Open prjemian opened 8 years ago
@carterbox: This could be complicated so some discussion before starting would be helpful. The unit tests could use any/all of the example data files (in ./punx/data
).
Unknown names can occur in at least two ways in NeXus data files:
In the NXDL files, flexible names are indicated as all upper case. This is a fairly recent NeXus decision so uniform adoption may not be consistent.
The challenge here is to decide about any unexpected name, how might it match into the current context.
NeXus issues/PRs:
One easy case is with NXdata, if only one dataset is present, then whatever its name, that will be the default dataset to be plotted. (This is exactly the default structure of the EPICS area detector HDF5 plugin.) That is just one of many cases to be found.
https://manual.nexusformat.org/datarules.html#naming-conventions
reserved prefix and suffixes
in validate.review_with_NXDL(), check if an otherwise unrecognized field or group name (not known to NXDL class) can fit some specification of the NXDL