| import copy |
| from functools import partial |
| from pathlib import Path |
| from zipfile import ZipFile |
|
|
| import h5py |
| import numpy as np |
|
|
|
|
| def _get_ds_dictionaries(ds_dict: dict[str, np.ndarray], _name, node): |
| |
| fullname = node.name |
| if isinstance(node, h5py.Dataset): |
| |
| |
| ds_dict[fullname] = np.array(node) |
| |
| |
| |
| |
|
|
|
|
| def get_results(results_zip_file: Path) -> dict[str, dict[str, np.ndarray]]: |
| results: dict[str, dict[str, np.ndarray]] = {} |
| with ZipFile(results_zip_file, "r") as the_zip: |
| for name in the_zip.namelist(): |
| if "reports.h5" in name: |
| with h5py.File(the_zip.open(name)) as h5: |
| ds_dict: dict[str, np.ndarray] = {} |
| ds_visitor = partial(_get_ds_dictionaries, ds_dict) |
| h5.visititems(ds_visitor) |
| results[name] = copy.deepcopy(ds_dict) |
| return results |
|
|
|
|
| def compare_arrays(arr1: np.ndarray, arr2: np.ndarray) -> bool: |
| if type(arr1[0]) == np.float64: |
| max1 = max(arr1) |
| max2 = max(arr2) |
| atol = max(1e-3, max1*1e-5, max2*1e-5) |
| return np.allclose(arr1, arr2, rtol=1e-4, atol=atol) |
| for n in range(len(arr1)): |
| if not compare_arrays(arr1[n], arr2[n]): |
| return False |
| return True |
|
|
|
|
| def compare_datasets(results1: dict[str, dict[str, np.ndarray]], results2: dict[str, dict[str, np.ndarray]]) -> bool: |
| for h5_file_path in results1: |
| if h5_file_path not in results2: |
| return False |
| for dataset_name in results1[h5_file_path]: |
| if dataset_name not in results2[h5_file_path]: |
| return False |
| arr1 = results1[h5_file_path][dataset_name] |
| arr2 = results2[h5_file_path][dataset_name] |
| if arr1.shape != arr2.shape: |
| return False |
| if not compare_arrays(arr1, arr2): |
| return False |
| return True |
|
|
