Live Chat

Installation

Clone the repository:

git clone [project url]

Install dependencies:

poetry install

Set up environment variables: Create a .env file at the root of your project with the following variables:

PINECONE_API_KEY=your-pinecone-api-key
PINECONE_INDEX_NAME=your-pinecone-index-name
OPENAI_API_KEY=your-openai-api-key

Run the application: Start the server with:

poetry run python man.py

The server will run on http://localhost:8000.

API Endpoints

WebSocket Endpoint

/chat: The WebSocket endpoint for the live chat. Users can send messages and receive responses in real-time.

Supported Commands

/reset: Resets the chat context. This clears all previous interactions and starts a new session.

/nocontext: Sends a query without using the stored context. The response is generated solely based on the current query.

REST API Endpoints

POST /upload: Uploads and processes a document to store in the Pinecone vector database.
Accepted File Types: application/pdf, application/vnd.openxmlformats-officedocument.wordprocessingml.document, text/plain

Response:

{
    "filename": "your_file.pdf",
    "message": "File uploaded and processed successfully"
}

Usage

Chat Interaction

Connect to the WebSocket: Use a WebSocket client (or a custom client) to connect to ws://localhost:8000/chat.
Send Messages:

Send any text message to interact with the AI assistant.
Use /reset to clear the session context.
Use /nocontext your_message to bypass the stored context.

Receive Responses: The AI assistant will respond to your queries in real-time, utilizing both the current context and the RAG mechanism.

Uploading Documents

Upload a document: Use the /upload endpoint to upload a document for processing.

curl -X POST "http://localhost:8000/upload" -F "file=@your_file.pdf"

Process the document: The document is split into chunks, embedded using a transformer model, and stored in Pinecone for future retrieval during chat interactions.

CodeGeneration-2020/test-task-ruslan_07.08