fix(cli): improve TTS and model download reliability

packages/cli/src/tts/manager.ts

@@ -1,8 +1,47 @@
-import { existsSync, mkdirSync } from "node:fs";
+import { existsSync, mkdirSync, writeFileSync, unlinkSync, readFileSync, statSync } from "node:fs";
 import { homedir } from "node:os";
 import { join } from "node:path";
 import { downloadFile } from "../utils/download.js";
 
+const LOCK_STALE_MS = 300_000; // 5 min — generous to cover large model downloads
+
+function acquireLock(lockPath: string): boolean {
+  try {
+    writeFileSync(lockPath, String(process.pid), { flag: "wx" });
+    return true;
+  } catch {
+    // Lock exists — check if it's stale
+    try {
+      const stat = statSync(lockPath);
+      if (Date.now() - stat.mtimeMs > LOCK_STALE_MS) {
+        unlinkSync(lockPath);
+        writeFileSync(lockPath, String(process.pid), { flag: "wx" });
+        return true;
+      }
+    } catch {
+      // Race — another process grabbed it
+    }
+    return false;
+  }
+}
+
+function releaseLock(lockPath: string): void {
+  try {
+    const content = readFileSync(lockPath, "utf-8").trim();
+    if (content === String(process.pid)) unlinkSync(lockPath);
+  } catch {
+    // Already cleaned up
+  }
+}
+
+async function waitForLock(lockPath: string, timeoutMs: number = 300_000): Promise<void> {
+  const deadline = Date.now() + timeoutMs;
+  while (Date.now() < deadline) {
+    if (!existsSync(lockPath)) return;
+    await new Promise((r) => setTimeout(r, 1_000));
+  }
+}
+
 const CACHE_DIR = join(homedir(), ".cache", "hyperframes", "tts");
 const MODELS_DIR = join(CACHE_DIR, "models");
 const VOICES_DIR = join(CACHE_DIR, "voices");
@@ -114,8 +153,20 @@
   }
 
   mkdirSync(MODELS_DIR, { recursive: true });
-  options?.onProgress?.(`Downloading TTS model ${model} (~311 MB)...`);
-  await downloadFile(url, modelPath);
+  const lockPath = `${modelPath}.lock`;
+  if (!acquireLock(lockPath)) {
+    options?.onProgress?.("Another process is downloading the model, waiting...");
+    await waitForLock(lockPath);
+    if (existsSync(modelPath)) return modelPath;
+  }
+
+  try {
+    if (existsSync(modelPath)) return modelPath;
+    options?.onProgress?.(`Downloading TTS model ${model} (~311 MB)...`);
+    await downloadFile(url, modelPath);
+  } finally {
+    releaseLock(lockPath);
+  }
 
   if (!existsSync(modelPath)) {
     throw new Error(`Model download failed: ${model}`);
@@ -135,8 +186,20 @@
   if (existsSync(voicesPath)) return voicesPath;
 
   mkdirSync(VOICES_DIR, { recursive: true });
-  options?.onProgress?.("Downloading voice data (~27 MB)...");
-  await downloadFile(VOICES_URL, voicesPath);
+  const lockPath = `${voicesPath}.lock`;
+  if (!acquireLock(lockPath)) {
+    options?.onProgress?.("Another process is downloading voice data, waiting...");
+    await waitForLock(lockPath);
+    if (existsSync(voicesPath)) return voicesPath;
+  }
+
+  try {
+    if (existsSync(voicesPath)) return voicesPath;
+    options?.onProgress?.("Downloading voice data (~27 MB)...");
+    await downloadFile(VOICES_URL, voicesPath);
+  } finally {
+    releaseLock(lockPath);
+  }
 
   if (!existsSync(voicesPath)) {
     throw new Error("Voice data download failed");

packages/cli/src/tts/synthesize.ts

@@ -48,7 +48,7 @@ function hasPythonPackage(python: string, pkg: string): boolean {
   try {
     execFileSync(python, ["-c", `import ${pkg}`], {
       stdio: ["pipe", "pipe", "pipe"],
-      timeout: 10_000,
+      timeout: 30_000, // ONNX runtime import can take 10+ seconds on cold start
     });
     return true;
   } catch {
@@ -203,6 +203,7 @@ export async function synthesize(
       {
         encoding: "utf-8",
         timeout: 300_000,
+        maxBuffer: 10 * 1024 * 1024, // 10 MB — ONNX runtime prints verbose warnings
         stdio: ["pipe", "pipe", "pipe"],
       },
     );

packages/cli/src/utils/download.ts

@@ -1,28 +1,52 @@
-import { createWriteStream, renameSync, unlinkSync } from "node:fs";
+import { createWriteStream, renameSync, unlinkSync, existsSync } from "node:fs";
 import { get as httpsGet } from "node:https";
 import { pipeline } from "node:stream/promises";
 
-/**
- * Download a file from a URL, following redirects.
- * Uses atomic write (download to .tmp, rename on success) to prevent
- * corrupt partial files from persisting in the cache on interruption.
- */
-export function downloadFile(url: string, dest: string): Promise<void> {
+const SOCKET_TIMEOUT_MS = 30_000;
+const RESPONSE_TIMEOUT_MS = 60_000;
+const MAX_RETRIES = 3;
+const RETRY_BASE_MS = 1_000;
+const MAX_REDIRECTS = 10;
+
+function attempt(url: string, dest: string): Promise<void> {
   const tmp = `${dest}.tmp`;
   return new Promise((resolve, reject) => {
+    let redirects = 0;
+
     const follow = (u: string) => {
-      httpsGet(u, (res) => {
+      if (++redirects > MAX_REDIRECTS) {
+        reject(new Error("Download failed: too many redirects"));
+        return;
+      }
+      const req = httpsGet(u, (res) => {
         if (res.statusCode === 301 || res.statusCode === 302) {
           const location = res.headers.location;
           if (location) {
+            res.resume();
             follow(location);
             return;
           }
         }
+        if (res.statusCode === 403 || res.statusCode === 429) {
+          res.resume();
+          reject(
+            new Error(
+              `Download failed: HTTP ${res.statusCode} (rate limited). ` +
+                `GitHub throttles unauthenticated release downloads. Retry in a moment.`,
+            ),
+          );
+          return;
+        }
         if (res.statusCode !== 200) {
+          res.resume();
           reject(new Error(`Download failed: HTTP ${res.statusCode}`));
           return;
         }
+
+        res.setTimeout(RESPONSE_TIMEOUT_MS, () => {
+          res.destroy(new Error("Download stalled: response timeout"));
+        });
+
         const file = createWriteStream(tmp);
         pipeline(res, file)
           .then(() => {
@@ -37,9 +61,13 @@ export function downloadFile(url: string, dest: string): Promise<void> {
             }
             reject(err);
           });
-      }).on("error", (err) => {
+      });
+      req.setTimeout(SOCKET_TIMEOUT_MS, () => {
+        req.destroy(new Error("Download failed: connection timeout"));
+      });
+      req.on("error", (err) => {
         try {
-          unlinkSync(tmp);
+          if (existsSync(tmp)) unlinkSync(tmp);
         } catch {
           // ignore cleanup failure
         }
@@ -49,3 +77,25 @@ export function downloadFile(url: string, dest: string): Promise<void> {
     follow(url);
   });
 }
+
+/**
+ * Download a file from a URL with retry, following redirects.
+ * Uses atomic write (download to .tmp, rename on success) to prevent
+ * corrupt partial files from persisting in the cache on interruption.
+ */
+export async function downloadFile(url: string, dest: string): Promise<void> {
+  let lastErr: Error | undefined;
+  for (let i = 0; i <= MAX_RETRIES; i++) {
+    try {
+      await attempt(url, dest);
+      return;
+    } catch (err) {
+      lastErr = err instanceof Error ? err : new Error(String(err));
+      if (i < MAX_RETRIES) {
+        const delay = RETRY_BASE_MS * 2 ** i;
+        await new Promise((r) => setTimeout(r, delay));
+      }
+    }
+  }
+  throw lastErr;
+}