feat(audio): add offline local neural voice via Piper (sv + en)

Announcements can use Piper, a free local neural TTS, instead of the browser's system
voice. No account, no paid service, and it keeps working with no internet at the track.

- npm run tts:install (or Settings -> Audio -> Install) downloads the standalone Piper
  release for Linux/Windows/macOS plus Swedish and English voices into piper/ once
- server/tts.js generates each line once, one at a time, and caches it as WAV under
  data/tts/; a new line takes ~0.3 s with the standalone build
- src/piper_speech.js plays it and falls back to the browser voice if Piper is missing,
  errors, or takes over 2.5 s, so no announcement is lost and timing is never held up
- "Pre-generate phrases and driver names" renders every fixed phrase and name in both
  languages before race day

Browser voice stays the default. Generation routes fall under the admin PIN.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

server/tts_install.js

@@ -0,0 +1,65 @@
+// Downloads the standalone Piper release for this OS and the Swedish + English voices
+// into <project>/piper/. Needs internet once; everything runs offline afterwards.
+const fs = require("fs");
+const os = require("os");
+const path = require("path");
+const { Readable } = require("stream");
+const { pipeline } = require("stream/promises");
+const { spawnSync } = require("child_process");
+
+const PIPER_RELEASE_BASE = "https://github.com/rhasspy/piper/releases/download/2023.11.14-2";
+const PIPER_ARCHIVES = {
+  "linux-x64": "piper_linux_x86_64.tar.gz",
+  "linux-arm64": "piper_linux_aarch64.tar.gz",
+  "win32-x64": "piper_windows_amd64.zip",
+  "darwin-x64": "piper_macos_x64.tar.gz",
+  "darwin-arm64": "piper_macos_aarch64.tar.gz",
+};
+const VOICE_BASE = "https://huggingface.co/rhasspy/piper-voices/resolve/main";
+const VOICES = [
+  { name: "sv_SE-nst-medium", path: "sv/sv_SE/nst/medium" },
+  { name: "en_US-lessac-medium", path: "en/en_US/lessac/medium" },
+];
+
+async function download(url, target) {
+  const response = await fetch(url, { redirect: "follow" });
+  if (!response.ok || !response.body) throw new Error(`Download failed (${response.status}): ${url}`);
+  const partial = `${target}.part`;
+  await pipeline(Readable.fromWeb(response.body), fs.createWriteStream(partial));
+  fs.renameSync(partial, target);
+}
+
+async function installPiperBinary({ projectDir, log }) {
+  const archive = PIPER_ARCHIVES[`${process.platform}-${process.arch}`];
+  if (!archive) throw new Error(`No standalone Piper release for ${process.platform}-${process.arch}`);
+  const tmpFile = path.join(os.tmpdir(), `racecore-${archive}`);
+  log(`Downloading Piper (${archive})...`);
+  await download(`${PIPER_RELEASE_BASE}/${archive}`, tmpFile);
+  log("Unpacking Piper...");
+  // tar ships with Linux, macOS and Windows 10+, and the Windows one also unpacks .zip.
+  const result = spawnSync("tar", ["-xf", tmpFile, "-C", projectDir], { encoding: "utf8", windowsHide: true });
+  fs.rmSync(tmpFile, { force: true });
+  if (result.status !== 0) throw new Error(`Unpacking failed: ${result.stderr || result.error?.message || "tar error"}`);
+}
+
+async function installVoices({ voicesDir, log }) {
+  fs.mkdirSync(voicesDir, { recursive: true });
+  for (const voice of VOICES) {
+    for (const file of [`${voice.name}.onnx.json`, `${voice.name}.onnx`]) {
+      const target = path.join(voicesDir, file);
+      if (fs.existsSync(target)) continue;
+      log(`Downloading voice ${file}...`);
+      await download(`${VOICE_BASE}/${voice.path}/${file}`, target);
+    }
+  }
+}
+
+// Skips anything already present, so re-running is cheap and safe.
+async function installAll({ projectDir, hasBinary, log = () => {} }) {
+  if (!hasBinary) await installPiperBinary({ projectDir, log });
+  else log("Piper already installed.");
+  await installVoices({ voicesDir: path.join(projectDir, "piper", "voices"), log });
+  log("Done.");
+}
+
+module.exports = { installAll, installVoices, installPiperBinary, VOICES, PIPER_ARCHIVES };