open_dataset improvements

[victoria-cherkas]

• Allow filtering of dataset via a variable name as argument in iconarray.open_dataset
• Allow passing standard xarray.open_dataset arguments through iconarray.open_dataset()

[victoria-cherkas]

launch jenkins

[victoria-cherkas]

launch jenkins

[victoria-cherkas]

launch jenkins

[victoria-cherkas]

https://github.com/C2SM/icon-vis/pull/36 PR for icon-vis using this branch of iconarray.

https://jenkins-mch.cscs.ch/job/IconVis/job/Iconvis_PR/10/

[victoria-cherkas]

for this file grib= '/scratch/vcherkas/iconarray/data/example_data/grib/lfff00010000_edgeplots':

Elapsed time GRIB (unfiltered): ds = iconarray.open_dataset(grib) New open_dataset(): 0:02:13 Old open_dataset(): 0:02:39

So not to worry about performance loss, it is clearly just a large file (that should probably have been cropped before uploading to the example data server).

[clairemerker]

Great! I think we should make it a task to reduce the test data size. We can for example crop the data to a smaller domain

[victoria-cherkas]

Great! I think we should make it a task to reduce the test data size. We can for example crop the data to a smaller domain

https://github.com/C2SM/iconarray/issues/57 Just added issue for it :)

[victoria-cherkas]

Ok thanks for the review @clairemerker we can wait for additional test, thanks for that

[clairemerker]

launch jenkins

[clairemerker]

launch jenkins

[clairemerker]

@victoria-cherkas did you already test this PR with icon-vis?

[victoria-cherkas]

@victoria-cherkas did you already test this PR with icon-vis?

Yes its passing on icon-vis :)