Skip to content
  • diagnostics

Diagnostics API

AMMM diagnostics follow a stage-gated workflow. Each diagnostic writes artefacts into a canonical stage folder under the run root (for example, 10_pre_diagnostics/ or 50_diagnostics/).

Machine-readable outcomes such as converged and well_calibrated are consumed downstream to condition whether decomposition, optimisation, and interpretation outputs should be trusted.

Data-load Validation (Pre-stage)

Section titled “Data-load Validation (Pre-stage)”

Stage 10_pre_diagnostics/

Section titled “Stage 10_pre_diagnostics/”

Stage 20_model_fit/ and 30_model_assessment/

Section titled “Stage 20_model_fit/ and 30_model_assessment/”

Stage 50_diagnostics/

Section titled “Stage 50_diagnostics/”