Cloud development environments

Cloud development environments are remote virtual machines that allow you to code, write, and otherwise run computational tasks remotely, often using a browser-based interface. Examples of such environments are Gitpod and GitHub Codespaces. Here are a few situations when they can be helpful:

• when the development set up for your project is too complex for newcomers, or requires a particular OS or machine type;
• when the time required to do a development set up would be impractical (for example, when running sprints);
• when you need to review a Pull Request without cloning the repository and setting up the full development environment.

You can have a default cloud development environment that reproduces the machine type you want to work on, or you can configure a pre-set environment that comes ready-to-use. For each of these tools, there are free plans including a number of computational hours, memory, or disk-usage available.

Here are the main such tools available today and a few examples of how they can be set up.

Gitpod

Gitpod can be used from several different code repository services including GitHub, GitLab, and Bitbucket. To use it, you can pre-pend the repository URL with the string gitpod.io/#, or you can install a browser extension to add a Gitpod button to your GitHub interface.

Gitpod also supports a number of popular IDEs so you can have a personalized experience.

See the Gitpod documentation for more information.

Example: pandas

Check the pandas documentation on using Gitpod for development.

To configure a pre-set development environment on Gitpod, you can either build a custom docker image with the tools and dependencies you need, or use a .gitpod.yml file, which will live on the root of your repository. You can, for example, set the system dependencies necessary to build and test your software project, or pre-install tools and even IDE extensions or plugins you’d like your contributors to use.

GitHub Codespaces

GitHub Codespaces is, as the name suggests, limited to use with GitHub. Similarly to Gitpod, it can also be set up with appropriate customizations by the use of a devcontainer.json file.

In addition, you can also select from a list of pre-built features to add to your base image so you don’t have to build the custom docker image yourself.

Example: NumPy

Check the configuration for Codespaces on the NumPy repository.