Open benjaminhwilliams opened 3 years ago
Don't forget about cbf_writer. I just used it to dump 100 cbfs from a SwissFEL dataset (NeXus hdf5). I validated it by using dials.show to verify I got the same geometry understood by FormatNexus as by FormatCBFFull.
Caveats: it dumps full, multipanel cbfs, which are only useful if the downstream application can read them. It looks like dlsnxs2cbf
dumps mini cbfs. cbf_writer also doesn't write goniometer/scan info, but that could be added by someone.
Thanks @phyy-nx, I'm afraid that the goniometer and scan info are necessary in the case that prompted this issue, so I think the quickest fix will be to work with dlstbx2nxs
. I imagine (naively?) that it's as simple as reading the detector info, serial number, etc. from the input HDF5 file, and populating the above-mentioned strings dynamically.
dxtbx.dlsnxs2cbf
assumes input data are from an Eiger detector: https://github.com/cctbx/dxtbx/blob/bffa28ea73c9468d0c93c33a058822c96304ff5a/util/dlsnxs2cbf.py#L88 spcifically, this Eiger detector at Diamond: https://github.com/cctbx/dxtbx/blob/bffa28ea73c9468d0c93c33a058822c96304ff5a/util/dlsnxs2cbf.py#L97It would be nice to remove this assertion, if possible, to make
dlsnxs2cbf
useful for other sources of NXmx-flavour NeXus HDF5 files.