downloader.entsoe.compute_submission_dates

downloader.entsoe.compute_submission_dates(
    now,
    *,
    train_years,
    start_margin_days=45,
    data_end=None,
)

Compute all date bookmarks needed for one ENTSO-E submission run.

Encapsulates the compute_dates logic shared identically across chronos.py, team4_optuna_submit.py, team4_spotoptim_submit.py, and team4_4zones_submit.py. The download window (start_dl, end_dl) is derived by delegating to derive_download_window with the same arguments, guaranteeing that the ENTSO-E pull window is identical for all callers that pass the same train_years / start_margin_days / data_end.

Parameters

Name	Type	Description	Default
now	pd.Timestamp	Reference wall-clock timestamp (tz-aware, typically `pd.Timestamp.now(tz="UTC")`). Must be a `pd.Timestamp` with timezone; coercion is refused to prevent clock-skew bugs.	required
train_years	int	Number of full calendar years of history the downstream model requires. Forwarded to `derive_download_window`.	required
start_margin_days	int	Extra days prepended to the download window so lag features survive preprocessing. Defaults to `45`. Forwarded to `derive_download_window`.	`45`
data_end	str \| None	Optional upper bound for the download window in `'YYYYMMDDHHMM'` format. `None` (default) sets the end to `today + 2 days`. Forwarded to `derive_download_window`.	`None`

Returns

Name	Type	Description
SubmissionDates	SubmissionDates	Frozen dataclass with `now`, `today`,
	SubmissionDates	`yesterday`, `tomorrow`, `last_target`, `start_dl`, and
	SubmissionDates	`end_dl`.

Raises

Name	Type	Description
	TypeError	If `now` is not a tz-aware `pd.Timestamp`.
	ValueError	If `train_years` or `start_margin_days` violate the constraints of `derive_download_window`, or if `data_end` does not conform to `'YYYYMMDDHHMM'`.

Examples

import pandas as pd
from spotforecast2_safe.downloader.entsoe import compute_submission_dates

now = pd.Timestamp("2026-06-14 08:30", tz="UTC")
dates = compute_submission_dates(now, train_years=3)

print(dates.today)       # 2026-06-14 00:00:00+00:00
print(dates.yesterday)   # 2026-06-13 00:00:00+00:00
print(dates.tomorrow)    # 2026-06-15 00:00:00+00:00
print(dates.last_target) # 2026-06-15 23:00:00+00:00
print(dates.start_dl)    # 202305010000  (3 years + 45 days back)
print(dates.end_dl)      # 202606160000  (today + 2 days)

2026-06-14 00:00:00+00:00
2026-06-13 00:00:00+00:00
2026-06-15 00:00:00+00:00
2026-06-15 23:00:00+00:00
202305010000
202606160000