Open dvanic opened 4 years ago
SampleDB directs the user to a URL depending on the SAMPLEDB_JUPYTERHUB_URL
and the notebook settings in the action schema, containing selected metadata from the measurement.
What happens then is largely dependent on the configuration of the JupyterHub (e.g. which spawner is used, which authenticator is used, which volumes are mounted in the individual notebook server containers in case of a kubespawner for example) and storage system. As a result, it is difficult to generalize the webapp that inserts the metadata into a notebook and redirects the user to that notebook.
I'll document the general process and try to produce a small example for such a webapp, working with local files. However each use case will need alterations, e.g. if the JupyterHub mounts user's directories from one of several NFS servers, the webapp should write the notebook not to a local file but to the correct path on that NFS server, etc.
I've uploaded a Flask blueprint for building a notebook templating server. I've also added a page on JupyterHub support to the new Administrator guide, which goes through the steps of what's necessary for interfacing with a JupyterHub instance.
Is there documentation for interfacing to a JupyterHub instance to visualise the results of a specific set of measurements, for example?