Allow `X` in `Pipeline` steps to be not an ElmStore or xarray.Dataset

ContinuumIO / elm

Phase I & part of Phase II of NASA SBIR - Parallel Machine Learning on Satellite Data

http://ensemble-learning-models.readthedocs.io

43 stars 27 forks source link

Allow `X` in `Pipeline` steps to be not an ElmStore or xarray.Dataset #100

Open PeterDSteinberg opened 7 years ago

PeterDSteinberg commented 7 years ago

There are two issues:

[ ] one is the usage of es.get_shared_canvas() which prevents even the usage of xarray.Dataset in some cases. See the code es.get_shared_canvas() es is X and get_shared_canvas is only a method of ElmStores. This issue can be fixed by making get_shared_canvas a standalone func that works on xarray.Dataset and ElmStore.
[ ] the other issue is that we may want numpy arrays or dataframes to be able to move through the steps of the Pipeline

PeterDSteinberg commented 7 years ago

When addressing data structure flexibility as part of #149 and earthio issue 12, note the bug mentioned in the first bullet point of the preceding comment. Options may be:

Use get_shared_canvas() as a function rather than method of a class
Make get_shared_canvas a method in each of the backend xarray classes that @gbrener develops as part of earthio issue 12