def test_func(self):
if not os.path.exists("./test_cache"):
os.mkdir("./test_cache")
if not os.path.exists("./test_params"):
os.mkdir("./test_params")
df = pd.DataFrame([[1, 2, 3], [4, 5, 6]], columns=["a", "b", "c"])
prep = preprep.Preprep("./test_cache", "./test_params")
prep = prep.add(lambda x: x * 4)
prep = prep.add(lambda x: x * 2)
prep = prep.add(lambda x: x * 3)
ret1 = prep.fit_gene(df)
ret2 = prep.fit_gene(df)
df_true = pd.DataFrame([[24, 48, 72], [96, 120, 144]],
columns=["a", "b", "c"])
assert_frame_equal(ret1, ret2)
assert_frame_equal(ret1, df_true)
prep = preprep.Preprep("./test_cache", "./test_params")
prep = prep.add(f1, name="unit_1")
prep = prep.add(f1)
prep = prep.add(f1)
prep.fit_gene(df)
shutil.rmtree("./test_cache")
shutil.rmtree("./test_params")
def test_connect(self):
if not os.path.exists("./test_cache"):
os.mkdir("./test_cache")
if not os.path.exists("./test_params"):
os.mkdir("./test_params")
df = pd.DataFrame([[1, 2, 3], [4, 5, 6]], columns=["a", "b", "c"])
prep = preprep.Preprep("./test_cache", "./test_params")
prep1 = prep.add(lambda x: x * 4)
prep2 = prep.add(lambda x: x * 4)
prep = preprep.Connect(lambda x, y: x + y, [prep2, prep1],
cache_format="csv")
ret = prep.fit_gene(df)
df_true = pd.DataFrame([[1, 2, 3], [4, 5, 6]], columns=["a", "b", "c"
]) * 8
assert_frame_equal(ret, df_true)
prep = preprep.Preprep("./test_cache", "./test_params")
prep1 = prep.add(lambda x: x * 4)
prep = preprep.Connect(lambda x, y: x + y, [prep, prep1],
cache_format="csv")
ret = prep.fit_gene(df)
df_true = pd.DataFrame([[1, 2, 3], [4, 5, 6]], columns=["a", "b", "c"
]) * 5
assert_frame_equal(ret, df_true)
shutil.rmtree("./test_cache")
shutil.rmtree("./test_params")
def test_op(self):
if not os.path.exists("./test_cache"):
os.mkdir("./test_cache")
if not os.path.exists("./test_params"):
os.mkdir("./test_cache")
df = pd.DataFrame([[1, 2, 3], [4, 5, 6]], columns=["a", "b", "c"])
op = TestOp1()
io = StringIO()
sys.stdout = io
prep = preprep.Preprep("./test_cache", "./test_params")
prep = prep.add(op, name="1")
prep = prep.add(op, name="2")
prep = prep.add(op, name="3")
io = StringIO()
sys.stdout = io
ret1 = prep.fit_gene(df, verbose=True)
sys.stdout = sys.__stdout__
stdout_log1 = io.getvalue()
io = StringIO()
sys.stdout = io
ret2 = prep.fit_gene(df, verbose=True)
sys.stdout = sys.__stdout__
stdout_log2 = io.getvalue()
ret3 = prep.gene(df)
assert_frame_equal(ret1, ret2)
assert_frame_not_equal(ret1, ret3)
self.assertNotEqual(stdout_log1, stdout_log2)
op2 = TestOp2()
io = StringIO()
sys.stdout = io
prep = preprep.Preprep("./test_cache", "./test_params")
prep = prep.add(op2, name="1")
prep = prep.add(op2, name="2")
prep = prep.add(op2, name="3")
prep.fit_gene(df, verbose=True)
sys.stdout = sys.__stdout__
stdout_log3 = io.getvalue()
self.assertNotEqual(stdout_log1, stdout_log3)
op2 = TestOp2()
op2.a = 100
io = StringIO()
sys.stdout = io
prep = preprep.Preprep("./test_cache", "./test_params")
prep = prep.add(op2, name="1")
prep = prep.add(op2, name="2")
prep = prep.add(op2, name="3")
prep.fit_gene(df, verbose=True)
sys.stdout = sys.__stdout__
stdout_log4 = io.getvalue()
self.assertNotEqual(stdout_log3, stdout_log4)
shutil.rmtree("./test_cache")
shutil.rmtree("./test_params")
def test_multiinput(self):
if not os.path.exists("./test_cache"):
os.mkdir("./test_cache")
if not os.path.exists("./test_params"):
os.mkdir("./test_params")
df1 = pd.DataFrame([[2, 2, 3], [4, 5, 5]])
df2 = pd.DataFrame([[1, 2, 3], [4, 5, 5]])
prep1 = preprep.Preprep("./test_cache", "./test_params", "input_1")
prep1 = prep1.add(lambda x: x * 4)
prep2 = preprep.Preprep("./test_cache", "./test_params", "input_2")
prep2 = prep2.add(lambda x: x * 4)
prep = preprep.Connect(lambda x, y: x + y, [prep1, prep2],
cache_format="csv")
prep.fit_gene({"input_1": df1, "input_2": df2})
#prep.fit_gene({"input_1":df1,"input_2":df2})
shutil.rmtree("./test_cache")
def test_multivalue_output(self):
if not os.path.exists("./test_cache"):
os.mkdir("./test_cache")
if not os.path.exists("./test_params"):
os.mkdir("./test_parms")
df = pd.DataFrame([[1, 2, 3], [4, 5, 6]], columns=["a", "b", "c"])
prep = preprep.Preprep("./test_cache", "./test_params")
prep = prep.add(f2)
ret11, ret12 = prep.fit_gene(df)
ret21, ret22 = prep.fit_gene(df)
assert_frame_equal(ret11, ret12)
assert_frame_equal(ret11, ret21)
prep = preprep.Preprep("./test_cache", "./test_params")
prep = prep.add(lambda df: "b")
prep.fit_gene("a")
shutil.rmtree("./test_cache")
def test_train_pred(self):
if not os.path.exists("./test_cache"):
os.mkdir("./test_cache")
if not os.path.exists("./test_params"):
os.mkdir("./test_parms")
df = pd.DataFrame([[1, 2, 3], [4, 5, 6]], columns=["a", "b", "c"])
raw_df = df.copy()
prep = preprep.Preprep("./test_cache", "./test_params")
prep = prep.add(TestOp3(), name="test1")
prep = prep.add(TestOp3(), name="test2")
self.assertRaises(KeyError, lambda: prep.gene(df))
# In case cache file doesn't exist.
ret1 = prep.fit_gene(df)
# In case cache file exists.
prep.fit_gene(raw_df)
ret2 = prep.gene(df)
assert_frame_equal(ret1, ret2)
shutil.rmtree("./test_cache")
shutil.rmtree("./test_params")
def test_feather_support(self):
if not os.path.exists("./test_cache"):
os.mkdir("./test_cache")
if not os.path.exists("./test_params"):
os.mkdir("./test_params")
df = pd.DataFrame([[1, 2, 3], [4, 5, 6]], columns=["a", "b", "c"])
prep = preprep.Preprep("./test_cache", "./test_params")
prep = prep.add(lambda x: x * 4,
name="feather_test",
cache_format="feather")
ret1 = prep.fit_gene(df)
ret2 = prep.fit_gene(df)
df_true = pd.DataFrame([[1, 2, 3], [4, 5, 6]], columns=["a", "b", "c"
]) * 4
assert_frame_equal(ret1, ret2)
assert_frame_equal(ret1, df_true)
shutil.rmtree("./test_cache")
shutil.rmtree("./test_params")