Failing spotfinder test - issue with h5 file assumptions

The test at tests/test_spotfinder.py::test_dispersion_gridscan is failing following the h5 updates PR (https://github.com/DiamondLightSource/fast-feedback-service/pull/86)

Specifically I think this line is a problem considering how I03 write grid scan data:
https://github.com/DiamondLightSource/fast-feedback-service/blob/7f1f16af908b1f583acffe9925fc792b50d829c8/h5read/src/h5read.c#L357

I03 write a nxs files for each grid scan when doing orthogonal grid scans, but the data for both is written to common _000001.h5, _000002.h5 etc files.
So the first grid scan dataset (420 images) points to the first 420 images in _000001.h5, and then the next grid scan points to the following block of images. So we can't make this assertion.