Prophet Adapter API

This section documents the Prophet adapter used to integrate prophet.Prophet into the eb-adapters package.

All content below is generated automatically from NumPy-style docstrings in the source code.

Prophet Adapter Module

eb_adapters.prophet

ProphetAdapter

Bases: BaseAdapter

Adapter for prophet.Prophet.

This adapter enables Prophet models to be used inside ElectricBarometer or other CWSL-based evaluation workflows by exposing a scikit-learn-like API.

Parameters:

Name	Type	Description	Default
model	Any | None	Optional pre-configured prophet.Prophet instance. If None, a default Prophet model is constructed. If the prophet package is not installed, constructing a default model will raise ImportError.	None

Notes

Input conventions:

• X encodes the time index as either:
• shape (n_samples,) of datetime-like values, or
• shape (n_samples, n_features) where the first column is datetime-like
• y is a one-dimensional array-like of numeric targets.

At fit time, the adapter constructs a DataFrame with columns:

• ds: timestamps parsed from X
• y: targets from y

and calls Prophet.fit(df).

At predict time, the adapter constructs a DataFrame with column ds and returns the yhat predictions as a one-dimensional numpy array.

Examples:

>>> from prophet import Prophet
>>> base = Prophet()
>>> model = ProphetAdapter(model=base)
>>> # X contains datetimes, y contains numeric targets
>>> # model.fit(X, y).predict(X)

fit(X, y, sample_weight=None)

Fit the underlying Prophet model.

Parameters:

Name	Type	Description	Default
X	ndarray	Time index values. Accepted forms are: - shape (n_samples,) of datetime-like values, or - shape (n_samples, n_features) where the first column is datetime-like	required
y	ndarray	Target vector of shape (n_samples,).	required
sample_weight	ndarray | None	Accepted for API compatibility but ignored by this adapter.	None

Returns:

Type	Description
ProphetAdapter	The fitted adapter (self), allowing method chaining.

Notes

This method imports pandas locally to avoid making pandas a hard dependency at module import time.

predict(X)

Predict using the fitted Prophet model.

Parameters:

Name	Type	Description	Default
X	ndarray	Time index values in the same format accepted by fit.	required

Returns:

Type	Description
ndarray	Predicted values of shape (n_samples,), taken from Prophet's yhat output column.

Raises:

Type	Description
RuntimeError	If the Prophet forecast output does not contain the yhat column.

Notes

This method imports pandas locally to avoid making pandas a hard dependency at module import time.