v1.8.0

Changelog

1.8.0

✨ New Features

• Lidarr download-client bridge — the app can now be configured inside Lidarr as a native Newznab indexer + SABnzbd download client, so Lidarr searches, grabs and imports automatically. Includes retry-cooldown protection against infinite re-grab loops, and the job registry is persisted to SQLite so downloads survive a restart.
• Automatic YouTube PO tokens via a bundled bgutil provider sidecar (helps with "Sign in to confirm you're not a bot" / format-unavailable), plus an optional manual po_token field and a Test button for the provider URL.
• Loudness normalization (EBU R128, −14 LUFS) as an opt-in download option.
• "Unban All" button on the logs page (URL Banned filter).
• Health endpoint /api/health + Docker HEALTHCHECK.
• GitHub Actions CI running pyflakes + pytest on every push/PR.

🧹 Improved

• YouTube matching now prefers the audio/Topic upload over a music video.
• AcoustID verification accepts a high-score recording from the expected release group when the exact recording id isn't matched (fewer false rejections of the right song).
• Clearer errors when DOWNLOAD_PATH is unset or the Lidarr library path isn't mounted/writable.

🐛 Bug Fixes

• Tagging crash when Lidarr returns a null releaseDate/trackCount.
• V5→V6 DB migration made idempotent (could block startup on some DBs).
• SSE progress stream no longer holds the global lock across blocking Lidarr HTTP calls, and snapshots track state safely.
• Guards against malformed/missing JSON and Lidarr error responses on the queue-reorder, album-details and add-to-queue endpoints.
• SQLite connection leak in the background sync worker.
• TOCTOU race when starting manual/playlist downloads.
• Numeric config values from config.json are coerced with safe fallbacks.
• Download-client "Busy" result re-queues instead of failing the grab.