preprocessing.coverage.SubmissionCoverage
preprocessing.coverage.SubmissionCoverage(last_full_hour, first_pred, n_steps)Coverage bookmarks computed by assert_submission_coverage.
These three values are the minimal output the downstream forecasting step needs: where context ends, where predictions start, and how many hourly steps to produce.
Attributes
| Name | Type | Description |
|---|---|---|
| last_full_hour | pd.Timestamp | Latest hour having a complete set of intra-hour samples (and passing the value-sanity check when corruption_mode="authoritative"). This is the context end / end_train anchor. |
| first_pred | pd.Timestamp | last_full_hour + 1 h. First prediction target. |
| n_steps | int | Number of hourly forecast steps from first_pred to last_target (inclusive), computed as (last_target - first_pred).total_seconds() / 3600 + 1. |