DocChatAI: Intelligent Document Assistant for Efficient Information Retrieval

DocChatAI is an advanced streamlit based web application that leverages Retrieval-Augmented Generation (RAG) to deliver intelligent, context-aware responses to user queries based on the content of uploaded PDF documents.

Features

• Document Upload: Users have the option to upload PDF documents which the app will utilize as additional context for generating answers.
• Chat Interface: Ask questions and receive contexually relevant & accurate answers, with insights pulled directly from the uploaded PDFs.
• Download Conversation: Before resetting the session, users can download a PDF of the conversation, including all interactions and responses generated from the uploaded documents. This feature ensures that users can retain a complete record of the session's context for future reference.
• Reset Button: This feature clears the chat history and removes all uploaded documents from backend storage. To start a new session, users will need to re-upload the documents.

Getting Started

These instructions will walk you through the process of setting up DocChatAI on your local machine for development and testing.

Prerequisites, Installations & Usage

1. Set up project directory
2. Install python 3.8 or higher (Used 3.10.9)
3. Clone the repository:

   git clone https://github.com/yourusername/DocChatAI.git

4. Navigate to the project directory:

   cd DocChatAI

5. Install necessary dependencies:

   pip install -r requirements.txt

6. Create a virtual environment to isolate dependencies for your project. For Windows:

   python -m venv venv
   .\venv\Scripts\activate

7. Create a "config.toml" file to configure global settings for Streamlit app, including parameters such as the application theme, server port, and other essential options.
8. Run the Streamlit app:

   streamlit run app.py
   

For Docker Usage

1. Build the docker image:

   docker build -t project-image .

2. Run the docker container:

   docker run -d -p 8501:8501 --name project-container project-image
   

Note: If requirements.txt doesn't have all the dependencies installed before making the docker image, use the following command to put all dependencies from the virtual environment to requirements.txt

pip freeze > requirements.txt