v0.8 : 19-06-2026
- relase 0.8
- Add support for pd.DataFrame()
- improve function map_dict_to_object
- Fix : return in hdf5_dataset_creator after parsing an unknown object
- Add function to map a dictionary to an object with a given structure

v0.7 : 12-03-2026
- Exclude some python object when parsing a dict or an object. For instance Geopandas and shapely are excluded, because pyhdf5_handler crash when handled these complex python object. Other module can be exclude using the global variable   EXCLUDE_PYTHON_OBJ.
- Fix: Max recursion counter was set to zeros at the begining of the recursive function.

v0.6 : 11-02-2026
- Fix : handle np.int64/np.float64 type like. Pyhdf5_handler will correctly return unique float or int stored in a dataset.

v0.5 : 15-12-2025
- Fix : return np.int() if the original data was an integer 

v0.4 : 25-10-2025
- Update documentation (rework)
- Add complementary function savec_object_to_hdf5()
- Small fixes
- Fix hdf5_view() does not print recursive view
- Add complementary tests

v0.3 : 10-09-2025
- Better check numpy, pandas and datetime data type
- Stop recursion to 100 (add a recsursion_limit parameter) (avoid infinite recursion on complex object)
- Fix added to save_object_to_hdf5file function, generate_object_structure and read_object_as_dict

v0.2 : 23-05-2025
- Special attributs of every datase now start with '_'
- Hdf5 Attributes are now read and stored in the dictonnary as {attribute_name : value}
- It is possible to read special attribute with read_dataset_attributes=True when using function hdf5_read_dataset()
- Rename functions : dump_dict_to_hdf5 -> save_dict_to_hdf5  and  save_dict_to_hdf5 -> save_dict_to_hdf5file
- Only import usefull user functions when loading the module.
- Add a release note


