Python Client for Joystick Remote Config
Use Joystick to easily manage and integrate JSON remote configs with any of your Python projects. Joystick is a modern remote config platform to manage all of your configurable parameters. We are natively multi-environment, preserve your version history, have advanced work-flow & permissions management, and much more.
This is a library helps your Python project communicate with the Joystick API for using JSON remote configs with your Python project.
Installation
You can install the package via Pip:
Usage
We provide two types of clients: asynchronous and synchronous. They have exactly the same interfaces, the only difference is how you import them.
Async / Sync
Sync
import os
from joystick import Client
joystick_api_key = os.getenv("JOYSTICK_API_KEY")
if joystick_api_key is None:
raise ValueError("Please set JOYSTICK_API_KEY environment variable.")
client = Client(
api_key=joystick_api_key,
)
response = client.get_contents({"cid1", "cid2"})
print(f'First content: {response["cid1"]}')
print(f'Second content: {response["cid2"]}')
Async
import asyncio
import os
from joystick import AsyncClient
async def main():
joystick_api_key = os.getenv("JOYSTICK_API_KEY")
if joystick_api_key is None:
raise ValueError("Please set JOYSTICK_API_KEY environment variable.")
client = AsyncClient(
api_key=joystick_api_key,
)
response = await client.get_contents({"cid1", "cid2"})
print(f'First content: {response["cid1"]}')
print(f'Second content: {response["cid2"]}')
asyncio.run(main())
All examples below will be provided for the
asyncversion of the client. For sync version, import the sync client, and drop theawaitkeyword.
Requesting config content by ContentId
Get a single config
Get multiple configs at the same time
Specifying Additional Parameters
When creating the Client/AsyncClient instance, you can specify additional parameters that will be used by all API calls from the client. These params can be used for ab testing and/or segmentation; different audiences can get customized responses. For more details see API documentation:
client = AsyncClient(
api_key=joystick_api_key,
cache_expiration_seconds=60,
serialized=True,
params={
"param1": "value1",
"param2": "value2",
},
sem_ver="0.0.1",
user_id="user-id-1",
)
Options
full_response
The full response from Joystick includes additional meta-data about segmentation and ab test groups. By default, you will not get the full response. For more details see Dynamic Config Content documentation.
If you want the full response, specify fullResponse option to the client methods.
get_content_response = await client.get_content('cid1', full_response=True)
# OR
get_contents_response = await client.get_contents({'cid1'} , full_response=True)
serialized
You can get your configuration content as a serialized JSON string.
When true, we will pass the query parameter responseType=serialized to Joystick API.
get_content_response = await client.get_content('cid1', serialized=True)
# OR
get_contents_response = await client.get_contents({'cid1'} , serialized=True)
This option can be set for every API call from the client by setting serialized as true via the constructor, or via property setter.
refresh
If you want to ignore the existing cache and request the new config – pass this option as true.
get_content_response = await client.get_content('cid1', refresh=True)
# OR
get_contents_response = await client.get_contents({'cid1'} , refresh=True)
Caching
By default, the client uses in-memory caching, which means that if you build the distributed application, every instance will go to the Joystick API for at least first call and the cache will be erased after the application is closed.
You can specify your cache implementation which implements either AsyncCacheInterface if you use AsyncClient, or SyncCacheInterface if you use SyncClient.
Async support
We rely on library httpx to make requests to Joystick API and we support the same platforms as httpx.
Clear the cache
If you want to clear the cache – run await client.clear_cache().
Library development
We use the pyenv to install multiple versions of Python on the developer's machine and venv to create the virtual environment for these versions:
pyenv global 3.5.10
rm -rf ./venv # This one might fail, if it's the first time you create `venv` in this proj.
python3 -m venv venv
source ./venv/bin/activate
pip install -e '.[dev]'