Arsenix: The Async-First Personalization Toolkit

Arsenix is a modern, lightweight, and async-first Python library designed for building sophisticated personalization and recommendation systems with ease. It provides a flexible and modular toolkit for tracking user behavior, managing data, and generating real-time recommendations.

Why Arsenix?

In a world of monolithic and complex recommendation frameworks, Arsenix stands out by being:

• Async-First by Design: Built from the ground up with asyncio, Arsenix is perfect for high-performance, I/O-bound applications, ensuring your recommendation engine never blocks your main thread.
• Pluggable Caching: Swap caching backends on the fly. Start with a simple in-memory cache, then scale up to DiskCache or Redis without changing your application logic.
• Declarative Algorithm Builder: Construct complex recommendation pipelines with a simple, chainable API. The FYPBuilder lets you define how to filter, sort, and rank items declaratively, making your code more readable and maintainable.
• Lightweight and Zero-Dependency (Core): The core library is dependency-free, keeping your environment clean. Optional features like advanced caching can be installed as needed, so you only pay for what you use.

Core Features

• Real-time Pattern Learning: Automatically learn user interest patterns from their interactions.
• Multi-User Recommendations: Generate personalized "For You Page" (FYP) style recommendations for each user.
• Flexible Data Store: A simple dictionary-based data store with asynchronous getters and setters.
• Built-in Persistence: Easily save and load your data store to and from JSON or YAML files.
• Pluggable Cache Engines:
  • LocalCache: In-memory dictionary cache (default).
  • DiskCache: File-based cache for persistence between sessions.
  • RedisCache: High-performance caching using a Redis server.

Getting Started

Installation

Install the core library via pip:

pip install arsenix

To include support for DiskCache and RedisCache, install the optional extras:

pip install arsenix[diskcache,redis]

Quick Example

Here's how easy it is to get personalized recommendations:

import asyncio
from arsenix.server import ArsenixServer

async def main():
    # 1. Initialize the server with some items
    data_store = {
        'item1': {'id': 'item1', 'tags': ['tech', 'python']},
        'item2': {'id': 'item2', 'tags': ['funny', 'cats']},
        'item3': {'id': 'item3', 'tags': ['tech', 'ai']},
    }
    server = ArsenixServer(data_store)

    # 2. Learn a user's interests
    user_id = 'user_123'
    await server.pattern.learn(user_id, ['tech', 'python'])

    # 3. Get recommendations for that user
    recommendations = await server.get_recommendations(user_id)
    print(recommendations)
    # Output: [{'id': 'item1', 'tags': ['tech', 'python']}]

if __name__ == "__main__":
    asyncio.run(main())

Documentation

For a deep dive into the API, advanced usage, and more examples, check out our full documentation in the Documentation.

License

Arsenix is licensed under the MIT License. See the LICENSE file for more details.