curl "http://localhost:8000/customer/01FM2G8EP38AVMH7PMTAJ123TA"
{"pk":"01FM2G8EP38AVMH7PMTAJ123TA","first_name":"Andrew","last_name":"Brookins","email":"a@example.com","join_date":"2020-01-02","age":38,"bio":""}

You can also get a list of all customer primary keys:

$ curl "http://localhost:8000/customers"
{"customers":["01FM2G8EP38AVMH7PMTAJ123TA"]}

Redis OM with Asyncio

Redis OM is designed to work with asyncio, so you can use Redis OM models asynchronously within FastAPI applications.

The only difference is that you import the Redis OM models from the aredis_om module instead of the redis_om module.

Here is the previous FastAPI app, but using asyncio-compatible Redis OM code:

import datetime
from typing import Optional

import redis

from fastapi import FastAPI, HTTPException
from starlette.requests import Request
from starlette.responses import Response

from fastapi_cache import FastAPICache
from fastapi_cache.backends.redis import RedisBackend
from fastapi_cache.decorator import cache

from pydantic import EmailStr

from aredis_om import HashModel, NotFoundError  # <- Notice, we import from aredis_om
from aredis_om import get_redis_connection

# This Redis instance is tuned for durability.
REDIS_DATA_URL = "redis://localhost:6380"

# This Redis instance is tuned for cache performance.
REDIS_CACHE_URL = "redis://localhost:6381"


class Customer(HashModel):
    first_name: str
    last_name: str
    email: EmailStr
    join_date: datetime.date
    age: int
    bio: Optional[str]


app = FastAPI()


@app.post("/customer")
async def save_customer(customer: Customer):
    # We can save the model to Redis by calling `save()`:
    return await customer.save()  # <- We use await here


@app.get("/customers")
async def list_customers(request: Request, response: Response):
    # To retrieve this customer with its primary key, we use `Customer.get()`:
    return {"customers": await Customer.all_pks()}  # <- We also use await here


@app.get("/customer/{pk}")
@cache(expire=10)
async def get_customer(pk: str, request: Request, response: Response):
    # To retrieve this customer with its primary key, we use `Customer.get()`:
    try:
        return await Customer.get(pk)  # <- And, finally, one more await!
    except NotFoundError:
        raise HTTPException(status_code=404, detail="Customer not found")


@app.on_event("startup")
async def startup():
    r = redis.asyncio.from_url(REDIS_CACHE_URL, encoding="utf8",
                          decode_responses=True)
    FastAPICache.init(RedisBackend(r), prefix="fastapi-cache")

    # You can set the Redis OM URL using the REDIS_OM_URL environment
    # variable, or by manually creating the connection using your model's
    # Meta object.
    Customer.Meta.database = get_redis_connection(url=REDIS_DATA_URL,
                                                  decode_responses=True)

NOTE: The modules redis_om and aredis_om are identical in almost every way. The only difference is that the aredis_om returns coroutines that you must await.