Create an intake catalog for data

[stevehadd]

create an intake catalog to describe the various stages of data.

[stevehadd]

This could be a good way to make azureML code portable, by having an intake driver for azureml, which just makes use of AzML datasets/datastores. This makes permission much easier (everyone in workspace can see workspace datasets/datastores), but means the code that reads the data will be more prtable if it uses intake as the interface to the data.

This should be a fairly lightweight driver. More info on creating drivers:

• https://intake.readthedocs.io/en/latest/making-plugins.html
• https://github.com/intake/intake-examples
• https://martindurant.github.io/blog/flurry-of-drivers/
• https://github.com/intake/intake-xarray

lab example

• https://github.com/informatics-lab/intake-iris