Add ability to set retry on Redis connection

[alxwrd]

First check

• I added a very descriptive title to this issue.
• I used the GitHub search to find a similar issue and didn't find it.
• I searched the Pyle38 documentation.
• I already read and followed all the tutorial in the docs and didn't find an answer.
• I already checked if it is not related to Pyle38 but to Pydantic.
• I already checked if it is not related to FastAPI but to aioredis.

Example

Here's a self-contained, minimal, reproducible, example with my use case:

import os

import pytest
import subprocess

from pyle38.client import Client


@pytest.mark.asyncio
async def test_client_handles_tile38_restart():
    client = Client(os.getenv("TILE38_LEADER_URI") or "redis://localhost:9851")

    response = await client.command("SET", ["fleet", "truck", "POINT", 1, 1])
    assert response["ok"]

    subprocess.run(["docker", "restart", "tile38"])

    response = await client.command("SET", ["fleet", "truck", "POINT", 1, 1])
    assert response["ok"]

    await client.quit()

Description

As discussed in #527 - if the connection is reset while you have a long running connection open, while the connection will self heal, the first request after network issues or a Tile38 restart will cause a redis.ConnectionError to be thrown.

Adding retry to the redis client will cause this error to be retried meaning it is never raised.

diff --git a/pyle38/client.py b/pyle38/client.py
index 94d89ca..7980f1f 100644
--- a/pyle38/client.py
+++ b/pyle38/client.py
@@ -2,6 +2,8 @@ from enum import Enum
 from typing import Dict, Sequence, Union

 import redis.asyncio as redis
+from redis.asyncio.retry import Retry
+from redis.backoff import ExponentialBackoff
 from redis.asyncio.connection import parse_url

 from .parse_response import parse_response
@@ -157,6 +159,8 @@ class Client:
                 single_connection_client=True,
                 decode_responses=True,
                 redis_connect_func=self.__on_connect,
+                retry=Retry(ExponentialBackoff(), 10),  # about 3 seconds
+                retry_on_error=[redis.ConnectionError, redis.TimeoutError],
             )

             self.__redis = r

It would be great to be able to configure this from the pyle38 side. Either by having the retry configurable, or by being able to pass an entire redis client to pyle38.

from pyle38 import Tile38
import redis.asyncio as redis

tile38 = Tile38(retry=...)
tile38 = Tile38(redis.Redis(retry=...))