sqlalchemy-tenant-wiper 0.1.0

A flexible SQLAlchemy-based library for tenant data deletion in multi-tenant applications

SQLAlchemy Tenant Wiper

A flexible SQLAlchemy-based library for tenant data deletion in multi-tenant applications. Supports custom SQLAlchemy Base configurations, dynamic tenant filtering using lambda expressions, and handles complex relationship paths with robust validations.

Features

• Flexible Base Configuration: Works with any SQLAlchemy declarative Base
• Dynamic Tenant Filtering: Lambda expressions adapt to each table's columns
• Relationship Path Support: Handle indirect tenant relationships through join paths
• Composite Primary Key Support: Properly handles tables with composite primary keys
• Configuration-Time Validation: Comprehensive validation at config creation
• Two-Phase Deletion: Safe deletion order respecting foreign key constraints
• Dry Run Mode: Preview what would be deleted before execution
• Batched Deletions: Efficient deletion in configurable batch sizes

Installation

Using pip

pip install sqlalchemy-tenant-wiper

Using Poetry

poetry add sqlalchemy-tenant-wiper

Quick Start

from sqlalchemy_tenant_wiper import TenantWiperConfig, TenantDeleter
from uuid import UUID

# Configure your tenant deletion settings
config = TenantWiperConfig(
    base=YourSQLAlchemyBase,
    tenant_filters=[
        lambda table: table.c.tenant_id == your_tenant_uuid,
        lambda table: table.c.org_id.in_(your_org_uuids)
    ],
    relationships=[
        'products__id=product_id__order_items__order_id=id__orders'
    ],
    excluded_tables=['audit_logs', 'system_logs'],
    batch_size=500
)

# Create deleter and execute
deleter = TenantDeleter(config)
deleter.delete(session, dry_run=True)  # Preview first
deleter.delete(session, commit=True)   # Execute deletion

Usage Examples

Basic Tenant Filtering

from sqlalchemy_tenant_wiper import TenantWiperConfig, TenantDeleter
from uuid import UUID

# Simple tenant filtering by tenant_id
tenant_uuid = UUID('12345678-1234-5678-9012-123456789abc')

config = TenantWiperConfig(
    base=Base,
    tenant_filters=[
        lambda table: table.c.tenant_id == str(tenant_uuid)
    ]
)

deleter = TenantDeleter(config)
deleter.delete(session, dry_run=True, commit=False)

Multiple Tenant Filters

# Multiple ways to identify tenant data
config = TenantWiperConfig(
    base=Base,
    tenant_filters=[
        lambda table: table.c.tenant_id == str(tenant_uuid),
        lambda table: table.c.organization_id.in_(org_uuids),
        lambda table: table.c.customer_code == 'ACME_CORP'
    ]
)

Relationship Paths for Indirect Tables

# Handle tables without direct tenant columns
config = TenantWiperConfig(
    base=Base,
    tenant_filters=[
        lambda table: table.c.tenant_id == str(tenant_uuid)
    ],
    relationships=[
        # Products -> OrderItems -> Orders (with tenant_id)
        'products__id=product_id__order_items__order_id=id__orders',
        
        # Audit logs -> Users (with tenant_id)  
        'audit_logs__user_id=id__users',
        
        # Complex multi-hop relationship
        'categories__id=category_id__products__id=product_id__order_items__order_id=id__orders'
    ]
)

Configuration Validation

# Validate configuration at startup/in tests
config = TenantWiperConfig(
    base=Base,
    tenant_filters=[...],
    relationships=[...],
    validate_on_init=True  # Default: validates on creation
)

# Or validate explicitly
config = TenantWiperConfig(base=Base, validate_on_init=False)
config.validate()  # Call when ready

Excluding Tables

# Skip certain tables from deletion
config = TenantWiperConfig(
    base=Base,
    tenant_filters=[...],
    excluded_tables=[
        'audit_logs',
        'system_configs', 
        'migration_history'
    ]
)

Configuration Reference

TenantWiperConfig

Parameter	Type	Description
base	SQLAlchemy Base	Your declarative base class
tenant_filters	List[Callable[[Table], Any]]	Lambda functions for tenant filtering
relationships	List[str]	Relationship path strings for indirect tables
excluded_tables	List[str]	Table names to exclude from deletion
validate_on_init	bool	Whether to validate config on creation (default: True)

TenantDeleter.delete()

Parameter	Type	Description
session	SQLAlchemy Session	Database session for operations
dry_run	bool	If True, only report what would be deleted
commit	bool	If True, commit the transaction

Relationship Path Syntax

Relationship paths use double underscore (__) separators:

table1__from_key=to_key__table2__from_key2=to_key2__table3

Example:

products__id=product_id__order_items__order_id=id__orders

This creates joins:

1. products.id = order_items.product_id
2. order_items.order_id = orders.id

The final table (orders) must have tenant filter columns.

Requirements

• Python 3.8+
• SQLAlchemy 1.4+

Development

Setup

git clone https://github.com/yourusername/sqlalchemy-tenant-wiper.git
cd sqlalchemy-tenant-wiper
poetry install

Testing

poetry run pytest

License

This project is licensed under the MIT License - see the LICENSE file for details.

TODO

[ ] Make enforcing listing of all tables as optional