symop.polynomial.kernels.measurements.registry

Registration of polynomial measurement kernels.

This module registers polynomial-representation measurement kernels with a MeasurementKernelRegistry.

The registered kernels implement number measurement for the DeviceKind.NUMBER_DETECTOR device, covering the following measurement intents:

  • observe:

    Compute full probability distributions without modifying the state.

  • detect:

    Sample an outcome and return a post-measurement state.

  • postselect:

    Condition on a specified outcome and return the corresponding branch.

Both ket and density representations are supported where applicable. For operations that require state collapse (detect and postselect), only density-state kernels are registered.

Notes

This function wires semantic measurement actions to concrete backend implementations for the polynomial representation (rep=POLY).

The registry dispatch key consists of:

  • device kind (e.g. NUMBER_DETECTOR)

  • measurement intent (observe, detect, postselect)

  • representation kind (POLY)

  • input state kind (KET or DENSITY)

This allows the runtime to resolve the correct kernel dynamically based on the input state and requested operation.

Functions

register_polynomial_measurement_kernels(*, ...)

Register polynomial number-measurement kernels.
register_polynomial_measurement_kernels(*, measurement_registry: MeasurementKernelRegistry) → None

Register polynomial number-measurement kernels.

Parameters:

measurement_registry (MeasurementKernelRegistry) – Registry into which the polynomial measurement kernels are registered.

Return type:

None

Notes

The following registrations are performed for DeviceKind.NUMBER_DETECTOR:

  • observe:

    • ket input → observe_number_detector_poly_ket()

    • density input → observe_number_detector_poly_density()

  • detect:

    • density input → detect_number_detector_poly_density()

  • postselect:

    • density input → postselect_number_detector_poly_density()

Ket-based kernels are provided only for observation, since detection and postselection require state collapse, which is canonically handled at the density-state level.

This function should be called during initialization of the polynomial backend to ensure that measurement actions are correctly dispatched.