Probably Genetic - Coding Challenge

See challenge.md for the challenge description.

Solution

Hi! You can see the solution here. Click on symptoms to see a page which allows you to search for symptoms, select them, and then compute the most likely diagnosis. Click on disorders to see a flat list of all the disorders available in the database.

Edge Cases and Future Directions

• Add provisions around data quality issues
• Add more data sources (e.g. SNOMED)
• In the UI, make it more apparent why a particular disease was in the returned list, and give a probability estimate
• Use Machine Learning to build a better algorithm for computing the most likely diagnosis
• Use a proper database more suitable for large applications (Postgres, MySQL, etc.)
• Pull more descriptive data into the database from a third party API for each disorder
• Make the disorders and symptoms more searchable - use a search bar which leverages NLP for more approximate query matching
• Use caching to improve load times, especially since the data is relatively static
• Fix the favicon. Haven't figured out how to get it to refresh.
• Use a single docker-compose file for setup and deployment

Run Unit Tests

python3 -m pytest

Local Development

If using Docker Compose (preferred)

Prerequisites

• sudo apt install docker-compose ; Ensure you have docker-compose installed.

Open two separate terminals, one to run the server and one to run the client.

Run the client

cd disease-finder/client && docker-compose up

Run the server

cd disease-finder/server && docker-compose up

If using local development (for Ubuntu 22.04)

Prerequisites

• node --version > v16.15.1
• yarn --version > v1.22.15
• sudo apt install python3.10-venv # Ensure you have python3.10-venv installed to properly setup the virtual environment.

Run the client

1. yarn install
2. yarn start

Run the server

1. chmod +x run.bash
2. ./run.bash

If you're testing in this way and need to re-start the server, please comment out the db initialization step in run.bash to save yourself 30 minutes 😊.

Stack

• React - frontend
• Flask - API
• sqlite - database
• SQLAlchemy - ORM