openstef-models 4.1.0__tar.gz → 4.1.1.dev0__tar.gz

{openstef_models-4.1.0 → openstef_models-4.1.1.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openstef-models
-Version: 4.1.0
+Version: 4.1.1.dev0
 Summary: Core models for OpenSTEF
 Project-URL: Documentation, https://openstef.github.io/openstef/index.html
 Project-URL: Homepage, https://lfenergy.org/projects/openstef/

{openstef_models-4.1.0 → openstef_models-4.1.1.dev0}/pyproject.toml

@@ -7,7 +7,7 @@ requires = [ "hatchling" ]
 
 [project]
 name = "openstef-models"
-version = "4.1.0"
+version = "4.1.1.dev0"
 description = "Core models for OpenSTEF"
 readme = "README.md"
 keywords = [ "energy", "forecasting", "machinelearning" ]

{openstef_models-4.1.0 → openstef_models-4.1.1.dev0}/src/openstef_models/models/forecasting/median_forecaster.py

@@ -214,13 +214,11 @@ class MedianForecaster(Forecaster, ExplainableForecaster, ContributionsMixin):
         for time_step in range(lag_array.shape[0]):
             # Get the lag features for the current time step.
             current_lags = lag_array[time_step]
-            # Calculate the median of the available lag features, ignoring NaNs.
-            median = np.nanmedian(current_lags)
-            # If the median calculation resulted in NaN (e.g., all lags were NaN), skip the autoregression step.
-            if not np.isnan(median):
-                median = float(median)
-            else:
+            # If all lags are NaN there is no value to predict, so skip the autoregressive step.
+            if np.all(np.isnan(current_lags)):
                 continue
+            # Calculate the median of the available lag features, ignoring NaNs.
+            median = float(np.nanmedian(current_lags))
 
             # Store the calculated median in the prediction array.
             prediction[time_step] = median

{openstef_models-4.1.0 → openstef_models-4.1.1.dev0}/tests/unit/integrations/mlflow/test_mlflow_storage_callback.py

@@ -279,6 +279,7 @@ def test_mlflow_storage_callback__model_selection__skips_on_tag_change(
     new_model = ForecastingModel(
         forecaster=SimpleTestForecaster(horizons=[LeadTime(timedelta(hours=6))], quantiles=[Q(0.5)]),
         tags={"version": "2.0"},
+        data_splitter=DataSplitter(test_fraction=0.0, val_fraction=0.0),
     )
     new_workflow = CustomForecastingWorkflow(model_id="test_model", model=new_model)
     new_result = new_model.fit(sample_dataset)