Data wranglers should be able to create IPython ( now Jupyter ) notebooks in a bundle, and have the notebook stored as a file in the Bundle. Then, add a 'bambry notebook' command to start a Jupyter notebook.
( It looks like the Model in the ContentManager API can be directly mapped to the File record in the Ambry bundle. )
These files should all be synced to and from the file system using the bambry sync command, so notebooks can be stored in Github with other bundle files.
The bambry notebook command starts a Jupyter server running on the bundle, configured to read and write notebooks to the bundle.
The bambry notebook should have a --docker command that runs the jupyter process inanother container. There are several implementations for jupyter in docker, but also see this note about pid reaping if you create your own.
Data wranglers should be able to create IPython ( now Jupyter ) notebooks in a bundle, and have the notebook stored as a file in the Bundle. Then, add a 'bambry notebook' command to start a Jupyter notebook.
First, implement a [ContentManager)[http://jupyter-notebook.readthedocs.org/en/latest/extending/contents.html] that will read and write notebooks and text files using the File records in a bundle. The context manager does not need to support directories, but it may be nice if it does.
( It looks like the Model in the ContentManager API can be directly mapped to the File record in the Ambry bundle. )
These files should all be synced to and from the file system using the
bambry synccommand, so notebooks can be stored in Github with other bundle files.The
bambry notebookcommand starts a Jupyter server running on the bundle, configured to read and write notebooks to the bundle.The
bambry notebookshould have a--dockercommand that runs the jupyter process inanother container. There are several implementations for jupyter in docker, but also see this note about pid reaping if you create your own.