This project is a Chatbot AI designed for FPT University's FSB Master Software Engineering program. It utilizes a Large Language Model (LLM) named Gemini, integrated with Retrieval-Augmented Generation (RAG) techniques to enhance user interaction and provide accurate, context-aware responses. The chatbot is built using the Langchain framework and is served via a Flask web application.
- Generative AI Chatbot: Leverages Gemini LLM to understand and generate human-like text responses.
- RAG Integration: Enhances responses by retrieving relevant information from external data sources.
- Meta Messenger Integration: Allows users to interact with the chatbot through Meta Messenger.
- LLM: Gemini
- Framework: Langchain
- Library: Flask
- Dependencies:
requests
langchain-google-genai
langchain
langchain-community
pypdf
pdf2image
tiktoken
unstructured[local-inference]
sentence_transformers
chromadb
- Python 3.x
- Meta Messenger Developer Account
- Google API Key
- Vercel KV (Key-Value Store) Account
Before running the app, ensure that the following environment variables are set:
export VERIFY_TOKEN=<your_verify_token>
export PAGE_ACCESS_TOKEN=<your_page_access_token>
export GOOGLE_API_KEY=<your_google_api_key>
export VERCEL_KV_URL=<your_vercel_kv_url>
export VERCEL_KV_TOKEN=<your_vercel_kv_token>
export PORT=<your_port>
export DEBUG=<your_debug_mode>
Clone the repository:
git clone https://github.com/your-repo/chatbot-ai-fpt-fsb.git
cd chatbot-ai-fpt-fsb
Install dependencies:
pip install -r requirements.txt
Run the following command to start the Flask application:
python3 index.py
Ensure your Flask application is running. Open Meta Messenger and start a conversation with your chatbot. Interact with the chatbot to test its response capabilities and contextual accuracy.
We welcome contributions to enhance the functionality and performance of this chatbot. Please fork the repository and submit a pull request with your proposed changes.
This project is licensed under the MIT License. See the LICENSE file for more details.
For any queries or support, please contact the project maintainer at [chaubao.work@gmail.com].