JSON Schema

The JSON Schema adaptor reads a dict with "type": "object" and a "properties" mapping, and its members are dicts. Membership is the jsonschema package's validator for the schema's own draft, and the explanation path is the validator's absolute_path, so a nested field reads .ship.city. Deciding whether a dict is a schema needs no import; validating one does, so a schema with jsonschema not installed is refused with the extra to install. It needs jsonschema 4.22 or later, pip install "mathema-language[jsonschema]", and is tested at 4.22 and at the latest release.

What it reads

JSON Schema keyword	Neutral model	Notes
type (integer, number, string, boolean, array, object)	the matching base type	a list of types with "null" is nullable
required	the named fields are required
additionalProperties: false	the exact column policy
minimum, maximum, exclusiveMinimum, exclusiveMaximum	the matching bounds
minLength, maxLength, minItems, maxItems	min_len, max_len
pattern	regex
enum, const, multipleOf, default	the matching constraint	an enum with no type is a categorical
anyOf, oneOf with a "null" branch	nullable	other unions read as any
a nested object	a struct
$ref	refused	inline the definition
format	not read	the validator decides it when a format checker is configured

A worked claim

The payment provider's charge events and the shop's router, from examples/shop/webhooks.py:

CHARGE_EVENT = {
    "type": "object",
    "properties": {
        "type": {"enum": ["charge.succeeded", "charge.failed", "charge.refunded"]},
        "amount": {"type": "integer", "minimum": 0},
    },
    "required": ["type", "amount"],
    "additionalProperties": False,
}

QUEUES = {"charge.succeeded": "billing", "charge.failed": "alerts"}


def queue_for(event: dict) -> str:
    """The queue a charge event is routed to."""
    return QUEUES[event["type"]]


def processing_fee(event: dict) -> float:
    """The provider's fee on a charge, in cents: 2.9% plus 30."""
    return event["amount"] * 0.029 + 30

Function	Claim	Verdict	Why
processing_fee	for event in L[shop.webhooks.CHARGE_EVENT], f(event) >= 30	proven	The lift reads the amount's minimum off the schema.
queue_for	for event in L[shop.webhooks.CHARGE_EVENT], f(event) in {"billing", "alerts"}	falsified	A charge.refunded event has no queue, and the router raises KeyError.

Enforcing the schema

A function that guards a boundary should refuse a record its schema rejects. excluded_outside_domain feeds it records just outside the language, and the witness names the value and why it is outside, in the library's own words. This function trusts its input, so a record the schema rejects goes straight through:

f = queue_for
for event in L[shop.webhooks.CHARGE_EVENT], excluded_outside_domain(event)
    falsified   event = {'type': 'charge.succeeded', 'amount': None} (outside L[shop.webhooks.CHARGE_EVENT] at .amount: None is not of type 'integer')

The same claim over the function that loads the record, from a request or a database, is the one that should hold.