Refactor/torch datasets improvement

[dennisbader]

Checklist before merging this PR:

• Mentioned all issues that this PR fixes or addresses.
• Summarized the updates of this PR under Summary.
• Added an entry under Unreleased in the Changelog.

Fixes #2686.

Summary

• simplifies training and inference datasets for TorchForecastingModel
• all training datasets now have uniform output: Tuple[past target, past cov, historic future cov, future cov, static cov, sample weight, future target]
• all ineference datasets now have uniform output: Tuple[past target, past cov, future past cov, historic future cov, future cov, static cov, target TimeSeries, pred start time]
• instead of having covariates specific datasets, the base datasets can now handle of covariates together. The remaining datasets are:
  • Training:
    • ShiftedTrainingDataset (old GenericShiftedDataset adapted to handle all covariates)
    • SequentialTrainingDataset (replaces old *CovariatesSequantialDataset)
    • HorizonBasedTrainingDataset (replaces old HorizonBasedDataset)
  • Prediction:
    • SequentialInfereceDataset (replaces old *CovariatesInferenceDataset)
• simplified HorizonBasedDataset to use ShiftedTrainingDataset as parent
• I observed a 10-15% performance boost for prediction with a model that uses all covariates

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[codecov]

Codecov Report

Attention: Patch coverage is 98.20467% with 10 lines in your changes missing coverage. Please review.

Project coverage is 95.13%. Comparing base (2309556) to head (34b3c21).
Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
darts/utils/data/__init__.py	25.00%	6 Missing ⚠️
...arts/models/forecasting/torch_forecasting_model.py	94.80%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2798      +/-   ##
==========================================
+ Coverage   94.61%   95.13%   +0.51%     
==========================================
  Files         145      145              
  Lines       15458    15049     -409     
==========================================
- Hits        14626    14317     -309     
+ Misses        832      732     -100     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.