Re-introducing the NDDataArray? #4461
Replies: 2 comments
-
@registerrier Yes, we could certainly abstract away some more generalized functionality from the What we would need is a So right now I don't see any advantage in doing this, except if it happens in an even more generalized way to handle WCS and HEALPix axes as well. |
Beta Was this translation helpful? Give feedback.
-
Or more specifically: we would require an |
Beta Was this translation helpful? Give feedback.
-
The new data model of CTAO relies on a N-dimension data array concept to represent tabulated functions. It is very similar to what we use as a base class in gammapy.irf, but less specialized.
We used to have a class to represent this concept in Gammapy, the
NDDataArray
but it was removed some time ago. Should we re-introduce it to support more cases?This would be basically the
IRF
base class without the specific methods and functionality required for IRFs. IRF would inherit from it and re-implement their specific features. It would be aQuantity
array with associatedMapAxes
and interpolation, get_by_coord evaluation/integral methods.Opinions @adonath , @maxnoe , @kosack ?
Beta Was this translation helpful? Give feedback.
All reactions