Document Assistant Chat Application

This is a simple web application that allows you to interact with an OpenAI Assistant that has access to your uploaded documents through a vector store.

Prerequisites

• Python 3.8 or higher
• OpenAI API key
• Documents already uploaded and processed into a vector store

Setup

1. Clone this repository
2. Create a virtual environment:

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

3. Install dependencies:

   pip install -r requirements.txt
   

4. Create a .env file in the root directory with your OpenAI API key:

   OPENAI_API_KEY=your_api_key_here
   

5. Run the file upload script to upload your documents and create a vector store:

   python scripts/file-upload.py
   

   This will create a vector_store_info.json file with your vector store ID.

Running the Application

1. Start the Flask application:

   python app.py
   

2. Open your browser and navigate to http://localhost:5000
3. Start chatting with the assistant about your documents!

Vector Store Management

Listing Vector Stores

You can list all available vector stores using the provided script:

python scripts/list-vector-stores.py

This will display all vector stores in your account, including their IDs, names, descriptions, creation dates, and associated files.

Switching Vector Stores

The web application includes a dropdown menu in the header that allows you to:

1. View all available vector stores
2. Select a different vector store to use with the assistant
3. Refresh the list of vector stores

When you switch to a different vector store, the application will:

• Update the vector store ID in vector_store_info.json
• Create a new assistant with access to the selected vector store
• Start a new conversation thread

How It Works

1. The application creates an OpenAI Assistant with access to your vector store
2. When you send a message, it's added to a thread
3. The assistant processes your message and searches through your documents to provide relevant answers
4. The response is displayed in the chat interface

Files

• app.py: The Flask application
• templates/index.html: The HTML template for the chat interface
• static/css/style.css: CSS styles for the chat interface
• static/js/script.js: JavaScript for handling the chat functionality
• scripts/file-upload.py: Script for uploading files and creating a vector store
• scripts/list-vector-stores.py: Script for listing all vector stores

Notes

• The assistant and thread IDs are stored in assistant_info.json for reuse
• The vector store ID is stored in vector_store_info.json
• The application uses the OpenAI Assistants API v2 with the file_search tool