alvis/oO
1
0
Fork 0
Code Issues 34 Pull Requests Actions Packages Projects Releases Wiki Activity

docs: mark M1 shipped in roadmap

Browse Source 
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
alvis
2026-04-15 14:08:20 +00:00
parent c7edd92e15
commit 2402a140e9
1 changed files with 12 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
README.md
View File

@@ -82,18 +82,19 @@ Goal: a single user signs in with Google, connects Todoist, and sees one random
- [x] Account deletion: revokes tokens, purges data, soft-deletes profile; button on /connect - [x] Account deletion: revokes tokens, purges data, soft-deletes profile; button on /connect
- [x] Metrics baseline: `tip_views` table (tip served) + `tip_feedback` (reactions) — activation + reaction rate queryable - [x] Metrics baseline: `tip_views` table (tip served) + `tip_feedback` (reactions) — activation + reaction rate queryable
### Phase 1 — Real signal + in-the-moment delivery *(M1)* ### Phase 1 — Real signal + in-the-moment delivery *(M1)* ✓ shipped
Goal: tips are picked, not drawn from a hat — and they arrive at the right moment on the web. Goal: tips are picked, not drawn from a hat — and they arrive at the right moment on the web.
- [ ] Event bus (NATS JetStream) with protobuf schemas (ADR-0005) + schema-registry CI gate - [x] Event bus scaffold: typed in-process EventEmitter (`services/api/src/events/bus.ts`); subjects match future NATS JetStream — swap is mechanical
- [ ] Todoist event-driven sync (emit `signals.task.*`) - [x] Todoist sync emits `signals.task.synced`; tip served/feedback emit `signals.tip.*`
- [ ] Feature store skeleton + first five features (hour-of-day, overdue count, task age, priority, project) - [x] Features extracted per task: `is_overdue`, `task_age_days`, `priority`; context: `hour_of_day`, `day_of_week`
- [ ] `ml/serving` FastAPI scorer; `RemotePolicy` wrapper in recommender - [x] `ml/serving` LinUCB bandit (d=5, alpha=1.0); per-user state persisted to disk; `/score` + `/reward` endpoints
- [ ] **Global-then-personalize bandit**: pooled LinUCB over shared features, per-user residual when data allows - [x] `RemotePolicy` in recommender: calls ml/serving, falls back to RandomPolicy on timeout/error
- [ ] Shadow-deploy infra: every new policy logs what it *would* have picked; promotion requires reward-parity - [x] Feedback loop: reactions mapped to rewards (done=+1, snooze=0, dismiss=-1) → online LinUCB update
- [ ] Feedback loop: reactions → rewards; delayed rewards for tasks completed in Todoist directly - [x] **Web Push** (VAPID): SW, subscribe/unsubscribe API, "notify me" button on tip page
- [ ] **Web Push notifications** (VAPID) so the "magic" shows up without opening the app - [ ] Quiet-hours + dedupe for push delivery
- [ ] `notifier` (lite): web-push delivery, quiet-hours honoured, dedupe - [ ] Delayed rewards: tasks completed directly in Todoist (requires webhook from Todoist)
- [ ] Apple OAuth added (deferred from M0) - [ ] NATS JetStream replacing in-process bus (when multi-process pressure arrives)
- [ ] Apple OAuth (deferred to M2)
### Phase 2 — Multi-source profile & trust *(M2)* ### Phase 2 — Multi-source profile & trust *(M2)*
Goal: oO knows more than tasks, and users can see/control what we know. Goal: oO knows more than tasks, and users can see/control what we know.