Scalecast

Added
- added the `Pipeline.MVPipeline` object.
Changed
Fixed

Added
Changed
Fixed
- changed where import calls occur in the `SeriesTransformer` module to avoid a circular import that can happen
- added `AttributeError` to the list of exceptions to catch in `util.find_optimal_transformations()` function.
- fixed initiating MVForecaster() when `merge_Xvars=='i'`.

Added
- increased documentation around forecasting different model types.
Changed
- 'LevelY' passed to history in `util.break_mv_forecaster()`
- changed the optional dependency `pip intall fbprophet` to `pip install prophet` (18)
Fixed
- added `IndexError` to the list of exceptions to catch in `util.find_optimal_transformations()` function.
- convert values in `**kwargs` from `numpy.bool_` to `bool` type when forecasting with HWES (19).

Added
- added the `util.find_optimal_transformation()` function.
Changed
- changed setup.py so that version needs to be added manually to make an update (16).
- `Forecaster.integrate()` now has `**kwargs` that are passed to `Forecaster.adf_test()`.
- changed some data processing in `SeriesTransform.DiffRevert()` to accomodate a future change from pandas.
- changed the name of `util.find_series_transformation()` to `util.find_statistical_transformation()`. this function is new as of 0.15.0 so hopefully it has not been adopted widely yet and this change doesn't cause issues for users.
- added a try/except wrapper around the `self.f.typ_set()` code in `SeriesTransformer.DetrendTransform()`
Fixed
- drops null values before running though `Forecaster.normality_test()`.
- changed the import in `util` to `from scalecast import Pipeline` as the `util.find_series_transformation()` was not working because of it.

Added
- added the `SeriesTransformer.DetrendTransform()` and `SeriesTransformer.DetrendRevert()` functions.
Changed
Fixed
- all reverter functions in `SeriesTransformer` revert level confidence intervals.

Added
- added the `Pipeline` module.
- added `util.find_optimal_series_transformation()`
- added `Forecaster.normality_test()` function
Changed
- changed the default `full_res` arg in `Forecaster.adf_test()` to `True`
- changed links to some model documentation in readme
Fixed