Add copy_dim to QuantileForecast, change dim method for multiva…

…riate data (#2352)

src/gluonts/model/forecast.py

@@ -565,10 +565,9 @@ def mean(self) -> np.ndarray:
         """
         Forecast mean.
         """
-        if self._mean is not None:
-            return self._mean
-        else:
-            return np.mean(self.samples, axis=0)
+        if self._mean is None:
+            self._mean = np.mean(self.samples, axis=0)
+        return self._mean
 
     @property
     def mean_ts(self) -> pd.Series:
@@ -614,17 +613,16 @@ def copy_aggregate(self, agg_fun: Callable) -> "SampleForecast":
         )
 
     def dim(self) -> int:
-        if self._dim is not None:
-            return self._dim
-        else:
+        if self._dim is None:
             if len(self.samples.shape) == 2:
                 # univariate target
                 # shape: (num_samples, prediction_length)
-                return 1
+                self._dim = 1
             else:
                 # multivariate target
                 # shape: (num_samples, prediction_length, target_dim)
-                return self.samples.shape[2]
+                self._dim = self.samples.shape[2]
+        return self._dim
 
     def as_json_dict(self, config: "Config") -> dict:
         result = super().as_json_dict(config)
@@ -706,7 +704,7 @@ def __init__(
             f"The forecast_array (shape={shape} should have the same "
             f"length as the forecast_keys (len={len(self.forecast_keys)})."
         )
-        self.prediction_length = shape[-1]
+        self.prediction_length = shape[1]
         self._forecast_dict = {
             k: self.forecast_array[i] for i, k in enumerate(self.forecast_keys)
         }
@@ -744,6 +742,25 @@ def quantile(self, inference_quantile: Union[float, str]) -> np.ndarray:
         else:
             return linear_interpolation(inference_quantile)
 
+    def copy_dim(self, dim: int) -> "QuantileForecast":
+        if len(self.forecast_array.shape) == 2:
+            forecast_array = self.forecast_array
+        else:
+            target_dim = self.forecast_array.shape[2]
+            assert dim < target_dim, (
+                f"must set 0 <= dim < target_dim, but got dim={dim},"
+                f" target_dim={target_dim}"
+            )
+            forecast_array = self.forecast_array[:, :, dim]
+
+        return QuantileForecast(
+            forecast_arrays=forecast_array,
+            start_date=self.start_date,
+            forecast_keys=self.forecast_keys,
+            item_id=self.item_id,
+            info=self.info,
+        )
+
     @property
     def mean(self) -> np.ndarray:
         """
@@ -755,17 +772,16 @@ def mean(self) -> np.ndarray:
         return self.quantile("p50")
 
     def dim(self) -> int:
-        if self._dim is not None:
-            return self._dim
-        else:
-            if (
-                len(self.forecast_array.shape) == 2
-            ):  # 1D target. shape: (num_samples, prediction_length)
-                return 1
+        if self._dim is None:
+            if len(self.forecast_array.shape) == 2:
+                # univariate target
+                # shape: (num_samples, prediction_length)
+                self._dim = 1
             else:
-                # 2D target. shape: (num_samples, target_dim,
-                # prediction_length)
-                return self.forecast_array.shape[1]
+                # multivariate target
+                # shape: (num_samples, prediction_length, target_dim)
+                self._dim = self.forecast_array.shape[2]
+        return self._dim
 
     def __repr__(self):
         return ", ".join(

test/model/test_forecast.py

@@ -39,6 +39,18 @@
     ),
 }
 
+MULTIVARIATE_FORECASTS = {
+    "SampleForecast": SampleForecast(
+        samples=np.arange(160).reshape(8, 5, 4) / 100,
+        start_date=START_DATE,
+    ),
+    "QuantileForecast": QuantileForecast(
+        forecast_arrays=np.random.normal(size=(np.size(QUANTILES), 7, 3)),
+        forecast_keys=np.array(QUANTILES, str),
+        start_date=START_DATE,
+    ),
+}
+
 
 @pytest.mark.parametrize("name", FORECASTS.keys())
 def test_Forecast(name):
@@ -84,6 +96,29 @@ def test_forecast_multivariate(forecast, exp_index):
     assert np.all(forecast.index == exp_index)
 
 
+@pytest.mark.parametrize("name", MULTIVARIATE_FORECASTS.keys())
+def test_copy_dim(name):
+    forecast = MULTIVARIATE_FORECASTS[name]
+    for dim in range(forecast.dim()):
+        univariate_forecast = forecast.copy_dim(dim)
+
+        assert univariate_forecast.dim() == 1
+        assert univariate_forecast.start_date == forecast.start_date
+        assert univariate_forecast.item_id == forecast.item_id
+        assert univariate_forecast.info == forecast.info
+
+        if name == "SampleForecast":
+            assert np.array_equal(
+                univariate_forecast.samples,
+                MULTIVARIATE_FORECASTS[name].samples[:, :, dim],
+            )
+        else:
+            assert np.array_equal(
+                univariate_forecast.forecast_array,
+                MULTIVARIATE_FORECASTS[name].forecast_array[:, :, dim],
+            )
+
+
 def test_linear_interpolation() -> None:
     tol = 1e-7
     x_coord = [0.1, 0.5, 0.9]