Fix/add fit/predict_kwargs argument to historical_forecasts

[madtoinou]

Fixes #1767.

Summary

• Added fit_kwargs and predict_kwargs argument to historical_forecasts, backtest and gridsearch
• Added some logic in fit_predict and predict_wrapper to pass the arguments to fit()/predict() only when it's supported (based on method signature)

Other Information

Rely on python to complain when unsupported or duplicated arguments are provided

[dennisbader]

Wouldn't this be a great opportunity to add fit and predict kwargs to historical forecasts? :)

[codecov-commenter]

Codecov Report

Attention: 2 lines in your changes are missing coverage. Please review.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Files	Coverage Δ
darts/models/forecasting/forecasting_model.py	94.87% <100.00%> (+0.49%)	⬆️
darts/models/forecasting/regression_model.py	96.76% <ø> (ø)
...arts/models/forecasting/torch_forecasting_model.py	90.76% <ø> (-0.03%)	⬇️
...casts/optimized_historical_forecasts_regression.py	97.22% <ø> (ø)
..._forecasts/optimized_historical_forecasts_torch.py	94.87% <100.00%> (+0.27%)	⬆️
darts/utils/historical_forecasts/utils.py	93.85% <93.33%> (-0.11%)	⬇️

... and 5 files with indirect coverage changes

📢 Thoughts on this report? Let us know!

[dennisbader]

Looks like a great start @madtoinou 🚀

I added a couple of suggestions on how we could simplify the kwargs handling and the fit/predict wrappers.
We can let python handle some of edge cases, and potentially, we can adapt the fit/predict wrappers to generalize over all models (and reduce code).

[dennisbader]

Looking really good now, thanks @madtoinou 🚀
Just some minor suggestions

[madtoinou]

Updated the PR, handling of the fit/predict_kwargs was improved and some tests were added.

[dennisbader]

Very nice, this looks awesome 🚀
This time really only minimal suggestions, we can merge soon.

[dennisbader]

Everything looks good now, great job @madtoinou 🚀

[pfk-beta]

Can't wait for release