New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Use timely-beliefs tooling to create forecast model specs #154
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I approve of the logic (looks very nice), but have some wished about documentation and an investigation into a parameter.
Does this deserve a changelog entry? I think it might. It would also force you to formulate the benefit of this PR from a higher perspective. I believe that with the level of timely beliefs integration we already have, this PR is necessary to achieve forecasting, or some part of that functionality. But I think you know this more precisely. |
One more thought: can you check if the tutorial on creating forecasts would require adaptations after this PR? |
# Conflicts: # documentation/changelog.rst
This PR introduces
TBSeriesSpecs
(subclassed fromSeriesSpecs
) that makes use of thecollect
functionality introduced earlier in FlexMeasures. We need such timely-beliefs tooling to ensure our beliefs data is properly mapped to a time series data structure that ML models can work with (i.e. feature frames). Currently, thecollect
functionality makes sure that the regressor data is deterministic and single-sourced.Also introduces the possibility to set the desired resolution of forecasts, plus some minor refactoring.