Context-Aware-RAG-bot

This code is an attempt of me trying to implement Retrieval Augmented Generation using the Gemini 1.5 flash model.

Progress

• parse through pdfs and chunk them
• vectorise the data onto Chroma DB
• have the llm cite information from uploaded pdf

Ongoing

• implement conversational buffer (memory)
• implement persistent file model, making it suitable for deployment
• prevent the model from accessing previously uploaded data
• implement a method to handle concurrent access by multiple users
• implement some form of rate limiting and access (preferably without a database)