pydantic-graph-persistance-redis 0.8.1

State persistance using Redis as backend

pydantic-graph-redis-persistance

Installation

The package is available on PyPi and can be install via pip or any other package manager:

pip install pydantic-graph-redis-persistance

Usage

The responsability of creating a proper asynchronous Redis client is yours, then you can simply create a state persistance instance for your graph run:

from pydantic_graph.persistance.redis import RedisStatePersistance
from redis.asyncio import Redis

redis = Redis(...)
run_id = "my_unique_run_id"
persistance = RedisStatePersistance(redis, run_id)

# You can now use your persistance through any graph run :).

Locking

To provide a quick and simple implementation, the current version use a deprecated Redis locking mecanism through SETNX operation. If you want you can implement your own locking mecanism (for instance redlock) or add a custom timeout to the lock you can provide you own AbstractRedisStateLock instance:

from pydantic_graph.persistance.redis import NXRedisStateLock
from pydantic_graph.persistance.redis import RedisStatePersistance

redis = Redis(...)
lock = NXRedisStateLock(
    lock_id="my_custom_lock_id",
    redis=redis,
    timeout=10000.0,
)
run_id = "my_unique_run_id"
persistance = RedisStatePersistance(redis, run_id, redis_state_lock=lock)