import { expect, test, type Page } from "@playwright/test"; import { spawn, type ChildProcessWithoutNullStreams } from "node:child_process"; import { existsSync } from "node:fs"; import http, { type IncomingMessage, type ServerResponse } from "node:http"; import net from "node:net"; interface TestServer { url: string; requests: () => string[]; close: () => Promise; } interface SpawnedApi { baseURL: string; logs: () => string; stop: () => Promise; } interface VoiceProbe { code: number; messages: string[]; binaryChunks: number; } // This fixture intentionally starts a DB-offline API with ALLOW_SEED_PERSONA_FALLBACK=true // so the voice provider cascade can be exercised without a Postgres dependency. const SEEDED_VOICE_PERSONA_CODE = "P1"; function readBody(req: IncomingMessage): Promise { return new Promise((resolve, reject) => { const chunks: Buffer[] = []; req.on("data", (chunk) => chunks.push(Buffer.isBuffer(chunk) ? chunk : Buffer.from(chunk))); req.on("end", () => resolve(Buffer.concat(chunks))); req.on("error", reject); }); } async function freePort(): Promise { return new Promise((resolve, reject) => { const server = net.createServer(); server.unref(); server.on("error", reject); server.listen(0, "127.0.0.1", () => { const address = server.address(); const port = typeof address === "object" && address ? address.port : 0; server.close(() => resolve(port)); }); }); } async function startHttpServer( handler: (req: IncomingMessage, res: ServerResponse) => void | Promise, ): Promise { const port = await freePort(); const requests: string[] = []; const server = http.createServer((req, res) => { requests.push(`${req.method ?? "?"} ${req.url ?? "?"}`); void Promise.resolve(handler(req, res)).catch((err) => { res.writeHead(500, { "content-type": "text/plain" }); res.end(String(err)); }); }); await new Promise((resolve) => server.listen(port, "127.0.0.1", resolve)); return { url: `http://127.0.0.1:${port}`, requests: () => requests, close: () => new Promise((resolve) => server.close(() => resolve())), }; } async function startFakeOpenAI(): Promise { return startHttpServer(async (req, res) => { await readBody(req); if (req.method === "POST" && req.url === "/v1/audio/transcriptions") { res.writeHead(200, { "content-type": "application/json" }); res.end(JSON.stringify({ text: "요즘 잠을 잘 못 자요.", language: "ko", duration: 1.2 })); return; } if (req.method === "POST" && req.url === "/v1/audio/speech") { res.writeHead(200, { "content-type": "audio/mpeg" }); res.end(Buffer.alloc(8192, 128)); return; } res.writeHead(404, { "content-type": "application/json" }); res.end(JSON.stringify({ error: "not found" })); }); } async function startFakeEngine(): Promise { return startHttpServer(async (req, res) => { await readBody(req); if (req.method === "GET" && req.url === "/health") { res.writeHead(200, { "content-type": "application/json" }); res.end(JSON.stringify({ ok: true, engine: "fake" })); return; } if (req.method === "POST" && req.url === "/v1/generate") { res.writeHead(200, { "content-type": "application/json" }); res.end( JSON.stringify({ text: "괜찮아요. 천천히 말해볼게요.", model: "fake-client", provider: "e2e", tokens_in: 1, tokens_out: 1, cost_usd: 0, inference_geo: "us", structured: null, }), ); return; } res.writeHead(404, { "content-type": "application/json" }); res.end(JSON.stringify({ error: "not found" })); }); } async function waitForApi(baseURL: string, proc: ChildProcessWithoutNullStreams): Promise { const started = Date.now(); let lastError = ""; while (Date.now() - started < 20_000) { if (proc.exitCode !== null) { throw new Error(`API exited early with code ${proc.exitCode}: ${lastError}`); } try { const response = await fetch(`${baseURL}/health`); if (response.ok) return; lastError = await response.text(); } catch (err) { lastError = err instanceof Error ? err.message : String(err); } await new Promise((resolve) => setTimeout(resolve, 250)); } throw new Error(`Timed out waiting for API ${baseURL}: ${lastError}`); } async function startApi({ engineURL, openAIBaseURL, }: { engineURL: string; openAIBaseURL: string; }): Promise { const port = await freePort(); const baseURL = `http://127.0.0.1:${port}`; const localPython311 = process.env.USERPROFILE ? `${process.env.USERPROFILE}\\AppData\\Local\\Programs\\Python\\Python311\\python.exe` : ""; const python = process.env.E2E_PYTHON ?? process.env.PYTHON311 ?? (localPython311 && existsSync(localPython311) ? localPython311 : (process.env.PYTHON ?? "python")); const proc = spawn( python, [ "-m", "uvicorn", "app.main:app", "--host", "127.0.0.1", "--port", String(port), "--log-level", "debug", ], { cwd: "../api", env: { ...process.env, PYTHONUNBUFFERED: "1", ENVIRONMENT: "dev", AUTH_DEV_LOGIN_ENABLED: "true", AUTH_ALLOWED_EMAIL_DOMAINS: '["hs.ac.kr","twentyoz.kr"]', ALLOW_SEED_PERSONA_FALLBACK: "true", DATABASE_URL: "postgresql://user:pass@127.0.0.1:1/vignette", DB_POOL_MIN_SIZE: "0", DB_COMMAND_TIMEOUT: "1", ENGINE_URL: engineURL, ENGINE_MODE: "claude_api", ENGINE_TIMEOUT: "10", ENGINE_CONNECT_TIMEOUT: "2", OPENAI_API_KEY: "e2e-fake-key", OPENAI_BASE_URL: `${openAIBaseURL}/v1`, FRONTEND_BASE_URL: "http://localhost:5173", CORS_ORIGINS: '["http://localhost:5173"]', }, windowsHide: true, }, ); let logs = ""; proc.stdout.on("data", (chunk) => { logs += String(chunk).slice(-4000); }); proc.stderr.on("data", (chunk) => { logs += String(chunk).slice(-4000); }); await waitForApi(baseURL, proc).catch((err) => { proc.kill(); throw new Error(`${err instanceof Error ? err.message : String(err)}\n${logs}`); }); return { baseURL, logs: () => logs, stop: async () => { if (proc.exitCode === null) proc.kill(); await new Promise((resolve) => { if (proc.exitCode !== null) { resolve(); return; } proc.once("exit", () => resolve()); setTimeout(resolve, 3000); }); }, }; } async function probeVoiceCascade(page: Page, apiBaseURL: string, sessionId: string): Promise { return page.evaluate( ({ apiBase, sid }) => new Promise((resolve) => { const wsURL = new URL(`/voice/ws?session_id=${encodeURIComponent(sid)}`, apiBase); wsURL.protocol = "ws:"; const ws = new WebSocket(wsURL.href); ws.binaryType = "arraybuffer"; const messages: string[] = []; let binaryChunks = 0; let sawTtsEnd = false; const timeout = window.setTimeout(() => { ws.close(); resolve({ code: -1, messages, binaryChunks }); }, 20_000); ws.onopen = () => { ws.send(JSON.stringify({ type: "audio_start", format: "webm" })); ws.send(new Uint8Array([1, 2, 3, 4, 5, 6]).buffer); ws.send(JSON.stringify({ type: "audio_end", format: "webm" })); }; ws.onmessage = (event) => { if (typeof event.data === "string") { messages.push(event.data); try { const parsed = JSON.parse(event.data) as { type?: string; state?: string }; if (parsed.type === "tts_end") sawTtsEnd = true; if (sawTtsEnd && parsed.type === "state" && parsed.state === "idle") { ws.send(JSON.stringify({ type: "close" })); } } catch { messages.push(JSON.stringify({ type: "error", detail: "invalid json from ws" })); } } else { binaryChunks += 1; } }; ws.onerror = () => { messages.push(JSON.stringify({ type: "error", detail: "browser websocket error" })); }; ws.onclose = (event) => { window.clearTimeout(timeout); resolve({ code: event.code, messages, binaryChunks }); }; }), { apiBase: apiBaseURL, sid: sessionId }, ); } test.describe("voice cascade success path", () => { test("runs STT, client turn, TTS, and audio chunks against controlled providers @single-run", async ({ page, }, testInfo) => { test.setTimeout(60_000); const openai = await startFakeOpenAI(); const engine = await startFakeEngine(); const api = await startApi({ engineURL: engine.url, openAIBaseURL: openai.url }); try { const health = await page.request.get(`${api.baseURL}/voice/health`); expect(health.ok(), await health.text()).toBeTruthy(); await expect(await health.json()).toMatchObject({ status: "ok", available: true }); await page.goto(`${api.baseURL}/health`); const browserSetup = await page.evaluate(async ({ apiBase, seededPersonaCode, workerIndex }) => { const login = await fetch(`${apiBase}/auth/dev-login`, { method: "POST", credentials: "include", headers: { "content-type": "application/json" }, body: JSON.stringify({ email: `voice-success.${workerIndex}@hs.ac.kr`, role: "learner", display_name: "Voice Success", }), }); const loginBody = await login.text(); if (!login.ok) { return { ok: false, step: "login", status: login.status, body: loginBody }; } const me = await fetch(`${apiBase}/auth/me`, { credentials: "include" }); const meBody = await me.text(); if (!me.ok) { return { ok: false, step: "me", status: me.status, body: meBody }; } const start = await fetch(`${apiBase}/sessions`, { method: "POST", credentials: "include", headers: { "content-type": "application/json" }, body: JSON.stringify({ persona_code: seededPersonaCode, theory_mode: "humanistic" }), }); const startBody = await start.text(); if (!start.ok) { return { ok: false, step: "sessions", status: start.status, body: startBody }; } return { ok: true, me: JSON.parse(meBody) as unknown, started: JSON.parse(startBody) as { session_id: string }, }; }, { apiBase: api.baseURL, seededPersonaCode: SEEDED_VOICE_PERSONA_CODE, workerIndex: testInfo.workerIndex, }); expect(browserSetup, api.logs()).toMatchObject({ ok: true }); if (!browserSetup.ok) throw new Error(JSON.stringify(browserSetup)); const started = browserSetup.started; const result = await probeVoiceCascade(page, api.baseURL, started.session_id); const events = result.messages.map((message) => JSON.parse(message) as { type: string; [key: string]: unknown }); expect( result.code, [ JSON.stringify(result, null, 2), `fakeOpenAI=${JSON.stringify(openai.requests())}`, `fakeEngine=${JSON.stringify(engine.requests())}`, api.logs(), ].join("\n\n"), ).toBe(1000); expect(events.some((event) => event.type === "degraded")).toBe(false); expect(events.some((event) => event.type === "error")).toBe(false); expect(events).toEqual( expect.arrayContaining([ expect.objectContaining({ type: "ready", session_id: started.session_id }), expect.objectContaining({ type: "state", state: "listening" }), expect.objectContaining({ type: "state", state: "thinking" }), expect.objectContaining({ type: "transcript", text: "요즘 잠을 잘 못 자요." }), expect.objectContaining({ type: "reply", text: "괜찮아요. 천천히 말해볼게요." }), expect.objectContaining({ type: "state", state: "speaking" }), expect.objectContaining({ type: "tts_chunk", seq: 0 }), expect.objectContaining({ type: "tts_end" }), expect.objectContaining({ type: "state", state: "idle" }), ]), ); expect(result.binaryChunks).toBeGreaterThan(0); } finally { await api.stop(); await engine.close(); await openai.close(); } }); });