Update backlog statuses; strengthen over-clustering issue with new evidence
Mark each item open/mitigated/done, fold completed work into Shipped, and upgrade the clustering bug with the on-device Jackson/airport over-merge proof and a concrete TF-IDF/embeddings fix. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Robin Kutesa
102
docs/BACKLOG.md
102
docs/BACKLOG.md
@@ -1,67 +1,69 @@
|
||||
# Backlog
|
||||
# Backlog & status
|
||||
|
||||
Open issues, seeding the Gitea tracker via `scripts/create-issues.sh` (run once the
|
||||
repo exists). `[client]` = iOS, `[backend]` = the platform service on the Docker host,
|
||||
`[ci]` = build/test.
|
||||
Source-of-truth tracker. The Gitea issue tracker is currently **empty (0 issues)** —
|
||||
run `scripts/create-issues.sh` with a token to populate it. `[client]` = iOS,
|
||||
`[backend]` = the platform service on the Docker host, `[ci]` = build/test.
|
||||
|
||||
Status: 🔴 open · 🟡 mitigated client-side (real fix still pending) · 🟢 done.
|
||||
|
||||
## Backend — data quality (highest priority)
|
||||
|
||||
1. **[backend] F1 stories are mis-classified into tech/finance/politics.** The topic
|
||||
classifier has no sport category, so F1 (with 100% F1 sources) scatters across topics.
|
||||
Evidence from the live feed: "Monaco Grand Prix stewards…" → `tech`, "F1 reports 35%
|
||||
CO2 reduction…" → `finance`, "McLaren launch appeal against Gasly penalty" → `politics`.
|
||||
Add a `sport` (or `f1`) topic and route motorsport sources there.
|
||||
2. **[backend] Over-clustering merges unrelated articles.** Stories pair a headline with a
|
||||
summary about something else, e.g. an Ebola/Bunia (DR Congo) headline carried a
|
||||
South-Africa summary, and a CRISPR headline carried Drakensberg/SA text. Raise the
|
||||
TF-IDF similarity threshold (~0.22 → 0.30–0.40) and make the canonical `headline` and
|
||||
`summary` come from the **same** representative article. This also inflates
|
||||
`sourceCount` / `crossSourceConfirmation` and the signal score.
|
||||
3. **[backend] Add accurate per-story tags (topic + region).** So the client can *read*
|
||||
tags (sport, uganda, east-africa, south-africa, …) instead of keyword-guessing. Unblocks
|
||||
moving the filter pills to server-side. Pairs with #1/#2.
|
||||
4. **[backend] Topic-classifier coverage for the 47 feeds.** Ensure science/world/tech map
|
||||
sensibly into the supported topics.
|
||||
1. 🟡 **[backend] F1 mis-classified into tech/finance/politics.** No sport category, so
|
||||
F1 (100% F1 sources) scatters across topics: "Monaco Grand Prix…" → `tech`,
|
||||
"F1 reports 35% CO2…" → `finance`, "McLaren launch appeal…" → `politics`. Add a
|
||||
`sport`/`f1` topic and route motorsport sources there.
|
||||
*Client mitigation shipped:* F1 pill matches clear F1 headlines + a majority of F1
|
||||
sources, and F1 is excluded from All/Tech/Finance. Backend still mis-tags.
|
||||
2. 🔴 **[backend] Catastrophic over-clustering — unrelated articles merged on a single
|
||||
shared word.** Newly confirmed on-device: a story merged a Russian-film-director
|
||||
article + "Who's afraid of **Michael Jackson**?" + "Chelsea **Jackson** Roberts"
|
||||
(wellness) — grouped on the surname *Jackson*. Another merged "Gunfire at Niger
|
||||
**airport**" + "L.A. County Whiteman **Airport** crash" + "**Niger** fintech". Result:
|
||||
wrong headline, summary, topic, and inflated signal. **Raise the TF-IDF threshold hard
|
||||
(~0.22 → 0.45–0.55) and require multiple shared distinctive terms / named-entity
|
||||
overlap, with a tight time window. If still over-merging, move to sentence embeddings
|
||||
(all-MiniLM-L6-v2, cosine ≥ ~0.6).** Make canonical headline + summary come from the
|
||||
same representative article. *No client fix is possible — the story object itself is
|
||||
wrong.*
|
||||
3. 🔴 **[backend] Add accurate per-story tags (topic + region).** So the client can read
|
||||
tags (sport, uganda, south-africa, …) instead of keyword-guessing. Unblocks #6.
|
||||
4. 🔴 **[backend] Topic-classifier coverage for the 47 feeds.** science/world/tech mapping.
|
||||
|
||||
## Backend — features
|
||||
|
||||
5. **[backend] Server-side push for breaking alerts (APNs or ntfy).** Needed for
|
||||
breaking-story notifications while the app is fully closed; the backend already computes
|
||||
the stories. Client currently does local notifications + background refresh only.
|
||||
5. 🔴 **[backend] Server-side push for breaking alerts (APNs or ntfy).** For
|
||||
breaking-story notifications while the app is closed. Client does local notifications +
|
||||
background refresh only.
|
||||
|
||||
## Client
|
||||
|
||||
6. **[client] Move interest filters server-side once tags exist (#3).** Replace the
|
||||
client-side keyword filters (F1 / Uganda / East Africa / South Africa) — which only see
|
||||
loaded stories and match the headline — with backend tag filtering for full coverage.
|
||||
7. **[client] SSID-based trusted networks.** Pin trusted Wi-Fi by name. Needs the Access
|
||||
WiFi Information entitlement (real device + paid account); no-ops in the simulator.
|
||||
Currently approximated by LAN reachability.
|
||||
8. **[client] Optional in-app WireGuard tunnel.** NetworkExtension packet-tunnel so Jarvis
|
||||
can bring up a WireGuard tunnel itself (WireGuard/Netmaker configs only). Entitlement +
|
||||
real device. Deferred.
|
||||
9. **[client] App icon + launch assets.** `AppIcon` is an empty placeholder.
|
||||
10. **[client] Flesh out Latest / Search tabs.** (Saved is now an explicit bookmarks list.)
|
||||
Add proper empty/loading states and dedicated behavior.
|
||||
11. **[client] Verify hero image loading.** `AsyncImage` against real feed `imageUrl`s
|
||||
(ATS, redirects, grey fallback) in the article reader.
|
||||
12. **[client] Offline-first polish.** Pre-cache top stories for true offline-first; today
|
||||
the green "cached" dot depends on opening an article.
|
||||
6. 🔴 **[client] Move interest filters server-side once tags (#3) exist.** Today's pills
|
||||
are client-side keyword/headline matches on loaded stories. *Refined since:* regions
|
||||
match headline-only (no polluted summaries/outlets), regions exclusive, F1 by source
|
||||
majority, East Africa removed.
|
||||
7. 🔴 **[client] SSID-based trusted networks.** Needs Access WiFi Information entitlement
|
||||
(device + paid account). Approximated by LAN reachability today.
|
||||
8. 🔴 **[client] Optional in-app WireGuard tunnel.** NetworkExtension; entitlement + device. Deferred.
|
||||
9. 🔴 **[client] App icon + launch assets.** `AppIcon` is an empty placeholder.
|
||||
10. 🔴 **[client] Flesh out Latest / Search tabs.** Saved is now bookmarks; Latest/Search
|
||||
still minimal.
|
||||
11. 🔴 **[client] Verify hero image loading.** `AsyncImage` vs real feed `imageUrl`s.
|
||||
12. 🔴 **[client] Offline-first polish.** Pre-cache top stories.
|
||||
|
||||
## CI
|
||||
|
||||
13. **[ci] Add unit/UI tests + a test matrix leg.** The build matrix compiles but runs no
|
||||
tests yet.
|
||||
14. **[ci] Backend CI build matrix.** Python 3.11–3.13 + lint + Docker image build/health
|
||||
probe, for the backend repo (template lives in chat/handoff).
|
||||
13. 🔴 **[ci] Add unit/UI tests + a test matrix leg.** Build matrix compiles, runs no tests.
|
||||
14. 🔴 **[ci] Backend CI build matrix.** Python 3.11–3.13 + lint + Docker build/health probe.
|
||||
|
||||
---
|
||||
|
||||
## Recently shipped (this iteration)
|
||||
## 🟢 Shipped
|
||||
|
||||
Signal feed + detail + reader + feed manager; LAN⇄Tailscale connectivity; notifications
|
||||
(morning/evening briefings + breaking alerts) with Open-Meteo weather; device code signing;
|
||||
feed resilience (no blank-on-refresh, ignore cancelled, coalesced WS refresh); clearer
|
||||
empty states; deterministic feed order; read/unread (drop stripe on open); 6-line summary
|
||||
preview; pinned mutually-exclusive filters (F1 split out, regions non-overlapping,
|
||||
Finance/Tech by topic); swipe actions (left = save, right = share + mark read).
|
||||
Signal feed + detail + reader + feed manager · LAN⇄Tailscale connectivity · notifications
|
||||
(morning/evening briefings + breaking alerts) with Open-Meteo weather · device code signing
|
||||
· feed resilience (no blank-on-refresh, ignore `cancelled`, coalesced WS refresh) · clearer
|
||||
empty states (polling / unreachable+host / offline / caught-up) · deterministic feed order ·
|
||||
6-line summary preview · pinned filters (F1 split out, regions exclusive & headline-matched,
|
||||
Finance/Tech by topic) · swipe actions (short→Save / long→Mark read, right→Share) ·
|
||||
read/seen handling (seen-tracking with per-load snapshot, read stories collapse into a
|
||||
"Read" shelf with compact rows, "all caught up" state).
|
||||
|
||||
@@ -25,8 +25,8 @@ mk() {
|
||||
# --- Backend: data quality ---
|
||||
mk "[backend] F1 stories mis-classified into tech/finance/politics" \
|
||||
"The topic classifier has no sport category, so F1 (100% F1 sources) scatters across topics. Live evidence: 'Monaco Grand Prix stewards…' -> tech, 'F1 reports 35% CO2 reduction…' -> finance, 'McLaren launch appeal against Gasly penalty' -> politics. Add a sport/f1 topic and route motorsport sources there."
|
||||
mk "[backend] Over-clustering merges unrelated articles" \
|
||||
"Stories pair a headline with an unrelated summary: an Ebola/Bunia (DR Congo) headline carried a South-Africa summary; a CRISPR headline carried Drakensberg/SA text. Raise the TF-IDF similarity threshold (~0.22 -> 0.30-0.40) and make canonical headline + summary come from the SAME representative article. Also inflates sourceCount/crossSourceConfirmation/score."
|
||||
mk "[backend] Catastrophic over-clustering: unrelated articles merged on one shared word" \
|
||||
"Confirmed on-device: a story merged a Russian-film-director article + 'Who is afraid of Michael Jackson?' + 'Chelsea Jackson Roberts' (wellness), grouped on the surname Jackson. Another merged 'Gunfire at Niger airport' + 'L.A. County Whiteman Airport crash' + 'Niger fintech', grouped on airport/Niger. Wrong headline/summary/topic + inflated signal. Raise TF-IDF threshold hard (~0.22 -> 0.45-0.55), require multiple shared distinctive terms / named-entity overlap + tight time window; if still over-merging move to sentence embeddings (all-MiniLM-L6-v2, cosine >= 0.6). Canonical headline + summary must come from the SAME article. No client fix possible."
|
||||
mk "[backend] Add accurate per-story tags (topic + region)" \
|
||||
"Expose accurate tags (sport, uganda, east-africa, south-africa, ...) so the client reads them instead of keyword-guessing. Unblocks server-side filter pills. Pairs with the F1 and over-clustering fixes."
|
||||
mk "[backend] Topic-classifier coverage for the 47 feeds" \
|
||||
|
||||
Reference in New Issue
Block a user