FIX-2633: Use correct time indices when running historical forecasts with output_chunk_shift in regression models

[MattiasDC]

…on regression models with 'output_chunk_shift > 0' and 'output_chunk_length == 1'. Extended unit tests to cover this

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 #2633.

Summary

When creating historical forecasts using the optimized version, use the correct time indices when output_chunk_shift > 0 and output_chunk_length == 1. Extended the unit tests to cover this case

Other Information

[MattiasDC]

This if-check is extended with an additional condition, which fixes the issue.

[dennisbader]

Thanks @MattiasDC for finding and fixing this bug! It's greatly appreciated :)
PR also looks great. I made some changes to the tests (the bounds were not correct, but we didn't properly check that before) and updated the changelog.

I'll merge once all tests have passed. Thanks again 🚀

[MattiasDC]

I'm on my phone now, but you have a typo in test comments 'sampels' iso 'samples'

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 94.19%. Comparing base (aad1440) to head (d37c0d5).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2634      +/-   ##
==========================================
- Coverage   94.24%   94.19%   -0.06%     
==========================================
  Files         141      141              
  Lines       15463    15466       +3     
==========================================
- Hits        14573    14568       -5     
- Misses        890      898       +8     

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