-Benjamin Ward
-Andrew Park
-Jeremy Gan
-Farbod Nematifar
-Cooper Webb
This is a web application that assesses novice-level code comprehension using generative AI. After selecting a question, students explain what sample code does in plain English. This text is then transformed into code by the gen AI and tested against test-cases to determine accuracy.
The frontend is organized around src/components. Each section of the app has its own folder of related components.
To navigate through the frontend was use React's routing system (see App.tsx)
Backend has four main sections: controllers, db/models, llm, routes (we left the test section in, but most of these tests have been migrated to the frontend - see testing section below)
controllers/routes: controllers handles incoming requests and processes data logic. Routes handles the actual endpoint/routes to connect our frontend to our backend db/models: db contains our sqlite database and db creation scripts and models is how we interact with the database llm: this includes our setup and api to run ollama tests: this includes our old testing suite before we switched to frontend
docker compose up -dNote: This will also build the images if not already built
To view the associated web pages navigate to localhost:4000
Note: any modifications made to TypeScript source files will be automatically detected and applied without the need to manually restart the service.
docker compose downOnce docker is running, all tests can be viewed on the browser using this URL: http://localhost:4000/test