Simon Heybrock
Simon Heybrock
Yes it does, you are right. So maybe the only missing bit is the extra lib for the allocation proxy?
@MridulS the point here is that this is not available since there is an apparently optional extra lib as part of TBB, which is not installed with `pip`.
> Just in case this is useful we do use different TBBs. In the pypi wheel it comes from conan ( > > https://github.com/scipp/scipp/blob/2bc44c832daf1a1362915e02ae740bc8655dda51/lib/cmake/scipp-conan.cmake#L31 > ) and the conda one...
`sc.scalar(1, unit='s')` is a time-difference, why do you expect that you could cast it to a (date)time?
Well, it works for Numpy, but I am not a fan (not every user knows what epoch is): ``` import numpy as np datetimes = np.array(['2024-06-01T12:00:00', '2024-06-01T13:00:00', '2024-06-01T14:00:00'], dtype='datetime64') (datetimes...
Probably. In the meantime, note also [DataGroup.apply](https://scipp.github.io/generated/classes/scipp.DataGroup.html#scipp.DataGroup.apply).
Using `linspace` for vectors is indeed not supported. As a workaround, can you create a linspace for a scale factor? ```python diagonal = start + linspace('diag', 0, 1, num=10) *...
No, I think it probably makes sense! It was just not thought of in the implementation.