CLI Reference

python -m calibration [OPTIONS]

Options

Option	Default	Description
--scenario	baseline	Scenario: baseline, growth_plus, buffer_stock
--phase	all	Phase: sensitivity, morris, grid, stability, pairwise, rescreen, cost, cross-eval, sweep
--method	morris	Sensitivity method: morris or oat
--morris-trajectories	10	Number of Morris trajectories
--workers	10	Parallel workers
--periods	1000	Simulation periods
--sensitivity-threshold	0.02	Minimum Δ for INCLUDE in grid search
--pruning-threshold	auto	Max score gap for keeping values: auto, none, or float
--sensitivity-seeds, --seeds	3	Seeds per sensitivity evaluation
--stability-tiers	100:10,50:20,10:100	Tiers as configs:seeds,...
--resume	false	Resume from checkpoint
--grid	(none)	Load custom grid from YAML/JSON file
--fixed	(none)	Fix a parameter: KEY=VALUE (repeatable)
--rank-by	combined	Ranking strategy: combined, stability, mean, stability-first, score-first, balanced
--k-factor	1.0	k in mean−k×std formula (combined mode)
--output-dir	(timestamped)	Custom output directory
--fix-from	(none)	Load fixed params from stability JSON (rescreen)
--params	(none)	Parameter group or comma-separated names (rescreen)
--base	(none)	Base config from stability JSON or YAML (cost, sweep)
--swaps	(none)	Swap values: param=v1,v2 (repeatable, cost)
--combo-grid	false	Run combo grid of cheap swaps (cost)
--scenarios	(none)	Comma-separated scenario list (cross-eval)
--configs	(none)	Configs from result JSON (cross-eval)
--stages	(none)	Stage definitions: LABEL:param=v1,v2 (sweep)
--cross-scenario	(none)	Cross-evaluate against this scenario (sweep)

Output Structure

Results are saved to timestamped directories:

output/2026-02-24_143052_baseline/
├── sensitivity.json          # Sensitivity analysis results
├── sensitivity_report.md     # Sensitivity markdown report
├── morris.json               # Morris method results
├── screening.json            # Grid screening results
├── screening_report.md       # Screening markdown report
├── stability.json            # Stability-tested results
├── stability_report.md       # Stability markdown report
├── full_report.md            # Combined report
├── best_config.yml           # Best config as YAML
└── pairwise.json             # Pairwise interactions

Common Workflows

# Full pipeline with defaults
python -m calibration --scenario baseline --workers 10

# Morris screening only
python -m calibration --phase sensitivity --workers 10

# OAT instead of Morris
python -m calibration --phase sensitivity --method oat --workers 10

# Custom grid input
python -m calibration --phase grid --grid custom_grid.yaml

# Custom stability tiers
python -m calibration --phase stability --stability-tiers "50:10,20:30,5:100"

# Growth+ scenario
python -m calibration --scenario growth_plus --workers 10

# Second-pass Morris screening
python -m calibration --phase rescreen --scenario baseline \
  --fix-from output/baseline_stability.json --params behavioral

# Targeted cost analysis
python -m calibration --phase cost --scenario baseline \
  --base output/baseline_stability.json --swaps "beta=2.5" --seeds 20

# Cross-scenario evaluation
python -m calibration --phase cross-eval \
  --scenarios baseline,growth_plus --configs output/stability.json

# Structured parameter sweep
python -m calibration --phase sweep --scenario baseline \
  --base output/stability.json \
  --stages "A:beta=0.5,1.0" "B:max_M=2,4" --cross-scenario growth_plus