TEST-modin-project#2033: speed up test_series.py (modin-project#2034)

* TEST-modin-project#2033: speed up test_series.py Signed-off-by: Anatoly Myachev <[email protected]> * TEST-modin-project#2033: fix kurtosis test Signed-off-by: Anatoly Myachev <[email protected]> * TEST-modin-project#2033: refactor test for modin-project#1953 Signed-off-by: Anatoly Myachev <[email protected]>
aregm · Sep 16, 2020 · d4d42b4 · d4d42b4
1 parent 4f88ac5
commit d4d42b4
Showing 1 changed file with 59 additions and 109 deletions.
diff --git a/modin/pandas/test/test_series.py b/modin/pandas/test/test_series.py
@@ -61,6 +61,8 @@
     test_data_small_keys,
     test_data_categorical_values,
     test_data_categorical_keys,
+    generate_multiindex,
+    test_data_diff_dtype,
 )
 
 pd.DEFAULT_NPARTITIONS = 4
@@ -1850,10 +1852,22 @@ def test_keys(data):
     df_equals(modin_series.keys(), pandas_series.keys())
 
 
-@pytest.mark.parametrize("data", test_data_values, ids=test_data_keys)
-@pytest.mark.parametrize("axis", axis_values, ids=axis_keys)
+def test_kurtosis_alias():
+    # It's optimization. If failed, Series.kurt should be tested explicitly
+    # in tests: `test_kurt_kurtosis`, `test_kurt_kurtosis_level`.
+    assert pd.Series.kurt == pd.Series.kurtosis
+
+
+@pytest.mark.parametrize("axis", [0, 1])
 @pytest.mark.parametrize("skipna", bool_arg_values, ids=bool_arg_keys)
-@pytest.mark.parametrize("level", [None, -1, 0, 1])
+def test_kurtosis(axis, skipna):
+    eval_general(
+        *create_test_series(test_data["float_nan_data"]),
+        lambda df: df.kurtosis(axis=axis, skipna=skipna),
+    )
+
+
+@pytest.mark.parametrize("axis", ["rows", "columns"])
 @pytest.mark.parametrize(
     "numeric_only",
     [
@@ -1867,20 +1881,26 @@ def test_keys(data):
         None,
     ],
 )
-@pytest.mark.parametrize("method", ["kurtosis", "kurt"])
-def test_kurt_kurtosis(data, axis, skipna, level, numeric_only, method):
-    func_kwargs = {
-        "axis": axis,
-        "skipna": skipna,
-        "level": level,
-        "numeric_only": numeric_only,
-    }
-    modin_series, pandas_series = create_test_series(data)
+def test_kurtosis_numeric_only(axis, numeric_only):
+    eval_general(
+        *create_test_series(test_data_diff_dtype),
+        lambda df: df.kurtosis(axis=axis, numeric_only=numeric_only),
+    )
+
+
+@pytest.mark.parametrize("level", [-1, 0, 1])
+def test_kurtosis_level(level):
+    data = test_data["int_data"]
+    modin_s, pandas_s = create_test_series(data)
+
+    index = generate_multiindex(len(data.keys()))
+    modin_s.columns = index
+    pandas_s.columns = index
 
     eval_general(
-        modin_series,
-        pandas_series,
-        lambda df: df.kurtosis(**func_kwargs),
+        modin_s,
+        pandas_s,
+        lambda s: s.kurtosis(axis=1, level=level),
     )
 
 
@@ -2015,16 +2035,18 @@ def test_median(data, skipna):
     modin_series, pandas_series = create_test_series(data)
     df_equals(modin_series.median(skipna=skipna), pandas_series.median(skipna=skipna))
 
-    # test for issue #1953
+
+@pytest.mark.parametrize("method", ["median", "skew", "std", "sum", "var", "prod"])
+def test_median_skew_std_sum_var_prod_1953(method):
+    # See #1953 for details
+    data = [3, 3, 3, 3, 3, 3, 3, 3, 3]
     arrays = [
         ["1", "1", "1", "2", "2", "2", "3", "3", "3"],
         ["1", "2", "3", "4", "5", "6", "7", "8", "9"],
     ]
-    modin_series = pd.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    pandas_series = pandas.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    modin_result = modin_series.median(level=0)
-    pandas_result = pandas_series.median(level=0)
-    df_equals(modin_result, pandas_result)
+    modin_s = pd.Series(data, index=arrays)
+    pandas_s = pandas.Series(data, index=arrays)
+    eval_general(modin_s, pandas_s, lambda s: getattr(s, method)(level=0))
 
 
 @pytest.mark.parametrize("data", test_data_values, ids=test_data_keys)
@@ -2206,66 +2228,38 @@ def test_pow(data):
     inter_df_math_helper(modin_series, pandas_series, "pow")
 
 
-@pytest.mark.parametrize(
-    "data",
-    test_data_values + test_data_small_values,
-    ids=test_data_keys + test_data_small_keys,
-)
-@pytest.mark.parametrize("axis", axis_values, ids=axis_keys)
+def test_product_alias():
+    assert pd.Series.prod == pd.Series.product
+
+
+@pytest.mark.parametrize("axis", [0, 1])
 @pytest.mark.parametrize(
     "skipna", bool_arg_values, ids=arg_keys("skipna", bool_arg_keys)
 )
+def test_prod(axis, skipna):
+    eval_general(
+        *create_test_series(test_data["float_nan_data"]),
+        lambda s: s.prod(axis=axis, skipna=skipna),
+    )
+
+
 @pytest.mark.parametrize(
     "numeric_only",
     [
         None,
         False,
-        pytest.param(
-            True,
-            marks=pytest.mark.xfail(
-                reason="numeric_only not implemented for pandas.Series"
-            ),
-        ),
+        pytest.param(True, marks=pytest.mark.xfail(reason="didn't raise Exception")),
     ],
 )
 @pytest.mark.parametrize(
     "min_count", int_arg_values, ids=arg_keys("min_count", int_arg_keys)
 )
-@pytest.mark.parametrize(
-    "operation",
-    [
-        "prod",
-        pytest.param(
-            "product",
-            marks=pytest.mark.skipif(
-                pandas.Series.product == pandas.Series.prod
-                and pd.Series.product == pd.Series.prod,
-                reason="That operation was already tested.",
-            ),
-        ),
-    ],
-)
-def test_prod(data, axis, skipna, numeric_only, min_count, operation):
+def test_prod_specific(min_count, numeric_only):
     eval_general(
-        *create_test_series(data),
-        lambda df, *args, **kwargs: type(df)([getattr(df, operation)(*args, **kwargs)]),
-        axis=axis,
-        skipna=skipna,
-        numeric_only=numeric_only,
-        min_count=min_count,
+        *create_test_series(test_data_diff_dtype),
+        lambda df: df.prod(min_count=min_count, numeric_only=numeric_only),
     )
 
-    # test for issue #1953
-    arrays = [
-        ["1", "1", "1", "2", "2", "2", "3", "3", "3"],
-        ["1", "2", "3", "4", "5", "6", "7", "8", "9"],
-    ]
-    modin_series = pd.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    pandas_series = pandas.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    modin_result = modin_series.prod(level=0)
-    pandas_result = pandas_series.prod(level=0)
-    df_equals(modin_result, pandas_result)
-
 
 @pytest.mark.parametrize("data", test_data_values, ids=test_data_keys)
 @pytest.mark.parametrize("q", quantiles_values, ids=quantiles_keys)
@@ -2784,17 +2778,6 @@ def test_skew(data, skipna):
     modin_series, pandas_series = create_test_series(data)
     df_equals(modin_series.skew(skipna=skipna), pandas_series.skew(skipna=skipna))
 
-    # test for issue #1953
-    arrays = [
-        ["1", "1", "1", "2", "2", "2", "3", "3", "3"],
-        ["1", "2", "3", "4", "5", "6", "7", "8", "9"],
-    ]
-    modin_series = pd.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    pandas_series = pandas.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    modin_result = modin_series.skew(level=0)
-    pandas_result = pandas_series.skew(level=0)
-    df_equals(modin_result, pandas_result)
-
 
 @pytest.mark.parametrize("data", test_data_values, ids=test_data_keys)
 @pytest.mark.parametrize("index", ["default", "ndarray"])
@@ -2913,17 +2896,6 @@ def test_std(request, data, skipna, ddof):
         modin_result = modin_series.std(skipna=skipna, ddof=ddof)
         df_equals(modin_result, pandas_result)
 
-    # test for issue #1953
-    arrays = [
-        ["1", "1", "1", "2", "2", "2", "3", "3", "3"],
-        ["1", "2", "3", "4", "5", "6", "7", "8", "9"],
-    ]
-    modin_series = pd.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    pandas_series = pandas.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    modin_result = modin_series.std(level=0)
-    pandas_result = pandas_series.std(level=0)
-    df_equals(modin_result, pandas_result)
-
 
 @pytest.mark.parametrize("data", test_data_values, ids=test_data_keys)
 def test_sub(data):
@@ -2972,17 +2944,6 @@ def test_sum(data, axis, skipna, numeric_only, min_count):
         min_count=min_count,
     )
 
-    # test for issue #1953
-    arrays = [
-        ["1", "1", "1", "2", "2", "2", "3", "3", "3"],
-        ["1", "2", "3", "4", "5", "6", "7", "8", "9"],
-    ]
-    modin_series = pd.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    pandas_series = pandas.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    modin_result = modin_series.sum(level=0)
-    pandas_result = pandas_series.sum(level=0)
-    df_equals(modin_result, pandas_result)
-
 
 @pytest.mark.parametrize("data", test_data_values, ids=test_data_keys)
 @pytest.mark.parametrize("axis1", [0, 1, "columns", "index"])
@@ -3361,17 +3322,6 @@ def test_var(data, skipna, ddof):
         modin_result = modin_series.var(skipna=skipna, ddof=ddof)
         df_equals(modin_result, pandas_result)
 
-    # test for issue #1953
-    arrays = [
-        ["1", "1", "1", "2", "2", "2", "3", "3", "3"],
-        ["1", "2", "3", "4", "5", "6", "7", "8", "9"],
-    ]
-    modin_series = pd.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    pandas_series = pandas.Series([3, 3, 3, 3, 3, 3, 3, 3, 3], index=arrays)
-    modin_result = modin_series.var(level=0)
-    pandas_result = pandas_series.var(level=0)
-    df_equals(modin_result, pandas_result)
-
 
 def test_view():
     modin_series = pd.Series([-2, -1, 0, 1, 2], dtype="int8")