vyperlang / try.vyperlang.org

source code for hosting the try.vyperlang.org website

Vyper JupyterHub

Hosted at try.vyperlang.org

This project uses Docker to manage its services. The docker-compose.yaml file defines the services that make up your app so they can be run together in an isolated environment.

Docker compose services

The project consists of two main services:

1. notebook-image: This service is built from the notebook.Dockerfile and is used to create a Docker image named titanoboa-notebook. This image is not run as a service but is used by the vypyter service.

2. vypyter: This service is built from the jupyterhub.Dockerfile and depends on the notebook-image service. It exposes port 8000 and uses environment variables defined in the .env file and additional ones defined in the docker-compose.yaml file.

Getting Started

Running locally

To run the project locally, you need to follow these steps:

1. Ensure Docker and Docker Compose are installed on your machine. If not, you can download them from the official Docker website.

2. Clone the project repository to your local machine using Git.

3. Create a GitHub OAuth app and add to a new .env file:

   • GITHUB_CLIENT_ID=...
     GITHUB_CLIENT_SECRET=...`
     GITHUB_CALLBACK_URL=http://localhost:8000/hub/oauth_callback
     

4. You can start the JupyterHub using Docker Compose. Run the command docker compose up to start it in the foreground, or docker compose up --detach to start in the background.

   • This command automatically builds the necessary images if they don't exist.
   • To build them manually, you can run docker compose build.
   • To build and run immediately, you can run docker compose up --build.
   • The FORK and REF environment variables can be used to specify the Titanoboa fork and branch/commit to use. The default is to use the master branch of the titanoboa repository.
   • The PORT environment variable can be used to specify the port to expose. The default is 8000.

5. The vypyter service should now be running and accessible at http://localhost:8000.

Deploying to a server

If you want to deploy the JupyterHub to a server, please read the following article: Deploy your own Vyper JupyterLab notebook server .

GitHub Actions

In order to run the GitHub Actions, you need to add the following secrets to your repository:

• AUTH_SSH_USER: The username used to authenticate with the server.
• AUTH_SSH_KEY: The private SSH key used to authenticate with the server.

Also, you need to add the following (environment) action variables to your repository:

• HOST: The hostname of the server.
• HOST_PUBLIC_KEY: The public SSH key used to authenticate with the server. This is used to verify the server's identity when connecting via SSH. You can find the public key by calling ssh-keyscan.
• REPO_DIR: The directory where the repository is be cloned on the server. For example, ~/try.vyperlang.org.
• BOA_COMMIT_ISH: The commit-ish to checkout on the server. By default, master.
• PORT: The port to expose the JupyterHub on the server. By default, 8000.
• JUPYTERHUB_IMAGE_NAME: Optional environment variable to specify the name of the JupyterHub image. By default, jupyterhub.
• NOTEBOOK_IMAGE_NAME: Optional environment variable to specify the name of the notebook image. By default, titanoboa-notebook.