Overview

Oncora has a concept of case objects that represent a single course of radiation treatment. Users may associate one or more prescription objects with a given case.

This project contains an API to store and retrieve those prescription objects, and to retrieve risk_predictions for a case based on its latest prescription. These are the available endpoints:

Endpoint	Method	Description	Code
/prescription/<case_id>	GET	Get the latest prescription for a given case	resources/prescription.py
/prescription/<case_id>	POST	Submit a prescription to associate with a case	resources/prescription.py
/prediction/<case_id>	GET	Retrieve risk predictions for a case based on its most recent prescription	resources/risk_prediction.py

All data is stored in MongoDB. Case documents are stored in the case collection, and prescription documents are stored in the prescription collection.

Goals

1. When a user submits a prescription to be saved we want to go ahead and generate risk predictions for that prescription, store them in the database along with the prescription and case used to generate them, and include the predictions in the response to the client.
   • The predictions should be stored in a collection called predictions
   • The predictions should be returned from the API as an attribute named risk_prediction
2. Add unit tests to cover any modified code, or any other appropriate test cases you happen to find
   • Refactor the existing code as much as you like to accomplish this

Setup

1. Install pipenv if it isn't already installed.
2. Clone this repo
3. Create virtual environment and install dependencies: PIPENV_VENV_IN_PROJECT=1 pipenv install -d
   • Pipenv will create a virtual environment in the project directory, so when you activate the virtual environment using pipenv shell, all the dependencies for the project are readily available in an isolated environment.
4. Start the API: docker-compose up -d
   • This will run a REST API listening on port 5000 and a MongoDB instance listening on port 27017.
5. Add some initial seed data to the database: docker-compose exec api python -m seed_data
   • Use this same command anytime to replace any data added to the database during development/testing with the seed data
6. Open the project in your editor of choice
   • Note that this project already includes some configuration to simplify debugging in VS Code, but you're welcome to use whatever editor/environment you like

Testing

1. The automated tests can be run with pytest: pipenv run pytest
   • You should initially see one passing and one failing test
   • Feel free to use any testing tools in your editor, too. If using VS Code, this should already work
2. Some sample api requests are included in two formats:
   1. sample_requests.curl.sh has a few curl commands
   2. sample_requests.postman_collection.json has the same requests as a collection that you can import into Postman