def handle_data():
json_request = request.get_json(cache=False)
email, dataset = get_email_and_dataset(json_request)
return json_response(
data_processing.handle_data(dataset_api=dataset_api,
dataset=dataset,
embedding_list=json_request.get('embedding'),
values=json_request.get('values'),
stats=json_request.get('stats'),
grouped_stats=json_request.get('groupedStats'),
selection=json_request.get('selection')))
def test_no_binning(dataset_api, input_dataset, test_data, measures,
dimensions, continuous_obs, basis):
obsm_field = basis
embedding_list = [dict(dimensions=2, name=basis)]
values = dict(dimensions=dimensions,
measures=measures +
list(map(lambda x: "obs/" + x, continuous_obs)))
results = handle_data(dataset_api=dataset_api,
dataset=input_dataset,
values=values,
embedding_list=embedding_list)
for key in measures:
v = results["values"][key] # dict of indices, values
if isinstance(v, dict):
values = np.zeros(test_data.shape[0])
values[v["indices"]] = v["values"]
else:
values = v
X = test_data[:, key].X
if scipy.sparse.issparse(X):
X = X.toarray()
X = X.flatten()
np.testing.assert_array_equal(values, X, err_msg=key)
for key in dimensions:
val = results["values"][key]
if pd.api.types.is_categorical_dtype(test_data.obs[key]):
val = pd.Categorical.from_codes(val["values"], val["categories"])
np.testing.assert_array_equal(val,
test_data.obs[key].values,
err_msg="obs field {}".format(key))
obsm = test_data.obsm[obsm_field]
coords = results["embeddings"][0]["coordinates"]
for i in range(obsm.shape[1]):
np.testing.assert_array_equal(coords["{}_{}".format(basis, i + 1)],
obsm[:, i],
err_msg="obsm {}".format(key))