GE RDF HDF5 remaining issues

[KrisThielemans]

After #390 there's a few more things to do.

• [x] various hard-wired numbers for the Signa. After fixing those, we should rename Signa files to something more generic (done in #601)
• [x] GESignaListmodeInputFileFormat only checks if it's an HDF5 file (done in #601)
• [x] ProjData should call ProjDataGEHDF5 (done in #601)
• [x] add tests (needs data)
• [x] merge #181 (incompatible results for some scanners, so probably needs v5.0)
• [ ] in ProjDataGEHDF5 avoid reading all views up-front. could read them when necessary
• [ ] image-flipping see #675

[evgueni-ovtchinnikov]

Another one:

hsize_t dims_out[dataset_list_Ndims];

is wrong (dataset_list_Ndims is not a constant), should be

hsize_t* dims_out = new hsize_t[dataset_list_Ndims];

(and delete[] dims_out; before return).

[AnderBiguri]

* [ ]  `SinglesRatesFromGEHDF5`
  
  * expose `GEHDF5Wrapper.get_exam_info_sptr()`

You mean creating SinglesRatesFromGEHDF5::get_exam_info_sptr()? GEHDF5Wrapper.get_exam_info_sptr() is exposed and read in that class.

* [ ]  in `ProjDataGEHDF5` avoid reading all views up-front. could read them when necessary

You mean read them in get_viewgram()/get_sinogram() instead of in initialise_viewgram_buffer?

[KrisThielemans]

You mean creating SinglesRatesFromGEHDF5::get_exam_info_sptr()? GEHDF5Wrapper.get_exam_info_sptr() is exposed and read in that class.

not sure what I was thinking at the time. Possibly because we were (erronuously) using TimeFrameDefinitions then, which sit in exam_info. Let's ignore this. (please edit the first box and remove from the list)

you mean read them in get_viewgram()/get_sinogram() instead of in initialise_viewgram_buffer?

yes. it could be done by making tof_data mutable`, and fill it when necessary. However, this will have little impact, and likely will conflict with updates on the TOF branch, so I'd postpone this.

[KrisThielemans]

on

construct_randoms_from_GEsingles.cxx needs to reorder data. This should be moved to this class this is about https://github.com/KrisThielemans/STIR-1/blob/7c5ca0aa99b60f01db35d3d3a0904694d3ffb0ed/src/utilities/construct_randoms_from_GEsingles.cxx#L244

This reversal of efficiencies should be moved to SinglesRatesFromGEHDF5. Once that's done, there might be very little GE specific in construct_randoms_from_GEsingles (I'll have to check).

[AnderBiguri]

I have now added some examples.

Missing from that work:

Any change related to the attenuation coefficients. I don't know where they are in the data I have nor how to read them.

Testing it with real data and real .fdef

[AnderBiguri]

Added tests (and tested them!)