Fix analysis.forecast.plots.metric_per_segment_distribution_plot to…

… handle `None` from metrics (#543)

CHANGELOG.md

@@ -42,7 +42,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Add parameter `missing_mode` into `R2` and `MedAE` metrics ([#537](https://github.com/etna-team/etna/pull/537))
 - Update `analysis.forecast.plots.plot_metric_per_segment` to handle `None` from metrics ([#540](https://github.com/etna-team/etna/pull/540))
 - Add parameter `missing_mode` into `RMSE` and `MSLE` metrics ([#542](https://github.com/etna-team/etna/pull/542))
-- 
+- Update `analysis.forecast.plots.metric_per_segment_distribution_plot` to handle `None` from metrics ([#543](https://github.com/etna-team/etna/pull/543))
 - 
 - 
 

etna/analysis/forecast/plots.py

@@ -726,7 +726,7 @@ def plot_metric_per_segment(
     Warnings
     --------
     UserWarning:
-        There are segments without non-missing metric values.
+        There are segments with all missing metric values.
     UserWarning:
         Some segments have different set of folds to be aggregated on due to missing values.
     """
@@ -803,7 +803,12 @@ def metric_per_segment_distribution_plot(
     seaborn_params: Optional[Dict[str, Any]] = None,
     figsize: Tuple[int, int] = (10, 5),
 ):
-    """Plot per-segment metrics distribution.
+    """Plot distribution of metric values over all segments.
+
+    If for some segment all metric values are missing, it isn't plotted, and the warning is raised.
+
+    If some segments have different set of folds with non-missing metrics,
+    it can lead to incompatible values between folds. The warning is raised in such case.
 
     Parameters
     ----------
@@ -831,6 +836,13 @@ def metric_per_segment_distribution_plot(
         if ``metric_name`` isn't present in ``metrics_df``
     NotImplementedError:
         unknown ``per_fold_aggregation_mode`` is given
+
+    Warnings
+    --------
+    UserWarning:
+        There are segments with all missing metric values.
+    UserWarning:
+        Some segments have different set of folds to be aggregated on due to missing values.
     """
     if seaborn_params is None:
         seaborn_params = {}
@@ -844,6 +856,9 @@ def metric_per_segment_distribution_plot(
     if metric_name not in metrics_df.columns:
         raise ValueError("Given metric_name isn't present in metrics_df")
 
+    _check_metrics_df_empty_segments(metrics_df=metrics_df, metric_name=metric_name)
+    _check_metrics_df_same_folds_for_each_segment(metrics_df=metrics_df, metric_name=metric_name)
+
     # draw plot for each fold
     if per_fold_aggregation_mode is None and "fold_number" in metrics_df.columns:
         if plot_type_enum == MetricPlotType.hist:

etna/analysis/forecast/utils.py

@@ -145,6 +145,6 @@ def _check_metrics_df_same_folds_for_each_segment(metrics_df: pd.DataFrame, metr
     df = metrics_df[["segment", "fold_number", metric_name]]
     # we don't take into account segments without any non-missing metrics, they are handled by other check
     df = df.dropna(subset=[metric_name])
-    num_unique = df.groupby("segment")["fold_number"].apply(frozenset).nunique()
+    num_unique = df.groupby("segment", group_keys=False)["fold_number"].apply(frozenset).nunique()
     if num_unique > 1:
         warnings.warn("Some segments have different set of folds to be aggregated on due to missing values.")

tests/test_analysis/test_forecast/test_plots.py

@@ -1,6 +1,7 @@
 import pandas as pd
 import pytest
 
+from etna.analysis import metric_per_segment_distribution_plot
 from etna.analysis import plot_metric_per_segment
 from etna.analysis import plot_residuals
 from etna.analysis.forecast.plots import _get_borders_comparator
@@ -80,7 +81,7 @@ def metrics_df_no_folds(metrics_df_with_folds) -> pd.DataFrame:
     "df_name, metric_name",
     [
         ("metrics_df_with_folds", "MAE"),
-        ("metrics_df_no_folds", "MSE"),
+        ("metrics_df_no_folds", "MAE"),
         ("metrics_df_no_folds", "MSE"),
     ],
 )
@@ -112,3 +113,49 @@ def test_plot_metric_per_segment_warning_non_comparable_segments(df_name, metric
     metrics_df = request.getfixturevalue(df_name)
     with pytest.warns(UserWarning, match="Some segments have different set of folds to be aggregated on"):
         plot_metric_per_segment(metrics_df=metrics_df, metric_name=metric_name)
+
+
+@pytest.mark.parametrize("plot_type", ["hist", "box", "violin"])
+@pytest.mark.parametrize(
+    "df_name, metric_name, per_fold_aggregation_mode",
+    [
+        ("metrics_df_with_folds", "MAE", None),
+        ("metrics_df_with_folds", "MAE", "mean"),
+        ("metrics_df_with_folds", "MAE", "median"),
+        ("metrics_df_no_folds", "MAE", None),
+        ("metrics_df_no_folds", "MSE", None),
+    ],
+)
+def test_plot_metric_per_segment_ok(df_name, metric_name, per_fold_aggregation_mode, plot_type, request):
+    metrics_df = request.getfixturevalue(df_name)
+    metric_per_segment_distribution_plot(
+        metrics_df=metrics_df,
+        metric_name=metric_name,
+        per_fold_aggregation_mode=per_fold_aggregation_mode,
+        plot_type=plot_type,
+    )
+
+
+@pytest.mark.parametrize(
+    "df_name, metric_name",
+    [
+        ("metrics_df_with_folds", "MAPE"),
+        ("metrics_df_no_folds", "RMSE"),
+    ],
+)
+def test_plot_metric_per_segment_warning_empty_segments(df_name, metric_name, request):
+    metrics_df = request.getfixturevalue(df_name)
+    with pytest.warns(UserWarning, match="There are segments with all missing metric values"):
+        metric_per_segment_distribution_plot(metrics_df=metrics_df, metric_name=metric_name)
+
+
+@pytest.mark.parametrize(
+    "df_name, metric_name",
+    [
+        ("metrics_df_with_folds", "MSE"),
+    ],
+)
+def test_plot_metric_per_segment_warning_non_comparable_segments(df_name, metric_name, request):
+    metrics_df = request.getfixturevalue(df_name)
+    with pytest.warns(UserWarning, match="Some segments have different set of folds to be aggregated on"):
+        metric_per_segment_distribution_plot(metrics_df=metrics_df, metric_name=metric_name)