#!/usr/bin/env python3
"""scripts/clean_csv.py

Usage:
    python3 scripts/clean_csv.py [input_dir] [--out-dir OUT] [--dry-run]

Ce script nettoie les fichiers CSV (trim, suppression de 'NULL', correction du
nombre de colonnes) et réalise des validations simples (colonnes ressemblant
à des UUID, colonnes date).
"""
import csv
import sys
from pathlib import Path
from datetime import datetime
import argparse
import re

UUID_RE = re.compile(r'^[0-9a-fA-F-]{36}$')


def is_uuid(s: str) -> bool:
    return bool(UUID_RE.match(s))


def looks_like_date(s: str) -> bool:
    if not s:
        return False
    # try ISO-like detection
    try:
        datetime.fromisoformat(s)
        return True
    except Exception:
        return False


def clean_csv_file(file_path: Path, out_path: Path = None, dry_run: bool = False):
    cleaned_rows = []
    errors = []

    with file_path.open("r", encoding="utf-8-sig", newline="") as infile:
        reader = csv.reader(infile)
        try:
            header = next(reader)
        except StopIteration:
            return {'file': str(file_path), 'error': 'empty', 'errors': []}

        nb_cols = len(header)
        cleaned_rows.append([h.strip() for h in header])

        # Heuristiques pour les validations
        uuid_cols = [i for i, h in enumerate(header) if h.lower() == 'id' or h.lower().endswith('_id')]
        date_cols = [i for i, h in enumerate(header) if any(k in h.lower() for k in ('date', '_le', '_at'))]

        for i, row in enumerate(reader, start=2):
            if len(row) < nb_cols:
                row.extend([""] * (nb_cols - len(row)))
            elif len(row) > nb_cols:
                row = row[:nb_cols]

            # Nettoyage cellule par cellule
            row = [cell.strip().replace('NULL', '') for cell in row]

            # Validations simples
            for ci in uuid_cols:
                if ci < len(row) and row[ci]:
                    if not is_uuid(row[ci]):
                        errors.append(f"Ligne {i} : colonne {header[ci]} ne ressemble pas à un UUID : {row[ci]!r}")
            for ci in date_cols:
                if ci < len(row) and row[ci]:
                    if not looks_like_date(row[ci]):
                        errors.append(f"Ligne {i} : colonne {header[ci]} n'est pas une date ISO-like : {row[ci]!r}")

            cleaned_rows.append(row)

    # Écrire la sortie si non --dry-run
    if not dry_run:
        target = out_path if out_path else file_path
        with target.open('w', encoding='utf-8', newline='') as outfile:
            writer = csv.writer(outfile)
            writer.writerows(cleaned_rows)

    return {'file': str(file_path), 'errors': errors}


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument('input_dir', nargs='?', default='bdd/data_test')
    parser.add_argument('--out-dir', '-o', help='Optional output directory for cleaned files')
    parser.add_argument('--dry-run', action='store_true', help='Do not write files, just report')
    args = parser.parse_args()

    base_dir = Path(args.input_dir)
    out_dir = Path(args.out_dir) if args.out_dir else None
    if out_dir and not out_dir.exists():
        out_dir.mkdir(parents=True)

    results = []
    for file_path in sorted(base_dir.glob('*.csv')):
        out_path = out_dir / file_path.name if out_dir else None
        res = clean_csv_file(file_path, out_path, args.dry_run)
        results.append(res)

    # Rapport
    any_errors = False
    for r in results:
        if 'error' in r:
            print(f"[AVERT] {r['file']}: {r['error']}")
        if r.get('errors'):
            any_errors = True
            print(f"[ERR] {r['file']} ->")
            for e in r['errors'][:20]:
                print('  -', e)
        else:
            print(f"[OK] {r['file']}")

    if any_errors:
        print('\nCertains fichiers présentent des problèmes de validation. Corrigez-les ou lancez avec --dry-run pour inspecter.')
        sys.exit(2)
    else:
        print('\nTous les fichiers ont été nettoyés avec succès.')


if __name__ == '__main__':
    main()