Fix Develop

darts/models/exponential_smoothing.py

@@ -9,15 +9,16 @@
 from .forecasting_model import UnivariateForecastingModel
 from ..logging import get_logger
 from ..timeseries import TimeSeries
+from .. import ModelMode
 
 logger = get_logger(__name__)
 
 
 class ExponentialSmoothing(UnivariateForecastingModel):
     def __init__(self,
-                 trend: Optional[str] = 'additive',
+                 trend: Optional[ModelMode] = ModelMode.ADDITIVE,
                  damped: Optional[bool] = False,
-                 seasonal: Optional[str] = 'additive',
+                 seasonal: Optional[ModelMode] = ModelMode.ADDITIVE,
                  seasonal_periods: Optional[int] = 12,
                  **fit_kwargs):
         """ Exponential Smoothing
@@ -28,14 +29,22 @@ def __init__(self,
 
         We refer to this link for the original and more complete documentation of the parameters.
 
+        `model_mode` must be a ModelMode Enum member.
+        You can access the Enum with `from darts import ModelMode`.
+
+        `ExponentialSmoothing(trend=None, seasonal=None)` corresponds to a single exponential smoothing.
+        `ExponentialSmoothing(trend=ModelMode.ADDITIVE, seasonal=None)` corresponds to a Holt's exponential smoothing.
+
         Parameters
         ----------
         trend
-            Type of trend component
+            Type of trend component. Either ModelMode.ADDITIVE or ModelMode.MULTIPLICATIVE.
+            Defaults to `ModelMode.ADDITIVE`.
         damped
-            Should the trend component be damped.
+            Should the trend component be damped. Defaults to False.
         seasonal
-            Type of seasonal component
+            Type of seasonal component. Either ModelMode.ADDITIVE or ModelMode.MULTIPLICATIVE.
+            Defaults to `ModelMode.ADDITIVE`.
         seasonal_periods
             The number of periods in a complete seasonal cycle, e.g., 4 for quarterly data or 7 for daily
             data with a weekly cycle.
@@ -60,9 +69,9 @@ def fit(self, series: TimeSeries, component_index: Optional[int] = None):
         super().fit(series, component_index)
         series = self.training_series
         hw_model = hw.ExponentialSmoothing(series.values(),
-                                           trend=self.trend,
+                                           trend=self.trend.value,
                                            damped=self.damped,
-                                           seasonal=self.seasonal,
+                                           seasonal=self.seasonal.value,
                                            seasonal_periods=self.seasonal_periods)
 
         hw_results = hw_model.fit(**self.fit_kwargs)

darts/models/theta.py

@@ -60,7 +60,8 @@ def __init__(self,
         self.theta = theta
         self.is_seasonal = False
         self.seasonality = None
-        self.season_period = seasonality_period
+        self.seasonality_period = seasonality_period
+        self.season_period = None
         self.season_mode = season_mode
 
         raise_if_not(season_mode in SeasonalityMode,
@@ -79,6 +80,8 @@ def fit(self, series: TimeSeries, component_index: Optional[int] = None):
         # or infers season_period from the TimeSeries itself.
         if self.season_mode is SeasonalityMode.NONE:
             self.season_period = 1
+        else:
+            self.season_period = self.seasonality_period
         if self.season_period is None:
             max_lag = len(ts) // 2
             self.is_seasonal, self.season_period = check_seasonality(ts, self.season_period, max_lag=max_lag)
@@ -195,7 +198,8 @@ def __init__(self,
         self.theta = theta
         self.is_seasonal = False
         self.seasonality = None
-        self.season_period = seasonality_period
+        self.seasonality_period = seasonality_period
+        self.season_period = None
         self.model_mode = model_mode
         self.season_mode = season_mode
         self.trend_mode = trend_mode
@@ -230,6 +234,8 @@ def fit(self, ts, component_index: Optional[int] = None):
         # or infers season_period from the TimeSeries itself.
         if self.season_mode is SeasonalityMode.NONE:
             self.season_period = 1
+        else:
+            self.season_period = self.seasonality_period
         if self.season_period is None:
             max_lag = len(ts) // 2
             self.is_seasonal, self.season_period = check_seasonality(ts, self.season_period, max_lag=max_lag)

darts/tests/test_4theta.py

@@ -41,16 +41,16 @@ def test_zero_mean(self):
             model.fit(sine_series)
 
     def test_theta(self):
-        random.seed(1)
+        np.random.seed(1)
         series = rt(length=50, mean=100)
-        theta_param = random.randrange(-5, 5)
+        theta_param = np.random.randint(1, 5)
         theta = Theta(theta_param)
         fourtheta = FourTheta(theta_param, normalization=False)
         theta.fit(series)
         fourtheta.fit(series)
         forecast_theta = theta.predict(20)
         forecast_fourtheta = fourtheta.predict(20)
-        self.assertTrue((forecast_theta - forecast_fourtheta <= 1e-12).all()[0])
+        self.assertTrue((forecast_theta - forecast_fourtheta <= 1e-11).all()[0])
 
     def test_best_model(self):
         random.seed(1)