nucleon_elastic_ff.data.h5io

Module provides h5 file interafaces.

nucleon_elastic_ff.data.h5io.assert_h5files_equal(actual: str, expected: str, atol: float = 0.0, rtol: float = 1e-07, group_actual: Optional[str] = None, group_expected: Optional[str] = None)

Reads to HDF5 files, compares if they have equal datasets.

Checks if for each entry |actual - expected| < atol + rtol * |expected| (uses numpy.testing.assert_allclose).

Arguments

actual: str

File name for actual input data.

expected: str

File name for expected input data.

atol: float = 0.0

Absolute error tolarance. See numpy assert_allcolse.

rtol: float = 1.0e-7

Relative error tolarance. See numpy assert_allcolse.

Raises

AssertionError:

If datasets are different (e.g., not present or actual data is different.)

nucleon_elastic_ff.data.h5io.create_dset(h5f: h5py._hl.files.File, key: str, data: Any, overwrite: bool = False)

Creates or overwrites (if requested) dataset in HDF5 file.

Arguments

h5f: h5py.File

The file to write to.

key: str

The name of the dataset.

data: Any

The data for the dataset.

overwrite: bool = False

Wether data shall be overwritten.

nucleon_elastic_ff.data.h5io.get_dset_chunks(dset: h5py._hl.dataset.Dataset, chunk_size: int) → Iterable[numpy.ndarray]

Returns components of data sliced in chunks determined by the chunk size.

This reduces the memory size when loading the array.

Argumets

dset: h5py.Dataset

Input data set to read.

chunk_size: int

Size of the chunks to load in. Slices the first dimension of the input dataset. Must be smaller or equal to the size of the first data set dimension.

nucleon_elastic_ff.data.h5io.get_dsets(container: Union[h5py._hl.files.File, h5py._hl.group.Group], parent_name: Optional[str] = None, load_dsets: bool = False, ignore_containers: Optional[List[str]] = None) → Dict[str, Union[h5py._hl.dataset.Dataset, numpy.ndarray]]

Access an HDF5 container and extracts datasets.

The method is iteratively called if the container contains further containers.

Arguments

container: Union[h5py.File, h5py.Group]

The HDF5 group or file to iteratetively search.

parent_name: Optional[str] = None

The name of the parent container.

load_dsets: bool = False

If False, data sets are not opened (lazy load). If True, returns Dict with numpy arrays as values.

ignore_containers: Optional[List[str]] = None

A list of HDF5 containers to ignore when iteratively solving. Can be regex expressions.

Returns

datasets: Dict[str, Union[h5py.Dataset, np.ndarray]]

A dictionary containing the full path HDF path (e.g., groupA/subgroupB) to the data set as keys and the unloaded values of the set as values.