tf_sharepoint_integration/clear_data.py

#!/usr/bin/env python3
"""
Clear all data from the SharePoint connector application.
This script will:
1. Drop all vector store tables (PostgreSQL)
2. Delete all DynamoDB tables (connections, tokens, etc.)
3. Clear stored credentials
"""

import os
import sys
import boto3
from storage.credentials_storage import get_credentials_storage

def clear_postgresql_vector_store():
    """Drop all vector store tables from PostgreSQL."""
    try:
        import psycopg2

        # Get connection params from environment
        db_host = os.getenv("POSTGRES_HOST", "localhost")
        db_port = os.getenv("POSTGRES_PORT", "5432")
        db_name = os.getenv("POSTGRES_DB", "sharepoint_vectors")
        db_user = os.getenv("POSTGRES_USER", "postgres")
        db_password = os.getenv("POSTGRES_PASSWORD", "postgres")

        print(f"📊 Connecting to PostgreSQL at {db_host}:{db_port}/{db_name}...")

        conn = psycopg2.connect(
            host=db_host,
            port=db_port,
            database=db_name,
            user=db_user,
            password=db_password
        )

        cursor = conn.cursor()

        # Get table prefix
        table_prefix = os.getenv("TABLE_PREFIX", "dev_")

        # Drop tables in correct order (child tables first)
        tables = [
            f"{table_prefix}document_tags",
            f"{table_prefix}document_chunks",
            f"{table_prefix}documents"
        ]

        for table in tables:
            try:
                cursor.execute(f"DROP TABLE IF EXISTS {table} CASCADE")
                print(f"  ✅ Dropped table: {table}")
            except Exception as e:
                print(f"  ⚠️  Could not drop {table}: {e}")

        conn.commit()
        cursor.close()
        conn.close()

        print("✅ PostgreSQL vector store cleared")
        return True

    except ImportError:
        print("⚠️  psycopg2 not installed, skipping PostgreSQL cleanup")
        return False
    except Exception as e:
        print(f"❌ Error clearing PostgreSQL: {e}")
        return False


def clear_dynamodb_tables():
    """Delete all DynamoDB tables."""
    try:
        # Determine if using local or AWS DynamoDB
        endpoint_url = os.getenv("DYNAMODB_ENDPOINT", "http://localhost:8000")
        region = os.getenv("AWS_REGION", "ap-southeast-2")
        table_prefix = os.getenv("TABLE_PREFIX", "dev_")

        if "localhost" in endpoint_url or "127.0.0.1" in endpoint_url:
            print(f"📊 Connecting to LOCAL DynamoDB at {endpoint_url}...")
        else:
            print(f"📊 Connecting to AWS DynamoDB in {region}...")

        dynamodb = boto3.client(
            'dynamodb',
            region_name=region,
            endpoint_url=endpoint_url if "localhost" in endpoint_url or "127.0.0.1" in endpoint_url else None
        )

        # List all tables
        response = dynamodb.list_tables()
        tables = response.get('TableNames', [])

        # Filter tables with our prefix
        our_tables = [t for t in tables if t.startswith(table_prefix)]

        if not our_tables:
            print(f"  ℹ️  No tables found with prefix '{table_prefix}'")
            return True

        print(f"  Found {len(our_tables)} tables to delete:")
        for table in our_tables:
            print(f"    - {table}")

        # Delete each table
        for table in our_tables:
            try:
                dynamodb.delete_table(TableName=table)
                print(f"  ✅ Deleted table: {table}")
            except Exception as e:
                print(f"  ⚠️  Could not delete {table}: {e}")

        print("✅ DynamoDB tables cleared")
        return True

    except Exception as e:
        print(f"❌ Error clearing DynamoDB: {e}")
        return False


def clear_credentials():
    """Clear all stored credentials."""
    try:
        print("📊 Clearing stored credentials...")

        storage = get_credentials_storage()

        # Get the storage file path
        if hasattr(storage, 'file_path'):
            file_path = storage.file_path
            if os.path.exists(file_path):
                os.remove(file_path)
                print(f"  ✅ Deleted credentials file: {file_path}")
            else:
                print(f"  ℹ️  No credentials file found at: {file_path}")
        else:
            print("  ℹ️  Credentials storage does not use file system")

        print("✅ Credentials cleared")
        return True

    except Exception as e:
        print(f"❌ Error clearing credentials: {e}")
        return False


def main():
    """Main function to clear all data."""
    print("=" * 60)
    print("🧹 CLEARING ALL SHAREPOINT CONNECTOR DATA")
    print("=" * 60)
    print()

    # Check for --force flag
    force = '--force' in sys.argv or '-f' in sys.argv

    if not force:
        try:
            response = input("⚠️  This will DELETE ALL data. Are you sure? (yes/no): ")
            if response.lower() not in ['yes', 'y']:
                print("❌ Aborted")
                return
        except EOFError:
            print("\n❌ No input received. Use --force to skip confirmation.")
            return

    print()
    print("Starting cleanup...")
    print()

    # Clear PostgreSQL vector store
    print("1️⃣  Clearing PostgreSQL Vector Store...")
    clear_postgresql_vector_store()
    print()

    # Clear DynamoDB tables
    print("2️⃣  Clearing DynamoDB Tables...")
    clear_dynamodb_tables()
    print()

    # Clear credentials
    print("3️⃣  Clearing Stored Credentials...")
    clear_credentials()
    print()

    print("=" * 60)
    print("✅ ALL DATA CLEARED SUCCESSFULLY")
    print("=" * 60)
    print()
    print("You can now start fresh by running the application:")
    print("  python app_dev.py")
    print()


if __name__ == "__main__":
    main()