"""Починка справочника: заменить повреждённые Excel-ошибкой service_id (#REF!) на чистые.

В исходном справочнике у части услуг код тарификатора пришёл как Excel `#REF!`, и это
протекло в канонический service_id (напр. `#REF!-#REF!#459`) — видно в живом /search.
Имена услуг при этом корректны. Скрипт генерирует чистые уникальные id и каскадно
обновляет ссылки в price_item (service_id и suggested_service_id) и learned_synonym.

Запуск: python scripts/fix_ref_ids.py <db_path>
"""

import sqlite3
import sys


def fix(db_path: str) -> dict:
    conn = sqlite3.connect(db_path)
    existing = {r[0] for r in conn.execute("SELECT service_id FROM service")}
    bad = [r[0] for r in conn.execute("SELECT service_id FROM service WHERE service_id LIKE '%#REF!%'")]

    mapping: dict[str, str] = {}
    i = 0
    for old in bad:
        while True:
            new = f"d{i:04d}"
            i += 1
            if new not in existing:
                break
        existing.add(new)
        mapping[old] = new

    # learned_synonym ссылается на service_id? обновим, если колонка есть.
    cols = {r[1] for r in conn.execute("PRAGMA table_info(learned_synonym)")}
    for old, new in mapping.items():
        conn.execute("UPDATE service SET service_id=? WHERE service_id=?", [new, old])
        conn.execute("UPDATE price_item SET service_id=? WHERE service_id=?", [new, old])
        conn.execute("UPDATE price_item SET suggested_service_id=? WHERE suggested_service_id=?", [new, old])
        if "service_id" in cols:
            conn.execute("UPDATE learned_synonym SET service_id=? WHERE service_id=?", [new, old])
    conn.commit()

    left = conn.execute("SELECT COUNT(*) FROM service WHERE service_id LIKE '%#REF!%'").fetchone()[0]
    left_pi = conn.execute("SELECT COUNT(*) FROM price_item WHERE service_id LIKE '%#REF!%'").fetchone()[0]
    matched = conn.execute("SELECT COUNT(*) FROM price_item WHERE service_id IS NOT NULL").fetchone()[0]
    conn.close()
    return {"fixed_services": len(mapping), "ref_left_service": left, "ref_left_price_item": left_pi, "matched_total": matched}


if __name__ == "__main__":
    print(fix(sys.argv[1] if len(sys.argv) > 1 else "data/medarchive.db"))