Documentation | Source Code | Issue Tracker

anyschema allows you to convert from type specifications (such as Pydantic models, SQLAlchemy tables, TypedDict, dataclasses, or plain Python dicts) to any dataframe schema (by "any" we intend those supported by Narwhals).

Let's see how it works in practice with an example:

from anyschema import AnySchema
from pydantic import BaseModel
from pydantic import PositiveInt


class Student(BaseModel):
    name: str
    age: PositiveInt
    classes: list[str]


schema = AnySchema(spec=Student)

# Convert to pyarrow schema
pa_schema = schema.to_arrow()

type(pa_schema)
# pyarrow.lib.Schema

pa_schema
# name: string
# age: uint64
# classes: list<item: string>
#   child 0, item: string

pl_schema = schema.to_polars()

type(pl_schema)
# polars.schema.Schema

pl_schema
# Schema([('name', String), ('age', UInt64), ('classes', List(String))])

To read more about anyschema functionalities and features consider checking out the documentation website.

anyschema is available on pypi, and it can be installed directly via any package manager. For instance:

uv pip install "anyschema[pydantic]"
# or with SQLAlchemy support
uv pip install "anyschema[sqlalchemy]"

To allow interoperability with Pydantic models or SQLAlchemy tables.

anyschema is designed for scenarios where type specifications (e.g., Pydantic models, SQLAlchemy tables) serve as a single source of truth for both validation and dataframe schema generation.

The typical use cases are: Data pipelines, database-to-dataframe workflows, API to database workflows, schema generation, type-safe data processing.

The project was inspired by a Talk Python podcast episode featuring the creator of LanceDB, who mentioned the need to convert from Pydantic models to PyArrow schemas.

This challenge led to a realization: such conversion could be generalized to many dataframe libraries by using Narwhals as an intermediate representation. anyschema makes this conversion seamless and extensible.