Commit graph

2 commits

Author SHA1 Message Date
ed2195d879 feat(meet-kevin): throttle inter-video LLM calls (30s) to stay under Anthropic RPM
First production run hit Anthropic's per-account rate_limit_error (429) trying
to burn through 16 backfill videos in seconds. The SDK's built-in retry can't
recover because the rate limit window resets slower than the 3 retry attempts.

Added meet_kevin_inter_video_sleep_seconds (default 30s) to PipelineDeps and
main's _process_pending_videos loop. 16 backfill videos now take ~8 min (16 * 30s
sleeps + ~30s per LLM call) instead of bursting into the rate limit.
2026-05-22 20:25:19 +00:00
8f5ee8f1c3 feat(meet-kevin): pipeline orchestrator + service main loop
Implements Task 8 of the Meet Kevin revival plan.

- pipeline.py: PipelineDeps dataclass (frozen, DI-friendly), process_one_video
  state machine (discovered→captioned→analyzed with retry/cost-cap logic),
  and daily_cost_used() SQL helper.
- main.py: async run() entry point with RSS poll loop, per-video pipeline
  processing, OTEL counters, SIGTERM/SIGINT shutdown, httpx client lifecycle,
  and clean Anthropic/DB teardown.
- tests: 5 pipeline unit tests (happy path, no captions, cost cap, retry
  increment, failed-after-3-retries) all passing; full watcher suite 56/56.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-21 19:48:43 +00:00