Welcome to Diver!

   This is an AI-powered chatbot that helps you navigate my tech blog https://naomilago.com. It can answer questions about my blog posts, my technology skills, and my projects. It can also help you find relevant questions about my own technology skills, by explaining complex concepts in simple terms and also in a friendly manner.

The most important thing is that this project is part of the Imersão IA, a project created by Alura and Google, a Brazilian online learning platform. The goal of this project is to help people learn how to get the most out of AI and explore Gemini's powerhouse.

Authorship

• @naomilago

   A curious Data Scientist who always like to explore new technology and is passionate about NLP, starting to dive into RAG systems now in 2024.

FAQ

What technologies were used?

• Diver was built using foundational tools, such as the Gemini model, LangChain, HuggingFace, ChromaDB, and Streamlit.

Can I trust this chatbot responses for complex questions?

• Well, as any RAG system, Diver is not perfect. The vector store used for retrieval is based on the current content within my blog, and the chatbot will always provide you with the best answer possible - including suggesting my contact if the answer is beyonde its knowledge.

What languages does it support?

• The Diver's core components includes Gemini, a multilingual model, so it can understand and answer questions in multiple languages - including English, Portuguese, Spanish, and many others.

Screenshots

Run Locally

   You can also run Diver locally! If you'd like to explore, just follow these steps and have fun :)

Clone the project

  git clone https://github.com/naomilago/diver.git

Go to the project directory

  cd diver

Install dependencies

  pip install -r requirements.txt

Create a /config/.secrets.json file with the following content:

  {
    "HF_TOKEN": "<YOUR-HUGGING-FACE-TOKEN>",
    "GEMINI_API_KEY": "<YOUR-GEMINI-API-KEY>",
    "LANGSMITH_API_KEY": "<YOUR-LANGSMITH-API-KEY>"
  }

Start the server

  python -m streamlit run src/app.py