Omega was (is?) wrong for multi-turn 360 degree scans

[jonwright]

In dataset we seem to have:

@property
def pk4d(self):
    if self._pk4d is None:
        self._pk4d = self.peaks_table.pk2dmerge(self.omega, self.dty)
    return self._pk4d

For a multi-turn example we appear to need:

self.peaks_table.pk2dmerge(self.omega_for_bins, self.dty)

Or something similar? Where omega_for_bins is omega%360. @haixing0a I think you tested this recently?

I see the problem in /data/visitor/ma5608/id11/20250212/Wenxi/PROCESSED_DATA/NS_Ti_1/NS_Ti_1_z600_3/tomo_1_index_JWtest.ipynb for Wenxi.

I have a feeling @jadball fixed this in git master recently. Perhaps I should just update but I am afraid to damage his other notebooks...

[jadball]

I believe this is still a problem. The only fix I did was in ds.sinohist - cc2c759

[jadball]

Before I forget, we should check the omega that ends up in the cf from frelon_segmenter...

[jonwright]

Seems @haixing0a had the rotation motor owisRz_cen360. This overcomes the problem, but not for our old difftomo360 scans.

I guess we need something like (angle-cut)%360+cut and perhaps define the 'cut-point' for scans that are -90 to 90 to be at 180 and not zero.

[jadball]

Fixed in #439 ? @jonwright

[jonwright]

In theory, yes, I it is fixed. I have not managed to make and run any test case yet. That is probably a separate issue, so we can close this one.