Parallelize Qobuz playlist track fetching by increasing the page size to 500 and processing remaining pages concurrently via `futures::try_join_all` with a configurable semaphore (default 3). Results are offset-sorted to preserve original order. Adds a `page_concurrency` configuration option, updates the API client initialization, and introduces the `futures` dependency. This reduces large playlist latency from ~1.6s to ~0.7s.
Replace the hardcoded semaphore capacity of 16 with a configurable `register_concurrency` setting (defaulting to 4). This mitigates SQLite write contention and optimizes concurrent API and network requests during parallel track caching.
Refactor Qobuz bundle fetching to improve resilience and performance. Add HTTP timeout and retry configuration, and implement a retry loop that returns version metadata. Introduce bundle version caching and persistence to skip redundant extractions when the bundle remains unchanged. Replace console logging with structured tracing macros and update the architectural tracking document to mark these improvements as complete.
