FlexMeasures / flexmeasures / 26493993418

81%

push

github

 Derive seasonal lags from sensor resolution (#2157)

* feat(forecasting): allow configurable seasonal lag steps

Context:
- Forecasting daily seasonality was hard-coded to 24 lag steps.
- That only represents one day for hourly sensors.

Change:
- Add a seasonal_lag_steps parameter to CustomLGBM.
- Keep the previous 24-step default for compatibility.

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

* feat(forecasting): derive seasonal lags from sensor resolution

Context:
- Daily lag steps depend on the target sensor resolution.
- PT15M sensors need 96 steps to represent one day, not 24.

Change:
- Compute one-day lag steps from the target sensor event resolution.
- Pass the derived value into CustomLGBM during training.

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

* fix(forecast): fall back for short seasonal histories

Context:
- Resolution-aware daily lags need enough training samples for the full requested horizon.
- Three days of PT15M data with a 48h horizon used to produce forecasts on main and should not become a hard failure.

Change:
- Fall back to the legacy 24-step lag pattern when the training window cannot support daily lags for the farthest horizon.
- Add regression coverage for under-sampled and sufficiently sampled PT15M histories.

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

* fix(forecast): validate seasonal lag steps

Context:
- Copilot noted externally configurable lag steps could be zero or negative.
- Invalid values would break modulo-based lag setup.

Change:
- Reject seasonal and fallback lag step values below 1.
- Fix the nearby seasonal-lag comment typo.

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

* fix(forecast): avoid truncated daily lag steps

Context:
- Dividing one day by sensor resolution with int() silently truncated non-divisor resolutions.
- That could train on an offset that was close to, but not exactly, one day.
... (continued)

38 of 41 new or added lines in 2 files covered. (92.68%)

12812 of 15914 relevant lines covered (80.51%)

0.81 hits per line