Add test for MADTransform x NaN

tests/test_transforms/test_math/test_statistics_transform.py

@@ -4,6 +4,7 @@
 import pandas as pd
 import pytest
 
+from etna.datasets import TSDataset
 from etna.transforms.math import MADTransform
 from etna.transforms.math import MaxTransform
 from etna.transforms.math import MeanTransform
@@ -19,11 +20,7 @@ def simple_df_for_agg() -> pd.DataFrame:
     df = pd.DataFrame({"timestamp": pd.date_range("2020-01-01", periods=n)})
     df["target"] = list(range(n))
     df["segment"] = "segment_1"
-
-    df = df.pivot(index="timestamp", columns="segment")
-    df = df.reorder_levels([1, 0], axis=1)
-    df = df.sort_index(axis=1)
-    df.columns.names = ["segment", "feature"]
+    df = TSDataset.to_dataset(df)
     return df
 
 
@@ -33,11 +30,17 @@ def df_for_agg() -> pd.DataFrame:
     df = pd.DataFrame({"timestamp": pd.date_range("2020-01-01", periods=n)})
     df["target"] = [-1, 1, 3, 2, 4, 9, 8, 5, 6, 0]
     df["segment"] = "segment_1"
+    df = TSDataset.to_dataset(df)
+    return df
+
 
-    df = df.pivot(index="timestamp", columns="segment")
-    df = df.reorder_levels([1, 0], axis=1)
-    df = df.sort_index(axis=1)
-    df.columns.names = ["segment", "feature"]
+@pytest.fixture
+def df_for_agg_with_nan() -> pd.DataFrame:
+    n = 10
+    df = pd.DataFrame({"timestamp": pd.date_range("2020-01-01", periods=n)})
+    df["target"] = [-1, 1, 3, None, 4, 9, 8, 5, 6, 0]
+    df["segment"] = "segment_1"
+    df = TSDataset.to_dataset(df)
     return df
 
 
@@ -197,6 +200,20 @@ def test_mad_transform(df_for_agg: pd.DataFrame, window: int, periods: int, fill
     np.testing.assert_array_almost_equal(expected, res["segment_1"]["result"])
 
 
+@pytest.mark.parametrize(
+    "window,periods,fill_na,expected",
+    ((3, 3, -17, [-17, -17, -17, 4 / 3, -17, -17, -17, 2, 14 / 9, 10 / 9]),),
+)
+def test_mad_transform_with_nans(
+    df_for_agg_with_nan: pd.DataFrame, window: int, periods: int, fill_na: float, expected: np.ndarray
+):
+    transform = MADTransform(
+        window=window, min_periods=periods, fillna=fill_na, in_column="target", out_column="result"
+    )
+    res = transform.fit_transform(df_for_agg_with_nan)
+    np.testing.assert_array_almost_equal(expected, res["segment_1"]["result"])
+
+
 @pytest.mark.parametrize(
     "transform",
     (