Biased cell distributions with xgandalf
This cell_explorer screenshot comes from running the tutorial dataset (A2a receptor data from SwissFEL, Jungfrau detector, orthorhombic C cell) using xgandalf-cell with --no-refine
:
The cell matching routine (compare_reindexed_cell_parameters
) should produce an unbiased cell no matter what the indexing routine prefers. However, this is obviously not working correctly.
Adding refinement normalises the distribution somewhat, but there is still some bias evident:
I reproduced the issue in a simulation, so it's not a geometry problem.