Flask-based Chat API with Document Embeddings

This project provides a Flask-based API for document embedding and chat functionality. It allows users to upload documents, generate embeddings, and interact with a chat system based on those embeddings.

Prerequisites

Before you begin, ensure you have met the following requirements:

• You have a working Python environment (3.7+)
• You have Conda installed on your system

Setting Up the Environment

1. Create a new Conda environment:

   conda create -n chat-api python=3.8
   

2. Activate the environment:

   conda activate chat-api
   

3. Install the required packages:

   pip install -r requirements.txt
   

Configuration

Set the following environment variables or update them in the script:

• JWT_SECRET_KEY: Secret key for JWT token generation (default: 'chatwithme')
• CHAT_URL: URL for the chat service (default: 'http://localhost:8000/query')
• MASTER_API_KEY: Master API key for authentication (default: '1234567890')

Running the Application

To run the application, execute:

python flask_server_docker.py

The server will start on port 5996 with SSL enabled. Make sure you have the certificate.crt and private.key files in the same directory.

API Endpoints

• /get_temp_api_key (POST): Get a temporary API key
• /api/start_chat (POST): Start a chat session
• /api/chat (POST): Send a chat message
• /make_embedding (POST): Generate embeddings for uploaded documents
• /start_chat (POST): Initiate a chat session
• /chat (POST): Send a chat message (alternative endpoint)

Security

This application uses JWT for authentication and HTTPS for secure communication. Ensure to keep your secret keys and SSL certificates secure.

embeddings.py

The embeddings.py file contains functions for generating embeddings from documents and converting document objects to dictionaries.

Functions

1. get_embeddings(directory): Generates embeddings for documents in the specified directory.
2. document_to_dict(doc): Converts a Document object to a dictionary.
3. get_documents(directory): Loads documents from a directory and converts them to a list of dictionaries.

API Testing (api_test.ipynb)

The api_test.ipynb Jupyter notebook demonstrates how to interact with the API endpoints. Here's a breakdown of the notebook:

1. Initiating a chat session:

   • Makes a POST request to the /api/start_chat endpoint with an API key.
   • Retrieves a JWT token and model URL from the response.

2. Sending a chat message:

   • Uses the JWT token to authenticate a POST request to the /api/chat endpoint.
   • Sends a query and receives a response from the chat model.

Example Usage

import requests

# Start chat session
api_key = "your_api_key_here"
start_chat_url = "https://your-api-url.com/api/start_chat"
response = requests.post(start_chat_url, headers={"X-API-Key": api_key})
jwt_token = response.json()["jwt_token"]
model_url = response.json()["url"]

# Send chat message
chat_url = "https://your-api-url.com/api/chat"
query = "Your question here"
response = requests.post(
    chat_url,
    headers={"Authorization": jwt_token},
    json={"query": query, "url": model_url}
)

print(response.json()["query"])
print(response.json()["answer"])

Ngrok Setup

As ICP needs IPV6 support, we can use ngrok to expose the local server to the internet. Here's how to set it up:

1. Download and install ngrok from the official website.

2. Run the following command to expose the local server:

   ngrok http 5996
   

3. Copy the HTTPS URL generated by ngrok and manually update the CHAT_URL environment variable in ICP_NFTs project.

Note

This is a development server. For production deployment, consider using a production-grade WSGI server and implement proper security measures.