empyrean.AdvancedIntegratorConfig

class AdvancedIntegratorConfig(integrator=IntegratorChoice.GR15, epsilon=1e-09, dt_initial=None, dt_min=None, encounter_timescale_divisor=1000.0, max_steps=10000000, max_dense_steps=100000, cache_integrator_steps=False, origin_switching=<factory>)[source]

Bases: object

Integrator-tuning knobs.

Defaults are calibrated for production. Most callers don’t touch this — PropagationConfig.advanced exists to make the surface complete and to enable bespoke runs (custom step bounds for tight encounters, dense output for visualization, etc.).

Methods

__init__([integrator, epsilon, dt_initial, ...])

Attributes

cache_integrator_steps

Enable dense-state interpolation between integration steps — used for light-time iteration, off-step state queries, and event refinement around close approaches.

dt_initial

Initial step size in days.

dt_min

Minimum allowed step size in days.

encounter_timescale_divisor

Divisor K for encounter dynamical-timescale step floor.

epsilon

Adaptive integrator truncation-error tolerance (relative b₆ for GR15, rtol for DOP853 paired with a fixed atol = 1e-14).

integrator

Integrator backend.

max_dense_steps

max_steps

origin_switching

Origin-switching trajectory splitting.
Parameters:

  • integrator (IntegratorChoice)

  • epsilon (float)

  • dt_initial (float | None)

  • dt_min (float | None)

  • encounter_timescale_divisor (float)

  • max_steps (int)

  • max_dense_steps (int)

  • cache_integrator_steps (bool)

  • origin_switching (OriginSwitchingConfig)

integrator: IntegratorChoice = 'gr15'

Integrator backend. Default IntegratorChoice.GR15.

epsilon: float = 1e-09

Adaptive integrator truncation-error tolerance (relative b₆ for GR15, rtol for DOP853 paired with a fixed atol = 1e-14).

dt_initial: float | None = None

Initial step size in days. None = auto from orbital timescale.

dt_min: float | None = None

Minimum allowed step size in days. None = auto.

encounter_timescale_divisor: float = 1000.0

Divisor K for encounter dynamical-timescale step floor.

max_steps: int = 10000000
max_dense_steps: int = 100000
cache_integrator_steps: bool = False

Enable dense-state interpolation between integration steps — used for light-time iteration, off-step state queries, and event refinement around close approaches.

origin_switching: OriginSwitchingConfig

Origin-switching trajectory splitting. Default enabled.