Welcome to the RAG (Retrieval-Augmented Generation) Workshop for Asian Women Advancing AI! This demo has been created by Zeba Karkhanawala to demonstrate the power of building your own Streamlit-based chatbot, leveraging Retrieval-Augmented Generation (RAG) techniques. The goal is to enable users to upload their documents, and interact with them using a simple yet effective AI-powered application that can answer any query related to the content of the uploaded document.
Retrieval-Augmented Generation (RAG) is an advanced AI technique that combines information retrieval (searching for relevant documents) with language generation (creating coherent, relevant answers from the retrieved documents). The RAG model improves the quality of responses from an AI system by integrating external documents or databases and using that information to answer user queries. This method can be particularly useful for tasks like question answering, summarization, and information extraction.
This demo demonstrates how you can build your very own Streamlit Chatbot that can query any document uploaded by the user. The chatbot will intelligently search through the uploaded document and return answers to your queries using RAG technology. The demo is powered by Streamlit, a powerful tool for building interactive web applications in Python.
- Streamlit: A Python framework to create interactive web apps for machine learning and data science projects.
- OpenAI GPT-4o: To generate responses to user queries using large language models.
- Chroma: A vector store used to store the document embeddings for quick retrieval.
- Langchain: A framework to simplify the process of combining LLMs and external tools like APIs, databases, and document loaders.
Follow these steps to set up and run the Streamlit app on your local machine:
To get started, clone the repository to your local machine using Git:
git clone https://github.com/yourusername/AWAAI-RAG-Demo.gitThe demo requires an OpenAI API Key to interact with the OpenAI models. To set up the API key, follow the steps mentioned in the 'How to get an OpenAI API key.pdf' document. Then follow the below steps.
- Create a new file named
.envin the root directory of the project. - Add the following line to the
.envfile, 7C40 replacingyour-openai-api-keywith your actual OpenAI API key:
OPENAI_API_KEY="your-openai-api-key"This project requires certain libraries to run. You can install them using the requirements.txt file included in the repository.
a) First, navigate to the project directory:
cd AWAAI-RAG-Demob) Then, install the required libraries using pip:
pip install -r requirements.txtOnce the dependencies are installed, you can run the Streamlit app with the following command:
streamlit run AWAAI_RAG_Demo_StreamlitApp.pyThis command will start the app and you can access it in your web browser at http://localhost:8501.
- Go to the GitHub repository.
- Click the Fork button at the top-right corner of the page.
- This creates a copy of the repository in your GitHub account.
- Go to Streamlit Cloud.
- Click "Sign in with GitHub".
- Connect your GitHub account.
- After logging in, click "Create app" from the top right corner.
- Under "Repository", select your forked repo.
- Choose the correct branch (usually
mainormaster). - Set the "Main file path" to your Streamlit script (here 'AWAAI_RAG_Demo_StreamlitApp.py').
- Click "Deploy".
- On your Streamlit Cloud app page, click on the three dots next to your app name.
- Go to "Settings"
- Click on "Secrets"
- Add your OpenAI API Key in this format:
OPENAI_API_KEY="your-openai-api-key"- Once the app is deployed, you can access it at the URL provided by Streamlit Cloud.
- You can also access it from the Streamlit Cloud dashboard.
- To update the app, you need to push the changes (that are made locally) to the repository and redeploy the app on Streamlit Cloud.