Fixtures and Mocks of the jupyterhub package

The MockHub Class

It is a subclass of JupyterHub that allows for easier testing of the JupyterHub class and other components that require a running instance of JupyterHub. It does this by providing mock implementations of specific functions of a JupyterHub instance, sometimes modified for testing purposes. For example, the init_db function initializes the MockHub database by first resetting it to a clean state before running tests, the initialize function initializes the MockHub application and adds an initial user to the database to run tests against, the stop function stops the MockHub application and performs any necessary cleanup of the testing environment etc

Example

The conftest module of the DockerSpawner repository imports and uses the JupyterHub MockHub class. This can be seen here.

It sets the hub_ip and hub_connect_ip attributes of the mocked instance. These attributes specify the IP addresses that the mock instance of JupyterHub will be listening in on and connecting to.

The App Fixture

The app fixture is used to provide a lightweight, running JupyterHub application for testing. It is responsible for starting the mock JupyterHub application instance, the MockHub and stopping it after all tests in the module have been completed.

Example

The conftest module of the DockerSpawner repository imports and uses the JupyterHub app fixture. This can be seen here.

It configures the mocked app instance to use the DockerSpawner class and its sub-classes: SwarmSpawner and SystemUserSpawner, to spawn single-user servers.

The MockHub class and the app fixture of the JupyterHub repository and the fixtures they rely on i.e: event_loop, io_loop, ssl_tmpdir, should be adapted for use in the JupyterHub Pytest Plugin to enable for testing of the DockerSpawner and other JupyterHub component implementations without having to import them from the JupyterHub repository.