def test_good_case(self):
# hypothetical data for two atoms
good_powder = {
"a_tensors": {
"0":
np.asarray([[[0.01, 0.02, 0.03], [0.01, 0.02, 0.03],
[0.01, 0.02, 0.03]],
[[0.01, 0.02, 0.03], [0.01, 0.02, 0.03],
[0.01, 0.02, 0.03]]])
},
"b_tensors": {
"0":
np.asarray([[[0.01, 0.02, 0.03], [0.01, 0.02, 0.03],
[0.01, 0.02, 0.03]],
[[0.01, 0.02, 0.03], [0.01, 0.02, 0.03],
[0.01, 0.02, 0.03]]])
}
}
good_tester = PowderData(num_atoms=2)
good_tester.set(items=good_powder)
extracted_data = good_tester.extract()
for key in good_powder:
for k_point in good_powder[key]:
self.assertEqual(
True,
np.allclose(good_powder[key][k_point],
extracted_data[key][k_point]))
def test_roundtrip(self):
initial_powderdata = PowderData(**self.good_items, num_atoms=2)
roundtrip_data = PowderData.from_extracted(initial_powderdata.extract())
for attr in 'get_a_tensors', 'get_b_tensors', 'get_frequencies':
for k_index in self.good_items['a_tensors']:
self.assertTrue(np.allclose(getattr(initial_powderdata, attr)()[k_index],
getattr(roundtrip_data, attr)()[k_index]))
def test_bad_items(self):
# wrong items: array instead of dict
bad_items = self.good_items.copy()
bad_items['a_tensors'] = bad_items['a_tensors'][0]
with self.assertRaises(TypeError):
PowderData(**bad_items, num_atoms=2)
# list instead of np array
bad_items = self.good_items.copy()
for key, value in bad_items.items():
bad_items[key][0] = value[0].tolist()
with self.assertRaises(TypeError):
PowderData(**bad_items, num_atoms=2)
# wrong size of items: data only for one atom ; should be for two atoms
bad_items = {
"a_tensors": {
0:
np.asarray([[[0.01, 0.02, 0.03], [0.01, 0.02, 0.03],
[0.01, 0.02, 0.03]]])
},
"b_tensors": {
0:
np.asarray([[[0.01, 0.02, 0.03], [0.01, 0.02, 0.03],
[0.01, 0.02, 0.03]]])
},
"frequencies": {
0: np.asarray([[[1.23, 4.56, 7.89]]])
}
}
with self.assertRaises(ValueError):
PowderData(**bad_items, num_atoms=2)
def test_getters(self):
good_powderdata = PowderData(**self.good_items, num_atoms=2)
for k_point in self.good_items["a_tensors"]:
self.assertTrue(np.allclose(self.good_items["a_tensors"][k_point],
good_powderdata.get_a_tensors()[k_point]))
self.assertTrue(np.allclose(self.good_items["b_tensors"][k_point],
good_powderdata.get_b_tensors()[k_point]))
self.assertTrue(np.allclose(self.good_items["frequencies"][k_point],
good_powderdata.get_frequencies()[k_point]))
def test_good_case(self):
good_powderdata = PowderData(**self.good_items, num_atoms=2)
extracted_data = good_powderdata.extract()
for key in self.good_items:
for k_index in self.good_items[key]:
self.assertTrue(np.allclose(self.good_items[key][k_index],
extracted_data[key][str(k_index)]))
# Should also work if num_atoms is not given
PowderData(**self.good_items)
def test_set(self):
poor_tester = PowderData(num_atoms=2)
# wrong items: list instead of numpy array
bad_items = {
"a_tensors": [[0.002, 0.001]],
"b_tensors": [[0.002, 0.001]]
}
with self.assertRaises(ValueError):
poor_tester.set(items=bad_items)
# wrong size of items: data only for one atom ; should be for two atoms
bad_items = {
"a_tensors": {
"0":
np.asarray([[[0.01, 0.02, 0.03], [0.01, 0.02, 0.03],
[0.01, 0.02, 0.03]]])
},
"b_tensors": {
"0":
np.asarray([[[0.01, 0.02, 0.03], [0.01, 0.02, 0.03],
[0.01, 0.02, 0.03]]])
}
}
with self.assertRaises(ValueError):
poor_tester.set(items=bad_items)
def test_input(self):
# wrong number of atoms
with self.assertRaises(ValueError):
_ = PowderData(num_atoms=-2)
def test_bad_num_atoms(self):
# wrong number of atoms
with self.assertRaises(ValueError):
PowderData(**self.good_items, num_atoms=-2)