Impossible to plot Insample Anomalies without a `forecast_df`

[omriKramer]

Right now it is Impossible to plot anomalies for insample data without passing a forecast_df even though it is a valid use case. Moreover, the prediction intervals for df are filtered out by df = _filter_series(...) at line 162 which prevents plotting the insample anomalies.

[jmoralez]

The following should work:

from utilsforecast.data import generate_series
from utilsforecast.plotting import plot_series

level = [95]
series = generate_series(2, n_models=1, level=level)
plot_series(series, series, level=level, plot_anomalies=True, max_insample_length=0)

[omriKramer]

You are right, didn't realize I can use max_insample_length thank you!

[jmoralez]

We can make the df optional and being able to provide just forecasts_df for this case to avoid having to repeat it

[jmoralez]

Fixed by #98