health-rec

Overview

Welcome to the health-rec project! This project is designed to develop a recommendation system for health and community services.

The project is built using a microservices architecture. It has the following components:

• Next.js frontend
• FastAPI backend
• Chroma vector database
• OpenAI for embedding and recommendation

Development

Prerequisites

• Docker
• Docker Compose
• Node.js
• Python
• pip
• npm

Setup

Clone the repository

Add API keys

In the .env.development file in the root of the project, set the following environment variables:

• OPENAI_API_KEY
• 211_API_KEY
• MAPBOX_API_KEY
• DATA_DIR
• COLLECTION_NAME

Make sure to set appropriate values for these variables. The CHROMA_HOST, CHROMA_PORT, and COLLECTION_NAME are already defined in the file, but you may need to adjust their values if necessary.

Install dependencies in a virtual environment for the backend

cd health_rec
python3 -m venv .venv
source .venv/bin/activate
poetry install --with test

Now you can run pre-commit checks and tests:

pre-commit run --all-files

Run the services

For development, we use docker compose to run the services.

To run only the backend and chroma db services, run the following command:

docker compose --env-file .env.development -f docker-compose.dev.yml up

To run the frontend and backend, run the following command:

docker compose --env-file .env.development --profile frontend -f docker-compose.dev.yml up

Download data

For example, to download GTA data, run the following command:

python scripts/download_data.py --api-key $YOUR_211_API_KEY --dataset on --is-gta --data-dir /mnt/data/211

To download Ontario-wide data, run the following command:

python scripts/download_data.py --api-key $YOUR_211_API_KEY --dataset on --data-dir /mnt/data/211

Upload data and embeddings

First we use an interactive container:

docker run -it --network health-rec_app-network -v <data_dir_with_json_files>:/data -v `pwd`:/workspace -w /workspace vectorinstitute/health-rec:backend-dev-latest bash

Then we can run the following commands to upload the data to the vector database:

python3 health_rec/manage_data.py create --name <collection_name>
OPENAI_API_KEY=$YOUR_OPENAI_API_KEY python3 health_rec/manage_data.py load --name <collection_name> --data_dir /data --load_embeddings
python3 health_rec/manage_data.py list

Careful while loading embeddings, it uses the OpenAI API, and hence make sure the data you want to use is correct. Test with a small amount of data first.

Navigate to the UI on the browser

http://localhost:<frontend_port>

In case, you wish to update frontend dependencies, run the following commands in the ui directory:

npm install <package_name>