fix(worker): run streamrip on one persistent event loop (fixes Qobuz download failures)

Each search/download called asyncio.run, creating and closing a fresh event loop.
streamrip's module-global download-concurrency Semaphore binds to the loop on first
use, so the 2nd+ download reused a Semaphore bound to a closed loop -> "Semaphore is
bound to a different event loop", tracks failed, album arrived short -> "incomplete
download" / needs_attention. Run all streamrip coroutines on one persistent loop.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

worker/lyra_worker/adapters/_streamrip.py

@@ -11,6 +11,21 @@ def _hashed_password(password: str) -> str:
 
 _AUDIO_EXT = {".flac", ".mp3", ".m4a", ".opus", ".ogg", ".aac", ".wav"}
 
+# streamrip holds module-global asyncio primitives (a download-concurrency Semaphore) that
+# bind to the running event loop on first use. `asyncio.run` creates a NEW loop per call and
+# closes it, orphaning that Semaphore, so the 2nd+ download fails with "Semaphore ... is bound
+# to a different event loop" and its tracks silently fail (-> incomplete download). Running every
+# streamrip coroutine on ONE persistent loop keeps those globals valid across downloads. The
+# worker processes one job at a time, so a single shared loop is safe.
+_loop: "asyncio.AbstractEventLoop | None" = None
+
+
+def _run(coro):
+    global _loop
+    if _loop is None or _loop.is_closed():
+        _loop = asyncio.new_event_loop()
+    return _loop.run_until_complete(coro)
+
 
 def _flatten_audio(dest: str) -> int:
     """streamrip writes into a nested `Artist - Album [...]/` subfolder. Move any audio
@@ -77,7 +92,7 @@ class StreamripClient:
         return cfg
 
     def search_album(self, artist: str, album: str) -> list[dict]:
-        return asyncio.run(self._search(f"{artist} {album}"))
+        return _run(self._search(f"{artist} {album}"))
 
     async def _search(self, query: str) -> list[dict]:
         from streamrip.client import QobuzClient
@@ -107,7 +122,7 @@ class StreamripClient:
         return results
 
     def download(self, source_ref: str, dest: str, on_progress: Callable[[float], None]) -> dict:
-        asyncio.run(self._download(source_ref, dest, on_progress))
+        _run(self._download(source_ref, dest, on_progress))
         # Count the files actually written and flatten streamrip's nested output into dest
         # so the tag stage can organize it. A partial download -> short count -> needs_attention.
         return {"track_count": _flatten_audio(dest), "path": dest}