This project involves a backend service primarily written in Go that integrates with the CanvasAPI and utilizes an AI model (from OpenAI) to enhance student productivity. The service fetches assignments and syllabi from Canvas courses and uses AI to prioritize tasks based on deadlines and workload. The Go backend will employ a SQLite database along with Redis for caching mechanisms.

Tech Stack

Go: our programming language of choice
Gin: http Web Go Framework, simplifies routing
sqlc: to help generate DB boilerplate from a SQL schema
Redis: utilized to cache DB queries
OpenAI: AI to prioritize certain assignments/projects

Project Structure

This project mainly consists of the pkgs/ and db/ directories:

The pkgs/ directory declares all the libraries that are used across the backend service including canvas/, redis/, and utils/. The canvas/ library contains most of the logic that interacts with the CanvasAPI, specifically fetching resources (such as assignments and syllabus) across multiple Canvas courses. It utilizes the user's Canvas API key to retrieve those resources. The redis/ package contains the functionality to interact with the Redis cache (improving performance of various queries).
The db/ directory combines the logic for generating SQL schemas and queries. It leverages the Go sqlc libary for generating Go structs used across the codebase.

Routing

The following routes are described as follows:

/:courseID/assignments/:assignmentID: Supports reading and deleting individual assignments based on their ID
/assignments: Handles data insertion into the SQLite database
/all-assignments: Retrieves all assignments from the database
/syllabus (Concept): A POST request that leverages the OpenAI model to summarize the syllabus and store that in the database

Caching

As mentioned before, we use Redis for caching. We cache any repeated queries to the database to improve performance. Below is a performance comparison between a query for an individual assignment with and without Redis caching:

Redis Performance

Future Works

Most of the AI implementation is missing as of our latest update (12/10/24). We would like to utilize the AI to prioritize assignments based on estimated time and difficulty to help students organize their workflow.
The service could also leverage a proper authentication system, not just utilizing the Canvas API as the way to identify each user.
We could create a frontend to provide users with a UI to utilize the tool.

Setup

Install Go
Get your Canvas token from your respective school Canvas site https://{somedomainhere}.instructure.com/profile/settings
Setup your .env file (see .env.example)

Optionally, install air for hot reloading support. No need to do this step if you're installing through Devbox.

How to Run

go run main.go

With air:

air

Updating SQL files with sqlc

If you've made changes to any .sql files, run go gen ./... to generate .sql.go files, which contain generated Go code for directly interacting with the database.