picks the most DOWNLOADABLE release, not just the biggest/fastest:
- peer_availability() mirrors the music side's scoring — free upload slot, tiered
upload speed, graduated queue penalty.
- group_video_files now captures queueLength, picks the best PEER per release by
(availability, then speed) instead of speed alone, and ranks hits by availability.
- _evaluate_hits sorts by (accepted, quality score, availability, size) so within a
quality tier a free-slot/empty-queue source wins over one stuck behind a 1500-deep
queue — fixes the downloads that sat at 0%.
queue/speed also surfaced on results for the UI. tested.
the real cause of the movie-search failures: slskd rate-limits search CREATION and
429s when exceeded — and the video path had NO throttle (the music side caps ~35/220s).
the thread-pool auto-grab fired searches in bursts (and each 429 returned instantly,
cascading into more), storming slskd into 429s; only a few slipped through.
add a shared throttle on start_search: min 2s between creations + a 35-per-220s window
cap (mirrors music) + a cooldown when a 429 IS hit (honors Retry-After) so it backs off
instead of hammering. pure reserve/cooldown logic tested.
The Soulseek ⌕ now actually queries your slskd instance instead of fabricating
results. core/video/slskd_search.py (isolated): build_query(scope,…) → POST
/api/v0/searches (shared soulseek.* URL+key) → poll /responses (bounded ~8s,
early-exit at 25) → keep video files → group_video_files() folds them into one hit
per release folder with a peer count + the fastest available source. Same
{title,size_bytes,…} shape, so parse→evaluate→rank/cards are unchanged.
Endpoint routes source=='soulseek' to slskd (torrent/usenet stay mocked), surfaces
'slskd not configured' / network errors, carries username/peers/slots/filename
through for the cards + a future real Grab, and caps to 40. Pure helpers tested (4);
isolation guard + ruff clean.
