Closed bayliffe closed 11 months ago
All modified lines are covered by tests :white_check_mark:
Comparison is base (
1c3b584
) 98.39% compared to head (7ba91ff
) 98.39%.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Loosen dz scaling application to use the longest forecast period scaling coefficient available if no scaling coefficient has a forecast period that is longer than the forecast itself.
This will enable us to e.g. use a T+7 day scaling coefficient when calibrating a T+8 days forecast. It also means we have no warning that such use if going on. So for example, once we implement 14-day forecasts if we fail to update this ancillary with a greater range of forecast periods it will silently use the longest available.
Testing: