def test_structured_data_from_csv_partial_col_type_classifier(tmp_dir):
clf = ak.StructuredDataClassifier(
column_types=common.PARTIAL_COLUMN_TYPES_FROM_CSV,
directory=tmp_dir,
max_trials=1)
clf.fit(x=common.TRAIN_FILE_PATH, y='survived', epochs=2,
validation_data=common.TEST_FILE_PATH)
def test_structured_data_get_col_names_from_df(fit, tmp_path):
clf = ak.StructuredDataClassifier(
directory=tmp_path,
seed=test_utils.SEED,
)
clf.fit(x=test_utils.TRAIN_CSV_PATH, y="survived")
assert nest.flatten(clf.inputs)[0].column_names[0] == "sex"
def test_structured_data_from_csv_classifier(tmp_dir):
clf = ak.StructuredDataClassifier(directory=tmp_dir, max_trials=1)
clf.fit(x=common.TRAIN_FILE_PATH,
y='survived',
epochs=2,
validation_data=common.TEST_FILE_PATH)
x_test = common.csv_test('classification')
assert clf.predict(x_test).shape == (len(x_test), 1)
def test_raise_error_unknown_str_in_col_type(tmp_path):
with pytest.raises(ValueError) as info:
ak.StructuredDataClassifier(
column_types={'age': 'num', 'parch': 'categorical'},
directory=tmp_path,
seed=utils.SEED)
assert 'Column_types should be either "categorical"' in str(info.value)
def test_structured_clf_fit_call_auto_model_fit(fit, tmp_path):
auto_model = ak.StructuredDataClassifier(directory=tmp_path, seed=utils.SEED)
auto_model.fit(
x=utils.generate_structured_data(num_instances=100),
y=utils.generate_one_hot_labels(num_instances=100, num_classes=3))
assert fit.is_called
def test_structured_data_from_numpy_classifier(tmp_dir):
num_data = 500
data = common.structured_data(num_data)
x_train = data
y = np.random.randint(0, 3, num_data)
y_train = y
clf = ak.StructuredDataClassifier(directory=tmp_dir, max_trials=1)
clf.fit(x_train, y_train, epochs=2, validation_data=(x_train, y_train))
def test_structured_clf_predict_csv_call_automodel_predict(
predict, fit, tmp_path):
auto_model = ak.StructuredDataClassifier(directory=tmp_path,
seed=utils.SEED)
auto_model.fit(x=utils.TRAIN_CSV_PATH, y="survived")
auto_model.predict(x=utils.TEST_CSV_PATH)
assert predict.is_called
def test_structured_clf_evaluate_call_automodel_evaluate(
evaluate, fit, tmp_path):
auto_model = ak.StructuredDataClassifier(directory=tmp_path,
seed=utils.SEED)
auto_model.fit(x=utils.TRAIN_CSV_PATH, y="survived")
auto_model.evaluate(x=utils.TRAIN_CSV_PATH, y="survived")
assert evaluate.is_called
def test_structured_data_classifier_transform_new_data(tmp_dir):
num_data = 200
num_train = 100
data = common.structured_data(num_data)
x_train, x_test = data[:num_train], data[num_train:]
y = np.random.randint(0, 3, num_data)
y_train, y_test = y[:num_train], y[num_train:]
clf = ak.StructuredDataClassifier(directory=tmp_dir, max_trials=1)
clf.fit(x_train, y_train, epochs=2, validation_data=(x_test, y_test))
def test_structured_data_from_csv_less_col_name_classifier(tmp_dir):
with pytest.raises(ValueError) as info:
clf = ak.StructuredDataClassifier(
column_names=common.LESS_COLUMN_NAMES_FROM_CSV,
directory=tmp_dir,
max_trials=1)
clf.fit(x=common.TRAIN_FILE_PATH, y='survived', epochs=2,
validation_data=common.TEST_FILE_PATH)
assert 'Expect column_names to have length' in str(info.value)
def test_structured_data_from_csv_false_col_type_classifier(tmp_dir):
with pytest.raises(ValueError) as info:
clf = ak.StructuredDataClassifier(
column_types=common.FALSE_COLUMN_TYPES_FROM_CSV,
directory=tmp_dir,
max_trials=1)
clf.fit(x=common.TRAIN_FILE_PATH, y='survived', epochs=2,
validation_data=common.TEST_FILE_PATH)
assert 'Column_types should be either "categorical"' in str(info.value)
def build_pipeline(self):
"""
Makes a pipeline based on data_config
"""
if self.problem_type == "classification":
automl_pipeline = ak.StructuredDataClassifier(**self.automl_settings)
elif self.problem_type == "regression":
automl_pipeline = ak.StructuredDataRegressor(**self.automl_settings)
return automl_pipeline
def test_structured_data_from_csv_col_type_mismatch_classifier(tmp_dir):
with pytest.raises(ValueError) as info:
clf = ak.StructuredDataClassifier(
column_types=common.COLUMN_TYPES_FROM_CSV,
directory=tmp_dir,
max_trials=1)
clf.fit(x=common.TRAIN_FILE_PATH, y='survived', epochs=2,
validation_data=common.TEST_FILE_PATH)
assert 'Column_names and column_types are mismatched.' in str(info.value)
def test_structured_data_from_numpy_classifier(tmp_dir):
num_data = 500
num_train = 400
data = common.structured_data(num_data)
x_train, x_test = data[:num_train], data[num_train:]
y = np.random.randint(0, 3, num_data)
y_train, y_test = y[:num_train], y[num_train:]
clf = ak.StructuredDataClassifier(directory=tmp_dir, max_trials=1)
clf.fit(x_train, y_train, epochs=2, validation_data=(x_train, y_train))
assert clf.predict(x_test).shape == (len(y_test), 1)
def test_structured_clf_fit_call_auto_model_fit(fit, tmp_path):
auto_model = ak.StructuredDataClassifier(directory=tmp_path,
seed=utils.SEED)
auto_model.fit(
x=pd.read_csv(utils.TRAIN_CSV_PATH).to_numpy().astype(
np.unicode)[:100],
y=utils.generate_one_hot_labels(num_instances=100, num_classes=3),
)
assert fit.is_called
def test_structured_data_from_numpy_col_name_classifier(tmp_dir):
num_data = 500
data = common.structured_data(num_data)
x_train = data
y = np.random.randint(0, 3, num_data)
y_train = y
clf = ak.StructuredDataClassifier(
column_names=common.COLUMN_NAMES_FROM_NUMPY,
directory=tmp_dir,
max_trials=1)
clf.fit(x_train, y_train, epochs=2, validation_data=(x_train, y_train))
def test_structured_data_clf_convert_csv_to_df_and_np(fit, tmp_path):
auto_model = ak.StructuredDataClassifier(directory=tmp_path, seed=utils.SEED)
auto_model.fit(x=utils.TRAIN_FILE_PATH,
y='survived',
epochs=2,
validation_data=(utils.TEST_FILE_PATH, 'survived'))
_, kwargs = fit.call_args_list[0]
assert isinstance(kwargs['x'], pandas.DataFrame)
assert isinstance(kwargs['y'], np.ndarray)
def test_titaninc_accuracy_over_77(tmp_path):
TRAIN_DATA_URL = "https://storage.googleapis.com/tf-datasets/titanic/train.csv"
TEST_DATA_URL = "https://storage.googleapis.com/tf-datasets/titanic/eval.csv"
train_file_path = tf.keras.utils.get_file("train.csv", TRAIN_DATA_URL)
test_file_path = tf.keras.utils.get_file("eval.csv", TEST_DATA_URL)
clf = ak.StructuredDataClassifier(max_trials=10, directory=tmp_path)
clf.fit(train_file_path, "survived")
accuracy = clf.evaluate(test_file_path, "survived")[1]
assert accuracy >= 0.77
def test_raise_error_unknown_str_in_col_type(tmp_path):
with pytest.raises(ValueError) as info:
ak.StructuredDataClassifier(
column_types={
"age": "num",
"parch": "categorical"
},
directory=tmp_path,
seed=test_utils.SEED,
)
assert 'column_types should be either "categorical"' in str(info.value)
def test_structured_data_classifier_from_csv(init, fit):
clf = ak.StructuredDataClassifier(directory=tmp_dir,
max_trials=1,
seed=common.SEED)
clf.fit(x=common.TRAIN_FILE_PATH, y='survived', epochs=2,
validation_data=common.TEST_FILE_PATH)
assert init.called
_, kwargs = fit.call_args_list[0]
assert isinstance(kwargs['x'], pandas.DataFrame)
assert isinstance(kwargs['y'], np.ndarray)
def main():
train_file_path = tf.keras.utils.get_file("train.csv", TRAIN_DATA_URL)
test_file_path = tf.keras.utils.get_file("eval.csv", TEST_DATA_URL)
clf = ak.StructuredDataClassifier(max_trials=10, directory='tmp_dir', overwrite=True)
start_time = timeit.default_timer()
clf.fit(train_file_path, 'survived')
stop_time = timeit.default_timer()
accuracy = clf.evaluate(test_file_path, 'survived')[1]
print('Accuracy: {accuracy}%'.format(accuracy=round(accuracy * 100, 2)))
print('Total time: {time} seconds.'.format(time=round(stop_time - start_time, 2)))
def test_structured_data_from_numpy_classifier(tmp_dir):
num_data = 500
num_train = 400
data = common.generate_structured_data(num_data)
x_train, x_test = data[:num_train], data[num_train:]
y = common.generate_one_hot_labels(num_instances=num_data, num_classes=3)
y_train, y_test = y[:num_train], y[num_train:]
clf = ak.StructuredDataClassifier(directory=tmp_dir,
max_trials=1,
seed=common.SEED)
clf.fit(x_train, y_train, epochs=2, validation_data=(x_train, y_train))
assert clf.predict(x_test).shape == (len(y_test), 3)
def test_structured_data_col_type_no_name_error(tmp_path):
with pytest.raises(ValueError) as info:
clf = ak.StructuredDataClassifier(
column_types={
"age": "numerical",
"parch": "categorical"
},
directory=tmp_path,
seed=test_utils.SEED,
)
clf.fit(x=np.random.rand(100, 30), y=np.random.rand(100, 1))
assert "column_names must be specified" in str(info.value)
def test_raise_error_unknown_name_in_col_type(tmp_path):
with pytest.raises(ValueError) as info:
ak.StructuredDataClassifier(
column_types={
"age": "numerical",
"parch": "categorical"
},
column_names=["age", "fare"],
directory=tmp_path,
seed=utils.SEED,
)
assert "Column_names and column_types are mismatched" in str(info.value)
def test_structured_data_from_numpy_col_type_classifier(tmp_dir):
num_data = 500
data = common.structured_data(num_data)
x_train = data
y = np.random.randint(0, 3, num_data)
y_train = y
with pytest.raises(ValueError) as info:
clf = ak.StructuredDataClassifier(
column_types=common.COLUMN_TYPES_FROM_NUMPY,
directory=tmp_dir,
max_trials=1)
clf.fit(x_train, y_train, epochs=2, validation_data=(x_train, y_train))
assert str(info.value) == 'Column names must be specified.'
def test_structured_data_classifier(tmp_path):
num_data = 500
num_train = 400
data = pd.read_csv(utils.TRAIN_CSV_PATH).to_numpy().astype(np.unicode)[:num_data]
x_train, x_test = data[:num_train], data[num_train:]
y = utils.generate_one_hot_labels(num_instances=num_data, num_classes=3)
y_train, y_test = y[:num_train], y[num_train:]
clf = ak.StructuredDataClassifier(
directory=tmp_path, max_trials=1, seed=utils.SEED
)
clf.fit(x_train, y_train, epochs=2, validation_data=(x_train, y_train))
clf.export_model()
assert clf.predict(x_test).shape == (len(y_test), 3)
def test_structured_classifier(init, fit):
num_data = 500
train_x = common.generate_structured_data(num_data)
train_y = common.generate_one_hot_labels(num_instances=num_data, num_classes=3)
clf = ak.StructuredDataClassifier(
column_names=common.COLUMN_NAMES_FROM_NUMPY,
directory=tmp_dir,
max_trials=1,
seed=common.SEED)
clf.fit(train_x, train_y, epochs=2, validation_data=(train_x, train_y))
assert init.called
assert fit.called
def test_structured_data_input_name_type_mismatch_error(tmp_path):
with pytest.raises(ValueError) as info:
clf = ak.StructuredDataClassifier(
column_types={
"_age": "numerical",
"parch": "categorical"
},
column_names=["age", "fare"],
directory=tmp_path,
seed=test_utils.SEED,
)
clf.fit(x=test_utils.TRAIN_CSV_PATH, y="survived")
assert "column_names and column_types are mismatched." in str(info.value)
def test_structured_data_clf_convert_csv_to_df_and_np(fit, tmp_path):
auto_model = ak.StructuredDataClassifier(directory=tmp_path,
seed=test_utils.SEED)
auto_model.fit(
x=test_utils.TRAIN_CSV_PATH,
y="survived",
epochs=2,
validation_data=(test_utils.TEST_CSV_PATH, "survived"),
)
_, kwargs = fit.call_args_list[0]
assert isinstance(kwargs["x"], pd.DataFrame)
assert isinstance(kwargs["y"], np.ndarray)
def train_autokeras(X, Y, x, y, modelfile, max_trials=10, epochs=600):
from sklearn import metrics
import autokeras as ak
from sklearn.preprocessing import MinMaxScaler
clf = ak.StructuredDataClassifier(overwrite=True, max_trials=max_trials)
clf.fit(X, Y, validation_data=(x, y), epochs=epochs)
model = clf.export_model()
model.save(modelfile)
akpred = clf.predict(x)
acc = getaccuracy(akpred, y)
return acc, model
