Create scenarios.py

Author: BryceWDesign

src/ohip_bench/scenarios.py

@@ -0,0 +1,144 @@
+"""
+Built-in deterministic benchmark scenarios for IX-HapticSight.
+
+This module provides a small catalog of reusable benchmark scenarios that can
+be run against the current RuntimeService + BenchmarkRunner stack.
+
+The intent is to make common evaluation paths explicit and discoverable:
+- consent happy path
+- consent denial path
+- safety denial path
+- execution capability mismatch path
+
+These scenarios are still repository-stage artifacts, not deployment evidence.
+"""
+
+from __future__ import annotations
+
+from .models import (
+    BenchmarkDomain,
+    BenchmarkExpectation,
+    BenchmarkScenario,
+)
+from .runner import make_consent_scenario
+
+
+def make_safety_red_scenario() -> BenchmarkScenario:
+    """
+    Contact request with explicit consent but a RED session safety level.
+
+    Expected result:
+    - DENIED
+    - not executable
+    - fault reason tied to session safety red
+    """
+    return BenchmarkScenario(
+        scenario_id="safety-red-001",
+        title="RED safety level blocks support contact",
+        domain=BenchmarkDomain.SAFETY,
+        description=(
+            "A contact request should be denied when the session starts in RED "
+            "safety state even if explicit consent exists."
+        ),
+        inputs={
+            "session": {
+                "session_id": "sess-1",
+                "subject_id": "person-1",
+                "interaction_state": "IDLE",
+                "execution_state": "IDLE",
+                "safety_level": "RED",
+                "consent_valid": False,
+                "consent_fresh": False,
+            },
+            "request": {
+                "request_id": "req-1",
+                "interaction_kind": "SUPPORT_CONTACT",
+                "source": "BENCHMARK",
+                "target_name": "shoulder",
+                "requested_scope": "shoulder_contact",
+                "requires_contact": True,
+                "requires_consent_freshness": True,
+            },
+            "consent": {
+                "grant_explicit": True,
+                "scopes": ["shoulder_contact"],
+                "source": "benchmark",
+            },
+            "nudge": {
+                "level": "GREEN",
+                "target": {
+                    "frame": "W",
+                    "xyz": [0.42, -0.18, 1.36],
+                    "rpy": [0.0, 0.0, 1.57],
+                },
+                "normal": [0.0, 0.8, 0.6],
+                "rationale": "benchmark shoulder support",
+                "priority": 0.9,
+                "expires_in_ms": 500,
+            },
+            "start_pose": {
+                "frame": "W",
+                "xyz": [0.10, 0.00, 1.00],
+                "rpy": [0.0, 0.0, 0.0],
+            },
+        },
+        expectation=BenchmarkExpectation(
+            expected_status="DENIED",
+            expected_executable=False,
+            expected_fault_reason="session_safety_red",
+            expected_execution_status="",
+        ),
+        tags=("safety", "red", "contact", "benchmark"),
+    )
+
+
+def make_consent_catalog() -> list[BenchmarkScenario]:
+    """
+    Standard consent-path scenarios for the current repo stage.
+    """
+    return [
+        make_consent_scenario(
+            scenario_id="consent-approved-001",
+            title="Explicit consent allows support contact",
+            explicit_consent=True,
+            expected_status="APPROVED",
+            expected_executable=True,
+            expected_execution_status="ACCEPTED",
+        ),
+        make_consent_scenario(
+            scenario_id="consent-denied-001",
+            title="Missing consent blocks support contact",
+            explicit_consent=False,
+            expected_status="DENIED",
+            expected_executable=False,
+            expected_fault_reason="consent_missing_or_invalid",
+        ),
+    ]
+
+
+def make_core_catalog() -> list[BenchmarkScenario]:
+    """
+    Core benchmark catalog for current repository maturity.
+
+    This intentionally stays small and deterministic. More scenarios can be
+    added later once the runtime, replay, and logging layers deepen.
+    """
+    catalog: list[BenchmarkScenario] = []
+    catalog.extend(make_consent_catalog())
+    catalog.append(make_safety_red_scenario())
+    return catalog
+
+
+def scenario_ids(scenarios: list[BenchmarkScenario]) -> list[str]:
+    """
+    Return the ordered scenario IDs from a benchmark catalog.
+    """
+    return [scenario.scenario_id for scenario in scenarios]
+
+
+__all__ = [
+    "make_safety_red_scenario",
+    "make_consent_catalog",
+    "make_core_catalog",
+    "scenario_ids",
+]