2.2 KiB
TODOS
Post-Faza 1 backlog. Priority tags: P1 (next sprint), P2 (after baseline proven), P3 (nice-to-have).
P2-ack-loop — Discord reaction feedback
Deferred from Faza 1. Operator reacts to alert with emoji (👍 = good call, 👎 = false positive, ⏭️ = phase-skip). Bot reads reactions via Discord Bot API (not webhook), appends to labels_live.json. Feeds precision tuning without manual labeling.
- Requires upgrading Discord integration from webhook to bot (token + gateway).
- Ack loop runs as separate async task; not in hot detection path.
- Start after 5+ sessions of stable Faza 1 baseline.
P2-screenshot-dump
Standalone helper atm dump --duration 4h --dir samples/ that just captures frames every N seconds to a directory without running detection. Used during early sessions to accumulate labeled corpus.
- Small module
src/atm/dumper.pyinvoked via main.py. - Separate from
runso operator can capture raw data even before calibration is dialed in.
P2-windows-scheduler-docs
Expand README with concrete Windows Task Scheduler XML export + import steps, example triggers for 16:30→18:30 and 21:00→23:00, and DST-change checklist reminder.
P3-faza2-exec
Auto-execution on TradeLocker. Blocked on TOS audit (see docs/phase2-prop-firm-audit.md). Not started until GO decision + 20+ Faza 1 sessions.
P3-multi-symbol
Monitor DIA and GLD concurrently (two separate charts). Current scope: one at a time, operator switches. Multi-symbol needs per-symbol config, separate state machines, and separate notification threads.
P3-web-dashboard
Read-only web view of today's audit JSONL + recent triggers. Useful for review after a session but not essential. Would share audit.py log format.
Quality debt
- Integration test for run_live loop: currently mocked at module level. Add a short-duration in-memory loop test that threads real detector/state_machine/audit together (no network).
- Coverage report: run
pytest --cov=atm --cov-report=term-missing, aim for ≥ 85% per module. - Typing strictness: run
pyright src/with strict mode, fix reported issues. - Perf baseline: profile one detection cycle on a representative frame; ensure < 100ms so 5s loop has ample headroom.