H-5641: Probability Distributions in Petrinaut

[kube]

Kapture.2026-02-21.at.03.40.43.mp4

This example is accessible in the menu (Probabilistic Satellites Launcher)

🌟 What is the purpose of this PR?

Adds probability distribution support to Petrinaut transition kernels. Users can now return Distribution.Gaussian(mean, deviation) or Distribution.Uniform(min, max) values from transition kernel code instead of plain numbers, enabling stochastic token generation.

It allows to define Transition Kernels using Distribution, like:

export default TransitionKernel((tokensByPlace, parameters) => {
  const distance = 80;
  const angle = Distribution.Uniform(0, Math.PI * 2);

  return {
    Space: [
      {
        // Here (x,y) are derived from the same Uniform `angle`
        x: angle.map(a => Math.cos(a) * distance),
        y: angle.map(a => Math.sin(a) * distance),
        direction: Distribution.Uniform(0, Math.PI * 2),
        velocity: Distribution.Gaussian(60, 20)
      }
    ],
  };
});

🔗 Related links

• H-5641 (internal)

🔍 What does this change?

• Extracts generateVirtualFiles from create-sdcpn-language-service.ts into its own module
• Adds a global sdcpn-lib.d.ts virtual file exposing Distribution type and Distribution.Gaussian/Distribution.Uniform namespace functions for type checking
• Adds Probabilistic<T> utility type so kernel output allows number | Distribution for numeric fields
• Creates distribution.ts runtime module with:
  • RuntimeDistribution tagged union type for runtime detection
  • distributionRuntimeCode injected into compiled user code context
  • sampleDistribution() using seeded RNG (Box-Muller transform for Gaussian)
• Updates compile-user-code.ts to inject the Distribution namespace into the execution environment
• Updates compute-possible-transition.ts to resolve Distribution values in kernel output by sampling, threading RNG state through all samples

Pre-Merge Checklist 🚀

🚢 Has this modified a publishable library?

This PR:

• modifies an npm-publishable library and I have added a changeset file(s)

📜 Does this require a change to the docs?

The changes in this PR:

• are in a state where docs changes are not yet required but will be

🕸️ Does this require a change to the Turbo Graph?

The changes in this PR:

• do not affect the execution graph

🛡 What tests cover this?

• Existing simulator tests continue to pass (no Distribution values = no behavior change)

❓ How to test this?

1. Checkout the branch
2. In a transition kernel editor, use Distribution.Gaussian(0, 1) or Distribution.Uniform(0, 10) as a value in the returned output tokens
3. Confirm that the type checker accepts Distribution where number is expected in kernel output
4. Run a simulation and confirm sampled values vary per the seeded RNG

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
hash	Error		Feb 21, 2026 3:01am
petrinaut	Ready	Preview	Feb 21, 2026 3:01am

2 Skipped Deployments

Project	Deployment	Actions	Updated (UTC)
hashdotdesign	Ignored	Preview	Feb 21, 2026 3:01am
hashdotdesign-tokens	Ignored	Preview	Feb 21, 2026 3:01am

[kube]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• H-5641: Probability Distributions in Petrinaut #8463 👈 (View in Graphite)
• H-5839, FE-488: Add LSP layer for Petrinaut Monaco editors #8300
• FE-223, H-5780, H-6227: Setup Petrinaut inside HASH + Build with Vite 8 and TypeScript 7 #8369
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.