nordabiz/scripts/import_ceidg_to_db.py

#!/usr/bin/env python3
"""
Import CEIDG data (PKD, owner, start date) to database.

Reads cached CEIDG JSON files and updates company records with:
- PKD code and description
- Business start date
- Owner name (for JDG without KRS)

Usage:
    python scripts/import_ceidg_to_db.py              # Dry run
    python scripts/import_ceidg_to_db.py --apply      # Apply changes
    python scripts/import_ceidg_to_db.py --nip 5881571773  # Single company
"""

import os
import sys
import json
import argparse
from pathlib import Path
from datetime import datetime

# Add parent directory to path
sys.path.insert(0, str(Path(__file__).parent.parent))

# Load environment
from dotenv import load_dotenv
load_dotenv(Path(__file__).parent.parent / '.env')

from database import SessionLocal, Company


def find_latest_ceidg_json() -> Path:
    """Find the most recent CEIDG JSON file."""
    json_dir = Path(__file__).parent.parent / "data" / "ceidg_json"
    if not json_dir.exists():
        return None

    json_files = list(json_dir.glob("ceidg_data_*.json"))
    if not json_files:
        return None

    # Sort by modification time, newest first
    return sorted(json_files, key=lambda f: f.stat().st_mtime, reverse=True)[0]


def load_ceidg_data(json_path: Path) -> list:
    """Load CEIDG data from JSON file."""
    with open(json_path, 'r', encoding='utf-8') as f:
        return json.load(f)


def import_ceidg_to_db(apply: bool = False, target_nip: str = None):
    """
    Import CEIDG data to database.

    Args:
        apply: If True, actually apply changes. If False, dry run.
        target_nip: If set, only process this NIP.
    """
    json_path = find_latest_ceidg_json()
    if not json_path:
        print("ERROR: No CEIDG JSON files found in data/ceidg_json/")
        return

    print(f"Loading CEIDG data from: {json_path.name}")
    ceidg_records = load_ceidg_data(json_path)
    print(f"Found {len(ceidg_records)} CEIDG records")

    db = SessionLocal()
    updated = 0
    skipped = 0
    not_found = 0

    try:
        for record in ceidg_records:
            nip = record.get('nip')
            if not nip:
                continue

            # Filter by target NIP if specified
            if target_nip and nip != target_nip:
                continue

            # Find company by NIP
            company = db.query(Company).filter(Company.nip == nip).first()
            if not company:
                not_found += 1
                if target_nip:
                    print(f"  NOT FOUND: NIP {nip}")
                continue

            # Extract CEIDG data
            pkd_code = record.get('pkd_glowny')
            pkd_description = record.get('pkd_opis')
            start_date_str = record.get('data_rozpoczecia')
            owner = record.get('wlasciciel', {})
            owner_first = owner.get('imie', '').title() if owner else None
            owner_last = owner.get('nazwisko', '').title() if owner else None

            # Parse start date
            business_start_date = None
            if start_date_str:
                try:
                    business_start_date = datetime.strptime(start_date_str, '%Y-%m-%d').date()
                except ValueError:
                    pass

            # Check if there are any changes
            changes = []
            if pkd_code and company.pkd_code != pkd_code:
                changes.append(f"PKD: {company.pkd_code} → {pkd_code}")
            if pkd_description and company.pkd_description != pkd_description:
                changes.append(f"PKD opis: {'set' if not company.pkd_description else 'update'}")
            if business_start_date and company.business_start_date != business_start_date:
                changes.append(f"Data rozpoczęcia: {company.business_start_date} → {business_start_date}")
            if owner_first and company.owner_first_name != owner_first:
                changes.append(f"Właściciel: {owner_first} {owner_last}")

            if not changes:
                skipped += 1
                continue

            # Print changes
            print(f"\n{company.name} (NIP: {nip}):")
            for change in changes:
                print(f"  • {change}")

            if apply:
                # Apply updates
                if pkd_code:
                    company.pkd_code = pkd_code
                if pkd_description:
                    company.pkd_description = pkd_description
                if business_start_date:
                    company.business_start_date = business_start_date
                if owner_first:
                    company.owner_first_name = owner_first
                if owner_last:
                    company.owner_last_name = owner_last

                updated += 1

        if apply:
            db.commit()
            print(f"\n✅ Zaktualizowano {updated} firm")
        else:
            print(f"\n🔍 DRY RUN: {updated} firm do aktualizacji")
            print("   Użyj --apply aby zapisać zmiany")

        print(f"   Pominięto (bez zmian): {skipped}")
        print(f"   Nie znaleziono w bazie: {not_found}")

    finally:
        db.close()


def main():
    parser = argparse.ArgumentParser(description='Import CEIDG data to database')
    parser.add_argument('--apply', action='store_true', help='Apply changes (default: dry run)')
    parser.add_argument('--nip', type=str, help='Process only this NIP')

    args = parser.parse_args()

    print("=" * 60)
    print("CEIDG → Database Import")
    print("=" * 60)

    import_ceidg_to_db(apply=args.apply, target_nip=args.nip)


if __name__ == '__main__':
    main()