def test_fit_csv_file_1(_):
constant.MAX_ITER_NUM = 1
constant.MAX_MODEL_NUM = 1
constant.EPOCHS_EACH = 1
constant.N_NEIGHBORS = 1
path = 'tests/resources'
clf = ImageClassifier(verbose=False, path=os.path.join(path, "temp"), resume=False)
clf.fit(csv_file_path=os.path.join(path, "images_test/images_name.csv"),
images_path=os.path.join(path, "images_test/Color_images"))
img_file_name, y_train = read_csv_file(csv_file_path=os.path.join(path, "images_test/images_name.csv"))
x_test = read_images(img_file_name, images_dir_path=os.path.join(path, "images_test/Color_images"))
results = clf.predict(x_test)
assert len(clf.load_searcher().history) == 1
assert len(results) == 5
clean_dir(os.path.join(path, "temp"))
clf = ImageClassifier(verbose=False, path=os.path.join(path, "temp"), resume=True)
clf.fit(csv_file_path=os.path.join(path, "images_test/images_name.csv"),
images_path=os.path.join(path, "images_test/Black_white_images"))
img_file_name, y_train = read_csv_file(csv_file_path=os.path.join(path, "images_test/images_name.csv"))
x_test = read_images(img_file_name, images_dir_path=os.path.join(path, "images_test/Black_white_images"))
results = clf.predict(x_test)
assert len(clf.load_searcher().history) == 1
assert len(results) == 5
clean_dir(os.path.join(path, "temp"))
def test_fit_predict_evalute_regression():
clean_dir(TEST_TEMP_DIR)
clf = TabularRegressor(path=TEST_TEMP_DIR)
nsample = 10000
[ntime, nnum, ncat, nmvc] = [3, 15, 3, 3]
datainfo = {'loaded_feat_types': [ntime, nnum, ncat, nmvc]}
x_num = np.random.random([nsample, nnum])
x_time = np.random.random([nsample, ntime])
x_cat = np.random.randint(0, 200, [nsample, ncat])
x_mvc = np.random.randint(0, 10, [nsample, nmvc])
train_x = np.concatenate([x_num, x_time, x_cat, x_mvc], axis=1)
train_y = train_x[:, 5]
clf.fit(train_x, train_y, datainfo=datainfo)
results = clf.predict(train_x)
assert len(results) == len(train_x)
score = clf.evaluate(train_x, train_y)
assert score >= 0.0
# test different model loading in predict
clf.clf = None
results = clf.predict(train_x)
assert len(results) == len(train_x)
clf.clf = None
clf.save_filename = None
results = clf.predict(train_x)
assert len(results) == len(train_x)
clean_dir(TEST_TEMP_DIR)
def test_fit_evalute_predict_binary_classification():
clean_dir(TEST_TEMP_DIR)
clf = TabularClassifier(path=TEST_TEMP_DIR)
nsample = 59
feat_exist_ind = [0, 5]
for ntime in feat_exist_ind:
for nnum in feat_exist_ind:
for ncat in feat_exist_ind:
datainfo = np.array(['TIME'] * ntime + ['NUM'] * nnum + ['CAT'] * ncat)
num_feat = ntime + nnum + ncat
x_num = np.random.random([nsample, nnum])
x_time = np.random.random([nsample, ntime])
x_cat = np.random.randint(0, 40, [nsample, ncat])
train_x = np.concatenate([x_num, x_time, x_cat], axis=1)
train_y = np.random.randint(0, 2, nsample)
print(datainfo)
if num_feat == 0:
with pytest.raises(ValueError):
clf.fit(train_x, train_y, data_info=datainfo)
else:
clf.fit(train_x, train_y, data_info=datainfo)
results = clf.predict(train_x)
assert all(map(lambda result: result in train_y, results))
score = clf.evaluate(train_x, train_y)
assert score >= 0.0
clean_dir(TEST_TEMP_DIR)
def test_fit_evalute_predict_binary_classification():
train_x = None
train_y = None
clean_dir(TEST_TEMP_DIR)
clf = TabularClassifier(path=TEST_TEMP_DIR)
nsample = 59
feat_exist_ind = [0, 5]
for ntime in feat_exist_ind:
for nnum in feat_exist_ind:
for ncat in feat_exist_ind:
datainfo = np.array(['TIME'] * ntime + ['NUM'] * nnum +
['CAT'] * ncat)
num_feat = ntime + nnum + ncat
x_num = np.random.random([nsample, nnum])
x_time = np.random.random([nsample, ntime])
x_cat = np.random.randint(0, 40, [nsample, ncat])
train_x = np.concatenate([x_num, x_time, x_cat], axis=1)
train_y = np.random.randint(0, 2, nsample)
print(datainfo)
if num_feat == 0:
with pytest.raises(ValueError):
clf.fit(train_x, train_y, data_info=datainfo)
else:
clf.fit(train_x, train_y, data_info=datainfo)
results = clf.predict(train_x)
assert all(map(lambda result: result in train_y, results))
score = clf.evaluate(train_x, train_y)
assert score >= 0.0
clean_dir(TEST_TEMP_DIR)
def test_timeout_resume(_, _1):
Constant.MAX_ITER_NUM = 1
# make it impossible to complete within 10sec
Constant.MAX_MODEL_NUM = 1000
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
test_x = np.random.rand(100, 25, 25, 1)
path = 'tests/resources/temp'
clean_dir(path)
clf = TextClassifier(path=path, verbose=False, resume=False)
clf.n_epochs = 100
clf.fit(train_x, train_y, time_limit=2)
history_len = len(clf.load_searcher().history)
assert history_len != 0
results = clf.predict(test_x)
assert len(results) == 100
clf = TextClassifier(path=path, verbose=False, resume=True)
assert len(clf.load_searcher().history) == history_len
Constant.MAX_MODEL_NUM = history_len + 1
clf.fit(train_x, train_y)
assert len(clf.load_searcher().history) == history_len + 1
results = clf.predict(test_x)
assert len(results) == 100
clean_dir(path)
def test_temp_folder_generator(_):
path = 'tests/resources/temp'
clean_dir(path)
path = temp_folder_generator()
assert path == "tests/resources/temp/autokeras"
path = 'tests/resources/temp'
clean_dir(path)
def test_object_detection():
detector = ObjectDetector()
detector.load()
img_path = 'tests/resources/images_test/od.JPG'
result = detector.predict(img_path, TEST_TEMP_DIR)
assert isinstance(result, list)
clean_dir(TEST_TEMP_DIR)
def test_temp_folder_generator(_):
path = 'tests/resources/temp'
clean_dir(path)
path = temp_folder_generator()
assert path == "tests/resources/temp/autokeras"
path = 'tests/resources/temp'
clean_dir(path)
def test_save_continue(_, _1):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 1
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
test_x = np.random.rand(100, 25, 25, 1)
clean_dir(TEST_TEMP_DIR)
clf = ImageClassifier(path=TEST_TEMP_DIR, verbose=False, resume=False)
clf.n_epochs = 100
clf.fit(train_x, train_y)
assert len(clf.cnn.searcher.history) == 1
Constant.MAX_MODEL_NUM = 2
clf = ImageClassifier(verbose=False, path=TEST_TEMP_DIR, resume=True)
clf.fit(train_x, train_y)
results = clf.predict(test_x)
assert len(results) == 100
assert len(clf.cnn.searcher.history) == 2
Constant.MAX_MODEL_NUM = 1
clf = ImageClassifier(verbose=False, path=TEST_TEMP_DIR, resume=False)
clf.fit(train_x, train_y)
results = clf.predict(test_x)
assert len(results) == 100
assert len(clf.cnn.searcher.history) == 1
clean_dir(TEST_TEMP_DIR)
def test_save_continue(_):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 1
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
test_x = np.random.rand(100, 25, 25, 1)
path = 'tests/resources/temp'
clean_dir(path)
clf = ImageClassifier(path=path, verbose=False, resume=False)
clf.n_epochs = 100
clf.fit(train_x, train_y)
assert len(clf.load_searcher().history) == 1
Constant.MAX_MODEL_NUM = 2
clf = ImageClassifier(verbose=False, path=path, resume=True)
clf.fit(train_x, train_y)
results = clf.predict(test_x)
assert len(results) == 100
assert len(clf.load_searcher().history) == 2
Constant.MAX_MODEL_NUM = 1
clf = ImageClassifier(verbose=False, path=path, resume=False)
clf.fit(train_x, train_y)
results = clf.predict(test_x)
assert len(results) == 100
assert len(clf.load_searcher().history) == 1
clean_dir(path)
def test_export_keras_model(_):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 1
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
test_x = np.random.rand(100, 25, 25, 1)
path = 'tests/resources/temp'
clean_dir(path)
clf = ImageClassifier(path=path, verbose=False, resume=False)
clf.n_epochs = 100
clf.fit(train_x, train_y)
score = clf.evaluate(train_x, train_y)
assert score <= 1.0
model_file_name = os.path.join(path, 'test_keras_model.h5')
clf.export_keras_model(model_file_name)
from keras.models import load_model
model = load_model(model_file_name)
results = model.predict(test_x)
assert len(results) == len(test_x)
del model, results, model_file_name
model_file_name = os.path.join(path, 'test_autokeras_model.pkl')
clf.export_autokeras_model(model_file_name)
from autokeras.utils import pickle_from_file
model = pickle_from_file(model_file_name)
results = model.predict(test_x)
assert len(results) == len(test_x)
score = model.evaluate(train_x, train_y)
assert score <= 1.0
clean_dir(path)
def test_save_continue(_, _1, _2):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 1
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
train_x = np.random.rand(100, 25, 25)
train_y = np.random.randint(0, 5, 100)
test_x = np.random.rand(100, 25, 25)
clean_dir(TEST_TEMP_DIR)
clf = TextClassifier(path=TEST_TEMP_DIR, verbose=False, resume=False)
clf.n_epochs = 100
clf.fit(train_x, train_y, time_limit=5)
assert len(clf.cnn.searcher.history) == 1
Constant.MAX_MODEL_NUM = 2
clf = TextClassifier(verbose=False, path=TEST_TEMP_DIR, resume=True)
clf.fit(train_x, train_y)
results = clf.predict(test_x)
assert len(results) == 100
assert len(clf.cnn.searcher.history) == 2
Constant.MAX_MODEL_NUM = 1
clf = TextClassifier(verbose=False, path=TEST_TEMP_DIR, resume=False)
clf.fit(train_x, train_y)
results = clf.predict(test_x)
assert len(results) == 100
assert len(clf.cnn.searcher.history) == 1
clean_dir(TEST_TEMP_DIR)
def test_timout():
path = 'tests/resources/temp'
clean_dir(path)
clf = ImageClassifier(path=path, verbose=False)
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y, time_limit=1)
clean_dir(path)
def test_greedy_searcher_sp(_, _1, _2, _3):
train_data, test_data = get_classification_data_loaders()
clean_dir(TEST_TEMP_DIR)
searcher = GreedySearcher(3, (28, 28, 3), verbose=False, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[CnnGenerator, CnnGenerator])
for _ in range(2):
searcher.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(searcher.history) == 2
def test_greedy_searcher_mlp(_, _1, _2):
train_data, test_data = get_classification_data_loaders_mlp()
clean_dir(TEST_TEMP_DIR)
generator = GreedySearcher(3, (28,), verbose=False, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[MlpGenerator, MlpGenerator])
for _ in range(2):
generator.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(generator.history) == 2
def test_greedy_searcher_mlp(_, _1, _2):
train_data, test_data = get_classification_data_loaders_mlp()
clean_dir(TEST_TEMP_DIR)
generator = GreedySearcher(3, (28,), verbose=False, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[MlpGenerator, MlpGenerator])
for _ in range(2):
generator.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(generator.history) == 2
def test_greedy_searcher_sp(_, _1, _2, _3):
train_data, test_data = get_classification_data_loaders()
clean_dir(TEST_TEMP_DIR)
searcher = GreedySearcher(3, (28, 28, 3), verbose=False, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[CnnGenerator, CnnGenerator])
for _ in range(2):
searcher.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(searcher.history) == 2
def test_bayesian_searcher(_, _1):
train_data, test_data = get_processed_data()
clean_dir(default_test_path)
generator = BayesianSearcher(3, (28, 28, 3), verbose=False, path=default_test_path)
Constant.N_NEIGHBOURS = 1
Constant.T_MIN = 0.8
for _ in range(2):
generator.search(train_data, test_data)
clean_dir(default_test_path)
assert len(generator.history) == 2
def test_model_trainer_classification():
model = CnnGenerator(3, (28, 28, 3)).generate().produce_model()
train_data, test_data = get_classification_data_loaders()
ModelTrainer(model,
train_data=train_data,
test_data=test_data,
metric=Accuracy,
loss_function=classification_loss,
verbose=True,
path=TEST_TEMP_DIR).train_model(max_iter_num=3)
clean_dir(TEST_TEMP_DIR)
def test_cross_validate(_):
constant.MAX_MODEL_NUM = 2
path = 'tests/resources/temp'
clean_dir(path)
clf = ImageClassifier(path=path, verbose=False, searcher_args={'trainer_args': {'max_iter_num': 0}})
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
results = clf.cross_validate(train_x, train_y, 2, trainer_args={'max_iter_num': 0})
clean_dir(path)
assert len(results) == 2
def test_out_of_memory(_, _1):
train_data, test_data = get_classification_data_loaders()
clean_dir(TEST_TEMP_DIR)
searcher = Searcher(3, (28, 28, 3), verbose=False, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss)
Constant.N_NEIGHBOURS = 1
Constant.T_MIN = 0.8
for _ in range(4):
searcher.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(searcher.history) == 0
def test_model_trainer_regression():
model = CnnGenerator(1, (28, 28, 3)).generate().produce_model()
train_data, test_data = get_regression_data_loaders()
ModelTrainer(model,
train_data=train_data,
test_data=test_data,
metric=MSE,
loss_function=regression_loss,
verbose=False,
path=TEST_TEMP_DIR).train_model(max_iter_num=3)
clean_dir(TEST_TEMP_DIR)
def test_bayesian_searcher(_, _1):
train_data, test_data = get_classification_data_loaders()
clean_dir(default_test_path)
generator = Searcher(3, (28, 28, 3), verbose=False, path=default_test_path, metric=Accuracy,
loss=classification_loss)
Constant.N_NEIGHBOURS = 1
Constant.T_MIN = 0.8
for _ in range(2):
generator.search(train_data, test_data)
clean_dir(default_test_path)
assert len(generator.history) == 2
def test_bayesian_searcher_mlp(_, _1, _2):
train_data, test_data = get_classification_data_loaders_mlp()
clean_dir(TEST_TEMP_DIR)
generator = Searcher(3, (28,), verbose=False, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[MlpGenerator, MlpGenerator])
Constant.N_NEIGHBOURS = 1
Constant.T_MIN = 0.8
for _ in range(2):
generator.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(generator.history) == 2
def test_bayesian_searcher_mlp(_, _1, _2):
train_data, test_data = get_classification_data_loaders_mlp()
clean_dir(TEST_TEMP_DIR)
generator = BayesianSearcher(3, (28,), verbose=False, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[MlpGenerator, MlpGenerator])
Constant.N_NEIGHBOURS = 1
Constant.T_MIN = 0.8
for _ in range(2):
generator.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(generator.history) == 2
def test_bayesian_searcher_sp(_, _1, _2, _3):
train_data, test_data = get_classification_data_loaders()
clean_dir(TEST_TEMP_DIR)
searcher = BayesianSearcher(3, (28, 28, 3), verbose=False, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[CnnGenerator, CnnGenerator])
Constant.N_NEIGHBOURS = 1
Constant.T_MIN = 0.8
for _ in range(2):
searcher.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(searcher.history) == 2
def test_grid_searcher(_, _1, _2):
train_data, test_data = get_classification_data_loaders()
clean_dir(TEST_TEMP_DIR)
searcher = GridSearcher(3, (28, 28, 3), verbose=True, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[CnnGenerator, CnnGenerator])
Constant.N_NEIGHBOURS = 1
Constant.T_MIN = 0.8
print(len(searcher.get_search_dimensions()))
for _ in range(len(searcher.get_search_dimensions())):
searcher.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(searcher.history) == len(searcher.search_dimensions)
def test_face_detector():
img_file, out_file = 'tests/resources/images_test/face_detector.jpg', os.path.join(
TEST_TEMP_DIR, 'output.jpg')
if os.path.exists(out_file):
os.remove(out_file)
face_detection = FaceDetector()
bboxs1, landmarks1 = face_detection.predict(img_file, out_file)
assert os.path.exists(out_file)
bboxs2, landmarks2 = face_detection.predict(img_file)
assert bboxs1.shape == bboxs2.shape == (
11, 5) and landmarks1.shape == landmarks2.shape == (11, 10)
clean_dir(TEST_TEMP_DIR)
def test_exception_handling(_, _2):
train_data, test_data = get_classification_data_loaders()
clean_dir(TEST_TEMP_DIR)
Constant.N_NEIGHBOURS = 2
Constant.T_MIN = 0.8
Constant.BETA = 1
generator = BayesianSearcher(3, (28, 28, 3), verbose=True, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[CnnGenerator, ResNetGenerator])
for _ in range(3):
generator.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(generator.history) == 0
def test_timeout(_):
# Constant.MAX_MODEL_NUM = 4
Constant.SEARCH_MAX_ITER = 1000
Constant.T_MIN = 0.0001
Constant.DATA_AUGMENTATION = False
clean_dir(TEST_TEMP_DIR)
clf = ImageClassifier(path=TEST_TEMP_DIR, verbose=False)
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
with pytest.raises(TimeoutError):
clf.fit(train_x, train_y, time_limit=0)
clean_dir(TEST_TEMP_DIR)
def test_timout(_, _1):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 4
Constant.SEARCH_MAX_ITER = 1
Constant.DATA_AUGMENTATION = False
path = 'tests/resources/temp'
clean_dir(path)
clf = ImageClassifier(path=path, verbose=False)
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y, time_limit=1)
clean_dir(path)
def test_fit_predict(_, _1):
constant.MAX_ITER_NUM = 2
constant.MAX_MODEL_NUM = 2
constant.EPOCHS_EACH = 1
path = 'tests/resources/temp'
clf = ImageClassifier(path=path, verbose=False)
train_x = np.array([[[1], [2]], [[3], [4]]])
train_y = np.array(['a', 'b'])
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert all(map(lambda result: result in np.array(['a', 'b']), results))
clean_dir(path)
def test_grid_searcher(_, _1, _2):
train_data, test_data = get_classification_data_loaders()
clean_dir(TEST_TEMP_DIR)
searcher = GridSearcher(3, (28, 28, 3), verbose=True, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[CnnGenerator, CnnGenerator])
Constant.N_NEIGHBOURS = 1
Constant.T_MIN = 0.8
print(len(searcher.get_search_dimensions()))
for _ in range(len(searcher.get_search_dimensions())):
searcher.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(searcher.history) == len(searcher.search_dimensions)
def test_fit_predict_save(_, _1):
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
for Model in [PredefinedResnet, PredefinedDensenet]:
clf = Model(verbose=True)
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert all(map(lambda result: result in train_y, results))
score = clf.evaluate(train_x, train_y)
assert score <= 1.0
clean_dir(TEST_TEMP_DIR)
def test_timeout(_, _1):
# Constant.MAX_MODEL_NUM = 4
Constant.SEARCH_MAX_ITER = 1000
Constant.T_MIN = 0.0001
Constant.DATA_AUGMENTATION = False
clean_dir(TEST_TEMP_DIR)
clf = TextClassifier(path=TEST_TEMP_DIR, verbose=False)
train_x = np.random.rand(100, 25, 25)
train_y = np.random.randint(0, 5, 100)
with pytest.raises(TimeoutError):
clf.fit(train_x, train_y, time_limit=0)
clean_dir(TEST_TEMP_DIR)
def test_extract_data_info():
clean_dir(TEST_TEMP_DIR)
tp = TabularPreprocessor()
nsample = 10000
half_sample = 5000
[tp.n_num, tp.n_cat] = [5, 1]
x_num = np.random.random([nsample, tp.n_num])
x_cat = np.array(['a'] * half_sample + ['b'] * half_sample).reshape([nsample, 1])
raw_x = np.concatenate([x_num, x_cat], axis=1)
data_info = tp.extract_data_info(raw_x)
ground_truth = np.array(['NUM'] * tp.n_num + ['CAT'] * tp.n_cat)
assert (data_info == ground_truth).all()
clean_dir(TEST_TEMP_DIR)
def test_timeout(_):
# Constant.MAX_MODEL_NUM = 4
Constant.SEARCH_MAX_ITER = 1000
Constant.T_MIN = 0.0001
Constant.DATA_AUGMENTATION = False
path = 'tests/resources/temp'
clean_dir(path)
clf = TextClassifier(path=path, verbose=False)
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
with pytest.raises(TimeoutError):
clf.fit(train_x, train_y, time_limit=0)
clean_dir(path)
def test_fit_predict(_, _1, _2):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 4
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
clean_dir(TEST_TEMP_DIR)
clf = TextClassifier(path=TEST_TEMP_DIR, verbose=True)
train_x = np.random.rand(100, 25, 25)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert all(map(lambda result: result in train_y, results))
clean_dir(TEST_TEMP_DIR)
def test_fit_predict2(_, _1):
path = 'tests/resources/temp'
clf = ImageClassifier(path=path, verbose=False)
constant.MAX_ITER_NUM = 1
constant.MAX_MODEL_NUM = 1
constant.EPOCHS_EACH = 1
train_x = np.random.rand(100, 25, 1)
test_x = np.random.rand(100, 25, 1)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
results = clf.predict(test_x)
assert len(results) == 100
clean_dir(path)
def test_out_of_memory(_, _2):
train_data, test_data = get_classification_data_loaders()
clean_dir(TEST_TEMP_DIR)
Constant.N_NEIGHBOURS = 2
Constant.SEARCH_MAX_ITER = 0
Constant.T_MIN = 0.8
Constant.BETA = 1
generator = BayesianSearcher(3, (28, 28, 3), verbose=True, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[CnnGenerator, ResNetGenerator])
for _ in range(3):
generator.search(train_data, test_data)
clean_dir(TEST_TEMP_DIR)
assert len(generator.history) == 0
def test_model_trainer_timout():
model = CnnGenerator(3, (28, 28, 3)).generate().produce_model()
timeout = 1
train_data, test_data = get_classification_data_loaders()
with pytest.raises(TimeoutError):
ModelTrainer(model,
train_data=train_data,
test_data=test_data,
metric=Accuracy,
loss_function=classification_loss,
verbose=True,
path=TEST_TEMP_DIR).train_model(max_iter_num=300, timeout=timeout)
clean_dir(TEST_TEMP_DIR)
def test_evaluate(_, _1, _2):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 1
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
train_x = np.random.rand(100, 25, 25)
train_y = np.random.randint(0, 5, 100)
clean_dir(TEST_TEMP_DIR)
clf = TextClassifier(path=TEST_TEMP_DIR, verbose=False, resume=False)
clf.n_epochs = 100
clf.fit(train_x, train_y)
score = clf.evaluate(train_x, train_y)
clean_dir(TEST_TEMP_DIR)
assert score <= 1.0
def test_fit_predict_regression(_, _1):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 4
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
Constant.DATA_AUGMENTATION = False
clean_dir(TEST_TEMP_DIR)
clf = ImageRegressor(path=TEST_TEMP_DIR, verbose=False)
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert len(results) == len(train_x)
clean_dir(TEST_TEMP_DIR)
def test_fit_generate(_):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 4
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
Constant.DATA_AUGMENTATION = False
image_path, size = TEST_TEMP_DIR, 32
clean_dir(image_path)
dcgan = DCGAN(gen_training_result=(image_path, size))
train_x = np.random.rand(100, 32, 32, 3)
dcgan.fit(train_x)
clean_dir(image_path)
noise = np.random.randn(32, 100, 1, 1).astype('float32')
dcgan.generate(noise)
def test_fit_csv_file(_, _1, _2):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 1
Constant.SEARCH_MAX_ITER = 1
path = 'tests/resources'
clf = ImageClassifier(verbose=False, path=os.path.join(path, "temp"), resume=False)
x_train, y_train = load_image_dataset(csv_file_path=os.path.join(path, "images_test/images_name.csv"),
images_path=os.path.join(path, "images_test/Color_images"))
clf.fit(x_train, y_train)
x_test, y_test = load_image_dataset(csv_file_path=os.path.join(path, "images_test/images_name.csv"),
images_path=os.path.join(path, "images_test/Color_images"))
results = clf.predict(x_test)
assert len(clf.cnn.searcher.history) == 1
assert len(results) == 5
clean_dir(os.path.join(path, "temp"))
def test_fit_predict_save(_):
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
for Model in [PredefinedResnet, PredefinedDensenet]:
clf = Model(verbose=True)
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert all(map(lambda result: result in train_y, results))
score = clf.evaluate(train_x, train_y)
assert score <= 1.0
model_path = os.path.join(TEST_TEMP_DIR, 'kerasmodel')
clf.save(model_path)
assert os.path.isfile(model_path)
clean_dir(TEST_TEMP_DIR)
def test_max_acq(_, _1):
train_data, test_data = get_classification_data_loaders()
clean_dir(default_test_path)
Constant.N_NEIGHBOURS = 2
Constant.SEARCH_MAX_ITER = 0
Constant.T_MIN = 0.8
Constant.BETA = 1
generator = Searcher(3, (28, 28, 3), verbose=False, path=default_test_path, metric=Accuracy,
loss=classification_loss)
for _ in range(3):
generator.search(train_data, test_data)
for index1, descriptor1 in enumerate(generator.descriptors):
for descriptor2 in generator.descriptors[index1 + 1:]:
assert edit_distance(descriptor1, descriptor2, 1) != 0
clean_dir(default_test_path)
def test_batch_dataset(_, _1):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 4
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
data_path = 'tests/resources'
clean_dir(TEST_TEMP_DIR)
csv_file_path = os.path.join(data_path, "images_test/images_name.csv")
image_path = os.path.join(data_path, "images_test/Color_images")
train_dataset = BatchDataset(csv_file_path, image_path, has_target=True)
test_dataset = BatchDataset(csv_file_path, image_path, has_target=True)
train_dataloader = DataLoader(train_dataset, batch_size=1, shuffle=False)
test_dataloader = DataLoader(test_dataset, batch_size=1, shuffle=False)
cnn = CnnModule(classification_loss, Accuracy, {}, TEST_TEMP_DIR, True)
cnn.fit(2, (4, 250, 250, 3), train_dataloader, test_dataloader, 12 * 60 * 60)
clean_dir(TEST_TEMP_DIR)
def test_max_acq(_, _2):
train_data, test_data = get_classification_data_loaders()
clean_dir(TEST_TEMP_DIR)
Constant.N_NEIGHBOURS = 2
Constant.SEARCH_MAX_ITER = 0
Constant.T_MIN = 0.8
Constant.BETA = 1
generator = BayesianSearcher(3, (28, 28, 3), verbose=False, path=TEST_TEMP_DIR, metric=Accuracy,
loss=classification_loss, generators=[CnnGenerator, ResNetGenerator])
for _ in range(3):
generator.search(train_data, test_data)
for index1, descriptor1 in enumerate(generator.descriptors):
for descriptor2 in generator.descriptors[index1 + 1:]:
assert edit_distance(descriptor1, descriptor2) != 0.0
clean_dir(TEST_TEMP_DIR)
def test_fit_predict_evalute_regression():
clean_dir(TEST_TEMP_DIR)
clf = TabularRegressor(path=TEST_TEMP_DIR)
nsample = 10000
[ntime, nnum, ncat] = [3, 15, 3]
datainfo = np.array(['TIME'] * ntime + ['NUM'] * nnum + ['CAT'] * ncat)
x_num = np.random.random([nsample, nnum])
x_time = np.random.random([nsample, ntime])
x_cat = np.random.randint(0, 200, [nsample, ncat])
train_x = np.concatenate([x_num, x_time, x_cat], axis=1)
train_y = train_x[:, 5]
clf.fit(train_x, train_y, data_info=datainfo)
results = clf.predict(train_x)
assert len(results) == len(train_x)
score = clf.evaluate(train_x, train_y)
assert score >= 0.0
clean_dir(TEST_TEMP_DIR)
def test_fit_predict_regression(_, _1, _2):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 4
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
Constant.DATA_AUGMENTATION = False
path = TEST_TEMP_DIR
print(os.getcwd())
# for f in os.listdir(path):
# print(f)
clean_dir(path)
clf = TextRegressor(path=path, verbose=False)
train_x = np.random.rand(100, 25, 25)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert len(results) == len(train_x)
clean_dir(path)
def test_final_fit(_, _1, _2, _3):
Constant.LIMIT_MEMORY = True
clean_dir(TEST_TEMP_DIR)
clf = TextClassifier(path=TEST_TEMP_DIR, verbose=False)
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 1
Constant.SEARCH_MAX_ITER = 1
Constant.N_NEIGHBOURS = 1
Constant.T_MIN = 0.8
train_x = np.random.rand(100, 25, 25)
train_y = np.random.randint(0, 5, 100)
test_x = np.random.rand(100, 25, 25)
test_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
clf.final_fit(train_x, train_y, test_x, test_y)
results = clf.predict(test_x)
assert len(results) == 100
clean_dir(TEST_TEMP_DIR)
def test_export_json(_, _1):
train_data, test_data = get_classification_data_loaders()
clean_dir(default_test_path)
generator = Searcher(3, (28, 28, 3), verbose=False, path=default_test_path, metric=Accuracy,
loss=classification_loss)
Constant.N_NEIGHBOURS = 1
Constant.T_MIN = 0.8
for _ in range(3):
generator.search(train_data, test_data)
file_path = os.path.join(default_test_path, 'test.json')
generator.export_json(file_path)
import json
data = json.load(open(file_path, 'r'))
assert len(data['networks']) == 3
assert len(data['tree']['children']) == 2
clean_dir(default_test_path)
assert len(generator.history) == 3
def test_fit_evalute_predict_multiclass_classification():
clean_dir(TEST_TEMP_DIR)
clf = TabularClassifier(path=TEST_TEMP_DIR)
clf.verbose = True
nsample = 10000
[ntime, nnum, ncat] = [11, 15, 13]
datainfo = np.array(['TIME'] * ntime + ['NUM'] * nnum + ['CAT'] * ncat)
x_num = np.random.random([nsample, nnum])
x_time = np.random.random([nsample, ntime])
x_cat = np.random.randint(0, 200, [nsample, ncat])
train_x = np.concatenate([x_num, x_time, x_cat], axis=1)
train_y = np.random.randint(0, 3, nsample)
clf.fit(train_x, train_y, data_info=datainfo)
results = clf.predict(train_x)
assert all(map(lambda result: result in train_y, results))
score = clf.evaluate(train_x, train_y)
assert score >= 0.0
clean_dir(TEST_TEMP_DIR)
def test_fit_predict(_, _1):
Constant.MAX_ITER_NUM = 1
Constant.MAX_MODEL_NUM = 4
Constant.SEARCH_MAX_ITER = 1
Constant.T_MIN = 0.8
Constant.DATA_AUGMENTATION = False
clf = ImageClassifier(path=TEST_TEMP_DIR, verbose=True)
train_x = np.random.rand(100, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert all(map(lambda result: result in train_y, results))
clf = ImageClassifier1D(path=TEST_TEMP_DIR, verbose=True)
train_x = np.random.rand(100, 25, 1)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert all(map(lambda result: result in train_y, results))
clf = ImageClassifier3D(path=TEST_TEMP_DIR, verbose=True)
train_x = np.random.rand(100, 25, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert all(map(lambda result: result in train_y, results))
clf = ImageRegressor1D(path=TEST_TEMP_DIR, verbose=True)
train_x = np.random.rand(100, 25, 1)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert len(results) == len(train_y)
clf = ImageRegressor3D(path=TEST_TEMP_DIR, verbose=True)
train_x = np.random.rand(100, 25, 25, 25, 1)
train_y = np.random.randint(0, 5, 100)
clf.fit(train_x, train_y)
results = clf.predict(train_x)
assert len(results) == len(train_y)
clean_dir(TEST_TEMP_DIR)
