Welcome to RAGlepedia, your cutting-edge virtual Agent powered by Generative AI (GenAI) models and the innovative Retrieval-Augmented Generation (RAG) approach!
RAGlepedia is designed to be your ultimate companion, providing precise and relevant answers to your queries using context received from VectorDB that contains Wikipedia articles, through state-of-the-art AI technology. With the integration of the RAG model, RAGlepedia ensures that you receive accurate information tailored to your needs.
- Advanced AI Capabilities: Leveraging Generative AI (GenAI) models for intelligent responses.
- Retrieval-Augmented Generation (RAG): Incorporating the RAG model for precise and relevant answers. Used Pinecone for Vector Database.
- Web API: Simple FastAPI interface for user-agent interaction.
- Add static templates and use FastAPI framework to load them
- Add RAG control
- Create a more friendly UI
To start using RAGlepedia, follow these simple steps:
Follow these steps to set up your environment:
- Clone the Repository:
git clone https://github.com/zaaachos/RAGlepedia.git
- Install Dependencies:
It is highly recommended, to use conda as your virtual enviroment:
conda create -n wikienv python=3.9
conda activate wikienv
Install the necessary dependencies by running:
pip install -r requirements.txt
You will also need to have an Azure subscription, and create an .env file having the following variables:
AZURE_OPENAI_API_KEY=<YOUR_OPENAI_KEY>
OPENAI_MODEL_NAME=<YOUR_OPENAI_MODEL>
OPENAI_MODEL_VERSION=<YOUR_VERSION>
OPENAI_MODEL_DEPLOYMENT_VERSION=<YOUR_OPENAI_DEPLOYMENT_MODEL>
AZURE_OPENAI_ENDPOINT=<YOUR_OPENAI_ENDPOINT>
OPENAI_API_TYPE=azure
OPENAI_API_VERSION=2023-07-01-preview
PINECONE_API_KEY=<YOUR_PINECONE_KEY>
PINECONE_INDEX_NAME=<YOUR_PINECONE_INDEX>
EMBEDDINGS_MODEL_NAME=<YOUR_OPENAI_EMBEDDING>
Run the Application Locally. Once dependencies are installed, you can run the FastAPI application locally by executing:
uvicorn main:app --reload
This will start the uvicorn
server, and you can access the application at http://localhost:8000 in your web browser.