fraudtwin.extensions

Stable extension protocols and local entry-point discovery.

Status: Stable

Classes

fraudtwin.extensions.BehaviorModel	Port for a deterministic customer-behavior model.
fraudtwin.extensions.DataFaultInjector	Port for deterministic observable-data mutations.
fraudtwin.extensions.ExtensionMetadata	Identity recorded in manifests for one installed extension.
fraudtwin.extensions.ExtensionRegistry	Explicit, deterministic registry of extension implementations.
fraudtwin.extensions.FraudScenario	Port for a deterministic fraud-scenario implementation.
fraudtwin.extensions.OutputSink	Port for a chunk-aware output adapter.
fraudtwin.extensions.PaymentRail	Port for a payment rail that preserves FraudTwin lifecycle semantics.

Functions

fraudtwin.extensions.discover_extensions

Load installed entry points in stable name order.

Constants and protocols

Name	Reference
Extension	fraudtwin.extensions.Extension

Detailed API

Stable extension contracts for FraudTwin.

Extensions are deliberately small ports. They describe the boundary between the deterministic simulator and user-owned behavior; they do not provide a second configuration system or a plugin runtime. Applications may register implementations explicitly, while package distributions can expose the same implementations through the fraudtwin.extensions entry-point group.

class fraudtwin.extensions.BehaviorModel(*args, **kwargs)

Bases: Protocol

Port for a deterministic customer-behavior model.

profile(entity, *, seed)

Return one behavior profile without mutating the source entity.

Return type:

Mapping[str, Any]

Parameters:

• entity (Mapping[str, Any])

• seed (int)

class fraudtwin.extensions.DataFaultInjector(*args, **kwargs)

Bases: Protocol

Port for deterministic observable-data mutations.

apply(record, *, seed)

Return the mutated record and preserve an audit-friendly identity.

Return type:

Mapping[str, Any]

Parameters:

• record (Mapping[str, Any])

• seed (int)

class fraudtwin.extensions.ExtensionMetadata(extension_id, extension_version, distribution='local', distribution_version='local')

Bases: object

Identity recorded in manifests for one installed extension.

Parameters:

• extension_id (str)

• extension_version (str)

• distribution (str)

• distribution_version (str)

class fraudtwin.extensions.ExtensionRegistry(extensions=())

Bases: object

Explicit, deterministic registry of extension implementations.

Parameters:

extensions (Iterable[FraudScenario | PaymentRail | BehaviorModel | DataFaultInjector | OutputSink])

class fraudtwin.extensions.FraudScenario(*args, **kwargs)

Bases: Protocol

Port for a deterministic fraud-scenario implementation.

generate(context, *, seed)

Yield oracle records or mutations for a bounded generation context.

Return type:

Iterable[Mapping[str, Any]]

Parameters:

• context (Mapping[str, Any])

• seed (int)

class fraudtwin.extensions.OutputSink(*args, **kwargs)

Bases: Protocol

Port for a chunk-aware output adapter.

write(records, *, run_id)

Persist a bounded batch and return checkpoint metadata.

Return type:

Mapping[str, Any]

Parameters:

• records (Iterable[Mapping[str, Any]])

• run_id (str)

class fraudtwin.extensions.PaymentRail(*args, **kwargs)

Bases: Protocol

Port for a payment rail that preserves FraudTwin lifecycle semantics.

lifecycle(payment, *, seed)

Yield lifecycle events for one payment.

Return type:

Iterable[Mapping[str, Any]]

Parameters:

• payment (Mapping[str, Any])

• seed (int)

fraudtwin.extensions.discover_extensions(*, group='fraudtwin.extensions')

Load installed entry points in stable name order.

Entry points are loaded from the local environment only. A package must therefore be installed explicitly; configuration never executes an arbitrary import path or downloads code.

Return type:

ExtensionRegistry

Parameters:

group (str)