gateway/tool_db_export_migration.py

# Copyright (c) 2025 Patrick Motsch
# All rights reserved.
"""
Datenbank Export-Tool für Migration.

Dieses Script exportiert alle Daten aus ALLEN PowerOn PostgreSQL-Datenbanken
in eine JSON-Datei, die als Migrationsdatensatz verwendet werden kann.
Zusätzlich wird eine separate JSON-Datei mit nur den Strukturen (ohne Daten)
erstellt: <dateiname>_structure.json

Datenbanken:
    - poweron_app        (User, Mandate, RBAC, Features, etc.)
    - poweron_chat       (Chat-Konversationen und Nachrichten)
    - poweron_management (Workflows, Prompts, Connections, etc.)
    - poweron_realestate (Real Estate Daten)
    - poweron_trustee    (Trustee Daten)

Verwendung:
    python tool_db_export_migration.py [--output <pfad>] [--pretty]

Optionen:
    --output, -o    Pfad zur Ausgabedatei (Standard: migration_export_<timestamp>.json)
                    Die Struktur-Datei wird automatisch als <dateiname>_structure.json erstellt
    --pretty, -p    JSON formatiert ausgeben (für bessere Lesbarkeit)
    --exclude       Komma-getrennte Liste von Tabellen, die ausgeschlossen werden sollen
    --include-meta  System-Metadaten (_createdAt, _modifiedAt, etc.) beibehalten
    --db            Nur bestimmte Datenbank(en) exportieren (komma-getrennt)
"""

import os
import sys
import json
import argparse
import logging
from datetime import datetime
from typing import Dict, List, Any, Optional
from pathlib import Path

# Add gateway to path for imports and set working directory
# Find gateway directory (could be in local/pending/ or gateway/)
scriptPath = Path(__file__).resolve()
gatewayPath = scriptPath.parent
# If we're in local/pending/, go up to find gateway/
if gatewayPath.name == "pending":
    gatewayPath = gatewayPath.parent.parent / "gateway"
elif gatewayPath.name == "local":
    gatewayPath = gatewayPath.parent / "gateway"
# If gateway doesn't exist, try current directory
if not gatewayPath.exists():
    gatewayPath = Path(__file__).parent.parent.parent / "gateway"
if gatewayPath.exists():
    sys.path.insert(0, str(gatewayPath))
    # Change working directory to gateway so APP_CONFIG can find .env file
    os.chdir(str(gatewayPath))
else:
    # Fallback: assume we're already in gateway/ or add parent
    sys.path.insert(0, str(Path(__file__).parent))
    # Try to change to gateway directory if it exists
    potentialGateway = Path(__file__).parent
    if potentialGateway.exists() and (potentialGateway / "modules" / "shared" / "configuration.py").exists():
        os.chdir(str(potentialGateway))

import psycopg2
import psycopg2.extras

# Use the real APP_CONFIG which handles encryption and environment-specific configs
from modules.shared.configuration import APP_CONFIG  # type: ignore

# Logging konfigurieren
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s',
    datefmt='%Y-%m-%d %H:%M:%S'
)
logger = logging.getLogger(__name__)

# Force APP_CONFIG to refresh after changing working directory
# This ensures .env file is loaded from the correct location
try:
    APP_CONFIG.refresh()
    envType = APP_CONFIG.get('APP_ENV_TYPE', 'unknown')
    keySysVar = APP_CONFIG.get('APP_KEY_SYSVAR', 'not_set')
    logger.debug(f"APP_CONFIG refreshed. Environment type: {envType}")
    logger.debug(f"APP_KEY_SYSVAR: {keySysVar}")
    logger.debug(f"Current working directory: {os.getcwd()}")
    
    # Check if master key is available (needed for decrypting secrets)
    if keySysVar != 'not_set':
        masterKeyEnv = os.environ.get(keySysVar)
        if masterKeyEnv:
            logger.debug(f"Master key found in environment variable: {keySysVar}")
        else:
            logger.warning(f"Master key not found in environment variable: {keySysVar}")
            logger.warning("Encrypted secrets may not be decryptable!")
except Exception as e:
    logger.warning(f"Could not refresh APP_CONFIG: {e}")

# Alle PowerOn Datenbanken
ALL_DATABASES = [
    "poweron_app",        # Haupt-App: User, Mandate, RBAC, Features
    "poweron_chat",       # Chat-Konversationen
    "poweron_management", # Workflows, Prompts, Connections
    "poweron_realestate", # Real Estate
    "poweron_trustee",    # Trustee
]

# Datenbank-Konfiguration: Mapping von DB-Name zu Config-Prefix
# Jede Datenbank hat ihre eigenen Variablen: DB_APP_HOST, DB_CHAT_HOST, etc.
DATABASE_CONFIG = {
    "poweron_app": "DB_APP",           # DB_APP_HOST, DB_APP_USER, DB_APP_PASSWORD_SECRET, etc.
    "poweron_chat": "DB_CHAT",         # DB_CHAT_HOST, DB_CHAT_USER, etc.
    "poweron_management": "DB_MANAGEMENT",
    "poweron_realestate": "DB_REALESTATE",
    "poweron_trustee": "DB_TRUSTEE",
}


def _getConfigValue(key: str, default: str = None) -> str:
    """Holt einen Konfigurationswert über APP_CONFIG (unterstützt Verschlüsselung)."""
    return APP_CONFIG.get(key, default)


def _getDbConfig(dbName: str) -> Dict[str, Any]:
    """
    Holt die Datenbankverbindungs-Konfiguration für eine spezifische Datenbank.
    Unterstützt sowohl DB-spezifische Variablen (DB_APP_HOST) als auch Fallback (DB_HOST).
    """
    prefix = DATABASE_CONFIG.get(dbName, "DB")
    
    # Versuche zuerst DB-spezifische Variablen, dann Fallback auf allgemeine
    host = _getConfigValue(f"{prefix}_HOST") or _getConfigValue("DB_HOST", "localhost")
    user = _getConfigValue(f"{prefix}_USER") or _getConfigValue("DB_USER")
    password = _getConfigValue(f"{prefix}_PASSWORD_SECRET") or _getConfigValue("DB_PASSWORD_SECRET")
    port = _getConfigValue(f"{prefix}_PORT") or _getConfigValue("DB_PORT", "5432")
    
    return {
        "host": host,
        "user": user,
        "password": password,
        "port": int(port) if port else 5432
    }


def _databaseExists(dbDatabase: str) -> bool:
    """Prüft ob eine Datenbank existiert."""
    config = _getDbConfig(dbDatabase)
    
    if not config["user"]:
        logger.warning(f"DB-User nicht gesetzt für Datenbank {dbDatabase}")
        return False
    if not config["password"]:
        logger.warning(f"DB-Password nicht gesetzt für Datenbank {dbDatabase}")
        return False
    
    try:
        # Verbinde zur postgres Datenbank um zu prüfen
        conn = psycopg2.connect(
            host=config["host"],
            port=config["port"],
            database="postgres",
            user=config["user"],
            password=config["password"]
        )
        conn.autocommit = True
        
        with conn.cursor() as cursor:
            cursor.execute(
                "SELECT 1 FROM pg_database WHERE datname = %s",
                (dbDatabase,)
            )
            exists = cursor.fetchone() is not None
        
        conn.close()
        return exists
    except Exception as e:
        logger.error(f"Fehler beim Prüfen der Datenbank {dbDatabase} auf {config['host']}:{config['port']}: {e}")
        return False


def _getDbConnection(dbDatabase: str):
    """Erstellt eine Verbindung zu einer spezifischen PostgreSQL-Datenbank."""
    config = _getDbConfig(dbDatabase)
    
    # Prüfe ob wichtige Konfigurationswerte fehlen
    if not config["user"]:
        logger.error(f"DB-User nicht gesetzt für {dbDatabase} - kann keine Verbindung herstellen")
        return None
    if not config["password"]:
        logger.error(f"DB-Password nicht gesetzt für {dbDatabase} - kann keine Verbindung herstellen")
        return None
    
    # Erst prüfen ob Datenbank existiert
    if not _databaseExists(dbDatabase):
        logger.warning(f"Datenbank '{dbDatabase}' existiert nicht auf {config['host']}:{config['port']} - übersprungen")
        return None
    
    try:
        conn = psycopg2.connect(
            host=config["host"],
            port=config["port"],
            database=dbDatabase,
            user=config["user"],
            password=config["password"],
            cursor_factory=psycopg2.extras.RealDictCursor
        )
        # Autocommit muss VOR set_client_encoding gesetzt werden, um Transaction-Konflikte zu vermeiden
        conn.autocommit = True
        conn.set_client_encoding('UTF8')
        logger.debug(f"Erfolgreich verbunden mit {config['host']}:{config['port']}/{dbDatabase}")
        return conn
    except Exception as e:
        logger.error(f"Datenbankverbindung zu {dbDatabase} auf {config['host']}:{config['port']} fehlgeschlagen: {e}")
        raise


def _getTables(conn) -> List[str]:
    """Gibt alle Tabellennamen in der Datenbank zurück."""
    with conn.cursor() as cursor:
        cursor.execute("""
            SELECT table_name 
            FROM information_schema.tables 
            WHERE table_schema = 'public' 
            AND table_type = 'BASE TABLE'
            ORDER BY table_name
        """)
        tables = [row["table_name"] for row in cursor.fetchall()]
    return tables


def _getTableData(conn, tableName: str, includeMeta: bool = False) -> List[Dict[str, Any]]:
    """Liest alle Daten aus einer Tabelle."""
    with conn.cursor() as cursor:
        cursor.execute(f'SELECT * FROM "{tableName}"')
        rows = cursor.fetchall()
        
        records = []
        for row in rows:
            record = dict(row)
            
            # Optional: System-Metadaten entfernen
            if not includeMeta:
                metaFields = ["_createdAt", "_modifiedAt", "_createdBy", "_modifiedBy"]
                for field in metaFields:
                    record.pop(field, None)
            
            # Konvertiere JSONB-Felder (sind bereits als Dict/List von psycopg2)
            for key, value in record.items():
                if isinstance(value, (int, float)):
                    record[key] = float(value) if isinstance(value, float) else int(value)
            
            records.append(record)
        
        return records


def _getTableRowCount(conn, tableName: str) -> int:
    """Zählt die Anzahl der Zeilen in einer Tabelle."""
    with conn.cursor() as cursor:
        cursor.execute(f'SELECT COUNT(*) as count FROM "{tableName}"')
        result = cursor.fetchone()
        return result["count"] if result else 0


def _getTableStructure(conn, tableName: str) -> Dict[str, Any]:
    """Holt die Struktur einer Tabelle (Spalten, Constraints, Indizes) ohne Daten."""
    structure = {
        "columns": [],
        "primaryKeys": [],
        "foreignKeys": [],
        "uniqueConstraints": [],
        "indexes": [],
        "checkConstraints": []
    }
    
    # Connection hat bereits autocommit = True, daher keine Transaction-Probleme
    with conn.cursor() as cursor:
            # Spalten-Informationen
            cursor.execute("""
                SELECT 
                    column_name,
                    data_type,
                    character_maximum_length,
                    numeric_precision,
                    numeric_scale,
                    is_nullable,
                    column_default,
                    udt_name
                FROM information_schema.columns
                WHERE table_schema = 'public' 
                AND table_name = %s
                ORDER BY ordinal_position
            """, (tableName,))
            
            for row in cursor.fetchall():
                colInfo = {
                    "name": row["column_name"],
                    "type": row["data_type"],
                    "udtName": row["udt_name"],
                    "nullable": row["is_nullable"] == "YES",
                    "default": row["column_default"]
                }
                
                if row["character_maximum_length"]:
                    colInfo["maxLength"] = row["character_maximum_length"]
                if row["numeric_precision"]:
                    colInfo["precision"] = row["numeric_precision"]
                if row["numeric_scale"]:
                    colInfo["scale"] = row["numeric_scale"]
                
                structure["columns"].append(colInfo)
            
            # Primary Keys
            cursor.execute("""
                SELECT 
                    kcu.column_name
                FROM information_schema.table_constraints tc
                JOIN information_schema.key_column_usage kcu 
                    ON tc.constraint_name = kcu.constraint_name
                WHERE tc.table_schema = 'public'
                AND tc.table_name = %s
                AND tc.constraint_type = 'PRIMARY KEY'
                ORDER BY kcu.ordinal_position
            """, (tableName,))
            
            structure["primaryKeys"] = [row["column_name"] for row in cursor.fetchall()]
            
            # Foreign Keys
            cursor.execute("""
                SELECT
                    kcu.column_name,
                    ccu.table_name AS foreign_table_name,
                    ccu.column_name AS foreign_column_name,
                    tc.constraint_name
                FROM information_schema.table_constraints AS tc
                JOIN information_schema.key_column_usage AS kcu
                    ON tc.constraint_name = kcu.constraint_name
                JOIN information_schema.constraint_column_usage AS ccu
                    ON ccu.constraint_name = tc.constraint_name
                WHERE tc.constraint_type = 'FOREIGN KEY'
                AND tc.table_schema = 'public'
                AND tc.table_name = %s
            """, (tableName,))
            
            for row in cursor.fetchall():
                structure["foreignKeys"].append({
                    "column": row["column_name"],
                    "referencesTable": row["foreign_table_name"],
                    "referencesColumn": row["foreign_column_name"],
                    "constraintName": row["constraint_name"]
                })
            
            # Unique Constraints - FIX: Tabellen-Aliase verwenden um ambiguous columns zu vermeiden
            cursor.execute("""
                SELECT 
                    kcu.column_name,
                    tc.constraint_name
                FROM information_schema.table_constraints tc
                JOIN information_schema.key_column_usage kcu 
                    ON tc.constraint_name = kcu.constraint_name
                    AND tc.table_schema = kcu.table_schema
                    AND tc.table_name = kcu.table_name
                WHERE tc.table_schema = 'public'
                AND tc.table_name = %s
                AND tc.constraint_type = 'UNIQUE'
                ORDER BY kcu.ordinal_position
            """, (tableName,))
            
            uniqueGroups = {}
            for row in cursor.fetchall():
                constraintName = row["constraint_name"]
                if constraintName not in uniqueGroups:
                    uniqueGroups[constraintName] = []
                uniqueGroups[constraintName].append(row["column_name"])
            
            structure["uniqueConstraints"] = [
                {"columns": cols, "constraintName": name}
                for name, cols in uniqueGroups.items()
            ]
            
            # Indizes (ohne Primary Key und Unique Constraints)
            cursor.execute("""
                SELECT
                    i.relname AS index_name,
                    a.attname AS column_name,
                    ix.indisunique AS is_unique
                FROM pg_class t
                JOIN pg_index ix ON t.oid = ix.indrelid
                JOIN pg_class i ON i.oid = ix.indexrelid
                JOIN pg_attribute a ON a.attrelid = t.oid AND a.attnum = ANY(ix.indkey)
                WHERE t.relkind = 'r'
                AND t.relname = %s
                AND NOT ix.indisprimary
                ORDER BY i.relname, a.attnum
            """, (tableName,))
            
            indexGroups = {}
            for row in cursor.fetchall():
                indexName = row["index_name"]
                if indexName not in indexGroups:
                    indexGroups[indexName] = {
                        "name": indexName,
                        "columns": [],
                        "unique": row["is_unique"]
                    }
                indexGroups[indexName]["columns"].append(row["column_name"])
            
            structure["indexes"] = list(indexGroups.values())
            
            # Check Constraints - FIX: Tabellen-Aliase verwenden
            cursor.execute("""
                SELECT
                    cc.constraint_name,
                    cc.check_clause
                FROM information_schema.check_constraints cc
                JOIN information_schema.constraint_column_usage ccu
                    ON cc.constraint_name = ccu.constraint_name
                WHERE ccu.table_schema = 'public'
                AND ccu.table_name = %s
            """, (tableName,))
            
            for row in cursor.fetchall():
                structure["checkConstraints"].append({
                    "constraintName": row["constraint_name"],
                    "checkClause": row["check_clause"]
                })
    
    return structure


def _exportSingleDatabaseStructure(
    dbDatabase: str,
    excludeTables: List[str]
) -> Optional[Dict[str, Any]]:
    """Exportiert nur die Struktur einer einzelnen Datenbank (ohne Daten)."""
    conn = _getDbConnection(dbDatabase)
    
    if conn is None:
        return None
    
    try:
        allTables = _getTables(conn)
        
        # System-Tabellen ausschliessen
        systemTables = ["_system"]
        tablesToExport = [
            t for t in allTables 
            if t not in systemTables and t not in excludeTables
        ]
        
        dbExport = {
            "tables": {},
            "summary": {},
            "tableCount": len(tablesToExport)
        }
        
        for tableName in tablesToExport:
            try:
                structure = _getTableStructure(conn, tableName)
                dbExport["tables"][tableName] = structure
                dbExport["summary"][tableName] = {
                    "columnCount": len(structure["columns"]),
                    "primaryKeyCount": len(structure["primaryKeys"]),
                    "foreignKeyCount": len(structure["foreignKeys"]),
                    "indexCount": len(structure["indexes"])
                }
                
                logger.info(f"    {tableName}: {len(structure['columns'])} Spalten")
                
            except Exception as e:
                logger.error(f"    Fehler bei Tabelle {tableName}: {e}")
                dbExport["tables"][tableName] = {}
                dbExport["summary"][tableName] = {"error": str(e)}
        
        return dbExport
        
    finally:
        conn.close()


def _exportSingleDatabase(
    dbDatabase: str,
    excludeTables: List[str],
    includeMeta: bool
) -> Optional[Dict[str, Any]]:
    """Exportiert eine einzelne Datenbank."""
    conn = _getDbConnection(dbDatabase)
    
    if conn is None:
        return None
    
    try:
        allTables = _getTables(conn)
        
        # System-Tabellen ausschliessen
        systemTables = ["_system"]
        tablesToExport = [
            t for t in allTables 
            if t not in systemTables and t not in excludeTables
        ]
        
        dbExport = {
            "tables": {},
            "summary": {},
            "tableCount": len(tablesToExport),
            "totalRecords": 0
        }
        
        for tableName in tablesToExport:
            try:
                records = _getTableData(conn, tableName, includeMeta)
                rowCount = len(records)
                dbExport["totalRecords"] += rowCount
                
                dbExport["tables"][tableName] = records
                dbExport["summary"][tableName] = {"recordCount": rowCount}
                
                if rowCount > 0:
                    logger.info(f"    {tableName}: {rowCount} Datensätze")
                
            except Exception as e:
                logger.error(f"    Fehler bei Tabelle {tableName}: {e}")
                dbExport["tables"][tableName] = []
                dbExport["summary"][tableName] = {"recordCount": 0, "error": str(e)}
                # Bei autocommit = True ist kein rollback() notwendig
        
        return dbExport
        
    finally:
        conn.close()


def exportDatabase(
    outputPath: Optional[str] = None,
    prettyPrint: bool = False,
    excludeTables: Optional[List[str]] = None,
    includeMeta: bool = False,
    onlyDatabases: Optional[List[str]] = None
) -> str:
    """
    Exportiert alle Datenbanken in eine JSON-Datei.
    Erstellt zusätzlich eine separate JSON-Datei mit nur den Strukturen (ohne Daten).
    
    Args:
        outputPath: Pfad zur Ausgabedatei (optional)
        prettyPrint: JSON formatiert ausgeben
        excludeTables: Liste von Tabellen, die ausgeschlossen werden sollen
        includeMeta: System-Metadaten beibehalten
        onlyDatabases: Nur diese Datenbanken exportieren
        
    Returns:
        Pfad zur erstellten Exportdatei
    """
    excludeTables = excludeTables or []
    
    # Welche Datenbanken exportieren?
    databasesToExport = onlyDatabases if onlyDatabases else ALL_DATABASES
    
    # Prüfe ob grundlegende Konfigurationswerte vorhanden sind
    # APP_CONFIG.get() entschlüsselt automatisch Werte, die mit _SECRET enden
    # Jede Datenbank hat ihre eigenen Variablen (DB_APP_USER, DB_CHAT_USER, etc.)
    missingConfigs = []
    for dbName in databasesToExport:
        config = _getDbConfig(dbName)
        if not config["user"] or not config["password"]:
            missingConfigs.append(f"{dbName}: User={'gesetzt' if config['user'] else 'FEHLT'}, Password={'gesetzt' if config['password'] else 'FEHLT'}")
    
    if missingConfigs:
        logger.error("WICHTIG: Einige Datenbank-Konfigurationen fehlen!")
        for missing in missingConfigs:
            logger.error(f"  {missing}")
        logger.error("  Bitte Umgebungsvariablen setzen oder .env Datei konfigurieren")
        logger.error("  Hinweis: Verschlüsselte Secrets benötigen APP_KEY_SYSVAR Umgebungsvariable!")
        logger.error("  Erwartete Variablen: DB_APP_USER, DB_CHAT_USER, DB_MANAGEMENT_USER, etc.")
    
    # Standard-Ausgabepfad generieren (im Log-Ordner)
    if not outputPath:
        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
        logDir = _getConfigValue("APP_LOGGING_LOG_DIR")
        if logDir and os.path.isabs(logDir):
            outputDir = logDir
        else:
            outputDir = os.path.join(os.path.dirname(__file__), "local", "logs")
        os.makedirs(outputDir, exist_ok=True)
        outputPath = os.path.join(outputDir, f"migration_export_{timestamp}.json")
    
    logger.info(f"Starte Export von {len(databasesToExport)} Datenbank(en)...")
    logger.info(f"Datenbanken: {', '.join(databasesToExport)}")
    
    # Export-Struktur erstellen
    exportData = {
        "meta": {
            "exportedAt": datetime.utcnow().isoformat() + "Z",
            "exportedFrom": _getConfigValue("APP_ENV_LABEL", "unknown"),
            "version": "1.0",
            "databaseCount": 0,
            "totalTables": 0,
            "totalRecords": 0,
            "excludedTables": excludeTables,
            "includesMeta": includeMeta
        },
        "databases": {}
    }
    
    # Struktur-Export erstellen
    structureData = {
        "meta": {
            "exportedAt": datetime.utcnow().isoformat() + "Z",
            "exportedFrom": _getConfigValue("APP_ENV_LABEL", "unknown"),
            "version": "1.0",
            "databaseCount": 0,
            "totalTables": 0,
            "excludedTables": excludeTables,
            "note": "Nur Strukturen, keine Daten"
        },
        "databases": {}
    }
    
    # Jede Datenbank exportieren
    for dbName in databasesToExport:
        logger.info(f"Exportiere Datenbank: {dbName}")
        
        # Daten exportieren
        dbExport = _exportSingleDatabase(dbName, excludeTables, includeMeta)
        
        # Struktur exportieren
        logger.info(f"Exportiere Struktur für Datenbank: {dbName}")
        dbStructure = _exportSingleDatabaseStructure(dbName, excludeTables)
        
        if dbExport is not None:
            exportData["databases"][dbName] = dbExport
            exportData["meta"]["databaseCount"] += 1
            exportData["meta"]["totalTables"] += dbExport["tableCount"]
            exportData["meta"]["totalRecords"] += dbExport["totalRecords"]
            logger.info(f"  -> {dbExport['tableCount']} Tabellen, {dbExport['totalRecords']} Datensätze")
        else:
            logger.info(f"  -> Übersprungen (existiert nicht)")
        
        if dbStructure is not None:
            structureData["databases"][dbName] = dbStructure
            structureData["meta"]["databaseCount"] += 1
            structureData["meta"]["totalTables"] += dbStructure["tableCount"]
    
    # JSON-Datei mit Daten schreiben
    logger.info(f"Schreibe Exportdatei: {outputPath}")
    
    with open(outputPath, "w", encoding="utf-8") as f:
        if prettyPrint:
            json.dump(exportData, f, indent=2, ensure_ascii=False, default=str)
        else:
            json.dump(exportData, f, ensure_ascii=False, default=str)
    
    # JSON-Datei mit Strukturen schreiben
    structurePath = outputPath.replace(".json", "_structure.json")
    logger.info(f"Schreibe Struktur-Exportdatei: {structurePath}")
    
    with open(structurePath, "w", encoding="utf-8") as f:
        if prettyPrint:
            json.dump(structureData, f, indent=2, ensure_ascii=False, default=str)
        else:
            json.dump(structureData, f, ensure_ascii=False, default=str)
    
    # Dateigrössen berechnen
    fileSize = os.path.getsize(outputPath)
    fileSizeStr = _formatFileSize(fileSize)
    
    structureFileSize = os.path.getsize(structurePath)
    structureFileSizeStr = _formatFileSize(structureFileSize)
    
    logger.info(f"Export abgeschlossen!")
    logger.info(f"  Datenbanken: {exportData['meta']['databaseCount']}")
    logger.info(f"  Tabellen: {exportData['meta']['totalTables']}")
    logger.info(f"  Datensätze: {exportData['meta']['totalRecords']}")
    logger.info(f"  Daten-Export: {fileSizeStr} - {outputPath}")
    logger.info(f"  Struktur-Export: {structureFileSizeStr} - {structurePath}")
    
    return outputPath


def _formatFileSize(sizeBytes: int) -> str:
    """Formatiert Dateigrösse in lesbares Format."""
    for unit in ['B', 'KB', 'MB', 'GB']:
        if sizeBytes < 1024:
            return f"{sizeBytes:.2f} {unit}"
        sizeBytes /= 1024
    return f"{sizeBytes:.2f} TB"


def printDatabaseSummary():
    """Zeigt eine Zusammenfassung aller Datenbanken an."""
    print("\n" + "=" * 70)
    print("DATENBANK ZUSAMMENFASSUNG - ALLE POWEREON DATENBANKEN")
    print("=" * 70)
    print(f"Umgebung: {_getConfigValue('APP_ENV_LABEL', 'unknown')}")
    print(f"Host: {_getConfigValue('DB_HOST', 'localhost')}")
    print("=" * 70)
    
    grandTotalRecords = 0
    grandTotalTables = 0
    
    for dbName in ALL_DATABASES:
        print(f"\n{dbName}")
        print("-" * 70)
        
        conn = _getDbConnection(dbName)
        if conn is None:
            print("  (Datenbank existiert nicht)")
            continue
        
        try:
            tables = _getTables(conn)
            dbTotalRecords = 0
            
            print(f"  {'Tabelle':<45} {'Datensätze':>15}")
            print(f"  {'-' * 45} {'-' * 15}")
            
            for tableName in tables:
                if tableName.startswith("_"):
                    continue  # System-Tabellen überspringen
                count = _getTableRowCount(conn, tableName)
                dbTotalRecords += count
                if count > 0:  # Nur nicht-leere Tabellen anzeigen
                    print(f"  {tableName:<45} {count:>15}")
            
            print(f"  {'-' * 45} {'-' * 15}")
            print(f"  {'Gesamt':<45} {dbTotalRecords:>15}")
            
            grandTotalRecords += dbTotalRecords
            grandTotalTables += len([t for t in tables if not t.startswith("_")])
            
        finally:
            conn.close()
    
    print("\n" + "=" * 70)
    print(f"GESAMTÜBERSICHT")
    print(f"  Datenbanken: {len(ALL_DATABASES)}")
    print(f"  Tabellen: {grandTotalTables}")
    print(f"  Datensätze: {grandTotalRecords}")
    print("=" * 70 + "\n")


def main():
    parser = argparse.ArgumentParser(
        description="Exportiert alle PowerOn Datenbank-Daten für Migration",
        formatter_class=argparse.RawDescriptionHelpFormatter,
        epilog="""
Datenbanken:
  poweron_app        - User, Mandate, RBAC, Features
  poweron_chat       - Chat-Konversationen
  poweron_management - Workflows, Prompts, Connections
  poweron_realestate - Real Estate Daten
  poweron_trustee    - Trustee Daten

Beispiele:
  python tool_db_export_migration.py
  python tool_db_export_migration.py --pretty
  python tool_db_export_migration.py -o backup.json --pretty
  python tool_db_export_migration.py --db poweron_app,poweron_chat
  python tool_db_export_migration.py --exclude Token,AuthEvent --include-meta
  python tool_db_export_migration.py --summary
        """
    )
    
    parser.add_argument(
        "-o", "--output",
        help="Pfad zur Ausgabedatei",
        type=str,
        default=None
    )
    
    parser.add_argument(
        "-p", "--pretty",
        help="JSON formatiert ausgeben",
        action="store_true"
    )
    
    parser.add_argument(
        "--exclude",
        help="Komma-getrennte Liste von Tabellen zum Ausschliessen",
        type=str,
        default=""
    )
    
    parser.add_argument(
        "--include-meta",
        help="System-Metadaten (_createdAt, etc.) beibehalten",
        action="store_true"
    )
    
    parser.add_argument(
        "--db",
        help="Nur bestimmte Datenbank(en) exportieren (komma-getrennt)",
        type=str,
        default=""
    )
    
    parser.add_argument(
        "--summary",
        help="Nur Zusammenfassung anzeigen (kein Export)",
        action="store_true"
    )
    
    args = parser.parse_args()
    
    # Nur Zusammenfassung anzeigen
    if args.summary:
        printDatabaseSummary()
        return
    
    # Exclude-Liste parsen
    excludeTables = []
    if args.exclude:
        excludeTables = [t.strip() for t in args.exclude.split(",") if t.strip()]
    
    # Datenbank-Liste parsen
    onlyDatabases = None
    if args.db:
        onlyDatabases = [db.strip() for db in args.db.split(",") if db.strip()]
    
    # Export durchführen
    try:
        outputPath = exportDatabase(
            outputPath=args.output,
            prettyPrint=args.pretty,
            excludeTables=excludeTables,
            includeMeta=args.include_meta,
            onlyDatabases=onlyDatabases
        )
        print(f"\n Export erfolgreich: {outputPath}\n")
        
    except Exception as e:
        logger.error(f"Export fehlgeschlagen: {e}")
        sys.exit(1)


if __name__ == "__main__":
    main()