Refine physical industry classifications and preserve pivot revenue limits

apps/docs/tests/research.test.ts

@@ -24,8 +24,8 @@ test("research revisions share the authentication database and preserve replay l
     assert.equal(consumeCode(config, "invalid"), "limited");
     const initial = readResearchSnapshot();
     assert.equal(initial.companies.length, 1039);
-    assert.equal(initial.evidence.length, 1445);
-    assert.equal(initial.observations.length, 202);
+    assert.equal(initial.evidence.length, 1457);
+    assert.equal(initial.observations.length, 207);
     assert.deepEqual(initial.candidateBatches, seed.candidateBatches, "candidate gaps persist in the same database metadata");
     assert.deepEqual(initial.directoryRechecks, seed.directoryRechecks, "follow-up classifications persist without rewriting the discovery audit");
     assert.deepEqual(readResearchSnapshot(), initial, "reopening does not duplicate records");
@@ -64,7 +64,7 @@ test("research validation rejects broken evidence and unsafe source links", () =
 
 test("reviews link explicit companies without turning global or excluded cases into companies", () => {
   const snapshot = structuredClone(seed) as ResearchSnapshot;
-  assert.equal(snapshot.reviews.filter((r) => r.scope === "companies").length, 85);
+  assert.equal(snapshot.reviews.filter((r) => r.scope === "companies").length, 87);
   assert.equal(snapshot.reviews.filter((r) => r.scope === "dataset").length, 10);
   assert.equal(snapshot.reviews.filter((r) => r.scope === "excluded-candidate").length, 38);
   const review = snapshot.reviews.find((r) => r.id === "R4")!;
@@ -179,8 +179,14 @@ test("directory coverage reconciles every candidate and keeps profile-only incom
     assert.ok(company.profileOnly || company.directoryOrigin === "yc-2025-2026");
     assert.equal(company.url, row.source);
     if (company.profileOnly) assert.equal(company.revenueStatus, "unknown");
-    assert.equal(company.year, row.foundedYear);
-    assert.ok([2025, 2026].includes(company.year!));
+    if (company.id === "YC31370") {
+      assert.equal(row.foundedYear, 2025, "original discovery remains unchanged");
+      assert.equal(company.year, null, "a pivot year cannot settle the company founding year");
+      assert.ok(snapshot.reviews.some((review) => review.id === "DAYJOB-R-COHORT" && review.companyIds?.includes(company.id)));
+    } else {
+      assert.equal(company.year, row.foundedYear);
+      assert.ok([2025, 2026].includes(company.year!));
+    }
   }
   const invalid = structuredClone(snapshot);
   invalid.coverage!.newProfiles += 1;
@@ -194,6 +200,20 @@ test("directory coverage reconciles every candidate and keeps profile-only incom
   assert.throws(() => validateResearch(lostOrigin), /coverage accounting/);
 });
 
+test("business pivots and procurement volume do not become comparable startup ARR", () => {
+  const snapshot = seed as ResearchSnapshot;
+  const dayjob = snapshot.companies.find((company) => company.id === "YC31370")!;
+  assert.equal(dayjob.year, null);
+  assert.equal(dayjob.directoryOrigin, "yc-2025-2026");
+  const milestones = snapshot.observations.filter((observation) => observation.company === dayjob.id);
+  assert.deepEqual(milestones.filter((observation) => observation.selection === "crosscheck").map((observation) => observation.value).sort(), [496000, 653000]);
+  assert.ok(milestones.every((observation) => observation.comparisonGroup === null && !observation.metric && !observation.observedAt));
+  assert.ok(milestones.filter((observation) => [150000, 800000].includes(observation.value)).every((observation) => observation.selection === "excluded"));
+  const procurement = snapshot.observations.find((observation) => observation.id === "SAUDARA-O-SOURCING")!;
+  assert.equal(procurement.selection, "excluded");
+  assert.equal(procurement.comparisonGroup, null);
+});
+
 test("income descriptions report sample dispersion and maximum-value sensitivity", () => {
   const stats = describeIncome([100, 200, 300])!;
   assert.equal(stats.mean, 200);