Define your JSON schema as Python dataclasses

Last update: Sep 20, 2022

Overview

nvelope

Define your JSON schema as Python dataclasses

Installation

pip install nvelope

The problem it solves

This is basically sommething like JSON-schema, but it works with static type checking, since the classes you define are just regular python dataclasses which can (and should) be type checked with mypy library.

It also lets not to just define the structure of your JSON data in a single place in your python code, but also to define custom checks and conversions from/to JSON for any type you want.

Original use case

Say you have two microservices communicating via JSON messages, both written in python.

You may define a shared package with the messages definition and use the model's .as_json() method on one end to serialize the message and .from_json() on the other to convert it into a DTO, checking and modifying the fields and their values along the way exactly how you defined it in a single place.

Combining this with static type checking (and maybe some unit tests) you can ensure that any message one microservice can send, the other can read as long as they use the same model to pack/unpack their JSONs.

Usage

Say you have a JSON representing a user in your app looking something like this

{
    "id": 530716139,
    "username": "johndoe",
    "language_code": "en"
}

You define an envelope for it

from dataclasses import dataclass
from typing import Optional

from nvelope import (Obj, int_conv, string_conv)

@dataclass      # note the @dataclass decorator, it is important
class User(Obj):
    _conversion = {
        "id": int_conv,
        "language_code": string_conv,
        "username": string_conv,
    }

    id: int
    language_code: str
    username: str

Now you have a model that knows how to read data from the JSON (not the raw string, actually, but to the types that are allowed by the standard json.dumps function e.g. dict, list, str, int, float, bool, None ) ...

user = User.from_json(
    {
        "id": 530716139,
        "username": "johndoe",
        "language_code": "en"
    }
)

... and knows how to convert itself into JSON

User(
    id=530716139,
    username="johndoe",
    language_code="en",
).as_json() 

# returns a dictionary {
#     "id": 530716139,
#     "username": "johndoe",
#     "language_code": "en"
# }

Compound envelopes

You can also define compound envelopes.

Say we want to define a message and include info about the sender. Having defined the User envelope, we can do it like this:

from nvelope import CompoundConv

@dataclass
class Message(Obj):
    _conversion = {
        "message_id": int_conv,
        "from_": CompoundConv(User),
        "text": string_conv,
    }

    from_: User
    text: str
    message_id: int

and use it the same way:

# reading an obj from json like this

Message.from_json(
    {
        "message_id": 44,
        "text": "hello there",
        "from_": {
            "id": 530716139,
            "username": "johndoe",
            "language_code": "en"
        }
    }
)

# and dumping an object to json like this
Message(
    message_id=44,
    text="whatever",
    from_=User(
        id=530716139,
        username="johndoe",
        language_code="en",
    )
).as_json()

Arrays

This is how you define arrays:

from nvelope import Arr


class Users(Arr):
    conversion = CompoundConv(User)


# Same API inherited from nvelope.Compound interface

Users.from_json(
    [
        {
            "id": 530716139,
            "username": "johndoe",
            "language_code": "en",
        },
        {
            "id": 452341341,
            "username": "ivandrago",
            "language_code": "ru",
        }
    ]
)

Users(
    [
        User(
            id=530716139,
            username="johndoe",
            language_code="en",
        ),
        User(
            id=452341341,
            username="ivandrago",
            language_code="ru",
        ),
    ]
).as_json()

Field aliases

At some point you may need to define an envelope for an API containing certain field names which cannot be used in python since they are reserved keywords.

There's a solution for this:

from nvelope import ObjWithAliases

@dataclass
class Comment(ObjWithAliases):
    _conversion = {
        "text": string_conv,
        "from_": int_conv,
    }
    
    
    _alias_to_actual = {
        "from_": "from",
    }
    
    text: str
    from_: User

In this case from key gets replaced by from_ in the python model.

Missing and optional fields

There's a difference between fields that can be set to None and fields which may be missing in the JSON at all.

This is how you specify that a some field may be missing from the JSON and that's OK:

from typing import Optional

from nvelope import MaybeMissing
from nvelope import OptionalConv

@dataclass
class Comment(ObjWithAliases):
    _alias_to_actual = {
        "from_": "from",
    }
    
    text: str
    img: Optional[str]          # this field can be set to None (null), but is must always be present in the JSON
    from_: MaybeMissing[User]   # this field can be missing from JSON body

    _conversion = {
        "text": string_conv,
        "img": OptionalConv(string_conv),   # note the wrapping with OptionalConv
        "from_": int_conv,
    }

This is how you check if the MaybeMissing field is actually missing

comment.from_.has()     # returns False if the field is missing

and this is how you get the value:

comment.value()     # raises an error if there's no value, 
                    # so it is recommended to check the output of .has()
                    #  before calling .value()

Custom conversions

You may define a custom conversions inheriting from nvelope.nvelope.Conversion abstract base class or using nvelope.nvelope.ConversionOf class.

For example, this is how datetime_iso_format_conv is defined:

from nvelope import WithTypeCheck, ConversionOf

datetime_iso_format_conv = WithTypeCheck(
    datetime.datetime,
    ConversionOf(
        to_json=lambda v: v.isoformat(),
        from_json=lambda s: datetime.datetime.fromisoformat(s),
    ),
)

Say we want to jsonify a datetime field as POSIX timestamp, instead of storing it in ISO string format.

datetime_timestamp_conv = ConversionOf(
    to_json=lambda v: v.timestamp(),
    from_json=lambda s: datetime.datetime.fromtimestamp(s),
)

We could also add WithTypeCheck wrapper in order to add explicit check that the value passed into .from_json() is indeed float.

datetime_timestamp_conv = WithTypeCheck(
    float,
    ConversionOf(
        to_json=lambda v: v.timestamp(),
        from_json=lambda s: datetime.datetime.fromtimestamp(s),
    )
)

simplejson is a simple, fast, extensible JSON encoder/decoder for Python

simplejson simplejson is a simple, fast, complete, correct and extensible JSON http://json.org encoder and decoder for Python 3.3+ with legacy suppo

1.5k Jan 5, 2023

jq for Python programmers Process JSON and HTML on the command-line with familiar syntax.

3 Jan 9, 2022

A JSON utility library for Python featuring Django-style queries and mutations.

JSON Enhanced JSON Enhanced implements fast and pythonic queries and mutations for JSON objects. Installation You can install json-enhanced with pip:

4 Aug 22, 2022

Python script for converting .json to .md files using Mako templates.

Install Just install poetry and update script dependencies Usage Put your settings in settings.py and .json data (optionally, with attachments) in dat

6 Dec 7, 2021

json|dict to python object

Pyonize convert json|dict to python object Setup pip install pyonize Examples from pyonize import pyonize

45 Nov 25, 2022

Editor for json/standard python data

1 Dec 7, 2021

Python script to extract news from RSS feeds and save it as json.

14 Dec 22, 2022

A Python tool that parses JSON documents using JsonPath

8 Dec 18, 2022

Simple Python Library to convert JSON to XML

json2xml Simple Python Library to convert JSON to XML

79 Nov 11, 2022

Comments

Recursive definition

Is there a way for recursive definitions? So for example let's say we want to consider sections in an article. A section have a title and maybe some subsections. The corresponding json could look like:

{
    "title": "This is a really nice title",
    "sections": [
        {
            "title": "Oh this title is even nicer"
            "sections: [
                {
                     "title: "Not so nice title, no subsections"
                }
            ]
        },
        {
             "title": "Section without subsection"
        }
    ]

So we could start with:

@dataclass
class Section(Obj):
    _conversion = {"title": string_conv}

    title: str

But obviously the maybe subsections are missing. Is there a way to model that? Thanks.

opened by donpatrice 4

Descriptors

From readme I see

@dataclass    
class User(Obj):
    _conversion = {
        "id": int_conv,
        "language_code": string_conv,
        "username": string_conv,
    }

    id: int
    language_code: str
    username: str

Which makes me curious on why to define an argument named _conversion, is this a design decision for some reason?

What about implementing that using the descriptor protocol?

@dataclass    
class User(Obj):
    id: int = IntField()
    language_code: str = StringField()
    username: str = StringField()

And having all *Field to be implementation of descriptors such as:

class IntField:

    def __set_name__(self, owner, name):
        self.public_name = name
        self.private_name = '_' + name

    def __get__(self, obj, objtype=None):
        value = getattr(obj, self.private_name)
        logging.info('Accessing %r giving %r', self.public_name, value)
        return int(value)  # conversion here on reading

    def __set__(self, obj, value):
        logging.info('Updating %r to %r', self.public_name, value)
        setattr(obj, self.private_name, int(value))  # conversion here on writing

opened by rochacbruno 1

Releases(v1.1.0)

v1.1.0(May 30, 2022)
EnumConversion class

Source code(tar.gz)
Source code(zip)
v1.0.0(Feb 3, 2022)
JSON schema generation in compound objects via .schema() method;

Conversion interface now requires .schema() method returning a JSON schema;

moved functionality of ObjWithAliases to Obj;

added possibility of storing undefined JSON fields in a model instance;

validated class decorator checking the correctness of a Obj and Arr subclass;

datetime_timestamp_conv to store a datetime as POSIX timestamp float.

Source code(tar.gz)
Source code(zip)
v0.4.0(Dec 10, 2021)

In 0.4.0 NvelopeError is thrown if something goes wrong. NvelopeError contains path to the exact field that caused the error.
Source code(tar.gz)
Source code(zip)
v0.3.1(Nov 23, 2021)

Better support for string type annotations
Source code(tar.gz)
Source code(zip)