projet_gestion_commande/scripts/import_clients.py

#!/usr/bin/env python3
"""
Import clients depuis data.txt (format psql pipe-séparé) vers PostgreSQL.
Usage : python3 import_clients.py [--dry-run]
"""

import os
import sys

import psycopg2

# ── Connexion ──────────────────────────────────────────────────────────────────
DB_CONFIG = {
    "host": "localhost",
    "port": 5432,
    "user": "",
    "password": "",
    "dbname": "",
}

DATA_FILE = os.path.join(os.path.dirname(__file__), "data.txt")
DRY_RUN = "--dry-run" in sys.argv

# ── Ordre des colonnes dans data.txt (DB production, construite incrémentalement)
# None = colonne inconnue/obsolète à ignorer
SOURCE_COLUMNS = [
    "id",
    "username",
    "password",
    "nom",
    "prenom",
    "telephone",
    "command",
    "amende",
    "cancel_commande",
    "referral_balance",  # position 9  : NUMERIC(10,2) → "0.00"
    "cancellations_count",  # position 10 : INTEGER       → "0"
    None,  # position 11 : colonne obsolète/inconnue, ignorée
    "last_penalty_reason",  # position 12 : TEXT
    "updated_at",
    "created_at",
    "must_change_password",  # position 15 : BOOLEAN → "f"/"t"
]


def parse_value(v: str):
    """Convertit une valeur brute psql en type Python (None si vide)."""
    v = v.strip()
    if v == "" or v.lower() == "null":
        return None
    if v.lower() == "t":
        return True
    if v.lower() == "f":
        return False
    return v


def load_data(path: str) -> list[list]:
    rows = []
    with open(path, encoding="utf-8") as f:
        for line in f:
            line = line.rstrip("\n")
            if not line.strip():
                continue
            # Lignes séparateur psql (ex: "----+----")
            if (
                set(
                    line.replace("|", "")
                    .replace("+", "")
                    .replace("-", "")
                    .replace(" ", "")
                )
                == set()
            ):
                continue
            parts = line.split("|")
            values = (
                [parse_value(p) for p in parts[:-1]]
                if parts[-1].strip() == ""
                else [parse_value(p) for p in parts]
            )
            if values:
                rows.append(values)
    return rows


def main():
    rows = load_data(DATA_FILE)
    if not rows:
        print("❌ Aucune ligne trouvée dans data.txt")
        sys.exit(1)

    print(f"📂 {len(rows)} lignes lues depuis {DATA_FILE}")

    data_col_count = len(rows[0])
    print(
        f"📋 Colonnes data.txt : {data_col_count}  |  SOURCE_COLUMNS définis : {len(SOURCE_COLUMNS)}"
    )

    if data_col_count != len(SOURCE_COLUMNS):
        print(
            f"⚠️  Ajuste SOURCE_COLUMNS dans le script ({len(SOURCE_COLUMNS)} définis, {data_col_count} dans le fichier)"
        )
        sys.exit(1)

    # Colonnes à insérer (on filtre les None = colonnes ignorées)
    insert_cols = [col for col in SOURCE_COLUMNS if col is not None and col != "id"]
    # On conserve aussi l'id pour FORCE_ID
    insert_cols_with_id = ["id"] + insert_cols

    placeholders = ", ".join(["%s"] * len(insert_cols_with_id))
    col_names = ", ".join(insert_cols_with_id)
    update_expr = ", ".join(f"{c} = EXCLUDED.{c}" for c in insert_cols)

    sql = f"""
        INSERT INTO clients ({col_names})
        VALUES ({placeholders})
        ON CONFLICT (username) DO UPDATE SET {update_expr}
    """

    conn = psycopg2.connect(**DB_CONFIG)
    conn.autocommit = False
    cur = conn.cursor()

    inserted = 0
    updated = 0
    errors = 0

    for i, row in enumerate(rows, 1):
        # Construire un dict {col_name: value} depuis SOURCE_COLUMNS
        row_dict = {}
        for pos, col in enumerate(SOURCE_COLUMNS):
            if col is None:
                continue
            row_dict[col] = row[pos] if pos < len(row) else None

        values = [row_dict.get(c) for c in insert_cols_with_id]
        username = row_dict.get("username", "?")

        try:
            if DRY_RUN:
                print(f"  [DRY-RUN] ligne {i} : {username} → {row_dict}")
                inserted += 1
                continue

            cur.execute("SELECT 1 FROM clients WHERE username = %s", (username,))
            exists = cur.fetchone() is not None

            cur.execute(sql, values)

            if exists:
                updated += 1
            else:
                inserted += 1

        except Exception as e:
            print(f"  ❌ Erreur ligne {i} ({username}) : {e}")
            conn.rollback()
            errors += 1
            continue

    if not DRY_RUN:
        cur.execute(
            "SELECT setval('clients_id_seq', COALESCE((SELECT MAX(id) FROM clients), 1))"
        )
        conn.commit()

    cur.close()
    conn.close()

    print(f"\n{'🔍 DRY-RUN — ' if DRY_RUN else ''}✅ Terminé")
    print(f"   Insérés    : {inserted}")
    print(f"   Mis à jour : {updated}")
    print(f"   Erreurs    : {errors}")
    print(f"   ℹ️  Séquence clients_id_seq resynchronisée sur MAX(id)")


if __name__ == "__main__":
    main()