Issue with HDF5 dataset having missing values.

If a dataset from 'positioners' get less value than there is frame then we can get the following error:

----------------------------- TypeError Exception -----------------------------
Traceback (most recent call last):
  File "/home/payno/dev/esrf/darkfield/darfix/src/darfix/gui/dimensionsWidget.py", line 286, in _find_dimensions
    self._dataset.find_dimensions(
  File "/home/payno/dev/esrf/darkfield/darfix/src/darfix/core/dataset.py", line 610, in find_dimensions
    unique_values, unique_counts = array_utils.unique(
                                   ^^^^^^^^^^^^^^^^^^^
  File "/home/payno/dev/esrf/darkfield/darfix/src/darfix/core/array_utils.py", line 13, in unique
    unique_values, unique_indices, *other_indices = numpy.unique(
                                                    ^^^^^^^^^^^^^
  File "/home/payno/dev/esrf/darkfield/venv/lib/python3.11/site-packages/numpy/lib/arraysetops.py", line 274, in unique
    ret = _unique1d(ar, return_index, return_inverse, return_counts,
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/payno/dev/esrf/darkfield/venv/lib/python3.11/site-packages/numpy/lib/arraysetops.py", line 333, in _unique1d
    perm = ar.argsort(kind='mergesort' if return_index else 'quicksort')
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: '>' not supported between instances of 'int' and 'NoneType'
-------------------------------------------------------------------------------

When the "dimensiondefinition" widget/find dimension is called.

Today the _HDF5MetadataReader get_value always return a value

class _HDF5MetadataReader:
    """
    Equivalent of silx.io.fabioh5.FabioReader to give access to HDF5 metadata

    FIXME: remove this class. Storage and reading of metadata has been based on EDF. This class is a work-around
    Design of handle metadata shouldn't be based on the data type but more abstracted
    """

    def get_value(self, kind, name):
        return self._metadata.get(name, None)

Which is not the case for EDF. darfix.core.dataset.extract_metadata_values has a specific handling of the missing keys to define a default value directly at this level.