def test_adagrad_mse_batch(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "adagrad_mse_batch.csv"),
range(1))
import adagrad_mse_batch as ex
ex.read_csv = self.read_csv
result = ex.main()
self.assertTrue(np.allclose(result.minimum, testdata))
def test_svm_multiclass_batch(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "svm_multiclass_batch.csv"),
range(1))
import svm_multiclass_batch as ex
ex.read_csv = self.read_csv
(predict_result, _) = ex.main()
self.assertTrue(np.allclose(predict_result.prediction, testdata))
def test_univariate_outlier_batch(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "univariate_outlier_batch.csv"),
range(1))
import univariate_outlier_batch as ex
ex.read_csv = self.read_csv
(_, result) = ex.main()
self.assertTrue(np.allclose(result.weights, testdata))
def test_pca_transform_batch(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "pca_transform_batch.csv"),
range(2))
import pca_transform_batch as ex
ex.read_csv = self.read_csv
_, result = ex.main()
self.assertTrue(np.allclose(result.transformedData, testdata))
def test_log_reg_dense_batch(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "log_reg_dense_batch.csv"),
range(1))
import log_reg_dense_batch as ex
ex.read_csv = self.read_csv
(_, predict_result, _) = ex.main()
self.assertTrue(np.allclose(predict_result.prediction, testdata))
def test_naive_bayes_stream(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "naive_bayes_batch.csv"),
range(1))
import naive_bayes_streaming as ex
ex.read_csv = self.read_csv
(predict_result, _) = ex.main()
self.assertTrue(np.allclose(predict_result.prediction, testdata))
def test_ridge_regression_stream(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "ridge_regression_batch.csv"),
range(2))
import ridge_regression_streaming as ex
ex.read_csv = self.read_csv
(predict_result, _) = ex.main()
self.assertTrue(np.allclose(predict_result.prediction, testdata))
def test_lbfgs_cr_entr_loss_batch(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "lbfgs_cr_entr_loss_batch.csv"),
range(1))
import lbfgs_cr_entr_loss_batch as ex
ex.read_csv = self.read_csv
result = ex.main()
self.assertTrue(np.allclose(result.minimum, testdata))
def test_decision_tree_classification_batch(self):
testdata = np_read_csv(
os.path.join(unittest_data_path,
"decision_tree_classification_batch.csv"), range(1))
import decision_tree_classification_batch as ex
ex.read_csv = self.read_csv
(_, predict_result, _) = ex.main()
self.assertTrue(np.allclose(predict_result.prediction, testdata))
def test_svm_batch(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "svm_batch.csv"), range(1))
import svm_batch as ex
ex.read_csv = self.read_csv
(predict_result, _) = ex.main()
self.assertTrue(
np.absolute(predict_result.prediction - testdata).max() <
np.absolute(predict_result.prediction.max() -
predict_result.prediction.min()) * 0.05)
def test_gradient_boosted_regression_batch(self):
testdata = np_read_csv(
os.path.join(unittest_data_path,
"gradient_boosted_regression_batch.csv"), range(1))
import gradient_boosted_regression_batch as ex
ex.read_csv = self.read_csv
(_, predict_result, _) = ex.main()
#MSE
self.assertTrue(
np.square(predict_result.prediction - testdata).mean() < 1e-2)
def test_svd_batch(self):
import svd_batch as ex
ex.read_csv = self.read_csv
(data, result) = ex.main()
self.assertTrue(
np.allclose(
data,
np.matmul(
np.matmul(result.leftSingularMatrix,
np.diag(result.singularValues[0])),
result.rightSingularMatrix)))
def test_low_order_moms_dense_stream(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "low_order_moms_dense_batch.csv"),
range(10))
import low_order_moms_streaming as ex
ex.read_csv = self.read_csv
res = ex.main()
r = np.vstack(
(res.minimum, res.maximum, res.sum, res.sumSquares,
res.sumSquaresCentered, res.mean, res.secondOrderRawMoment,
res.variance, res.standardDeviation, res.variation))
self.assertTrue(np.allclose(r, testdata))
def test_cosine_distance_batch(self):
testdata = np_read_csv(
os.path.join(unittest_data_path, "cosine_distance_batch.csv"),
range(1))
import cosine_distance_batch as ex
ex.read_csv = self.read_csv
result = ex.main()
r = result.cosineDistance
self.assertTrue(
np.allclose(
np.array([[np.amin(r)], [np.amax(r)], [np.mean(r)],
[np.average(r)]]), testdata))
def test_svd_stream(self):
import svd_streaming as ex
ex.read_csv = self.read_csv
result = ex.main()
data = np.loadtxt("./data/distributed/svd_1.csv", delimiter=',')
for f in [
"./data/distributed/svd_{}.csv".format(i) for i in range(2, 5)
]:
data = np.append(data, np.loadtxt(f, delimiter=','), axis=0)
self.assertTrue(
np.allclose(
data,
np.matmul(
np.matmul(result.leftSingularMatrix,
np.diag(result.singularValues[0])),
result.rightSingularMatrix)))
def call(self, ex):
return ex.main(readcsv=pd_read_csv)
def call(self, ex):
method = 'singlePassCSR' if any(x in ex.__name__ for x in ['low_order_moms', 'covariance']) else 'fastCSR'
if hasattr(ex, 'dflt_method'):
low_order_moms
method = ex.dflt_method.replace('defaultDense', 'fastCSR').replace('Dense', 'CSR')
return ex.main(readcsv=csr_read_csv, method=method)
def test_kdtree_knn_classification_batch(self):
import kdtree_knn_classification_batch as ex
ex.read_csv = self.read_csv
(_, predict_result, test_labels) = ex.main()
self.assertTrue(
np.count_nonzero(test_labels != predict_result.prediction) < 170)
