Merge pull request 'Fix #32: gestringifieerde arrays in onboarding-overview output normaliseren' (#33) from fix/issue-32-onboarding-schema into main

Reviewed-on: #33

src/lib/__tests__/llmSchemas.test.js

@@ -9,6 +9,7 @@ import {
   customTopicSchema,
   graphActionsSchema,
   proposeGraphDeltaSchema,
+  onboardingOverviewSchema,
 } from '../llmSchemas';
 
 const sampleTopic = {
@@ -192,3 +193,56 @@ describe('proposeGraphDeltaSchema', () => {
     ).toThrow();
   });
 });
+
+describe('onboardingOverviewSchema (issue #32 — fast-tier stringified arrays)', () => {
+  const base = {
+    title: 'Privacy',
+    what_it_is: 'How we handle data.',
+    why_it_matters: 'You touch personal data weekly.',
+  };
+  const points = ['Point one', 'Point two', 'Point three'];
+  const topics = [{ topic_id: 'avg', label: 'AVG' }];
+
+  it('accepts well-formed output', () => {
+    const r = onboardingOverviewSchema.parse({ ...base, key_points: points, topics_covered: topics });
+    expect(r.key_points).toEqual(points);
+    expect(r.topics_covered).toEqual(topics);
+  });
+
+  it('coerces JSON-stringified arrays back to arrays (the exact #32 failure)', () => {
+    const r = onboardingOverviewSchema.parse({
+      ...base,
+      key_points: JSON.stringify(points),
+      topics_covered: JSON.stringify(topics),
+    });
+    expect(r.key_points).toEqual(points);
+    expect(r.topics_covered).toEqual(topics);
+  });
+
+  it('splits a bullet/newline string of key points as a fallback', () => {
+    const r = onboardingOverviewSchema.parse({
+      ...base,
+      key_points: '- Point one\n• Point two\n3. Point three',
+      topics_covered: topics,
+    });
+    expect(r.key_points).toEqual(points);
+  });
+
+  it('keeps the first 5 key points instead of failing on overage', () => {
+    const seven = Array.from({ length: 7 }, (_, i) => `P${i + 1}`);
+    const r = onboardingOverviewSchema.parse({ ...base, key_points: seven, topics_covered: topics });
+    expect(r.key_points).toEqual(['P1', 'P2', 'P3', 'P4', 'P5']);
+  });
+
+  it('still rejects genuinely bad output', () => {
+    expect(() =>
+      onboardingOverviewSchema.parse({ ...base, key_points: ['only', 'two'], topics_covered: topics }),
+    ).toThrow();
+    expect(() =>
+      onboardingOverviewSchema.parse({ ...base, key_points: points, topics_covered: 'not json at all' }),
+    ).toThrow();
+    expect(() =>
+      onboardingOverviewSchema.parse({ ...base, key_points: 42, topics_covered: topics }),
+    ).toThrow();
+  });
+});

src/lib/llmSchemas.js

@@ -229,14 +229,43 @@ export const themeSessionSchema = z.object({
   keyTakeaways: z.array(z.string().min(1)).min(3),
 });
 
+// Fast-tier models occasionally emit array tool fields as a JSON-encoded
+// string ("[\"a\",\"b\"]") instead of a real array (issue #32). Parse those
+// back to arrays before validating; leave anything else untouched so real
+// type errors still fail validation.
+function coerceStringifiedArray(v) {
+  if (typeof v === 'string') {
+    const s = v.trim();
+    if (s.startsWith('[')) {
+      try { return JSON.parse(s); } catch { /* keep original, let Zod report */ }
+    }
+  }
+  return v;
+}
+
+const onboardingKeyPoints = z.preprocess((v) => {
+  let out = coerceStringifiedArray(v);
+  // Fallback: a bullet/newline list as one string → split into points.
+  if (typeof out === 'string' && out.includes('\n')) {
+    out = out
+      .split('\n')
+      .map((line) => line.replace(/^\s*[-•*\d.]+\s*/, '').trim())
+      .filter(Boolean);
+  }
+  // Overage is trivial — keep the first 5 rather than failing the user.
+  if (Array.isArray(out) && out.length > 5) out = out.slice(0, 5);
+  return out;
+}, z.array(z.string().min(1)).min(3).max(5));
+
 export const onboardingOverviewSchema = z.object({
   title: z.string().min(1),
   what_it_is: z.string().min(1),
   why_it_matters: z.string().min(1),
-  key_points: z.array(z.string().min(1)).min(3).max(5),
+  key_points: onboardingKeyPoints,
-  topics_covered: z
+  topics_covered: z.preprocess(
-    .array(z.object({ topic_id: z.string().min(1), label: z.string().min(1) }))
+    coerceStringifiedArray,
-    .min(1),
+    z.array(z.object({ topic_id: z.string().min(1), label: z.string().min(1) })).min(1),
+  ),
 });
 
 /**

src/lib/llmTools.js

@@ -470,11 +470,11 @@ export const EMIT_ONBOARDING_OVERVIEW_TOOL = {
         items: { type: 'string' },
         minItems: 3,
         maxItems: 5,
-        description: '3–5 short, concrete takeaways a newcomer should remember about this theme.',
+        description: '3–5 short, concrete takeaways a newcomer should remember about this theme. Must be a JSON array of strings — never a single string.',
       },
       topics_covered: {
         type: 'array',
-        description: 'The topics that make up this theme. Reuse the exact topic_id you were given so the UI can link back.',
+        description: 'The topics that make up this theme. Reuse the exact topic_id you were given so the UI can link back. Must be a JSON array of objects — never a string.',
         items: {
           type: 'object',
           properties: {