fix(transcript): collapse rolling segment echoes

Normalize incremental backend hypothesis chains before persistence and ignore stale or replayed webhook callbacks so duplicate transcript text does not survive ingest.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

src/tests/audio.test.ts

@@ -7,17 +7,18 @@ import { describe, it, expect, vi, afterEach } from 'vitest';
 
 const execFileMock = vi.hoisted(() => {
 	const fn = vi.fn();
+	type ExecFilePromisifyArgs = [string, string[]];
+	type ExecFileCallback = (err: Error | null, stdout: string, stderr: string) => void;
+	type ExecFileMock = (...args: [...ExecFilePromisifyArgs, ExecFileCallback]) => void;
+	const invoke = fn as unknown as ExecFileMock;
 	Object.defineProperty(fn, Symbol.for('nodejs.util.promisify.custom'), {
 		configurable: true,
-		value: (...args: unknown[]) =>
+		value: (...args: ExecFilePromisifyArgs) =>
 			new Promise<{ stdout: string; stderr: string }>((resolve, reject) => {
-				(fn as ReturnType<typeof vi.fn>)(
-					...args,
-					(err: Error | null, stdout: string, stderr: string) => {
-						if (err) reject(err);
-						else resolve({ stdout, stderr });
-					}
-				);
+				invoke(...args, (err: Error | null, stdout: string, stderr: string) => {
+					if (err) reject(err);
+					else resolve({ stdout, stderr });
+				});
 			})
 	});
 	return fn;

src/tests/postprocess.test.ts

@@ -66,6 +66,50 @@ describe('deduplicateSegments — mergeConsecutive', () => {
 	});
 });
 
+// ── rolling prefix/suffix chain collapse ───────────────────────────────────────
+
+describe('deduplicateSegments — rolling backend hypotheses', () => {
+	it('collapses prefix-growth chains from stored backend segments', () => {
+		const input = [
+			seg(0, 15.24, 16.6, 'Hello everyone.'),
+			seg(1, 16.6, 19.47, 'Hello everyone. Um, welcome to this talk.'),
+			seg(2, 19.47, 19.48, 'Um, welcome to this talk.'),
+			seg(3, 19.48, 21.67, "Um, welcome to this talk. I'll be speaking about small model"),
+			seg(4, 21.67, 21.68, "I'll be speaking about small model"),
+			seg(5, 21.68, 24.59, "I'll be speaking about small model inference and a gap that we've")
+		];
+
+		const result = deduplicateSegments(input);
+
+		expect(result).toHaveLength(2);
+		expect(result[0]).toMatchObject({
+			index: 0,
+			start: 15.24,
+			end: 19.48,
+			text: 'Hello everyone. Um, welcome to this talk.'
+		});
+		expect(result[1]).toMatchObject({
+			index: 1,
+			start: 19.48,
+			end: 24.59,
+			text: "I'll be speaking about small model inference and a gap that we've"
+		});
+	});
+
+	it('does not collapse similar phrases when there is a real timing gap', () => {
+		const input = [
+			seg(0, 0, 1, 'Hello everyone.'),
+			seg(1, 2, 4, 'Hello everyone. Welcome back.')
+		];
+
+		const result = deduplicateSegments(input);
+
+		expect(result).toHaveLength(2);
+		expect(result[0].text).toBe('Hello everyone.');
+		expect(result[1].text).toBe('Hello everyone. Welcome back.');
+	});
+});
+
 // ── ngramDedup ────────────────────────────────────────────────────────────────
 
 describe('deduplicateSegments — ngramDedup', () => {

src/tests/webhook.test.ts

@@ -132,6 +132,43 @@ describe('POST /api/webhook/[jobId] — locally cancelled job', () => {
 	});
 });
 
+// ── Duplicate / stale callback guards ──────────────────────────────────────────
+
+describe('POST /api/webhook/[jobId] — duplicate and stale callbacks', () => {
+	it('ignores replayed success callbacks after the transcript is already done', async () => {
+		mockGetJob.mockReturnValue({
+			...makeJob('job-done'),
+			status: 'done',
+			segmentsJson: JSON.stringify([makeSeg(0, 'Already saved.')]),
+			whisperJobId: 'whisper-id'
+		});
+
+		const res = await POST(makeEvent('job-done', makeWhisperJob()) as any);
+		expect(res.status).toBe(200);
+		expect(await res.json()).toEqual({ ok: true, ignored: 'duplicate_webhook' });
+		expect(mockSetJobStatus).not.toHaveBeenCalled();
+		expect(mockUpdateJob).not.toHaveBeenCalled();
+		expect(mockWriteOutputs).not.toHaveBeenCalled();
+	});
+
+	it('ignores stale callbacks from an older whisper job after retry', async () => {
+		mockGetJob.mockReturnValue({
+			...makeJob('job-stale'),
+			status: 'transcribing',
+			whisperJobId: 'current-whisper-job'
+		});
+
+		const res = await POST(
+			makeEvent('job-stale', makeWhisperJob({ id: 'old-whisper-job', segments: [makeSeg(0, 'stale')] })) as any
+		);
+		expect(res.status).toBe(200);
+		expect(await res.json()).toEqual({ ok: true, ignored: 'stale_whisper_job' });
+		expect(mockSetJobStatus).not.toHaveBeenCalled();
+		expect(mockUpdateJob).not.toHaveBeenCalled();
+		expect(mockWriteOutputs).not.toHaveBeenCalled();
+	});
+});
+
 // ── Whisper job failed / cancelled ───────────────────────────────────────────
 
 describe('POST /api/webhook/[jobId] — whisper failure', () => {