@torrinba I think we might have encountered a similar situation with the MSE. @jmcclena might remember.
The approach there was to do set the COCOS transformation to be None for those variables, and handle the transformation of the individual elements by hand.

@orso82 That's what I was thinking at first too. But it was a bit easier for MSE, as each row had a different cocos. So we were able to set it with an array of cocoss

_cocos_signals['mse.channel.:.active_spatial_resolution.:.geometric_coefficients']=[None, None, 'TOR', None, None, None, 'TOR', None, None]#[DEL?]# -1.000000 # geometric_coefficients

@jmcclena I had forgotten about the array of COCOS!

@torrinba please take a look if you can use the same functionality (or perhaps you can extend it to work with matrixes)

@orso82 the challenge with the em_coupling matrices is not that the rows or columns need to have different COCOs, it's that each dimension does. For example using the matrix A to compute b = A x, the vectors b and x have different COCOs so A needs to have a combination of those. Does OMAS support this?

@torrinba The logic for handling the array can be found here:

Perhaps we can add logic to split by *, and perform all transforms in the split for em_coupling?

Something like 1354 turns to:

norm = 1 
for transform in transform.split('*'):
    norm *= omas_physics.cocos_transform(self.cocosio, self.cocos)[transform]

Stale issue message

This issue has not seen any activity in the past 60 days. It is now marked as stale and will be closed in 7 days if no further activity is registered.