featureflow-python-sdk

Python SDK for the featureflow feature management platform

Get your Featureflow account at featureflow.io

Get Started

The easiest way to get started is to follow the Featureflow quick start guides

Installation

The SDK is available on PyPI as featureflow-sdk.

You can either add it as a dependency or install it globally.

python -m pip install featureflow-sdk

Usage

Create one Featureflow client for the lifetime of your application and share it. The constructor starts background threads that poll Featureflow for flag configuration and batch evaluation events back to the server. Constructing a client per request (or per evaluation) spawns new threads and re-downloads your flags every time — create it once at startup, then call evaluate() on that single instance as often as you like; evaluation is local and cheap.

import os

from featureflow import Featureflow, User

# Once, at application startup. Use your server environment key (sdk-srv-env-...).
featureflow = Featureflow(os.environ["FEATUREFLOW_SERVER_KEY"])


def show_my_feature(user_id):
    user = User(key=user_id, attributes={"tier": "gold"})

    if featureflow.evaluate("my-cool-feature", user).isOn():
        print("I'm enabled")

In a web application, create the client at startup and reuse it across requests — see examples/server for how the Flask example does exactly this.

For multivariate flags, check a specific variant or read the evaluated value:

evaluated = featureflow.evaluate("checkout-flow", user)

if evaluated.is_("variant-b"):
    ...

print(evaluated.value())  # e.g. "variant-b", or "off"

Naming your application

Optionally tag this workload with an application name so the Featureflow dashboard can attribute SDK usage and flag evaluations to it (Admin → SDKs, and the "Evaluated by" panel on each feature's statistics tab):

featureflow = Featureflow(
    os.environ["FEATUREFLOW_SERVER_KEY"],
    application="checkout-api",
)

The name is a slug — lowercase letters, numbers, ., _ and -, at most 64 characters. An invalid value is dropped with a warning and no tag is sent. The FEATUREFLOW_APPLICATION environment variable is used when the option is not set in code.

Example server

See examples/server for a small Flask app you can run locally to manually test evaluations against your Featureflow account from a browser.