#!/usr/bin/env python3
"""
Import scripts/generated-model-content.json into the model_content table.

Reads each row and upserts (brand_slug, model_slug) -> pdp_body, opener_variants,
closer_variants, style_tags, ai_model, ai_generated_at. Does NOT touch intro /
description / faq columns (owned by the template-based subtype page generator).

Run migration 010_model_content_ai.sql first.

Usage:
  python scripts/import-model-content.py              # import all rows
  python scripts/import-model-content.py --dry-run    # show what would change
  python scripts/import-model-content.py --brand nike # filter per merk
"""
import os
import sys
import json
import argparse
from pathlib import Path

import psycopg2
import psycopg2.extras
from dotenv import load_dotenv

load_dotenv()

DB_URL = os.environ["DATABASE_URL"]
HERE = Path(__file__).parent
INPUT = HERE / "generated-model-content.json"
AI_MODEL_TAG = "claude-opus-4-7-manual"


def load_rows():
    with open(INPUT, "r", encoding="utf-8") as f:
        return json.load(f)


def upsert(cur, row):
    cur.execute(
        """
        INSERT INTO model_content (
          brand_slug, model_slug, model_name,
          pdp_body, opener_variants, closer_variants, style_tags,
          ai_model, ai_generated_at
        )
        VALUES (%s, %s, %s, %s, %s::jsonb, %s::jsonb, %s::jsonb, %s, NOW())
        ON CONFLICT (brand_slug, model_slug) DO UPDATE SET
          model_name      = EXCLUDED.model_name,
          pdp_body        = EXCLUDED.pdp_body,
          opener_variants = EXCLUDED.opener_variants,
          closer_variants = EXCLUDED.closer_variants,
          style_tags      = EXCLUDED.style_tags,
          ai_model        = EXCLUDED.ai_model,
          ai_generated_at = EXCLUDED.ai_generated_at,
          updated_at      = NOW()
        """,
        (
            row["brand_slug"],
            row["model_slug"],
            row["model_name"],
            row["pdp_body"],
            json.dumps(row.get("opener_variants") or []),
            json.dumps(row.get("closer_variants") or []),
            json.dumps(row.get("style_tags") or []),
            AI_MODEL_TAG,
        ),
    )


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument("--dry-run", action="store_true")
    parser.add_argument("--brand", type=str, default=None)
    args = parser.parse_args()

    rows = load_rows()
    if args.brand:
        rows = [r for r in rows if r["brand_slug"] == args.brand]

    print(f"Importing {len(rows)} rows from {INPUT}")
    if args.dry_run:
        print("DRY RUN — no DB writes")

    conn = psycopg2.connect(DB_URL)
    conn.autocommit = False
    cur = conn.cursor()

    # Sanity check: the model_content table must have the new columns.
    cur.execute(
        """
        SELECT column_name FROM information_schema.columns
        WHERE table_name = 'model_content'
          AND column_name IN ('pdp_body','opener_variants','closer_variants','style_tags','ai_model','ai_generated_at')
        """
    )
    present = {r[0] for r in cur.fetchall()}
    expected = {"pdp_body", "opener_variants", "closer_variants", "style_tags", "ai_model", "ai_generated_at"}
    missing = expected - present
    if missing:
        print(f"ERROR: model_content is missing columns: {sorted(missing)}", file=sys.stderr)
        print("Run: psql -d $DATABASE -f db/migrations/010_model_content_ai.sql", file=sys.stderr)
        sys.exit(2)

    ok = 0
    fail = 0
    for idx, row in enumerate(rows, 1):
        label = f"{row['brand_slug']}/{row['model_slug']}"
        try:
            if not args.dry_run:
                upsert(cur, row)
            ok += 1
            print(f"[{idx}/{len(rows)}] OK   {label}")
        except Exception as e:
            fail += 1
            print(f"[{idx}/{len(rows)}] FAIL {label}: {e}", file=sys.stderr)
            conn.rollback()

    if args.dry_run:
        conn.rollback()
    else:
        conn.commit()
    conn.close()

    print(f"\nDone. ok={ok} fail={fail}")


if __name__ == "__main__":
    main()