AtanasGruev
commented
1 week ago Hello, I was able to reproduce the error. In your case, note that the problem arises from incorrectly passing DifferenceTransform instead of DifferenceTransform() as the transform argument in ProphetDetectorConfig on line 31.
Describe the bug Thanks for such great package!
I am following the jupyter notebook 1_AnomalyFeatures. I get the following error when the Prophet model is configured:
TypeError: to_dict() missing 1 required positional argument: 'self'
To Reproduce Just run the three first cells in the notebook
Expected behavior Have the Prophet model configured without error
Screenshots TypeError Traceback (most recent call last) Cell In [35], line 32 28 # Prophet is a popular forecasting algorithm. Here, we specify that we would like 29 # to pre-processes the input time series by applying a difference transform, 30 # before running the model on it. 31 config3= ProphetDetectorConfig(transform= DifferenceTransform) ---> 32 model3= ProphetDetector(config3)
File ~/python_3.9.0/lib/python3.9/site-packages/merlion/models/forecast/prophet.py:124, in Prophet.init(self, config) 123 def init(self, config: ProphetConfig): --> 124 super().init(config) 125 self.model = prophet.Prophet( 126 yearly_seasonality=self.yearly_seasonality, 127 weekly_seasonality=self.weekly_seasonality, (...) 131 holidays=None if self.holidays is None else pd.DataFrame(self.holidays), 132 )
File ~/python_3.9.0/lib/python3.9/site-packages/merlion/models/forecast/base.py:79, in ForecasterBase.init(self, config) 78 def init(self, config: ForecasterConfig): ---> 79 super().init(config) 80 self.target_name = None 81 self.exog_dim = None ... ---> 67 value = value.to_dict() 68 elif isinstance(value, Enum): 69 value = value.name # Relies on there being an appropriate getter/setter!
TypeError: to_dict() missing 1 required positional argument: 'self'
Desktop (please complete the following information):