Collaborators: Addison Thurston, Henry Beblo, Henry Morgan, Carter DeLargy
Creates assignments/quizzes on Canvas based off of images/text prompts and publish/unpublish/delete exisiting Canvas assignments.
If an image is being entered, it's first taken and transformed via our Geometric Filter so that it's prepared to be read for AI.
Then, the fitted photo is passed through to the Google Cloud Vision API.
This extracted text is then processed and formatted (with parameters separated with '|') with the Gemini Pro 2.5 API.
This formatted text is then parsed and passed through to the Canvas API, where it can be published on Canvas.
API keys for 'Canvas', 'Google Cloud Vision', and a Gemini model are needed (we used 'Gemini Pro 2.5' because it had the largest free quota).