[Bug]: NeXus parser (nomad) getting error while reading unit degree from pint units

[RubelMozumder]

Contact Details

[email protected]

What happened?

Error massage:

ERROR    nomad.processing     2024-07-29T09:46:26 could not normalize section
  - exception: Traceback (most recent call last):
      File "/home/rubel/NOMAD-FAIRmat/GH/pynxtools/src/pynxtools/dataconverter/writer.py", line 324, in _put_data_into_hdf5
        add_units_key(self.output_nexus[path_hdf5], path)
      File "/home/rubel/NOMAD-FAIRmat/GH/pynxtools/src/pynxtools/dataconverter/writer.py", line 264, in add_units_key
        dataset.attrs["units"] = self.data[units_key]
      File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
      File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
      File "/home/rubel/NOMAD-FAIRmat/GH/nomad-measurements/.PE9/lib/python3.9/site-packages/h5py/_hl/attrs.py", line 104, in __setitem__
        self.create(name, data=value)
      File "/home/rubel/NOMAD-FAIRmat/GH/nomad-measurements/.PE9/lib/python3.9/site-packages/h5py/_hl/attrs.py", line 182, in create
        htype = h5t.py_create(original_dtype, logical=True)
      File "h5py/h5t.pyx", line 1663, in h5py.h5t.py_create
      File "h5py/h5t.pyx", line 1687, in h5py.h5t.py_create
      File "h5py/h5t.pyx", line 1747, in h5py.h5t.py_create
    TypeError: Object dtype dtype('O') has no native HDF5 equivalent
    ```
    
To write a dataset in a hdf5 file and using `degree` as a unit for a quantity in dataset, the error can be reproduced.

Probably, the error is coming from hdf5 lib, as `pint` lib returns a degree symbol `o` instead of string and hdf5 lib unable to decode or unable detect common data type.

### Relevant log output

_No response_

[mkuehbach]

The proper usage of pint and h5py for °C is like this:
e.g.

import pint
from pint import UnitRegistry
ureg = UnitRegistry()
# assume above lines have been called somewhere else then you can do
T = ureg.Quantity(20.0, ureg.degC)
print(T.units)

T itself is an object, if you want to write the unit via h5py and say pynxtools'

template["/ENTRY[entry]/.../temperature/@units"] = T.units

Do these hints help @RubelMozumder ?

This way pint will also take care of the fact that temperatures are not multiplicative quantities:
https://pint.readthedocs.io/en/0.10.1/nonmult.html

Suggestions:

• Check pynxtools-side writing code
• Check where ureg is hooked in NOMAD
• Check which string will then be written to HDF
• Check how this is consumed by the HDF5 reader when a quantity is getting instantiated in NOMAD