jangada.serialization.Persistable._save_data_in_group

static Persistable._save_data_in_group(key: str, value: Any, group: Group) → None

Recursively save data to an HDF5 group.

This method handles the conversion of Python objects to HDF5 structures, mapping different types to appropriate HDF5 storage (attributes, datasets, or subgroups).

Parameters:

keystr

The name/key for this data in the HDF5 group.

valueAny

The value to save.

grouph5py.Group

The HDF5 group to save into.

Raises:

TypeError

If the value type is not supported for HDF5 storage.

Notes

Type mapping: - None → attribute: ‘NoneType:None’ - Path → attribute: ‘Path:/absolute/path’ - str/Number → attribute: direct storage - list → subgroup with __container_type__ = ‘list’ - dict → subgroup with __container_type__ = ‘dict’ - dataset types → HDF5 dataset with __dataset_type__ attribute

Datasets are created with resizable first dimension (maxshape=(None, …)) to support appending, except for scalar datasets which cannot be resized.