schemamodels 0.6.1

Dynamically create useful data classes from JSON schemas

Schema Models

Dynamically created data classes from JSON schemas

Use this library to quickly turn a JSON Schema into a Python dataclass that you can immediately consume.

Installation

Install this package using the usual suspects.

pip install schemamodels

Usage

Assuming you have a JSON schema like:

    {
        "$id": "https://schema.dev/fake-schema.schema.json",
        "$schema": "http://json-schema.org/draft-07/schema#",
        "title": "fake-schema",
        "description": "Blue Blah",
        "type": "object",
        "properties": {
            "property_a": {
              "default": 5,
              "type": "integer"
            },
            "property_b": {
              "type": "string"
            }
        }
    }

from schemamodels import SchemaModelFactory

schema_string = '..'
my_json_schema = json.loads(schema_string)

factory = SchemaModelFactory()
factory.register(my_json_schema)

Use your new dataclass

from schemamodels import exceptions
from schemamodels.dynamic import FakeSchema

your_data_instance = FakeSchema(property_a=2334)  # OK

with pytest.raises(exceptions.ValueTypeViolation):
  your_data_instance = FakeSchema(property_a="hello")

Rationale

The JSON Schema can come from anywhere

Regardless of where the JSON schema originated, it only needs to be valid for the Draft version you care about. There are a number of libraries better suited validating a JSON Schema document. A user of this library would obtain a JSON Schema document using their prefered method (filesystem, remote), then pass it to this library.

Just-enough validation

At this time, I'm not interested in validating a JSON Schema. However, there are some basic checks I'd like to have performed every time create a new instance of a object that's designed to hold my data. Also, questions about the quality of the data is out of scope.

I want to have the confidence that the data has a structure the adhears to the rules provided by a JSON Schema.

I want to be sure that the dictionary exported by these data classes would pass validation checks. The specific tool used to validate an instance of data against the original JSON Schema shouldn't matter.

I'm tired of writing Python classes by hand

While I like using Python-classes to write Python declaratively, I think letting JSON Schema drive the data models creates an opportunity to automate.

When I have a valid JSON Schema, I can create a new Python dataclass with one line of code.