test(experiment): ground RTO apparatus in real fixture

2026-08-18 12:39:19 +01:00
parent 25a88c5fc3
commit 4a34dcc361
2 changed files with 218 additions and 1 deletions
@@ -1,10 +1,41 @@
 /**
 * Expanded mock scenario library for the Confidence Engine workspace.
 * Each scenario produces a complete investigation journey through turns.
- * 
+ *
 * UI-only development work — no reasoning engine changes.
 */

+/* ── RTO.A4 real captured fixture (RTO.A1 superseded) ──────────── */
+
+import fixtureData from "../../tests/fixtures/live-product-launch-update-response.json";
+
+/**
+ * Thin projection of a real captured Engine update response into the
+ * shape already expected by the workspace/mock client.
+ *
+ * The real captured fixture remains the source of truth — no reasoning
+ * content is recreated or synthesized here.
+ */
+function projectRealUpdateFixture(fixture) {
+  return {
+    success: fixture.success,
+    stage: fixture.stage,
+    situationGraph: fixture.updatedSituationGraph,
+    updatedSituationGraph: fixture.updatedSituationGraph,
+    selectedQuestion: fixture.selectedQuestion,
+    newlySurfacedNodeIds: (fixture.proposal?.addedNodes || [])
+      .filter(node => node.kind === "unknown")
+      .map(node => node.id),
+    diagnostics: fixture.diagnostics,
+    noQuestionReason: null,
+    // Mark as experimental so buildScenarioFixture routes through the scenario branch
+    _experimental: { apparatusMode: "rto-real-fixture" },
+    _sourceFixturePath: "../../tests/fixtures/live-product-launch-update-response.json",
+  };
+}
+
+const __rtoRealScenario = projectRealUpdateFixture(fixtureData);
+
 /* ── Node / Edge factories ─────────────────────────────── */

 export function mkN(id, label, opts) {
@@ -389,6 +420,11 @@ export var MULTI_THREAD_ITEMS = multiThreadItems;
 /* ── Registry ─────────────────────────────────────── */

 var SCENARIOS = {
+  "rto-real-product-launch": {
+    scenario: __rtoRealScenario,
+    label: "[RTO.A4] Real captured product-launch update",
+    centralStatement: __rtoRealScenario?.updatedSituationGraph?.centralStatement || "",
+  },
  "default":       { turns: comparisonTurns,     label: "Comparison (product ratings)", centralStatement: comparisonTurns[0].centralStatement },
  "comparison":    { turns: comparisonTurns,     label: "Comparison (product ratings)", centralStatement: comparisonTurns[0].centralStatement },
  "contradictory": { turns: contradictoryTurns,  label: "Contradictory evidence",       centralStatement: contradictoryTurns[0].centralStatement },
@@ -409,6 +445,11 @@ export function buildScenarioFixture(scenarioName, turnIdx) {
  var s = SCENARIOS[scenarioName];
  if (!s) return null;

+  /* RTO.A4 real fixture — returned as-is (already projected) */
+  if (s.scenario && s.scenario._sourceFixturePath) {
+    return s.scenario;
+  }
+
  /* Experimental multi-thread — does not follow the turns model */
  if (s.scenario && s.scenario._experimental) {
    var mt = s.scenario;
@@ -0,0 +1,176 @@
+/**
+ * RTO.A4 — Real captured fixture apparatus verification.
+ *
+ * Proves the bounded projection boundary:
+ *   1. source fixture is the real captured JSON
+ *   2. projected situationGraph is the captured updatedSituationGraph
+ *   3. exactly 3 unresolved unknowns are derived from real graph state
+ *   4. their IDs/labels match the captured nodes
+ *   5. recommended item derives from selectedQuestion.nodeId
+ *   6. diagnostics recommendation agrees for this fixture
+ *   7. no synthetic availableThreads/recommendedThreadId data is required
+ *   8. any presentation selection remains outside situationGraph
+ *   9. active/reasoning ownership fields are not mutated
+ */
+
+import { describe, it, expect } from "vitest";
+import rawFixture from "../fixtures/live-product-launch-update-response.json";
+
+import { buildScenarioFixture } from "@/lib/mocks/scenarios.js";
+
+const projected = buildScenarioFixture("rto-real-product-launch", 0);
+
+describe("RTO.A4 real captured fixture apparatus", () => {
+  /* ── 1. Source fixture is the real captured JSON ─────────── */
+
+  it("uses the live-product-launch-update-response.json as source", () => {
+    expect(projected._sourceFixturePath).toBe("../../tests/fixtures/live-product-launch-update-response.json");
+  });
+
+  /* ── 2. Projected situationGraph is the captured updatedSituationGraph ─── */
+
+  it("projects situationGraph from fixture.updatedSituationGraph", () => {
+    // projected.situationGraph is the SAME object as projected.updatedSituationGraph (alias)
+    expect(projected.situationGraph).toBe(projected.updatedSituationGraph);
+  });
+
+  it("preserves centralStatement from captured graph", () => {
+    expect(projected.situationGraph.centralStatement.length).toBeGreaterThan(100);
+  });
+
+  /* ── 3. Exactly 3 unresolved unknowns derived from real graph state ─── */
+
+  it("derives exactly 3 unresolved unknowns", () => {
+    const unresolved = projected.situationGraph.nodes.filter(
+      (n) => n.kind === "unknown" && n.status !== "resolved"
+    );
+    expect(unresolved.length).toBe(3);
+  });
+
+  /* ── 4. IDs/labels match the captured nodes ──────────────── */
+
+  it("resolves to the three known unresolved node IDs", () => {
+    const resolved = JSON.stringify(rawFixture.updatedSituationGraph.nodes
+      .filter(n => n.kind === "unknown" && n.status !== "resolved")
+      .map(n => n.id)
+      .sort());
+    expect(resolved).toBe(JSON.stringify(["nR4vL9w", "ntpt9ki", "nxmeiab"].sort()));
+  });
+
+  it("labels come directly from captured nodes, not synthesized", () => {
+    const unresolved = projected.situationGraph.nodes.filter(
+      (n) => n.kind === "unknown" && n.status !== "resolved"
+    );
+    const labels = unresolved.map((n) => n.label);
+    expect(labels).toContain(
+      "Probability or current status of the large enterprise customer signing their contract before launch or within the year"
+    );
+    expect(labels).toContain(
+      "Whether competitors are actively developing similar products and how soon they might release them"
+    );
+    expect(labels).toContain(
+      "Sufficiency of committed or highly probable revenue from other customers to cover £300k cost and yield acceptable return"
+    );
+  });
+
+  /* ── 5. Recommended item derives from selectedQuestion.nodeId ─── */
+
+  it("derives recommendation from selectedQuestion.nodeId", () => {
+    expect(projected.selectedQuestion.nodeId).toBe("nR4vL9w");
+  });
+
+  it("recommendation matches one of the unresolved graph nodes", () => {
+    const resolved = projected.situationGraph.nodes.find(
+      (n) => n.id === projected.selectedQuestion.nodeId && n.kind === "unknown"
+    );
+    expect(resolved).toBeDefined();
+    expect(resolved.label).toContain("Sufficiency of committed or highly probable revenue");
+  });
+
+  /* ── 6. Diagnostics recommendation agrees for this fixture ── */
+
+  it("diagnostics.unknownSelectionExplanation.selectedNodeId agrees with selectedQuestion", () => {
+    expect(projected.diagnostics.unknownSelectionExplanation.selectedNodeId).toBe("nR4vL9w");
+    expect(projected.selectedQuestion.nodeId).toBe(
+      projected.diagnostics.unknownSelectionExplanation.selectedNodeId
+    );
+  });
+
+  /* ── 7. No synthetic availableThreads/recommendedThreadId required ─── */
+
+  it("does not require synthetic reasoning fields to establish the open set", () => {
+    // The open set comes from situationGraph.nodes, not from synthetic data
+    const graphUnknowns = projected.situationGraph.nodes.filter(
+      (n) => n.kind === "unknown" && n.status !== "resolved"
+    );
+    expect(graphUnknowns.length).toBe(3);
+    // No _experimental or synthetic thread fields on situationGraph
+    expect(projected.situationGraph._experimental).toBeUndefined();
+  });
+
+  it("does not introduce synthetic recommendationOrder, questionFrame, thread confidence, or priority", () => {
+    const unresolved = projected.situationGraph.nodes.filter(
+      (n) => n.kind === "unknown" && n.status !== "resolved"
+    );
+    for (const node of unresolved) {
+      expect(node).not.toHaveProperty("recommendationOrder");
+      expect(node).not.toHaveProperty("questionFrame");
+    }
+  });
+
+  /* ── 8. Presentation selection remains outside situationGraph ── */
+
+  it("keeps presentation-only state separate from situationGraph", () => {
+    expect(projected.situationGraph.selectedPresentationItemId).toBeUndefined();
+  });
+
+  it("activeUnknownNodeId is not mutated by the projection", () => {
+    expect(projected.situationGraph.activeUnknownNodeId).toBe(
+      rawFixture.updatedSituationGraph.activeUnknownNodeId
+    );
+  });
+
+  it("selectedQuestion is preserved as captured, not rewritten", () => {
+    // selectedQuestion comes directly from the fixture — check key fields
+    expect(projected.selectedQuestion.nodeId).toBe(rawFixture.selectedQuestion.nodeId);
+    expect(projected.selectedQuestion.question).toBe(rawFixture.selectedQuestion.question);
+  });
+
+  /* ── 9. No production reasoning or ownership mutation ─────── */
+
+  it("does not modify activeUnknownNodeId for presentation selection", () => {
+    const resolved = projected.situationGraph.resolvedNodeIds || [];
+    expect(resolved).toEqual([]);
+  });
+
+  it("resolvedNodeIds matches captured state (empty)", () => {
+    const resolved = projected.situationGraph.resolvedNodeIds || [];
+    expect(resolved.length).toBe(0);
+  });
+
+  /* ── Additional boundary checks ─────────────────────────── */
+
+  it("preserves success and stage from fixture", () => {
+    expect(projected.success).toBe(true);
+    expect(projected.stage).toBe("update_applied");
+  });
+
+  it("newlySurfacedNodeIds derives from proposal.addedNodes (unknown kind)", () => {
+    const expected = rawFixture.proposal.addedNodes
+      .filter(n => n.kind === "unknown")
+      .map(n => n.id);
+    expect(projected.newlySurfacedNodeIds).toEqual(expected);
+  });
+
+  it("selectedQuestion matches captured field exactly", () => {
+    expect(projected.selectedQuestion.nodeId).toBe(rawFixture.selectedQuestion.nodeId);
+    expect(projected.selectedQuestion.question).toBe(rawFixture.selectedQuestion.question);
+  });
+
+  it("diagnostics object matches captured diagnostics key fields", () => {
+    expect(projected.diagnostics.promptVersion).toBe(rawFixture.diagnostics.promptVersion);
+    expect(projected.diagnostics.modelName).toBe(rawFixture.diagnostics.modelName);
+    expect(projected.diagnostics.validationStatus).toBe(rawFixture.diagnostics.validationStatus);
+    expect(projected.diagnostics.nodeCount).toBe(rawFixture.diagnostics.nodeCount);
+  });
+});