bundle.hdf5¶

Submodules¶

bundle.hdf5.store

Classes¶

Store

Simple HDF5 store for reading and writing datasets and attributes.

Package Contents¶

class bundle.hdf5.Store(path: pathlib.Path | str, mode: str = 'r')¶

Simple HDF5 store for reading and writing datasets and attributes.

Usage:

with Store("data.h5", mode="w") as store:
    store.write_dataset("group/data", np.array([1, 2, 3]))
    store.write_attrs("group", {"version": "1.0"})

with Store("data.h5", mode="r") as store:
    arr = store.read_dataset("group/data")
    attrs = store.read_attrs("group")

path¶

mode = 'r'¶

property file: h5py.File¶

write_dataset(name: str, data: numpy.ndarray, **kwargs)¶: Write or overwrite a dataset at the given path.

read_dataset(name: str) → numpy.ndarray¶: Read a dataset and return as numpy array.

write_attrs(path: str, attrs: dict | bundle.core.data.Data)¶

Write attributes to a group or dataset (creating a group if path doesn’t exist).

Accepts a plain dict or a bundle.core.data.Data instance (uses model_dump()). Only scalar and string values are stored; complex nested objects are JSON-serialised.

read_attrs(path: str) → dict¶: Read all attributes from a group or dataset.

read_attrs_as(path: str, model: type[D]) → D¶

Read attributes and construct a Data (or any Pydantic model) instance.

Values that were JSON-serialised on write are automatically deserialised.

list_datasets(group_path: str = '/') → list[str]¶: List all dataset names under a group.

list_groups(group_path: str = '/') → list[str]¶: List all group names under a group.

has(name: str) → bool¶: Check if a dataset or group exists.