/gemini-langchain-rag

A demo project to build a Google Cloud Certifications Chatbot using Langchain, Google Cloud AI Platform, and Vertex AI.

Primary LanguagePythonMIT LicenseMIT

Gemini LangChain Rag

A demo project to build a Google Cloud Certifications Chatbot using Langchain, Google Cloud AI Platform, and Vertex AI. The project demonstrates how to:

  • Load and split certification study guides from URLs.
  • Create vector embeddings using Vertex AI.
  • Store and search document embeddings in BigQuery.
  • Interface with a generative AI model for answering user queries.
  • Run a web demo with Streamlit.

Gemini LangChain RAG Architecture

Project Structure

  • ai.py: Contains the AI helper function that integrates the language model and document similarity search to answer certification questions.
  • knowledge_base.py: Loads certification study guides, splits them into chunks, and adds them to the BigQuery vector store.
  • search.py: Demonstrates running a similarity search on the vector store.
  • vertexai_vectorsearch.py: An alternative example implementing vector search with Vertex AI’s matching engine.
  • streamlit.py: Runs a Streamlit web demo for interacting with the chatbot.
  • .env: Environment configuration file for setting project-specific variables.
  • requirements.txt: Lists the dependencies for the project.

Prerequisites

  • Python 3.9+
  • A Google Cloud Project with AI Platform and BigQuery enabled.
  • Proper service account credentials configured for accessing Google Cloud resources.

Setup

  1. Clone the Repository

    git clone https://github.com/rishabkumar7/gemini-langchain-rag
cd gemini-langchain-rag

  2. Configure Environment Variables

    • Copy the example file and modify the values: 
      cp .env.example .env
    • Update the .env file with your Google Cloud project ID, dataset, table, region, and API key.

  3. Install Dependencies

    pip install -r requirements.txt

Running the Demo

Command Line Execution

  • Knowledge Base Loading
    Run knowledge_base.py to load and add certification documents to the vector store.

    python knowledge_base.py

  • Vector Search Example
    Run search.py or vertexai_vectorsearch.py to test the vector search functionality.

    python search.py
# or
python vertexai_vectorsearch.py

  • Chatbot AI Helper
    The ai.py file invokes the AI helper function with a sample query. You can modify the code to test different questions.

Web Demo with Streamlit

Launch the web-based chatbot interface provided by streamlit.py:

streamlit run streamlit.py

Access the local URL provided by Streamlit to interact with the chatbot.

License

This project is licensed under the MIT License.

Additional Information

  • Ensure you have set up proper cloud credentials as outlined in Google Cloud’s documentation.
  • For more details on Langchain integration with Google Vertex AI, refer to Langchain’s official docs.

Enjoy exploring Google Cloud Certifications with gemini-langchain-rag!

Author