sqlalchemy-pydantic-orm

This library makes it a lot easier to do nested database operation with SQLAlchemy. With this library it is for example possible to validate, convert, and upload a 100-level deep nested JSON (dict) to its corresponding tables in a given database, within 3 lines of code.

Pydantic is used for creating the dataclass and validating it. Pydantic already has a function called .from_orm() that can do a nested get operation, but it only supports ORM -> Pydantic and not Pydantic -> ORM. That's exactly where this library fills in, with 2 specific functions .orm_create() and .orm_update(), and one general function .to_orm() that combines the functionality of the first 2, calling one or the other, depending on if there is an id provided.

Requirements

• Python 3.8+
• SQLAlchemy 1.4+
• Pydantic 1.8+

Installation

$ pip install AFAS-pydantic-orm

To tinker with the code yourself, install the full dependencies with:

$ pip install AFAS-pydantic-orm[dev]

Useful references

• https://pydantic-docs.helpmanual.io/usage/models/
• https://fastapi.tiangolo.com/tutorial/sql-databases/

Examples

Below 1 example is provided (more coming).

For a bigger and more detailed example you can look at the /examples/ folder.

Example 1 - Using manual created schemas

Create your own Pydantic schemas and link them to the SQLAlchemy ORM-models.

Create your SQLAlchemy ORM-models (one-to-one or one-to-many)

class Parent(Base):
    id = Column(Integer, primary_key=True, index=True, nullable=False)
    name = Column(String, nullable=False)
    car = relationship("Car", cascade="all, delete", uselist=False, back_populates="owner")
    children = relationship("Child", cascade="all, delete")
    
class Car(Base):
    id = Column(Integer, primary_key=True, index=True, nullable=False)
    color = Column(String, nullable=False)
    owner_id = Column(Integer, ForeignKey("parents.id"), nullable=False)
    owner = relationship("Parent", back_populates="car")

class Child(Base):
    id = Column(Integer, primary_key=True, index=True, nullable=False)
    name = Column(String, nullable=False)
    parent_id = Column(Integer, ForeignKey("parents.id"), nullable=False)

Create your Pydantic base and CRUD schemas using these ORM models, and the imported ORMBaseSchema

Base schemas

from sqlalchemy_pydantic_orm import ORMBaseSchema
from .models import Parent, Car, Child

class ParentBase(ORMBaseSchema):
    name: str
    _orm_model = PrivateAttr(Parent)

class CarBase(ORMBaseSchema):
    color: str
    _orm_model = PrivateAttr(models.Car)

class ChildBase(ORMBaseSchema):
    name: str
    _orm_model = PrivateAttr(models.Child)

GET schemas

class Parent(ParentBase):
    id: int
    children: List[Child]
    car: Car

class Car(CarBase):
    id: int

class Child(ChildBase):
    id: int

CREATE/UPDATE schemas

class ParentCreate(ParentBase):
    id: Optional[int]
    children: List[ChildCreate]
    car: CarCreate

class CarCreate(CarBase):
    id: Optional[int]

class ChildCreate(ChildBase):
    id: Optional[int]

Use your schemas to do nested CRUD operations.

with ConnectionDatabase() as db:
    create_schema = schemas.ParentCreate.parse_obj(create_dict)
    parent_db = create_schema.orm_create()
    db.add(parent_db)
    db.commit()

    db_create_schema = schemas.Parent.from_orm(parent_db)
    print(db_create_schema.dict())

    update_schema = schemas.ParentUpdate.parse_obj(update_dict)
    update_schema.to_orm(db)
    db.commit()

    db_update_schema = schemas.Parent.orm_update(parent_db)
    print(db_update_schema.dict())

Note: with .orm_create() you have to call db.add() before calling db.commit(). With orm_update you give the db session as parameter, and you only have to call db.commit().

Example 2 - Using generated schemas

TODO: Integrate with https://github.com/tiangolo/pydantic-sqlalchemy