title	emoji	colorFrom	colorTo	sdk	sdk_version	app_file	pinned
GBV Langchain Bot	🐢	green	gray	gradio	3.36.1	app.py	false

GBV in Emergencies Q&A Bot

Overview

This is a question-answering chatbot specialized in Gender-Based Violence (GBV) response in humanitarian settings. The bot uses LangChain and OpenAI's language models to provide accurate information by drawing from a specialized knowledge base while supplementing answers with Wikipedia data when necessary.

Features

• Interactive chat interface built with Gradio
• Retrieval-augmented generation using a vector database (Chroma)
• Conversational memory with summarization capability
• Multiple knowledge sources (specialized GBV documents + Wikipedia)
• Streaming responses for better user experience

Architecture

The application uses several key components:

1. Vector Database: Chroma DB storing embeddings of GBV-related documents
2. Language Model: OpenAI's GPT models (configurable between GPT-3.5-Turbo and GPT-4)
3. Retrieval System: LangChain's RetrievalQAWithSourcesChain for accurate information retrieval
4. Agent Framework: Conversational ReAct agent to intelligently route between knowledge sources
5. Memory System: ConversationSummaryBufferMemory to maintain context in long conversations

Usage

Users can ask questions related to GBV in humanitarian settings. Example questions:

• What are the GBV guiding principles?
• Which UN agency leads the GBV response in emergencies?
• How can we engage men and boys in GBV prevention and response?
• Please outline a strategy to minimize GBV risks in a temporary settlement
• What is the integration factor between GBV and SRH?

Setup and Deployment

1. Ensure you have Python installed
2. Install dependencies: pip install -r requirements.txt
3. Set up your OpenAI API key as an environment variable: OPENAI_TOKEN
4. Run the application: python app.py
5. Access the interface through your browser

Deployment on Hugging Face Spaces

This application is configured for deployment on Hugging Face Spaces using the Gradio SDK. The configuration is defined in the frontmatter of this README.

License

This project is intended for educational and humanitarian purposes. Please use responsibly.