RAG ChatBot
The Brief
What does this project do?
This project implements a RAG chatbot that uses an LLM model such as Ollama to answer questions based on previously uploaded PDF files.
What technologies does this project use?
- Python - Programming languages
- FastAPI - Back end system
- ChromaDB - DB vector to save all data embedding
- Docker - Container projects
- Ollama Mistral - LLM used to elaborate the answer
- Vue - Used to create an interface UI
Project flow
- User uploads a PDF file
- Back end receives the file and saves it in the folder
- Ask a question about the file
- Back end calculates the embeddings question
- Search the database for all chunks close to the question
- Create a prompt with the chunk question
- Ollama gives the answer
- Back end gives the answer
What did I learn from this project?
In this project, I learned the basics of how AI can really help in reading the data we upload.
I learned how to upload files, save them, and convert the data into vectors that AI can read.
I learned the concept of embeddings and vector databases, prompt insertion, and LLM processing to generate responses.