Hospital Chatbot

A full-stack hospital assistant project built with LangChain-style orchestration and graph-backed retrieval.

Repository Structure

• chatbot_api/
  • FastAPI-based backend for chatbot and agent endpoints
  • src/agents/hospital_rag_agent.py
  • src/chains/ for domain chains
  • src/tools/ for utility tools
• chatbot_frontend/
  • Frontend service for chat interaction
• hospital_neo4j_etl/
  • Data pipeline/service for loading hospital data into Neo4j
• data/
  • Source CSV datasets (hospitals, physicians, reviews, visits, etc.)
• tests/
  • Async and sync agent request tests
• docker-compose.yml
  • Local multi-service orchestration

Tech Stack

• Python 3.11+
• FastAPI
• LangChain ecosystem
• Neo4j
• Docker and Docker Compose

Quick Start

1. Configure environment variables in local .env files as needed.
2. Build and start services:

docker compose up --build

3. Access services after startup:

• API: as configured by chatbot_api service
• Frontend: as configured by chatbot_frontend service

Local Development

Backend:

cd chatbot_api
# install dependencies from pyproject.toml with your preferred tool

Frontend:

cd chatbot_frontend
# install dependencies from pyproject.toml with your preferred tool

ETL:

cd hospital_neo4j_etl
# run load scripts after configuring Neo4j connection

Notes

• Keep secrets out of source control.
• The project .gitignore excludes virtual environments and local .env files.