product_recommender

Product Recommender API

Overview

This project is a FastAPI-based service designed to assist users in finding and learning about products. It integrates a LangChain knowledge base with a Milvus vector store and OpenAI GPT model to deliver personalized product recommendations and summaries.

Features

• CSV-Based Knowledge Base: Dynamically loads product data from CSV files in the specified directory.
• Vector Store Integration: Uses Milvus for efficient storage and retrieval of vectorized product data.
• Embedding Models: Utilizes OpenAI embeddings for generating vector representations of product information.
• Agent Functionality: A custom agent, "Product Recommender," interacts with users, retrieves relevant products, and summarizes their descriptions.
• API Endpoint: Accepts user queries via a /query endpoint and provides detailed responses in JSON format.
• Personalized Recommendations: Offers tailored product suggestions based on user queries and retrieved knowledge.

Technologies Used

• FastAPI: Framework for building the REST API.
• Milvus: Vector database for similarity searches.
• Phi Framework: Used for creating the custom agent and integrating knowledge bases.
• LangChain: Facilitates document loading and vectorized queries.
• OpenAI GPT: Leverages the gpt-4o-mini model for natural language processing.
• Python Dotenv: Manages environment variables.

Directory Structure

.
├── app.py                   # Main API and agent logic
├── your_directory_path/     # Directory containing CSV files with product data
├── requirements.txt         # Python dependencies
├── .env                     # Environment variables (e.g., OpenAI API key)

Installation

1. Clone the repository:

   git clone https://github.com/your-repo/product-recommender-api.git
   cd product-recommender-api

2. Set up a Python virtual environment:

   python3 -m venv venv
   source venv/bin/activate  # For Linux/macOS
   venv\Scripts\activate     # For Windows

3. Install dependencies:

   pip install -r requirements.txt

4. Configure Environment Variables: Create a .env file in the project root with your OpenAI API key:

   OPENAI_API_KEY=your_openai_api_key

Usage

1. Prepare CSV Data:

   • Add product data as CSV files to the your_directory_path/ directory.
   • Ensure each file follows a consistent format, with headers like Index, Height, Weight, etc. Update the field names in the code to match your data structure.

2. Run the FastAPI server:

   uvicorn app:app --host 0.0.0.0 --port 8000

3. Query the Agent: Use the /query endpoint to interact with the product recommender:

   • Request Format:

     {
       "query": "Recommend a lightweight laptop for office use."
     }

   • Response Format:

     {
       "response": "Here are some recommended lightweight laptops for office use: ..."
     }

4. Customize Data:

   • Modify the your_directory_path directory and update the CSV loading logic in the code to fit your specific dataset.
   • Adjust the knowledge base configurations or embedding model as needed.

Configuration

• Knowledge Base Path: Update the directory_path variable to point to your dataset.
• Milvus Vector Store: Modify the URI for the database connection as required.
• CSV File Format: Ensure the field names in the csv_args match your dataset schema.
• Model and Embeddings: Replace OpenAIChat and OpenAIEmbeddings with alternative models or embeddings if needed.

Dependencies

• faiss
• fastapi
• phi
• langchain_community
• langchain_openai
• langchain_milvus
• uvicorn
• python-dotenv

Install all dependencies with:

pip install -r requirements.txt

Additional Notes

• Ensure that the Milvus vector store is properly set up and accessible.
• Customize the agent's instructions for specific interaction styles or use cases.
• Extend the system to support additional product categories or data sources.