andredias / codelab Goto Github PK

The Code Lab project allows you to program directly from the browser, without any other external dependency. It is interesting for cases where you want to do a quick experiment but there is no computer with the language installed and configured nearby.

This project was originally made as a proof of concept and also to help my students use different programming languages without needing to install anything on their machines.

1. run make dev to start the backend and frontend servers
2. open https://localhost in your browser

To see the backend API documentation, open https://localhost/api/docs in your browser.

Running code from an unknown source poses security risks because a malicious piece of code can try to take control of the system, access restricted areas and data or misuse available resources (memory, processes, disk, etc.). The way to mitigate those risks is to run the code in a sandbox, which is a type of virtualization of a restricted and controlled environment [1].

In the Code Lab, the sandbox is based on a container Docker called Codebox, which contains all the languages, libraries and tools offered by Code Lab, but that runs with limitations on user permissions, time, memory and network access. Even if there is a security breach, its effects will be contained and then eliminated when the container is destroyed.

The models used to exchange information are specified in the file backend/app/models.py.

Codebox is the most important part, but it doesn't work alone: It all starts with the web application, which follows the SPA standard (Single Page Application) and is built with Vue.js. The backend is hosted in a droplet of DigitalOcean. The application is served through a reverse proxy by Caddy and a Hypercorn server, which uses the ASGI (Asynchronous Server Gateway Interface) standard, and serves the application built with FastAPI.

The execution of the projects is not immediate. First, a check is made in the cache (Redis) and only if it is not there, the project is executed in Codebox.

In order to contribute to the project or to try it out from your local machine, read CONTRIBUTING.rst.