🚀 LLM Chat Backend (FastAPI + MongoDB)

This is a backend component that facilitates interaction with LLMs (OpenAI GPT-4o-mini). It provides:

• CRUD operations for conversations containing queries and responses.
• LLM query handling with conversation history as context.
• Anonymized query storage in MongoDB for auditing.
• REST API with OpenAPI docs for easy integration.

---

📌 Features

✅ FastAPI-based backend
✅ MongoDB for conversation storage
✅ Dockerized environment
✅ CRUD endpoints for managing conversations
✅ Send prompt queries & receive LLM responses
✅ Automatic logging & anonymization of chat history
✅ OpenAPI documentation for API usage

---

🛠 Tech Stack

Component	Technology
Backend	FastAPI (Python 3.10+)
Database	MongoDB (via Docker)
ORM	Beanie (MongoDB ODM)
LLM API	OpenAI Python Client
Validation	Pydantic
Containerization	Docker, Docker Compose

---

🚀 Setup and Running the Application

1️⃣ Clone the Repository

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

2️⃣ Configure Environment Variables

Duplicate the .env.example file and name it .env. It should look like the following:

MONGO_URI="mongodb://mongo:27017/launchpad_db"
OPENAI_API_KEY="INPUT YOUR OPENAI API KEY HERE"

Enter your OpenAI API Secret key as a string ✅ Notes:

• mongo is the service name inside Docker Compose.
• Do not use localhost inside Docker Compose (use mongo instead).

3️⃣ Run with Docker Compose

To build and start both FastAPI & MongoDB, run:

docker-compose up -d --build

✅ Notes:

• d → Runs in detached mode (in the background).
• -build → Ensures FastAPI is rebuilt.

Check running containers:

docker ps

4️⃣ Verify MongoDB is Running

To check if MongoDB is working inside the Docker container, enter:

docker exec -it mongodb mongosh

Then, inside MongoDB shell:

show dbs  # List databases
use launchpad_db  # Switch to project database
show collections  # Show collections

5️⃣ Access the FastAPI Application

If the two containers are running correctly, you should be able to access:

API Documentation (Swagger UI): http://localhost:8000/docs

You can test the endpoints here. Enter the request body and view the response body.

📌 API Endpoints

The endpoints were created based off the openai.yaml file provided by the team.

1️⃣ Create a Conversation

POST /conversations/

Example Request Body:

{
  "name": "My Conversation",
  "params": {
    "temperature": 0.7
  }
}

Example Response:

{
  "id": "656a21fb3b3f76f3a2e6b1c3",
  "name": "My Conversation",
  "params": {"temperature": 0.7}
}

2️⃣ Get All Conversations

GET /conversations/

Example Response:

[
  {
    "id": "656a21fb3b3f76f3a2e6b1c3",
    "name": "My Conversation",
    "params": {
      "temperature": 0.7
    }
  },
  {
    "id": "656a21fb3b3f76f3a2e6b1c4",
    "name": "Another Conversation",
    "params": {
      "temperature": 0.5
    }
  }
]

3️⃣ Get a Specific Conversation

GET /conversations/{conversation_id}

Example Response:

{
  "id": "656a21fb3b3f76f3a2e6b1c3",
  "name": "My Conversation",
  "params": {
    "temperature": 0.7
  },
  "tokens": 100,
  "pinned": false,
  "prompts": [
    {"role": "user", "content": "Hello!"},
    {"role": "assistant", "content": "Hi, how can I assist you?"}
  ],
  "modifications": {}
}

Error Response:

{
  "detail": {
    "code": 404,
    "message": "Specified resource(s) was not found"
  }
}

4️⃣ Update a Conversation

PUT /conversations/{conversation_id}

Example Request Body:

{
  "name": "Updated Conversation",
  "params": {
    "temperature": 0.9,
    "max_tokens": 200
  }
}

Example Response:

{
  "id": "656a21fb3b3f76f3a2e6b1c3",
  "name": "Updated Conversation",
  "params": {
    "temperature": 0.9,
    "max_tokens": 200
  }
}

5️⃣ Delete a Conversation

DELETE /conversations/{conversation_id}

6️⃣ Send a Prompt to LLM

POST /conversations/{conversation_id}

Example Request Body:

{
  "content": "Hello, how are you?"
}

Example Response:

{
  "message": "I'm doing well, thank you!"
}

📌 Stopping & Managing Containers

Stop the running contaainers

docker-compose down

Restart the application

docker-compose up -d

📌 Project Structure

/your-repo
│── /app                 # FastAPI source code
│   ├── main.py          # FastAPI entry point
│   ├── models/          # Pydantic models
│   ├── routes/          # FastAPI endpoints
│   ├── config/          # MongoDB connection
│── Dockerfile           # FastAPI Docker setup
│── docker-compose.yml   # Docker Compose for MongoDB + FastAPI
│── requirements.txt     # Python dependencies
│── .env                 # Environment variables
|── .env.example         # Example of .env file
|── .gitignore           # Contains untracked file names
│── README.md            # This file

📌 Troubleshooting

Error: Port 27017 is already allocated

Open docker terminal and enter this command to see the running containers:

docker ps

Check the name of the container that is running on Port 27017.

docker stop <name of container>

Then restart by running

docker-compose up -d