jacobs-json-schema

This package is yet another JSON Schema validator. I wrote it because I needed something small to run in Python 3.5.

Most data validation features are supported without any dependencies (see "Conformance").

Documentation

Is available at Read the Docs.

Usage

Before using this library, the schema dna data must already be parsed into a Python data structure. This can be as simple as using json.loads().

from jacobsjsonschema.draft7 import Validator

schema = { "type": "string" }
validator = Validator(schema)

data = "Hello world"
validator.validate(data)
# Will throw if there are any validation errors

Lazy error reporting is also supported. This means that as much of the data as possible is evaluated, and errors are collected instead of raising an exception.

schema = { "type": "string" }
validator = Validator(schema, lazy_error_reporting=True)

data = "Hello world"
if validator.validate(data):
    print("Validated")
else:
    for error in validator.get_errors():
        print(error)

Conformance

There are two ways of running the validator:

1. Passing a JSON-deserialized Python dictionary as the schema. There are no additional external dependencies needed. Straightforward $ref references (no usage of $id) within the same schema are supported.
2. Parse the schema using utility from the jacobs-json-doc python package, and pass the wrapped schema to the validator. Several external dependencies are required. Full $id and $ref functionality is supported.

Where "Mostly" is specified, it passes all tests excluding those from ref.json, id.json, defintions.json, refRemote.json and unknownKeyword.json.

Specification	Standalone	using jacobs-json-doc
Python Version	3.5+	3.7+
Draft-04	Mostly	Passed
Draft-06	Mostly	Passed
Draft-07	Mostly	Passed
Draft-2019-09	Untested	Untested
Draft-2020-12	Untested	Untested

License

MIT License. If you modify the source, please publish your modifications.