Skip to content

CLI Reference

This page summarizes the command-line interfaces shipped with the project and the artifact files they generate.

mot-uniform

mot-uniform runs a two-uniform experiment with a configurable payoff and writes the standard artifact set.

Usage

mot-uniform [options]

Main Options

Option Meaning Default
--n number of midpoint atoms in each marginal 50
--x-interval A B support interval for S1 1 3
--y-interval C D support interval for S2 0 4
--payoff built-in payoff name abs_spread
--strike strike for spread option payoffs 0.0
--eps regularization strengths 1.0 0.3 0.1 0.03 0.01
--output-dir directory for generated artifacts artifacts

Supported Payoffs

  • abs_spread
  • squared_distance
  • call_on_spread
  • put_on_spread
  • straddle_on_spread

Example

mot-uniform --n 60 --x-interval 1 3 --y-interval 0 4 --payoff call_on_spread --strike 0.25 --eps 0.3 0.1 --output-dir artifacts_call

Standard Output Files

  • exact_uniform_summary.png
  • regularization_path.png
  • stability_diagnostics.png
  • structural_diagnostics.png
  • summary.json
  • experiment_report.md

mot-gallery runs the curated experiment set used by the documentation and writes a gallery directory.

Usage

mot-gallery --output-dir gallery_artifacts

Output Structure

The gallery output contains:

  • one subdirectory per example
  • gallery_overview.png
  • gallery_summary.json
  • gallery_summary.md
  • gallery_casebook.md

Each example subdirectory contains the same standard files produced by mot-uniform.

For causal gallery entries, subdirectories contain causal artifact names such as causal_summary.json and causal_experiment_report.md. Continuous convergence entries contain continuous_limit.png and continuous_summary.json.

mot-causal

mot-causal runs a causal multi-period experiment on a chain of uniform marginals.

Usage

mot-causal [options]

Main Options

Option Meaning Default
--n atoms per time step 30
--intervals A,B ... one support interval per time step, formatted as comma-separated bounds 1,3 0.5,3.5 0,4
--payoff built-in adjacent-step payoff name abs_spread
--strike strike for spread option payoffs 0.0
--eps regularization strengths 0.3 0.1 0.03
--output-dir directory for generated causal artifacts causal_artifacts

Example

mot-causal --n 30 --intervals 1,3 0.5,3.5 0,4 --payoff abs_spread --eps 0.3 0.1 --output-dir causal_out

Output Files

  • causal_transport_chain.png
  • causal_bound_convergence.png
  • marginal_evolution.png
  • causal_vs_unconstrained.png
  • causal_summary.json
  • causal_experiment_report.md

Script Entry Points

The repository also includes script wrappers:

  • python scripts/run_uniform_abs_spread.py
  • python scripts/generate_example_gallery.py
  • python scripts/benchmark_causal.py

These call the same library entry points as the installed CLI commands.

Typical Workflow

  1. run mot-uniform for a single experiment
  2. inspect summary.json and the diagnostics plots
  3. run mot-causal for a multi-period chain when adapted constraints matter
  4. run mot-gallery for cross-example comparison
  5. review gallery_summary.md and gallery_casebook.md

This is usually sufficient for basic exploration and documentation generation.