Wave 5 — approachability & front door: CLI — discover --preset (physics/ecology/epidemiology/finance/general) + presets catalog; explicit flags override presets. library becomes a provenance-aware workspace: world/data content hashes, columns, config, tags captured on add; library search + compare over the workspace (backward-compatible index). SDK — lawsynth.recipes (parity with CLI presets) + Study.discover(recipe=..., **overrides); three runnable domain demos. Van der Pol shows the payoff: general recipe misses the cubic term, mechanics recipe recovers -x+v-x^2*v (R2 0.99995). Docs — a real product README (hero + Lorenz GIF, CLI+Python quickstart, feature matrix by product loop, surfaces + HTTP endpoints, honest tone); refreshed getting-started pages and a CHANGELOG Unreleased section. All flags verified against the shipped CLI usage. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
1.7 KiB
Your first World
There are two ways to get a .lsworld bundle to work with: discover one from
observations, or scaffold one from a built-in template.
Discover from observations
Follow the quickstart CSV route when the law should be inferred from data:
lawsynth discover observations.csv --time time --state x,y --output world.lsworld
The generated World is a canonical, validated archive — not a generic JSON document. The Rust bundle writer chooses the layout and validates the World before writing it.
Scaffold from a template
To explore the loop without your own data, generate a known world (and, optionally, sample observations from it):
lawsynth templates
lawsynth new lorenz --output lorenz.lsworld --data lorenz.csv --samples 2000
Built-in templates: lorenz, lotka-volterra, pendulum, van-der-pol, sir.
Inspect before reuse
lawsynth inspect world.lsworld
The inspector distinguishes continuous and discrete Worlds and reports state, variable, and parameter counts. A failed inspection means the artifact is not accepted by either bundle reader — a file extension is not proof of validity.
Where to go next
lawsynth explain world.lsworld— read the laws and their assumptions.lawsynth simulate/lawsynth forecast— run it forward and ask what-if.lawsynth validate world.lsworld --data obs.csv --holdout 0.2— score it on held-out data.lawsynth report world.lsworld --output report.html— share a self-contained report.
For a manually constructed World, use the Rust lawsynth-world API or the built
Python native API. Keep source values finite, identifiers valid, units compatible,
and exactly one law assigned to each state.