LLM Backend Application

Overview

This project is designed to provide a robust backend service for managing conversations and generating responses using OpenAI's GPT-4 model. The system is built with FastAPI and MongoDB, ensuring high performance and scalability.

System Design

Architecture

The system is composed of several key components:

1. FastAPI Application: The core of the backend service, handling HTTP requests and routing.
2. MongoDB: A NoSQL database used to store conversation data.
3. OpenAI Integration: Utilizes OpenAI's API to generate responses based on user queries.

Key Modules

• main.py: Entry point of the application. Configures and starts the FastAPI server.
• routes.py: Defines the API endpoints and their respective handlers.
• services.py: Contains business logic for creating and updating conversations.
• db.py: Manages database connections and collections.
• settings.py: Configuration settings for the application, including environment variables.
• generatefakedata.py: Utility script for generating and inserting sample data into the database.
• log.py: Configures logging for the application.

Data Models

The data models are defined in models.py and include:

• Conversation: Represents a conversation document.
• Message: Represents individual messages within a conversation.
• ChatCompletion: Represents the completion response from OpenAI.

Getting Started

Prerequisites

• Python 3.8+
• MongoDB
• OpenAI API Key

Installation

1. Clone the repository:

   git clone https://github.com/your-repo/project-name.git
   cd project-name

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install dependencies:

   pip install -r requirements.txt

4. Set up environment variables: Create a .env file in the root directory and add the following:

   OPENAI_API_KEY=your_openai_api_key
   OPENAI_MODEL=gpt-4o-mini
   DB_URL=your_mongodb_url
   DB_NAME=your_db_name
   BASE_URL=127.0.0.1
   PORT=8000

Starting the Server

1. Run the FastAPI server:

   ENVIRONMENT=dev python main.py

   This will start the server at http://127.0.0.1:8000.

Generating Sample Data

To generate and insert sample data into the MongoDB collection, run:

python generatefakedata.py

Logging

Logging is configured via logging.ini. Logs are written to a file named with the current date in the format YYYY-MM-DD_Uvicorn.log.

API Endpoints

• POST /conversations: Create a new conversation.
• GET /conversations/{item_id}: Retrieve a conversation by ID.
• GET /conversations/{item_id}/history: Retrieve the history of a conversation.
• PUT /conversations/{item_id}: Update a conversation by ID.
• DELETE /conversations/{item_id}: Delete a conversation by ID.
• POST /conversations/{item_id}/prompts: Add a prompt to an existing conversation.

Limitations

• The POST /conversations/{item_id}/prompts API does not work as expected due to an error {"detail":[{"type":"model_attributes_type","loc":["body","prompt"],"msg":"Input should be a valid dictionary or object to extract fields from","input":"{\"role\": \"user\", \"content\": \"help me to create a credit report based on the file attached\", \"additionalProp1\": {}}"}]}

Contributing

Please read CONTRIBUTING.md for details on our code of conduct and the process for submitting pull requests.

License

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