Skip to content

Your first scenario in 5 minutes

This page walks through running an existing Ophamin scenario against the bundled MockSubstrate — no external system required, runs in a few seconds, produces a signed proof record.

1. Confirm the install

ophamin scenario list

You should see 32 scenarios across five tiers. Pick one that doesn't need a real corpus on disk — BayesianPhiPosteriorScenario runs against a Family-L simulation built into the scenario itself:

ophamin scenario show bayesian-phi-posterior

2. Run it

The bundled MockSubstrate is the default substrate for the demo mode; for any scenario that doesn't require an external corpus or a real Kimera repo, no further wiring is needed. From a Python REPL:

from ophamin.measuring.scenarios import BayesianPhiPosteriorScenario
from ophamin.measuring.proof.codec import dump

scenario = BayesianPhiPosteriorScenario(
    simulate_from_family_l=True,
    sample_sizes=(20, 200),
    draws=1000, tune=500,
    seed=42,
)
proof = scenario.run()
dump(proof, "my_first_proof.json")
print(f"verdict: {proof.verdict.outcome}")
print(f"observed: {proof.verdict.observed_value:.4f}")
print(f"threshold: {proof.claim.threshold.describe()}")
print(f"proof_id: {proof.proof_id}")

Output:

verdict: VALIDATED
observed: 0.3162
threshold: hdi_contraction_ratio <= 0.40 fraction
proof_id: a3b9f8c2…

3. Verify the proof

The proof is HMAC-signed with the default sign key. Inspect or validate it:

ophamin schema info my_first_proof.json
ophamin schema validate my_first_proof.json

To verify the signature, pass the key (the framework's default sign key is ophamin-default-sign-key; in production you'd use your own):

ophamin schema validate my_first_proof.json --key ophamin-default-sign-key

4. Read the proof

The proof is a 9-section JSON record. Open it in your editor or:

python -m json.tool my_first_proof.json | less

The nine sections:

  1. Identity — proof_id, schema_version, ophamin_version, created_at
  2. Claim — falsifiable statement + operationalisation + threshold
  3. Pre-registration — config_hash, data_hash, analysis plan captured BEFORE the run
  4. Data — substrate name + commit + datasets used
  5. Evidence — per-pillar statistics with library attribution
  6. Verdict — VALIDATED / REFUTED / INCONCLUSIVE
  7. Reproduction — exact command + environment lock
  8. Provenance — W3C PROV-JSON graph
  9. Signature — HMAC-SHA256 over the canonical JSON form

See SCHEMAS.md for the full layout and versioning policy.

Next steps