Open forman opened 8 years ago
Similar comment as for #81. Wasn't this supposed to be a visualization issue only for this iteration? If it's possible in the light of uc11.
We should probably go for the 'usual suspect' when implementing this:
https://pypi.python.org/pypi/GDAL
But there is a need for discussion on how this affects the overall 'Cate' software. For example, what happens with the CDM? Do we have a 'default' projection that we expect the datasets to have? If a dataset is converted into polar stereographic coordinates, can the GUI handle this? Etc.
Relevant CF Conventions: http://cfconventions.org/Data/cf-conventions/cf-conventions-1.7/cf-conventions.html#grid-mappings-and-projections
Or consider using rasterio in combination with its xarray backend:
However that introduces a new dependency. I also belive rasterio uses a different GDAL binding, possibly adding another gigabyte package to our distribution. Will check...
Not necessary for UC11, re-projection is done as part of plotting.
Provide an reprojection operation to a create new, reprojected dataset from an existing one:
Especially the polar stereographic projection shall be considered early in order to support remapping of gridded ECVs (Ozone, Clouds, Aerolsols, ...) to the ones that use Shapefiles (Glaciers, Icesheets).
Note that this operation is not necessarily required to to create the scaleable maps requested by #79. Instead this operation is likely to be used to collocate datasets that don't share a common grid definition.