[BUG] Issue of the class ForecastingHorizon and related functions causing downstream issues with Prophet Ensemble

Describe the bug Since Prophet needs the index to be in pd.DatetimeIndex format, it does not work in ensemble mode with other models.

To Reproduce

y = load_airline()
y_train, y_test = temporal_train_test_split(y, test_size=36)
fh = ForecastingHorizon(y_test.index, is_relative=False)

forecaster = EnsembleForecaster([
    ("autoARIMA", AutoARIMA(sp=12, suppress_warnings=True)),
    ("autoETS", AutoETS(auto=True, sp=12, n_jobs=-1)),
    ("fbprophet", Prophet(seasonality_mode="multiplicative", add_country_holidays={"country_name": "US"}))
])
%time forecaster.fit(y_train)
y_pred = forecaster.predict(fh)

Error message is as follows

NotImplementedError: <class 'pandas.core.indexes.period.PeriodIndex'> is not supported. Please use type: <class 'pandas.core.indexes.datetimes.DatetimeIndex'> instead.

Expected behavior It would be ok to provide a working example of this use case. If it means I have to change data index to pd.DatetimeIndex for other models, I am ok doing that but I tried it and it gave an error as well

# Convert index to pd.DatetimeIndex
z = y.copy()
z.index = y.index.to_timestamp()
z_train, z_test = temporal_train_test_split(z, test_size=36)

forecaster = EnsembleForecaster([
    ("autoARIMA", AutoARIMA(sp=12, suppress_warnings=True)),
    ("autoETS", AutoETS(auto=True, sp=12, n_jobs=-1)),
    ("fbprophet", Prophet( seasonality_mode="multiplicative", add_country_holidays={"country_name": "US"}))
])
%time forecaster.fit(z_train)
y_pred = forecaster.predict(fh)

Error Message is as follows

sktime\forecasting\base\_fh.py in _check_cutoff(self, cutoff)
    388 
    389         if isinstance(index, pd.PeriodIndex):
--> 390             assert isinstance(cutoff, pd.Period)
    391         if isinstance(index, pd.DatetimeIndex):
    392             assert isinstance(cutoff, pd.Timestamp)

I tried another variant of the above (changes the arg to the pred function) but got a different error

forecaster = EnsembleForecaster([
    ("autoARIMA", AutoARIMA(sp=12, suppress_warnings=True)),
    ("autoETS", AutoETS(auto=True, sp=12, n_jobs=-1)),
    ("fbprophet", Prophet( seasonality_mode="multiplicative", add_country_holidays={"country_name": "US"}))
])
%time forecaster.fit(z_train)
y_pred = forecaster.predict(fh.to_relative(cutoff=y_train.index[-1]))

sktime\forecasting\base\_fh.py in _check_values(values)
    108     # check values does not contain duplicates
    109     if len(values) != values.nunique():
--> 110         raise ValueError("`values` must not contain duplicates.")
    111 
    112     # return sorted values

ValueError: `values` must not contain duplicates.

Versions System: python: 3.6.12 |Anaconda, Inc.| (default, Sep 9 2020, 00:29:25) [MSC v.1916 64 bit (AMD64)] executable: C:\Users\Nikhil.conda\envs\sktime_dev\python.exe machine: Windows-10-10.0.18362-SP0

Python dependencies: pip: 20.3.3 setuptools: 51.0.0.post20201207 sklearn: 0.24.0 sktime: 0.5.0 statsmodels: 0.12.1 numpy: 1.19.4 scipy: 1.5.4 Cython: 0.29.17 pandas: 1.1.5 matplotlib: 3.3.3 joblib: 1.0.0 numba: 0.52.0 pmdarima: 1.8.0 tsfresh: 0.17.0

[aiwalter]

@ngupta23 I was playing a bit around and was able to recreate this. This issues do not come from sktime.Prophet. Could you pls rename the this issue? It seems to be solely an issue of the class ForecastingHorizon and related functions.

[mloning]

Closed in favour or #534