This is a custom REST API that gets school information from the web, handles encryption with the local keychain, and interacts with firebase.
Languages: Python, Shell
DB: Firebase, Postgresql
API: FastAPI in python
Scraping: Selenium, BS4, Requests, JSON parsing
Scraping Drivers: Chrome (Headless), Firefox
Encryption: RSA
Deployment Strategy: Docker (Images built on Linux)
Supported: Veracross (Schedule + Events), Schoology, Flik
In-Development:
Planned Support: Google Classroom, Showbie
- With docker and docker compose installed run the command
sh docker.sh
- End server with
docker-compose down
- Server can also be run through (ensure venv and dependencies working/installed)
sh run.sh
- For Running options:
sh run.sh -h
- Found in scripts folder:
cd scripts
- Reinstall dependencies quickly
sh dependencies.sh
- Clear pycs
sh clear_pycs.sh
- Generate a new RSA key
sh newkey.sh
(All other scripts are not for direct user use)
To view documentation: run the server and visit /docs
- /scrape (scrape a user's information and save to firebase)
- /link (verify user's information and save/encrypt to firebase/local keychain)
- /docs (official documentation) = /ping (ping the server)
-
M1 chip not supported with Firebase functionality (workaround: run through a docker image built on linux instead)
- Ensure Docker, Docker Compose, and Docker Desktop are all installed
- Run Docker Desktop
- Run: (This will execute the docker file which will handle dependencies and file management)
docker-compose build- Run: (Launches the server by running the docker image and assigns the port)
docker-compose up- Navigate to localhost at port 80 (http://localhost:80)
-
psycopg2-binary installation on M1 Workarounds: brianmario/mysql2#795 (comment) psycopg/psycopg2#1286 (comment)
-
deployment Since, server now deploys on DO Container (built on heroku), m1 is not supported (use buildx when building docker images to ensure its' not build on m1 architecture)