#!/usr/bin/env python3
"""Enrich the verified NPPES institutional email list with CMS revalidation status.

The institutional list (npi,email,...) is email-rich but has no revalidation
data, so only the OIG-screening segment can run on it. This joins it to the CMS
"Revalidation Due Date List" bulk file (revalidation_base.csv, keyed by NPI) so
the revalidation_overdue / revalidation_due_soon / npi_reactivation segments can
ALSO run on the full 63k -- unlocking the flagship $599 revalidation pitch
against the whole institutional pool instead of the 826-row warmup master.

Output adds the columns the campaign cron's selectors need:
  reval_due_date, days_overdue, reval_status (overdue|upcoming|not_on_list)

Usage:
  python3 scripts/enrich_institutional_revalidation.py \
      INSTITUTIONAL.csv  REVALIDATION_BASE.csv  OUT.csv
"""
from __future__ import annotations
import csv
import sys
from datetime import date, datetime

csv.field_size_limit(10_000_000)

COL_NPI = "National Provider Identifier"
COL_DUE = "Revalidation Due Date"
COL_ADJ = "Adjusted Due Date"


def parse_due(s: str):
    s = (s or "").strip()
    if not s or s.upper() == "TBD":
        return "TBD" if s.upper() == "TBD" else None
    for fmt in ("%m/%d/%Y", "%Y-%m-%d"):
        try:
            return datetime.strptime(s, fmt).date()
        except ValueError:
            continue
    return None


def main() -> int:
    inst_f, reval_f, out_f = sys.argv[1:4]
    today = date.today()

    # 1) NPI -> due date from the CMS revalidation list (last/best wins).
    due_by_npi: dict[str, object] = {}
    with open(reval_f, newline="", encoding="latin-1") as f:
        r = csv.DictReader(f)
        for row in r:
            npi = (row.get(COL_NPI) or "").strip()
            if not npi or not npi.isdigit():
                continue
            d = parse_due(row.get(COL_ADJ) or "") or parse_due(row.get(COL_DUE) or "")
            if d is not None:
                due_by_npi[npi] = d  # date or "TBD"
    print(f"revalidation list: {len(due_by_npi):,} NPIs with a due date/TBD", file=sys.stderr)

    # 2) enrich the institutional rows.
    on_list = overdue = upcoming = 0
    with open(inst_f, newline="", encoding="utf-8") as fin, \
         open(out_f, "w", newline="", encoding="utf-8") as fout:
        reader = csv.DictReader(fin)
        fields = reader.fieldnames + ["reval_due_date", "days_overdue", "reval_status"]
        w = csv.DictWriter(fout, fieldnames=fields)
        w.writeheader()
        total = 0
        for row in reader:
            total += 1
            npi = (row.get("npi") or "").strip()
            d = due_by_npi.get(npi)
            if d is None:
                row["reval_due_date"] = ""
                row["days_overdue"] = ""
                row["reval_status"] = "not_on_list"
            elif d == "TBD":
                on_list += 1
                row["reval_due_date"] = ""
                row["days_overdue"] = "0"
                row["reval_status"] = "upcoming"
                upcoming += 1
            else:
                on_list += 1
                od = (today - d).days
                row["reval_due_date"] = d.strftime("%m/%d/%Y")
                row["days_overdue"] = str(od)
                row["reval_status"] = "overdue" if od > 0 else "upcoming"
                if od > 0:
                    overdue += 1
                else:
                    upcoming += 1
            w.writerow(row)

    print(f"institutional rows: {total:,}", file=sys.stderr)
    print(f"  on revalidation list: {on_list:,} "
          f"(overdue={overdue:,}, upcoming={upcoming:,})", file=sys.stderr)
    print(f"  -> {out_f}", file=sys.stderr)
    return 0


if __name__ == "__main__":
    raise SystemExit(main())