This package provides a client that integrates with Fastapi and provides a decorator to cache fastapi controllers responses.
$ pip install fastapi-redis
import redis_client from fastapi_redis
redis_client.set('some_key', 'some_data')
Models can be saved as well and the client will serialize them to json:
import redis_client from fastapi_redis
from pydantic.main import BaseModel
MyModel(BaseModel):
data: str
redis_client.set('some_key', MyModel(data='some_data'))
redis_client.set('some_key', MyModel(data='some_data_that_expires'), expiration=timedelta(days=1))
The data can be saved using the Fastapi Background Tasks so it wouldn't generate any delay on the response to the user:
import redis_client from fastapi_redis
@router.get('/')
async def my_controller_method(background_tasks: BackgroundTasks):
redis_client.set_in_background(background_tasks, 'some_key', 'some_data')
redis_client.set_in_background(background_tasks, 'some_key', MyModel(data='some_data_that_expires'), expiration=timedelta(days=1))
import redis_client from fastapi_redis
some_data = redis_client.get('some_key')
The @redis_cache
decorator can be used to cache the response from a controller. Any argument of the controller can be used to specify the key of the data to store and if background_tasks: BackgroundTasks
is defined as an argument it will store the data in the background as well:
import redis_client from fastapi_redis
@router.get('/resource/{id}')
@router_cache('resource_{id}_{user.id}', timedelta(days=1))
async def my_controller_method(id: str,
user: User = Depends(get_current_user)
background_tasks: BackgroundTasks):
return some_data_that_will_be_cached
Sometimes you want to reset what's on the cache and reprocess the response of the method/function/api without waiting for the timeout, these can be archieved sending the variable reset_cache
to the @router_cache()
decorator:
import redis_client from fastapi_redis
@router.get('/resource/{id}')
@router_cache('resource_{id}_{user.id}', timedelta(days=1))
async def my_controller_method(id: str,
reset_cache: str = None,
user: User = Depends(get_current_user)
background_tasks: BackgroundTasks):
return some_data_that_will_be_cached
This way when the user sends to this url the query parameter ?reset_cache=true
it will reprocess the data and cache it again.