Skip to content

Latest commit

 

History

History
91 lines (58 loc) · 2.7 KB

README.md

File metadata and controls

91 lines (58 loc) · 2.7 KB
Raw

Build Status PyPi

marshmallow-autoschema

Generate marshmallow schemas from code annotations and type hinting, converting object structures to serializable native types and back effortlessly.

from marshmallow_autoschema import autoschema, validate_field, Many, One
from marshmallow.validate import Range


@validate_field('word_count', Range(0, 10000))
@autoschema
class Page:

    def __init__(self, *,
                 word_count: int=0) -> None: pass


@autoschema
class Book:

    def __init__(self, *,
                 cover: One[Page],
                 pages: Many[Page]) -> None: pass

data = {
    "cover": {"word_count": 12},
    "pages": [
        {"word_count": 0},
        {"word_count": 12},
        {"word_count": 100},
    ],
}

book = Book.load(data)

assert isinstance(book, Book)
assert isinstance(book.cover, Page)

out = book.dump()

assert data == out

For the decorator to work properly, a few properties are required:

  • All properties must be listed as arguments in the constructor.
  • Only kwonly arguments are considered, the asterisk is required in the declaration.
  • Type annotations are required.
  • You can't mix traditionals marshmallow schemas with schemas generated by autoschema

Additionally:

  • Default values can be used and apply as normal normal function default values.
  • On a List[] type, =None will default to an empty list.
  • The constructor body can be specified and is called after all members have been assigned.

Annotation keywords specific to @autoschema

One[…]

Represent one instance of the type passed singly. This annotation is required for any schema type generated by autoschema.

Many[…]

Represent a List of instance of the type passed. This annotation is required for any schema type generated by autoschema.

Advanced Usage

@autoschema and @autoschema_camelcase are offered as a convenience for the most common cases. For other cases, schema_metafactory can be used.

Arguments:

  • schema_base_class: As a minimum, the base schema class must have a post_load handler to create the model object.
  • field_namer: Function to convert the class members into the serialized name.
  • extended_field_map: Include additional field type handlers for your definitions.

Additional Examples

License

Copyright (c) 2017- Delve Labs Inc.

This library is distributed under the MIT License. See joined LICENSE file for details.