empyrean.SessionDiff

class SessionDiff(reduced_chi2_delta, iterations_delta, n_observations_delta, update_norm_current, update_norm_prior)

Bases: object

Pairwise diagnostic between two fits in the same session.

Methods

__init__(reduced_chi2_delta, ...)

Attributes

reduced_chi2_delta	Δ reduced χ² (positive ⇒ current fit is worse).
iterations_delta	Δ iteration count between current and prior fits, over DetermineResult.iterations — a TOTAL across each fit's solves.
n_observations_delta	Δ number of observations used (negative when masked between refines).
update_norm_current	DetermineResult.update_norm on the current fit — the μ-DAMPED last accepted step, not a convergence metric and not comparable to ODConfig.convergence_tol.
update_norm_prior	The same quantity on the prior fit, with the same caveat.

Parameters:

• reduced_chi2_delta (float)

• iterations_delta (int)

• n_observations_delta (int)

• update_norm_current (float)

• update_norm_prior (float)

reduced_chi2_delta: float

Δ reduced χ² (positive ⇒ current fit is worse).

iterations_delta: int

Δ iteration count between current and prior fits, over DetermineResult.iterations — a TOTAL across each fit’s solves. Two fits that reached the same orbit by different ladders differ here without either being worse.

n_observations_delta: int

Δ number of observations used (negative when masked between refines).

update_norm_current: float

DetermineResult.update_norm on the current fit — the μ-DAMPED last accepted step, not a convergence metric and not comparable to ODConfig.convergence_tol. It is here as a damping-trajectory diagnostic between two fits of the same arc. For whether either fit converged, and on what, read that fit’s DetermineResult.termination / DetermineResult.gn_step_qnorm.

update_norm_prior: float

The same quantity on the prior fit, with the same caveat.