- lib/bio.js: refreshBio() pulls the article intro via the Wikipedia API
(exintro, descriptive UA) and upserts the bio cache; getBio() reads it
with a short accurate static fallback. Page loads NEVER hit Wikipedia.
- crawl route: refreshBio() runs on every 15-min tick (over daily) —
best-effort, a wiki hiccup can't fail the news crawl.
- /api/bio: read-only route.
- page.js: WhoBox renders the live text + 'Full article on Wikipedia' link
+ SOURCE: WIKIPEDIA · UPDATED <date> line.
- tests/bio-test.mjs: fallback + live-refresh + offline degradation.
Forked from Forthena (PERSONAL/forthena) and retargeted:
- NO login, NO LLM: the edition is public and the ranking is deterministic
(stories that name Sen. BAM lead; freshest first) — Daniel: no
like/dislike or store profile for visitors.
- Crawler scoped to BAM: Bing news keyword queries + Inquirer/Rappler
fallback wires, with a two-tier relevance gate (named-BAM always;
Aquino+legislative-context only from fallback) so op-eds and non-Bam
senators are dropped.
- Desks: BAM (lead), Gaming, Education, Politics, Energy, PH, Business.
- Community board /board: 'Thank you Senator BAM' — nickname-only posts,
HTML-stripped, URL-only rejected, IP rate-limited (3/10min, 20/day).
- Yellow-gamer theme (Orbitron/Rajdhani/JetBrains Mono, dark + gold,
HUD corners), supporter disclaimer in header and footer.
- Tests: category/date/relevance/text suites + Mongo board suite +
full HTTP e2e (all passing).
Infra: 66.181.46.132 (bidaangmamamayan.com), local Mongo 7.0,
nginx + certbot TLS, PM2 bama (:3000) + 15-min crawl cron.
