strainer-2020 1.4.0

Fast functional serializers

Strainer-2020 is a different take on serialization and validation in python. It utilizes a functional style over classes.

Strainer-2020 officially supports Python 3.6 and newer, and runs great on PyPy.

This is a fork of original Alex Kessinger pystrainer library with improvements that changed operation but not the usage paradigm. I’m trying to keep his commitment to efficiency.

Features

• Functional

• Complex Python object serialization

• Data deserialization into simple Python structures

• Data validation

Changes

• serialization is done by data serializers defined for fields

• validators perform data validation only

• basic field types have simplified interface functions tthat wrap generic field() function

• datetime/time serialization and deserialization preserves timezone information or lack of it; naive datetimes/times are serialized as naive and then deserialized as naive too

Serialization Example

import datetime
from strainer import (serializer, field, child,
                      formatters, validators,
                      ValidationException, fields)

artist_serializer = serializer(
    field('name', validators=[validators.required()])
)

album_schema = serializer(
    field('title', validators=[validators.required()]),
    fields.date('release_date', required=True),
    child('artist', serializer=artist_serializer, validators=[validators.required()])
)

class Artist(object):
    def __init__(self, name):
        self.name = name

class Album(object):
    def __init__(self, title, release_date, artist):
        self.title = title
        self.release_date = release_date
        self.artist = artist

bowie = Artist(name='David Bowie')
album = Album(
    artist=bowie,
    title='Hunky Dory',
    release_date=datetime.datetime(1971, 12, 17)
)

Now we can serialize, deserialize, and validate data

>>> album_schema.serialize(album)
{'artist': {'name': 'David Bowie'},
 'release_date': '1971-12-17',
 'title': 'Hunky Dory'}
>>> album_schema.deserialize(album_schema.serialize(album))
{'artist': {'name': 'David Bowie'},
 'release_date': datetime.date(1971, 12, 17),
 'title': 'Hunky Dory'}
>>> input = album_schema.serialize(album)
>>> del input['artist']
>>> album_schema.deserialize(input)
ValidationException: {'artist': ['This field is required']}

The example has been borrowed from Marshmallow and tweaked.

Installation

To install Strainer-2020, simply:

$ pip install strainer-2020
✨🍰✨

Satisfaction, guaranteed.

Release History

1.4.0 (2020)

• infrastructure refresh

• separate operation of validators and serializers

• change to datetime serialization that preserves timezone information or lack of it

• drop Python 2 support, limit support to modern Python 3 only

1.0.1

• refining validators

• added attr_getter to child, and many

1.0.0

• Updating docs

• Making it official

0.0.9

• Fixing python 3 comptatability issue

0.0.8

• Removed an errant print statement

0.0.7

• Fixed a bug with datetime validators

0.0.6

• Fixed a bug with multiple validation, pointing to the correct index

• Fixed a bug that applied vlaidation to entire array in multiple instead of elements

• Added a dict_field, if source is dict, instead of an object

• Added ability to pass validators to child, and many instances applying validation before moving to sub-element

• Added tests around catching nested validation errors

• Added formatters, so things can be formatted on the way out

• Got rid of encoders, not the domain of this project

• Everything can be imported from one namespace

• Changed the API from to_representation/to_internal to serialize/deserialize

0.0.5 (2016-11-29)

• Fleshed out docs

• Added datetime validator

• Increased speed bu reducing loops

0.0.4 (2016-11-23)

• Add some validators

0.0.1 (2016-11-23)

• Birth