You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While thinking about other packages that could use this functionality I had the rather later realization that trollimage is a really good candidate for this. xarray has a recommended way of extending xarray objects: http://xarray.pydata.org/en/stable/internals.html#extending-xarray
If we convert the XRImage logic (stretch methods, saving methods, etc) in to an accessor class then you could have all of this functionality available from my_data_arr.xrimage.rio_save(...). Using this with enhancements could really help with satpy's user experience with creating their own enhancements and making things more logical.
This also mostly satisfies my feeling that XRImage should be replaced by just using xarray objects since I didn't feel it completely deserved its own package outside of satpy. Main remaining issue is that satpy users who want to add enhancements to the official satpy release either add them to satpy or to trollimage. At least that is the way it is done now. I suppose a satpy accessor could access the trollimage accessor, but then we're back at the beginning with me wondering if trollimage helps development or hurts it by being its own package.
The text was updated successfully, but these errors were encountered:
Problem description
While thinking about other packages that could use this functionality I had the rather later realization that trollimage is a really good candidate for this. xarray has a recommended way of extending xarray objects: http://xarray.pydata.org/en/stable/internals.html#extending-xarray
If we convert the XRImage logic (stretch methods, saving methods, etc) in to an accessor class then you could have all of this functionality available from
my_data_arr.xrimage.rio_save(...)
. Using this with enhancements could really help with satpy's user experience with creating their own enhancements and making things more logical.This also mostly satisfies my feeling that XRImage should be replaced by just using xarray objects since I didn't feel it completely deserved its own package outside of satpy. Main remaining issue is that satpy users who want to add enhancements to the official satpy release either add them to satpy or to trollimage. At least that is the way it is done now. I suppose a satpy accessor could access the trollimage accessor, but then we're back at the beginning with me wondering if trollimage helps development or hurts it by being its own package.
The text was updated successfully, but these errors were encountered: