def setup_training_data(self, train_file, target):
# Loading training data
self.train_data = data_io.get_data(setup.get_datasets_path(),
train_file)
# Setting what to predict
self.target_column = target
def setUp(self):
self.data = data_io.get_data(setup.get_datasets_path(),
"titanic_train.csv")
self.data["Nonsense"] = np.nan
self.data = self.data.apply(pd.to_numeric, errors='coerce')
self.data.loc[self.data.shape[0]] = [np.nan] * self.data.shape[1]
self.target = 'Survived'
self.start = self.data.shape
def standard_pred_type(self, filename, target, goal, message):
# Arrange
data = data_io.get_data(setup.get_datasets_path(), filename)
# Act
ml_type = dataset_insight.get_prediction_type(data[target])
# Assert
self.assertEqual(ml_type, goal, msg=message)
def test_merge(self):
# Arrange
filepath = setup.get_datasets_path() + "titanic_train.csv"
full_csv = False
full_rds = False
# Act
merge = mg.Merger()
merge.merge(["Name", filepath, filepath, filepath])
for File in os.listdir(setup.get_datasets_path()):
if File.endswith("full.dataset.csv"):
full_csv = True
os.remove(setup.get_datasets_path() + File)
if File.endswith("full.dataset.rds"):
full_rds = True
os.remove(setup.get_datasets_path() + File)
# Assert
self.assertEqual(full_csv, True)
self.assertEqual(full_rds, True)
def test_scaling(self):
# Arrange
data = data_io.get_data(setup.get_datasets_path(),
"winequality-red.csv")
# Act
target = data_scaling.scale_numeric_data(data)
target = pd.DataFrame.as_matrix(target)
# Assert
self.assertAlmostEqual(target.mean(), 0, 5)
self.assertAlmostEqual(target.std(), 1, 5)
def test_data_split(self):
# Arrange
data = data_io.get_data(setup.get_datasets_path(),
"winequality-red.csv")
# Act
x_train, x_test, y_train, y_test = data_splitting.get_train_test(
data, "quality")
# Assert
self.assertEqual(len(x_train[:, 0]), len(y_train))
self.assertEqual(len(x_test[:, 0]), len(y_test))
self.assertAlmostEqual(len(y_train) / len(y_test), 4, 2)
class TestSplitter(TestCase):
# test_comma_split tests the functionality of the splitter, with the delimiter as a comma.
test_csv = False
train_csv = False
test_rds = False
train_rds = False
filepath = setup.get_datasets_path()
def helper(self):
self.test_csv = False
self.train_csv = False
self.test_rds = False
self.train_rds = False
for File in os.listdir(self.filepath):
if File.endswith(".testing.data.csv"):
self.test_csv = True
os.remove(self.filepath + File)
if File.endswith(".training.data.csv"):
self.train_csv = True
os.remove(self.filepath + File)
if File.endswith(".training.data.rds"):
self.train_rds = True
os.remove(self.filepath + File)
if File.endswith(".testing.data.rds"):
self.test_rds = True
os.remove(self.filepath + File)
def split_body(self, path, delim):
#Arrange
splitter = sp.Splitter()
splitter.Split(path, delim)
#Act
self.helper()
#Assert
self.assertEquals = (self.test_csv, True)
self.assertEquals = (self.train_csv, True)
self.assertEquals = (self.test_rds, True)
self.assertEquals = (self.train_rds, True)
def test_comma_split(self):
self.split_body(self.filepath + "titanic_train.csv", ",")
# test_semicolon_split tests the functionality of the splitter, with the delimiter as a semicolon.
def test_semicolon_split(self):
self.split_body(self.filepath + "winequality-red.csv", ";")
for item in dirs:
if imghdr.what(path +
item) in imgfiles: # if the file has the correct type
img = Image.open(path + item)
height, width = img.size
if height == size and width == size:
size_count = size_count + 1
else:
bad_count = bad_count + 1
return size_count, bad_count
# Performing Testing
path = get_datasets_path()
# img count
img, not_img = get_file_count(path)
print("This classifier only works on image files.")
print("You have " + str(img) + " image files and " + str(not_img) +
" non-image files.")
# img sizes
size = 28
good_size, bad_size = get_img_size(path, size)
print("Each image in the path must have same the length and width.")
print("You have " + str(img) + " files of size " + str(size) + " and " +
str(bad_size) + " files not of that size.")
# scoring
def setUp(self):
self.data = data_io.get_data(setup.get_datasets_path(),
"winequality-red.csv")
self.target = 'quality'
self.start = self.data.shape
def setUp(self):
self.data = data_io.get_data(setup.get_datasets_path(), "titanic_train.csv")
def setUp(self):
self.data = data_io.get_data(setup.get_datasets_path(),
"winequality-red.csv")
self.strategy = filler_strategy.FillerStrategy(self.data)
def standard_delim(self, filename, target_delim):
# Act
delim = dataset_insight.get_delimiter(setup.get_datasets_path() + filename)
# Assert
self.assertEqual(delim, target_delim, "Delimiter detected incorrectly.")
def setUp(self):
self.data = data_io.get_data(setup.get_datasets_path(), "titanic_train.csv")
self.data["Nonsense"] = np.nan
self.data["Filled"] = 1
self.data = self.data.apply(pd.to_numeric,errors='coerce')
def setup_test_data(self, test_file):
# Loading testing data
self.test_data = data_io.get_data(setup.get_datasets_path(), test_file)