fix(worker): smarter Soulseek peer selection + stall/progress-aware transfers

Draining the queue via Soulseek was stalling: slow peers (~100KB/s) never finished a
full FLAC album within the fixed 10-min timeout, got cancelled at ~89%, and the
fall-through restarted from scratch on the next peer — while single-song folders named
like the album were also being tried.

- Stall detection: abandon a peer with no byte progress for ~90s (queued/dead) instead
  of waiting the full timeout; a peer that IS progressing keeps its slot up to a 60-min
  backstop, so a slow-but-working transfer can finish.
- Rank search candidates by peer: free upload slot, then speed, then shortest queue.
- Drop candidates with far fewer tracks than the release (single-song folders) up front,
  so no download attempt is wasted on them.

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

worker/lyra_worker/pipeline.py

@@ -395,6 +395,12 @@ def run_pipeline(
     for adapter, results in per_adapter:
         for c in results:
             found.append(replace(c, source_tier=adapter.tier, confidence=score_confidence(target, c)))
+    # Drop candidates that can't be the full album — a result with far fewer tracks than the release
+    # has (e.g. a single-song Soulseek folder named like the album). Same threshold as the
+    # completeness 'too small' guard, applied up front so we never waste a download attempt on it.
+    # Unknown counts (0) and legitimately smaller editions (e.g. 11 vs 12) are kept.
+    if target.track_count:
+        found = [c for c in found if c.track_count == 0 or c.track_count * 2 > target.track_count]
     _persist_candidates(conn, job_id, found)
 
     # Qobuz pacing: when the budget gate is closed (off-hours / daily cap / spacing), drop Qobuz