autogluon.timeseries 1.1.0b20240409__tar.gz → 1.1.0b20240410__tar.gz

{autogluon.timeseries-1.1.0b20240409 → autogluon.timeseries-1.1.0b20240410}/PKG-INFO

@@ -1,7 +1,7 @@
 Metadata-Version: 2.1
 Name: autogluon.timeseries
-Version: 1.1.0b20240409
-Summary: AutoML for Image, Text, and Tabular Data
+Version: 1.1.0b20240410
+Summary: Fast and Accurate ML in 3 Lines of Code
 Home-page: https://github.com/autogluon/autogluon
 Author: AutoGluon Community
 License: Apache-2.0
@@ -14,7 +14,7 @@ Description:
         <div align="center">
         <img src="https://user-images.githubusercontent.com/16392542/77208906-224aa500-6aba-11ea-96bd-e81806074030.png" width="350">
         
-        ## AutoML for Image, Text, Time Series, and Tabular Data
+        ## Fast and Accurate ML in 3 Lines of Code
         
         [![Latest Release](https://img.shields.io/github/v/release/autogluon/autogluon)](https://github.com/autogluon/autogluon/releases)
         [![Conda Forge](https://img.shields.io/conda/vn/conda-forge/autogluon.svg)](https://anaconda.org/conda-forge/autogluon)

{autogluon.timeseries-1.1.0b20240409 → autogluon.timeseries-1.1.0b20240410}/src/autogluon/timeseries/models/autogluon_tabular/mlforecast.py

@@ -174,6 +174,21 @@ class AbstractMLForecastModel(AbstractTimeSeriesModel):
         """
         return df
 
+    def _add_scale_as_static_feature(self, data: TimeSeriesDataFrame) -> TimeSeriesDataFrame:
+        """Add mean/std of the target column for each series as a static feature."""
+        data = data.copy(deep=False)
+        scale_features = (
+            data[self.target]
+            .groupby(ITEMID, sort=False)
+            .agg(["mean", "std"])
+            .rename(columns={"mean": "__target_mean", "std": "__target_scale"})
+        )
+        if data.static_features is None:
+            data.static_features = scale_features
+        else:
+            data.static_features = pd.concat([data.static_features, scale_features], axis=1)
+        return data
+
     @staticmethod
     def _shorten_all_series(mlforecast_df: pd.DataFrame, max_length: int):
         logger.debug(f"Shortening all series to at most {max_length}")
@@ -252,6 +267,11 @@ class AbstractMLForecastModel(AbstractTimeSeriesModel):
         if static_features is not None:
             df = pd.merge(df, static_features, how="left", on=ITEMID, suffixes=(None, "_static_feat"))
 
+        for col in self.metadata.known_covariates_real:
+            # Normalize non-boolean features using mean_abs scaling
+            if not df[col].isin([0, 1]).all():
+                df[f"__scaled_{col}"] = df[col] / df[col].abs().groupby(df[ITEMID]).mean().reindex(df[ITEMID]).values
+
         # We assume that df is sorted by 'unique_id' inside `TimeSeriesPredictor._check_and_prepare_data_frame`
         return df.rename(columns=column_name_mapping)
 
@@ -265,6 +285,8 @@ class AbstractMLForecastModel(AbstractTimeSeriesModel):
     ) -> None:
         from mlforecast import MLForecast
 
+        train_data = self._add_scale_as_static_feature(train_data)
+
         self._check_fit_params()
         fit_start_time = time.time()
         self._train_target_median = train_data[self.target].median()
@@ -463,6 +485,7 @@ class DirectTabularModel(AbstractMLForecastModel):
         known_covariates: Optional[TimeSeriesDataFrame] = None,
         **kwargs,
     ) -> TimeSeriesDataFrame:
+        data = self._add_scale_as_static_feature(data)
         original_item_id_order = data.item_ids
         data, known_covariates, forecast_for_short_series = self._remove_short_ts_and_generate_fallback_forecast(
             data=data, known_covariates=known_covariates
@@ -479,7 +502,7 @@ class DirectTabularModel(AbstractMLForecastModel):
         # MLForecast raises exception of target contains NaN. We use inf as placeholder, replace them by NaN afterwards
         data_future[self.target] = float("inf")
         data_extended = pd.concat([data, data_future])
-        mlforecast_df = self._to_mlforecast_df(data_extended, data.static_features)
+        mlforecast_df = self._to_mlforecast_df(data_extended, data_extended.static_features)
         if self._max_ts_length is not None:
             # We appended `prediction_length` time steps to each series, so increase length
             mlforecast_df = self._shorten_all_series(mlforecast_df, self._max_ts_length + self.prediction_length)
@@ -594,6 +617,7 @@ class RecursiveTabularModel(AbstractMLForecastModel):
         known_covariates: Optional[TimeSeriesDataFrame] = None,
         **kwargs,
     ) -> TimeSeriesDataFrame:
+        data = self._add_scale_as_static_feature(data)
         original_item_id_order = data.item_ids
         data, known_covariates, forecast_for_short_series = self._remove_short_ts_and_generate_fallback_forecast(
             data=data, known_covariates=known_covariates

{autogluon.timeseries-1.1.0b20240409 → autogluon.timeseries-1.1.0b20240410}/src/autogluon/timeseries/models/local/abstract_local_model.py

@@ -185,9 +185,9 @@ class AbstractLocalModel(AbstractTimeSeriesModel):
         if end_time is not None and time.time() >= end_time:
             raise TimeLimitExceeded
 
+        model_failed = False
         if time_series.isna().all():
             result = self._dummy_forecast.copy()
-            model_failed = True
         else:
             try:
                 result = self._predict_with_local_model(
@@ -196,7 +196,6 @@ class AbstractLocalModel(AbstractTimeSeriesModel):
                 )
                 if not np.isfinite(result.values).all():
                     raise RuntimeError("Forecast contains NaN or Inf values.")
-                model_failed = False
             except Exception:
                 if self.use_fallback_model:
                     result = seasonal_naive_forecast(

{autogluon.timeseries-1.1.0b20240409 → autogluon.timeseries-1.1.0b20240410}/src/autogluon/timeseries/version.py

@@ -1,3 +1,3 @@
 """This is the autogluon version file."""
-__version__ = '1.1.0b20240409'
+__version__ = '1.1.0b20240410'
 __lite__ = False

{autogluon.timeseries-1.1.0b20240409 → autogluon.timeseries-1.1.0b20240410}/src/autogluon.timeseries.egg-info/PKG-INFO

@@ -1,7 +1,7 @@
 Metadata-Version: 2.1
 Name: autogluon.timeseries
-Version: 1.1.0b20240409
-Summary: AutoML for Image, Text, and Tabular Data
+Version: 1.1.0b20240410
+Summary: Fast and Accurate ML in 3 Lines of Code
 Home-page: https://github.com/autogluon/autogluon
 Author: AutoGluon Community
 License: Apache-2.0
@@ -14,7 +14,7 @@ Description:
         <div align="center">
         <img src="https://user-images.githubusercontent.com/16392542/77208906-224aa500-6aba-11ea-96bd-e81806074030.png" width="350">
         
-        ## AutoML for Image, Text, Time Series, and Tabular Data
+        ## Fast and Accurate ML in 3 Lines of Code
         
         [![Latest Release](https://img.shields.io/github/v/release/autogluon/autogluon)](https://github.com/autogluon/autogluon/releases)
         [![Conda Forge](https://img.shields.io/conda/vn/conda-forge/autogluon.svg)](https://anaconda.org/conda-forge/autogluon)

{autogluon.timeseries-1.1.0b20240409 → autogluon.timeseries-1.1.0b20240410}/src/autogluon.timeseries.egg-info/requires.txt

@@ -16,9 +16,9 @@ utilsforecast<0.0.11,>=0.0.10
 tqdm<5,>=4.38
 orjson~=3.9
 tensorboard<3,>=2.9
-autogluon.core[raytune]==1.1.0b20240409
-autogluon.common==1.1.0b20240409
-autogluon.tabular[catboost,lightgbm,xgboost]==1.1.0b20240409
+autogluon.core[raytune]==1.1.0b20240410
+autogluon.common==1.1.0b20240410
+autogluon.tabular[catboost,lightgbm,xgboost]==1.1.0b20240410
 
 [all]
 optimum[onnxruntime]<1.19,>=1.17