"""I/O Routines for HDF5 files."""
__all__ = ["read_hdf5", "write_hdf5"]
import copy
import h5py
import numpy as np
from .pathlib import get_filepath
dtypes = (
np.uint8,
np.uint16,
np.uint32,
np.uint64,
int,
np.int16,
np.int32,
np.int64,
float,
np.float16,
np.float32,
np.float64,
)
[docs]def read_hdf5(filepath):
"""
Read HDF5 file as a Python dictionary
Parameters
----------
filepath : str
Path to file on disk.
Returns
-------
dict
Deserialized HDF5 file.
Example
-------
Define an exemplary dictionary and save to file:
>>> import os
>>> import numpy as np
>>> import deepmr.io
>>> pydict = {'headerstr': 'someinfo', 'testdouble': np.ones(3, dtype=np.float32)}
>>> filepath = os.path.realpath('.')
>>> deepmr.io.write_hdf5(filepath)
Load from disk:
>>> loaded_dict = deepmr.io.read_hdf5(filepath)
Result is the same dictionary created before:
>>> loaded_dict.keys()
['headerstr', 'testdouble']
>>> matfile['testdouble']
array([1.0, 1.0, 1.0])
>>> matfile['headerstr']
'someinfo'
"""
filepath = get_filepath(filepath, True, "h5")
with h5py.File(filepath, "r") as h5file:
return _recursively_load_dict_contents_from_group(h5file, "/")
[docs]def write_hdf5(input, filepath):
"""
Write a given dictionary to HDF5 file.
Parameters
----------
input : dict
Input dictionary.
filepath : str
Path to file on disk.
Example
-------
Define an exemplary dictionary:
>>> import numpy as np
>>> pydict = {'headerstr': 'someinfo', 'testdouble': np.ones(3, dtype=np.float32)}
Save the dictionary to disk:
>>> import os
>>> import deepmr.io
>>> filepath = os.path.realpath('.')
>>> deepmr.io.write_hdf5(filepath)
"""
input = copy.deepcopy(input)
with h5py.File(filepath, "w") as h5file:
_recursively_save_dict_contents_to_group(h5file, "/", input)
def _recursively_load_dict_contents_from_group(h5file, path):
ans = {}
for key, item in h5file[path].items():
if isinstance(item, h5py._hl.dataset.Dataset):
tmp = item[()]
if isinstance(tmp, bytes):
tmp = tmp.decode()
if isinstance(tmp, np.ndarray):
tmp = tmp.squeeze()
ans[key] = tmp
elif isinstance(item, h5py._hl.group.Group):
ans[key] = _recursively_load_dict_contents_from_group(
h5file, path + key + "/"
)
return ans
def _recursively_save_dict_contents_to_group(h5file, path, dic):
for key, item in dic.items():
if isinstance(item, (*dtypes, str, bytes)):
h5file[path + key] = item
elif np.isscalar(item):
h5file[path + key] = (item,)
elif isinstance(item, (list, tuple)):
h5file[path + key] = np.asarray(item)
elif isinstance(item, np.ndarray):
h5file[path + key] = item
elif isinstance(item, dict):
_recursively_save_dict_contents_to_group(h5file, path + key + "/", item)
elif item is None:
pass
else:
raise ValueError(f"Cannot save {type(item)} type")
