This commit 608a0fd removed a consistency check in DataSetBlock which was asserting that the number of angles wasn't less than the number of projections. This turned out to not always be true (360 data which has been stitched together will have half the original number of angles, so the number of angles will be less than the number of projections in general for 360 pipelines).
However, it ma be sensible to have some kind of consistency check between the number of angles and the data, given that the angles held in AuxiliaryData is modifiable via a setter defined on DataSetBlock.
Alternatively, perhaps the angles numpy array could be locked/not writeable by default, to reduce the possibility of the number of angles going out of sync with the number of projections (by limiting the situations in which the angles array is modifiable).
This commit 608a0fd removed a consistency check in
DataSetBlockwhich was asserting that the number of angles wasn't less than the number of projections. This turned out to not always be true (360 data which has been stitched together will have half the original number of angles, so the number of angles will be less than the number of projections in general for 360 pipelines).However, it ma be sensible to have some kind of consistency check between the number of angles and the data, given that the angles held in
AuxiliaryDatais modifiable via a setter defined onDataSetBlock.Alternatively, perhaps the angles numpy array could be locked/not writeable by default, to reduce the possibility of the number of angles going out of sync with the number of projections (by limiting the situations in which the angles array is modifiable).