Execute Spotify/Apple jobs in parallel

[mre]

• Moved from sequential to parallel job processing using multiprocessing.Pool
• Grouped jobs by source to prevent rate limiting (same source runs sequentially,
  different sources run in parallel)
• Fixed multiprocessing pickling issues by moving worker functions from main.py
  to separate worker.py module
• Introduced a PodcastJob dataclass for workers
• Added if name == 'main': guard for proper multiprocessing behavior.
• Each worker process gets its own database connection to avoid shared state issues