marshmallow3-annotations

1.1.0 verified Mon Apr 27 auth: no python

A library that allows using Python type annotations to define marshmallow 3 schemas, automatically generating Schema classes from annotated dataclasses or models. Version 1.1.0 supports dicts; release cadence is sporadic.

pip install marshmallow3-annotations

Common errors

error ModuleNotFoundError: No module named 'marshmallow3_annotations' ↓

cause Incorrect import path (maybe using hyphens or different casing).

fix

Install with: pip install marshmallow3-annotations. Then import: from marshmallow3_annotations import annotation_schema

error TypeError: annotation_schema() missing 1 required positional argument: 'cls' ↓

cause Passing an instance instead of the class, or no argument.

fix

Call annotation_schema with the class itself: annotation_schema(MyClass)

error AttributeError: 'NoneType' object has no attribute 'fields' ↓

cause annotation_schema returned None or the generated schema is empty because the dataclass has no annotated fields.

fix

Ensure the class is a dataclass with type-annotated fields (e.g., name: str).

Warnings

gotcha annotation_schema expects a class, not an instance. Pass the class itself. ↓

fix UserSchema = annotation_schema(User) # not User()

gotcha Only simple types like str, int, float, bool, list, dict are supported. Custom types or nested dataclasses may not work. ↓

fix Use marshmallow fields directly for complex types, or define a nested schema manually.

deprecated The library may not be actively maintained; consider using marshmallow-dataclass or marshmallow-dataclass3 instead. ↓

fix Migrate to marshmallow-dataclass (more features) or marshmallow-dataclass3 (Python 3.7+).

Imports

annotation_schema

wrong

from marshmallow3_annotations import annotation_schema as old_name

correct

from marshmallow3_annotations import annotation_schema

Correct import uses underscore, no alias needed

AnnotationSchema

from marshmallow3_annotations import AnnotationSchema

Quickstart

Decorate a dataclass with type annotations, then call annotation_schema to generate a marshmallow Schema.

from marshmallow import Schema, fields
from marshmallow3_annotations import annotation_schema
from dataclasses import dataclass

@dataclass
class User:
    name: str
    email: str

UserSchema = annotation_schema(User)
# Use UserSchema like any marshmallow Schema
schema = UserSchema()
result = schema.dump(User(name='John', email='john@example.com'))
print(result)  # {'name': 'John', 'email': 'john@example.com'}