evOo-local/apps/web/e2e/voice.spec.ts

import { expect, type Page, test } from "@playwright/test";
import { captureScreenshot, completeOnboarding, openUserSettings, rpc, signup } from "./helpers";

async function turnOnSpeaking(page: Page) {
  const settings = await openUserSettings(page, "voice");
  const voice = settings.getByTestId("voice-settings");
  await expect(voice).toBeVisible();
  await expect(voice.getByRole("button", { name: "Other voice providers" })).toHaveCount(0);
  await expect(voice.getByLabel("API key", { exact: true })).toHaveCount(0);
  // The app turns the local voice on by itself the first time it starts.
  const speaking = voice.getByRole("switch", { name: "Speaking on or off" });
  // Clicking before the automatic start has settled would turn the voice off again.
  const startedByItself = await expect(speaking)
    .toHaveAttribute("aria-checked", "true", { timeout: 5_000 })
    .then(
      () => true,
      () => false,
    );
  if (!startedByItself) await speaking.click();
  await expect(speaking).toHaveAttribute("aria-checked", "true");
  await expect(voice.getByTestId("voice-carousel")).toBeVisible();
  return voice;
}

test("voice settings pick a voice from the carousel, speak a reply, and open a call", async ({
  page,
}, testInfo) => {
  const stamp = Date.now();
  await signup(page, `voice-${stamp}@rakazo.test`, "password12", `Voice ${stamp}`);
  await completeOnboarding(page);

  await expect(page.getByRole("button", { name: "Call", exact: true })).toHaveCount(0);
  await expect(
    page.getByTestId("composer-bar").getByRole("button", { name: "Voice", exact: true }),
  ).toHaveCount(1);
  await captureScreenshot(page, testInfo, "voice-composer");

  const voice = await turnOnSpeaking(page);
  const cards = voice.getByTestId("voice-card");
  // English UI lists only the English voices of the local server.
  await expect(cards.filter({ hasText: "Heart" })).toHaveCount(1);
  await expect(cards.filter({ hasText: "George" })).toHaveCount(1);
  await expect(cards.filter({ hasText: "Eva" })).toHaveCount(0);
  await expect(voice.getByRole("button", { name: "Heart", exact: true })).toHaveAttribute(
    "aria-pressed",
    "true",
  );

  await voice.getByRole("button", { name: "Bella", exact: true }).click();
  await expect(voice.getByRole("button", { name: "Bella", exact: true })).toHaveAttribute(
    "aria-pressed",
    "true",
  );
  await expect(voice.getByRole("button", { name: "Heart", exact: true })).toHaveAttribute(
    "aria-pressed",
    "false",
  );
  await expect
    .poll(async () => (await rpc<{ voiceId: string }>(page, "voice/status", {})).voiceId)
    .toBe("af_bella");

  const previewSpoken = page.waitForResponse(
    (response) => response.url().includes("/api/voice/speak") && response.ok(),
  );
  await voice.getByRole("button", { name: "Play Sky", exact: true }).click();
  await previewSpoken;
  // A preview only plays the voice; the chosen one stays.
  await expect(voice.getByRole("button", { name: "Bella", exact: true })).toHaveAttribute(
    "aria-pressed",
    "true",
  );

  for (const colorScheme of ["light", "dark"] as const) {
    await page.emulateMedia({ colorScheme });
    await captureScreenshot(page, testInfo, `voice-settings-carousel-${colorScheme}`);
  }
  await page.emulateMedia({ colorScheme: "light" });
  await page.getByRole("button", { name: "Close voice settings" }).click();
  await expect(page.getByTestId("user-settings")).toHaveCount(0);

  const composer = page.getByPlaceholder(/Message/);
  await composer.fill("say hello");
  await expect(
    page.getByTestId("composer-bar").getByRole("button", { name: "Voice", exact: true }),
  ).toHaveCount(0);
  await captureScreenshot(page, testInfo, "voice-composer-typing");
  await page.keyboard.press("Enter");
  // A reply can render more than one text bubble; speak the latest one.
  const speakReply = page.getByRole("button", { name: "Speak this reply" }).last();
  await expect(speakReply).toBeVisible({ timeout: 30_000 });
  const replySpoken = page.waitForResponse(
    (response) => response.url().includes("/api/voice/speak") && response.ok(),
  );
  await speakReply.click();
  await replySpoken;

  await page
    .getByTestId("composer-bar")
    .getByRole("button", { name: "Voice", exact: true })
    .click();
  await expect(page.getByTestId("call-view")).toBeVisible();
  await expect(page.getByTestId("voice-panel")).toBeVisible();
  await expect(page.getByRole("button", { name: "Hang up" })).toBeVisible();
  await captureScreenshot(page, testInfo, "voice-stage");
  await page.getByRole("button", { name: "Hang up" }).click();
  await expect(page.getByTestId("call-view")).toHaveCount(0);
});

test("German lists only German voices and switches the saved voice to one", async ({
  page,
}, testInfo) => {
  const stamp = Date.now();
  await signup(page, `voice-de-${stamp}@rakazo.test`, "password12", `Stimme ${stamp}`);
  await completeOnboarding(page);

  let voice = await turnOnSpeaking(page);
  await expect(voice.getByRole("button", { name: "Heart", exact: true })).toHaveAttribute(
    "aria-pressed",
    "true",
  );
  await page.getByRole("button", { name: "Close voice settings" }).click();

  // Switching the app language moves the saved voice to that language too.
  const settings = await openUserSettings(page);
  await settings.getByTestId("ui-locale-select").click();
  await settings.getByRole("option", { name: "Deutsch", exact: true }).click();
  await expect(settings.getByRole("heading", { name: "Sprache", exact: true })).toBeVisible();
  await expect
    .poll(async () => (await rpc<{ voiceId: string }>(page, "voice/status", {})).voiceId)
    .toBe("dm_thorsten");
  await settings.getByTestId("settings-nav-voice").click();
  voice = settings.getByTestId("voice-settings");
  const cards = voice.getByTestId("voice-card");
  await expect(cards.filter({ hasText: "Eva" })).toHaveCount(1);
  await expect(cards.filter({ hasText: "Thorsten" })).toHaveCount(1);
  await expect(cards.filter({ hasText: "Heart" })).toHaveCount(0);
  await expect(voice.getByRole("button", { name: "Thorsten", exact: true })).toHaveAttribute(
    "aria-pressed",
    "true",
  );
  await expect(voice.getByRole("button", { name: "Eva anhören", exact: true })).toBeVisible();

  for (const colorScheme of ["light", "dark"] as const) {
    await page.emulateMedia({ colorScheme });
    await captureScreenshot(page, testInfo, `voice-settings-carousel-de-${colorScheme}`);
  }
  await page.emulateMedia({ colorScheme: "light" });
  await page.setViewportSize({ width: 390, height: 844 });
  await expect(voice.getByTestId("voice-carousel")).toBeVisible();
  await captureScreenshot(page, testInfo, "voice-settings-carousel-de-phone");
});