Handle dataset from HDF5 files

TODO

• update Data selection widget to handle HDF5
• Make sure the 'dimension definition' widget works
• add unit-test
• test on a workflow
  • with dummy dataset
  • Doesn't handle the 'metadata widget' @rarodrig do you know how much used it is ? Is it a must have ?
  • provide HDF5 dataset to examples (by default). The
• update test to also test HDF5

Warning / limitation:

about the PR

This PR is not handling the 'metadata kind' that can be found in EDF-fabio (positioners, counters...). And only handle an HDF5 group as parent of all relevant metadata for darfix (such as obpitch...)

about magnification

It Seems that Dataset.compute_transformation is using some 'hardcoded' keys like 'ffz', 'mainx'... which can be an issue when moving to HDF5

rsm histogram example

Failing (but was already failing for EDF dataset)

extra

information

• close #89 (closed)
• close #13 (closed)

test datasets / scripts

Some datasetes have been converted from EDF to 'expected like HDF5'. They can be found in /scisoft/darfix/dataset. Thet have been converted using the following dummy script: convert_edf_to_hdf5.py

Edit @denolf For me it is: /data/scisoft/darfix/datasets

screenshot