Merge pull request #80 from ourownstory/forecast-in-focus

update name of forecast_in_focus to highlight_forecast_step_n

Author: ourownstory

neuralprophet/neural_prophet.py

@@ -195,7 +195,7 @@ def __init__(
         ## set during prediction
         self.future_periods = None
         ## later set by user (optional)
-        self.forecast_in_focus = None
+        self.highlight_forecast_step_n = None
         self.true_ar_weights = None
 
     def _init_model(self):
@@ -497,8 +497,8 @@ def _train(self, df, df_val=None):
         loader = self._init_train_loader(df)
         val = df_val is not None
         ## Metrics
-        if self.forecast_in_focus is not None:
-            self.metrics.add_specific_target(target_pos=self.forecast_in_focus - 1)
+        if self.highlight_forecast_step_n is not None:
+            self.metrics.add_specific_target(target_pos=self.highlight_forecast_step_n - 1)
         if self.normalize_y:
             self.metrics.set_shift_scale((self.data_params['y'].shift, self.data_params['y'].scale))
         if val:
@@ -528,12 +528,12 @@ def _train(self, df, df_val=None):
 
     def _eval_true_ar(self, verbose=False):
         assert self.n_lags > 0
-        if self.forecast_in_focus is None:
+        if self.highlight_forecast_step_n is None:
             if self.n_lags > 1:
                 raise ValueError("Please define forecast_lag for sTPE computation")
             forecast_pos = 1
         else:
-            forecast_pos = self.forecast_in_focus
+            forecast_pos = self.highlight_forecast_step_n
         weights = self.model.ar_weights.detach().numpy()
         weights = weights[forecast_pos - 1, :][::-1]
         sTPE = utils.symmetric_total_percentage_error(self.true_ar_weights, weights)
@@ -552,8 +552,8 @@ def _evaluate(self, loader, verbose=None):
         if self.fitted is False: raise Exception('Model object needs to be fit first.')
         if verbose is None: verbose = self.verbose
         val_metrics = metrics.MetricsCollection([m.new() for m in self.metrics.batch_metrics])
-        if self.forecast_in_focus is not None:
-            val_metrics.add_specific_target(target_pos=self.forecast_in_focus - 1)
+        if self.highlight_forecast_step_n is not None:
+            val_metrics.add_specific_target(target_pos=self.highlight_forecast_step_n - 1)
         ## Run
         val_metrics_dict = self._evaluate_epoch(loader, val_metrics)
 
@@ -840,16 +840,16 @@ def set_true_ar_for_eval(self, true_ar_weights):
         """
         self.true_ar_weights = true_ar_weights
 
-    def set_forecast_in_focus(self, forecast_number=None):
+    def highlight_nth_step_ahead_of_each_forecast(self, step_number=None):
         """Set which forecast step to focus on for metrics evaluation and plotting.
 
         Args:
-            forecast_number (int): i-th step ahead forecast to use for performance statistics evaluation.
-                Can also be None.
+            step_number (int): i-th step ahead forecast to use for statistics and plotting.
+                default: None.
         """
-        if forecast_number is not None:
-            assert forecast_number <= self.n_forecasts
-        self.forecast_in_focus = forecast_number
+        if step_number is not None:
+            assert step_number <= self.n_forecasts
+        self.highlight_forecast_step_n = step_number
         return self
 
     def add_covariate(self, name, regularization=None, normalize='auto', only_last_value=False):
@@ -1029,7 +1029,7 @@ def plot(self, fcst, ax=None, xlabel='ds', ylabel='y', figsize=(10, 6)):
                     include_previous_forecasts=num_forecasts - 1, plot_history_data=True)
         return plotting.plot(
             fcst=fcst, ax=ax, xlabel=xlabel, ylabel=ylabel, figsize=figsize,
-            highlight_forecast=self.forecast_in_focus
+            highlight_forecast=self.highlight_forecast_step_n
         )
 
     def plot_last_forecast(self, fcst, ax=None, xlabel='ds', ylabel='y', figsize=(10, 6),
@@ -1058,7 +1058,7 @@ def plot_last_forecast(self, fcst, ax=None, xlabel='ds', ylabel='y', figsize=(10
         fcst = utils.fcst_df_to_last_forecast(fcst, n_last=1 + include_previous_forecasts)
         return plotting.plot(
             fcst=fcst, ax=ax, xlabel=xlabel, ylabel=ylabel, figsize=figsize,
-            highlight_forecast=self.forecast_in_focus, line_per_origin=True,
+            highlight_forecast=self.highlight_forecast_step_n, line_per_origin=True,
         )
 
     def plot_components(self, fcst, figsize=None):
@@ -1077,7 +1077,7 @@ def plot_components(self, fcst, figsize=None):
             m=self,
             fcst=fcst,
             figsize=figsize,
-            forecast_in_focus=self.forecast_in_focus,
+            forecast_in_focus=self.highlight_forecast_step_n,
         )
 
     def plot_parameters(self, weekly_start=0, yearly_start=0, figsize=None):
@@ -1095,7 +1095,7 @@ def plot_parameters(self, weekly_start=0, yearly_start=0, figsize=None):
         """
         return plotting.plot_parameters(
             m=self,
-            forecast_in_focus=self.forecast_in_focus,
+            forecast_in_focus=self.highlight_forecast_step_n,
             weekly_start=weekly_start,
             yearly_start=yearly_start,
             figsize=figsize,

neuralprophet/test_debug.py

@@ -61,7 +61,7 @@ def test_ar_net(verbose=True):
         weekly_seasonality=False,
         daily_seasonality=False,
     )
-    m.set_forecast_in_focus(m.n_forecasts)
+    m.highlight_nth_step_ahead_of_each_forecast(m.n_forecasts)
     m.fit(df, validate_each_epoch=True)
     future = m.compose_prediction_df(df, n_historic_predictions=len(df))
     forecast = m.predict(df=future)
@@ -123,7 +123,7 @@ def test_lag_reg(verbose=True):
         m = m.add_covariate(name='A')
         m = m.add_regressor(name='B')
         m = m.add_regressor(name='C')
-        # m.set_forecast_in_focus(m.n_forecasts)
+        # m.highlight_nth_step_ahead_of_each_forecast(m.n_forecasts)
     m.fit(df, validate_each_epoch=True)
     future = m.compose_prediction_df(df, n_historic_predictions=365)
     forecast = m.predict(future)
@@ -214,7 +214,7 @@ def test_plot(verbose=True):
         # daily_seasonality=False,
     )
     m.fit(df)
-    m.set_forecast_in_focus(7)
+    m.highlight_nth_step_ahead_of_each_forecast(7)
     future = m.compose_prediction_df(df, n_historic_predictions=10)
     forecast = m.predict(future)
     # print(future.to_string())