code/pipeline/steps/s19_finalize_selection.py

"""
s19_finalize_selection.py
Finalize the benchmark testbed selection.

Reads the deduplicated strict testbeds (L3/l3_testbeds_strict_dedup.csv, produced by
s16_dedup_select.py with return-period columns re-merged by s17_return_period.py;
legacy ds_260603 numbering: s28 and s27) and keeps the curated final set:
rows that are both `selected` (>=1 NCEI flash-flood point in the watershed) AND have
a fitted LP3 return period >= config.RP_MIN_FINAL years.

All columns are preserved. Writes L3/l3_final_selected.csv and reports the number of
testbeds and unique ff_episode_id.

Run:  python s19_finalize_selection.py
"""

import sys, pathlib
sys.path.insert(0, str(pathlib.Path(__file__).resolve().parents[1]))
import config
from lib import io, geo, nwis, nldi

import pandas as pd

DEDUP = config.B_L3 / "l3_testbeds_strict_dedup.csv"
OUT   = config.B_L3 / "l3_final_selected.csv"


def main():
    df = pd.read_csv(DEDUP, low_memory=False)

    sel = df[(df["selected"] == True)  # noqa: E712
             & (df["lp3_return_period_yr"] >= config.RP_MIN_FINAL)].copy()

    sel.to_csv(OUT, index=False)
    print(f"Wrote {OUT}")
    print(f"  final testbeds        : {len(sel)}")
    print(f"  unique ff_episode_id  : {sel['ff_episode_id'].nunique()}")


if __name__ == "__main__":
    main()