db-to-sqlalchemy-generator

Provide the DB specification document path or paste the DDL, and the Skill parses each table's fields, constraints, indexes, and foreign keys to produce SQLAlchemy 2.0 models using Mapped type annotations. Output files follow the models/<table_name>.py convention.

The Skill generates four standard Pydantic 2.x DTOs per table: Base, Create, Update, and Read. Base holds required business fields, Update makes all fields optional for partial updates, and Read adds id and timestamps with from_attributes enabled.

Yes, SQL Server-specific column types are mapped using sqlalchemy.dialects.mssql types. Standard types like NVARCHAR, DECIMAL, and DATETIME map to String, Numeric, and DateTime(timezone=True) respectively.

Foreign keys become ForeignKey columns paired with relationship declarations using back_populates on both sides. Many-to-many associations generate an intermediate table model with bidirectional relationship(secondary=...) mappings.

Tables with is_deleted or deleted_at columns automatically use SoftDeleteMixin, and tables with created_at and updated_at use TimestampMixin. This avoids duplicating these system columns in every generated model.