def test_save_by_csv_using_kwargs(self):
df = pd.DataFrame({
"a": [1, 2],
"b": [2, 3],
"c": [3, 4]
}, index=[9, 10])
name_1 = "test_save_by_csv_using_kwargs_1.csv"
content_1 = '''9,1,2,3
10,2,3,4
'''
saver = AutoSaver(self.test_assets_dir, index=True, header=False)
saver.save(df, name_1, line_terminator="\n")
with open(os.path.join(self.test_assets_dir, name_1), "r", encoding="utf-8") as f:
content = f.read()
self.assertTrue(content, content_1)
name_2 = "test_save_by_csv_using_kwargs_2.csv"
content_2 = '''a;b;c
1;2;3
2;3;4
'''
saver = AutoSaver(index=False)
saver.save(df, os.path.join(self.test_assets_dir, name_2), sep=";")
with open(os.path.join(self.test_assets_dir, name_2), "r", encoding="utf-8") as f:
content = f.read()
self.assertTrue(content, content_2)
saver = AutoSaver(index=True, sep=",")
saver.save(df, os.path.join(self.test_assets_dir, name_2),
sep=";", index=False)
with open(os.path.join(self.test_assets_dir, name_2), "r", encoding="utf-8") as f:
content = f.read()
self.assertTrue(content, content_2)
def test_save_by_csv_param_error(self):
with self.assertRaises(ValueError):
AutoSaver(self.test_assets_dir,
example_path=r"tests/assets/data1.csv", encoding="utf-8")
with self.assertRaises(ValueError):
saver = AutoSaver(self.test_assets_dir,
example_path=r"tests/assets/data1.csv")
saver.save("Nothing", "some_name.csv", encoding="utf-8")
def test_GBK(self):
# REMINDER: do not use self.get_csv_content when testing GBK.
# It defaultly uses utf-8.
saver = AutoSaver(save_dir=self.test_assets_dir,
example_path=r"tests/assets/saver-example-GBK.csv")
target = '''index,value
0,1.0
1,0.0
2,1.5
3,0.5
'''
saver.save(self.np_1d, self.filename)
with open(os.path.join(self.test_assets_dir, self.filename), "r", encoding="GBK") as f:
content = f.read()
self.assertEqual(content, target)
saver.save(self.np_2d_1val, self.filename)
with open(os.path.join(self.test_assets_dir, self.filename), "r", encoding="GBK") as f:
content = f.read()
self.assertEqual(content, target)
saver.save(self.pd_df_2d, self.filename)
with open(os.path.join(self.test_assets_dir, self.filename), "r", encoding="GBK") as f:
content = f.read()
self.assertEqual(content, target)
def test_saving_two_value(self):
saver = AutoSaver(save_dir=self.test_assets_dir,
example_path=r"tests/assets/saver-example-two-value.csv")
arr = np.array([
[1, 1.5],
[0.5, 0.7],
[0.2, 0.3]
])
df = pd.DataFrame(arr)
target = '''ix,val1,val2
0,1.0,1.5
1,0.5,0.7
2,0.2,0.3
'''
saver.save(arr, self.filename)
with open(os.path.join(self.test_assets_dir, self.filename), "r", encoding="utf-8") as f:
content = f.read()
self.assertEqual(content, target)
saver.save(df, self.filename)
with open(os.path.join(self.test_assets_dir, self.filename), "r", encoding="utf-8") as f:
content = f.read()
self.assertEqual(content, target)
def test_saving_one_header_no_value(self):
saver = AutoSaver(save_dir=self.test_assets_dir,
example_path=r"tests/assets/saver-example-one-header-no-value.csv")
target = ''',val
0,1.0
1,0.0
2,1.5
3,0.5
'''
self.run_all_X_s(saver, target)
def test_saving_without_index(self):
saver = AutoSaver(save_dir=self.test_assets_dir,
example_path=r"tests/assets/saver-example-without-index.csv")
target = '''1.0
0.0
1.5
0.5
'''
self.run_all_X_s(saver, target)
def test_saving_without_head(self):
saver = AutoSaver(save_dir=self.test_assets_dir,
example_path=r"tests/assets/saver-example-without-head.csv")
target = '''0,1.0
1,0.0
2,1.5
3,0.5
'''
self.run_all_X_s(saver, target)
def test_ix_starts_1(self):
saver = AutoSaver(save_dir=self.test_assets_dir,
example_path=r"tests/assets/saver-example-ix-starts-1.csv")
target = '''index,value
1,1.0
2,0.0
3,1.5
4,0.5
'''
self.run_all_X_s(saver, target)
def test_basic(self):
saver = AutoSaver(save_dir=self.test_assets_dir,
example_path=r"tests/assets/saver-example-basic.csv")
target = '''index,value
0,1.0
1,0.0
2,1.5
3,0.5
'''
self.run_all_X_s(saver, target)
def test_no_filename(self):
old_files = os.listdir(self.test_assets_dir)
saver = AutoSaver(self.test_assets_dir)
saver.save(pd.DataFrame())
files = os.listdir(self.test_assets_dir)
self.assertEqual(len(files), len(old_files)+1)
new_file = [_ for _ in files if _ not in old_files][0]
self.assertTrue(
bool(re.compile(r"\d{4}-\d{6}\.csv").match(new_file))
)
def test_other_sep(self):
saver = AutoSaver(save_dir=self.test_assets_dir,
example_path=r"tests/assets/saver-example-other-sep.csv")
target = '''index;value
0;1.0
1;0.0
2;1.5
3;0.5
'''
self.run_all_X_s(saver, target)
def test_memo(self):
saver = AutoSaver(self.test_assets_dir)
saver.save(pd.DataFrame(), "test_memo_df.csv", "hello")
with open(os.path.join(self.test_assets_dir, "memo.txt"), "r", encoding="utf-8") as f:
content = f.read()
self.assertEqual(content.strip(), "test_memo_df.csv: hello")
saver.save(pd.DataFrame(), "test_memo_df_2.csv", "hello again")
with open(os.path.join(self.test_assets_dir, "memo.txt"), "r", encoding="utf-8") as f:
content = f.read()
self.assertEqual(content.strip(), "\n".join(
["test_memo_df.csv: hello", "test_memo_df_2.csv: hello again"]))
def test_saving_without_example_path(self):
saver = AutoSaver(self.test_assets_dir)
with self.assertRaises(TypeError):
saver.save(self.np_1d, "wrong.csv")
saver.save(self.pd_df_1d, "test_saving_without_example_path_1.csv")
self.pd_df_1d.to_csv(os.path.join(
self.test_assets_dir, "test_saving_without_example_path_2.csv"))
with open(os.path.join(self.test_assets_dir, "test_saving_without_example_path_1.csv"), "r", encoding="utf-8") as f:
content_1 = f.read()
with open(os.path.join(self.test_assets_dir, "test_saving_without_example_path_2.csv"), "r", encoding="utf-8") as f:
content_2 = f.read()
self.assertEqual(content_1, content_2)
def test_saving_strange_index(self):
saver = AutoSaver(save_dir=self.test_assets_dir,
example_path=r"tests/assets/saver-example-strange-index.csv")
target = '''strange,val
4,1.0
8,0.0
19,1.5
45,0.5
150,0.2
'''
arr = np.array([1.0, 0.0, 1.5, 0.5, 0.2])
saver.save(arr, self.filename)
self.assertEqual(self.get_csv_content(), target)
saver.save(arr.reshape((5, 1)), self.filename)
self.assertEqual(self.get_csv_content(), target)
saver.save(pd.Series(arr), self.filename)
self.assertEqual(self.get_csv_content(), target)
saver.save(pd.DataFrame(arr, columns=[""]), self.filename)
self.assertEqual(self.get_csv_content(), target)
def test_str_as_index(self):
saver = AutoSaver(save_dir=self.test_assets_dir,
example_path=r"tests/assets/saver-example-str-as-index.csv")
str_ix = ["strix1", "strix2", "strix3", "strix4"]
with self.assertRaises(SpeculationFailedError):
saver.save(self.np_1d, self.filename)
with self.assertRaises(SpeculationFailedError):
saver.save(self.pd_df_1d, self.filename)
saver.save(self.pd_df_2d, self.filename)
target = '''ix,val
0,1.0
1,0.0
2,1.5
3,0.5
'''
self.assertEqual(self.get_csv_content(), target)
saver.save(pd.Series(self.np_1d, index=str_ix), self.filename)
target = '''ix,val
strix1,1.0
strix2,0.0
strix3,1.5
strix4,0.5
'''
self.assertEqual(self.get_csv_content(), target)
saver.save(pd.DataFrame(self.np_1d, index=str_ix), self.filename)
self.assertEqual(self.get_csv_content(), target)
saver.save(pd.DataFrame({
"ix": str_ix,
"val": self.np_1d,
}), self.filename)
self.assertEqual(self.get_csv_content(), target)
def test_create_dir(self):
AutoSaver(save_dir=self.test_assets_dir)
self.assertTrue(os.path.isdir(self.test_assets_dir))