def test_output_h5_file_already_exists(self):
with h5py.File(image_path.replace('.png', '.h5')) as _:
pass
with self.assertRaises(FileExistsError):
translator = ImageTranslator()
_ = translator.translate(image_path)
def test_path_not_str(self):
with self.assertRaises(TypeError):
translator = ImageTranslator()
_ = translator.translate(np.arange(4))
def test_path_does_not_exist(self):
with self.assertRaises(FileNotFoundError):
translator = ImageTranslator()
_ = translator.translate('no_such_file.png')
def test_invalid_interp(self):
with self.assertRaises(ValueError):
translator = ImageTranslator()
_ = translator.translate(image_path,
bin_factor=2,
interp_func='dsdsdsd')
def test_invalid_h5_path(self):
with self.assertRaises(TypeError):
translator = ImageTranslator()
_ = translator.translate(image_path, h5_path=np.arange(4))
def test_invalid_dtype(self):
with self.assertRaises(TypeError):
translator = ImageTranslator()
_ = translator.translate(image_path, bin_factor=['dfrdd', True])
def test_float_parms(self):
with self.assertRaises(TypeError):
translator = ImageTranslator()
_ = translator.translate(image_path, bin_factor=1.34)
def test_neg_parms(self):
with self.assertRaises(ValueError):
translator = ImageTranslator()
_ = translator.translate(image_path, bin_factor=-2)
def test_too_many_dims(self):
with self.assertRaises(ValueError):
translator = ImageTranslator()
_ = translator.translate(image_path, bin_factor=(1, 2, 3))
def main_translate(self, **kwargs):
h5_path = kwargs.pop('h5_path', image_path.replace('.png', '.h5'))
delete_existing_file(h5_path)
input_image = rand_image.copy()
usize, vsize = input_image.shape[:2]
translator = ImageTranslator()
h5_path = translator.translate(image_path, **kwargs)
image_parms = dict()
if 'bin_factor' in kwargs.keys():
bin_factor = kwargs.pop('bin_factor')
if bin_factor is None:
_ = kwargs.pop('interp_func', None)
else:
if isinstance(bin_factor, int):
bin_factor = (bin_factor, bin_factor)
interp_func = kwargs.pop('interp_func', Image.BICUBIC)
image_parms.update({
'image_binning_size': np.array(bin_factor),
'image_PIL_resample_mode': interp_func
})
img_obj = Image.fromarray(input_image)
img_obj = img_obj.convert(mode="L")
img_obj = img_obj.resize(
(int(vsize / bin_factor[1]), int(usize / bin_factor[0])),
resample=interp_func)
input_image = np.asarray(img_obj)
image_parms.update({'normalized': False})
input_image = input_image.copy()
if 'normalize' in kwargs.keys():
normalize = kwargs.pop('normalize')
if normalize:
input_image -= np.min(input_image)
input_image = input_image / np.float32(np.max(input_image))
image_parms.update({'normalized': True})
image_parms.update({
'image_min': np.min(input_image),
'image_max': np.max(input_image)
})
with h5py.File(h5_path, mode='r') as h5_f:
self.basic_file_validation(h5_f)
h5_meas_grp = h5_f['Measurement_000']
h5_chan_grp = h5_meas_grp['Channel_000']
usid_main = USIDataset(h5_chan_grp['Raw_Data'])
# check the attributes under this group
for key, expected_val in image_parms.items():
self.assertTrue(
np.all(
hdf_utils.get_attr(h5_meas_grp, key) == expected_val))
one_d_image = input_image.T.reshape(-1, 1)
self.assertTrue(np.allclose(one_d_image, usid_main[()]))
# self.assertTrue(np.allclose(rand_image, np.reshape(usid_main[()], rand_image.shape)))
pos_data = np.vstack((np.tile(np.arange(input_image.shape[0]),
input_image.shape[1]),
np.repeat(np.arange(input_image.shape[1]),
input_image.shape[0]))).T
validate_aux_dset_pair(self,
h5_chan_grp,
usid_main.h5_pos_inds,
usid_main.h5_pos_vals, ['Y', 'X'],
['a.u.', 'a.u.'],
pos_data,
h5_main=usid_main,
is_spectral=False)
delete_existing_file(h5_path)