def run_training():
data = load_data(file_name='train_data.csv')
X, y = resample_classes(data)
pipeline.CHD_pipe.fit(X, y)
_logger.info(f'saving model version: {_version}')
save_pipeline(pipeline_to_save=pipeline.CHD_pipe)
def test_prediction_endpoint_val_200(flask_test_client):
test_data = load_data(file_name=config.TEST_DATA)
post_json = test_data.to_json(orient='records')
response = flask_test_client.post('/v1/predict/log_reg',
json=json.loads(post_json))
assert response.status_code == 200
response_json = json.loads(response.data)
assert response_json.get('errors') is None
assert len(response_json.get('predictions')) == 1
def test_prediction_endpoint(flask_test_client):
test_data = load_data(file_name=model_config.TEST_DATA)
post_json = test_data.to_json(orient='records')
response = flask_test_client.post('/v1/predict/log_reg',
json=json.loads(post_json))
assert response.status_code == 200
response_json = json.loads(response.data)
prediction = response_json['predictions']
response_version = response_json['version']
assert isinstance(prediction[0][1], float)
assert response_version == _version
def capture_predictions() -> None:
""" Save test data predictions to .csv file"""
new_dict = {}
saved_file = 'test_predictions.csv'
test_data = load_data(file_name='test_data.csv')
predictions = make_prediction(input_data=test_data)
array = predictions.get('predictions')[0]
version = predictions.get('version')
new_dict['predictions'] = array
new_dict['version'] = version
preds_df = pd.DataFrame(new_dict)
#save the file into package repo
preds_df.to_csv(f'{config.PACKAGE_ROOT}/{saved_file}', index=False)
def test_make_single_prediction():
test_data = load_data(file_name='test_data.csv')
subject = make_prediction(input_data=test_data)
assert subject is not None
assert isinstance(subject.get('predictions')[0], np.ndarray)