FigChain Python Client

Official Python client library for FigChain configuration management.

Features

• Real-time configuration updates - Subscribe to configuration changes with background polling
• Rule-based rollouts - Evaluate feature flags and configurations based on user context
• Type-safe models - Avro-based serialization for efficient data transfer with Python dataclasses
• Flexible storage - Thread-safe in-memory storage
• Python 3.7+ - Support for modern Python versions

Installation

Install using pip:

pip install figchain

Quick Start

from figchain import FigChainClient, Context

# Your generated config class (see "Generating Models")
from my_app.models import MyConfig 

# Initialize the client
client = FigChainClient(
    base_url="https://api.figchain.io",
    client_secret="your-client-secret",
    environment_id="your-environment-id",
    namespaces={"default"}
)

# Define context for evaluation (e.g., user properties for traffic splitting)
context: Context = {
    "userId": "user123",
    "plan": "premium"
}

# Fetch configuration safely
# returns Optional[MyConfig]
config = client.get_fig("your-fig-key", MyConfig, context=context)

if config:
    if config.enabled:
        print(f"Feature enabled with color: {config.backgroundColor}")
    else:
        print("Feature disabled")

# Clean up resources when done
client.close()

Generating Models

The Python client uses Avro schemas to generate type-safe dataclasses. You can use the included script to generate these models from your .avsc files:

# If installed from source in venv:
python3 scripts/generate_models.py path/to/schema.avsc path/to/output_models.py

Development

1. Setup Environment:

   python3 -m venv venv
   . venv/bin/activate
   pip install -e .[dev]

2. Run Tests:

   pytest

License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

Support

• Documentation
• Issues
• Contact