Public docs center: `/docs/intro` on the main site (bilingual product documentation, including intraday analysis, calibrated probability, model stack, TAF, settlement sources, history, and extension).
[](https://star-history.com/#yangyuan-zhen/PolyWeather&Date)
- Realtime terminal live: visible city charts subscribe through `/api/events?cities=...&since_revision=...`, receive `city_observation_patch.v1` SSE patches, and replay short gaps from Redis Stream in production or SQLite fallback in local/single-node mode.
- Chart refresh is observation-driven: live patches merge into the current chart without a loading overlay; only visible charts run a 60s no-patch fallback, and returning from a background browser tab triggers a foreground catch-up refresh.
- Temperature charts default to All Day, keep an optional Peak window derived from the DEB hourly path, and render all timestamps in the selected city's local time.
- The chart core has been split into focused logic/canvas/state modules; Recharts now receives explicit measured dimensions to avoid 0x0 rendering and disappearing curves.
- DEB hourly consensus (`deb_hourly_consensus.v1`) is now the preferred hourly forecast path for peak-window detection and chart overlays; DEB remains a forecast curve, never an observation source.
- Legacy Gaussian probability is rendered as horizontal probability bands and a `mu` reference line on the chart, rather than as a fake time-series curve.
- Settlement runway curves are visible by default for AMSC/AMOS cities; the configured settlement runway is highlighted and auxiliary runways are shown as secondary context.
- Hong Kong uses CoWIN station `6087` (Po Leung Kuk Choi Kai Yau School) as the 1-minute reference-station curve, with HKO 10-minute observations kept as the official meteorological layer.
- Telegram airport/runway pushes are bilingual by default and use settlement-endpoint runway temperatures for slope/current/summary copy.
- Runtime state, cache, and core offline training/backfill flows now use SQLite as the primary path; legacy JSON/JSONL files remain only for migration, export, and explicit fallback input.
- EMOS/CRPS calibration is wired and trainable, but production should stay on `legacy` or `emos_shadow`; `emos_primary` is only for candidates that pass local offline evaluation and manual rollout.
- Intraday analysis is now positioned as a professional meteorology read: headline, confidence, base/upside/downside paths, next observation point, evidence chain, failure modes, and confirmation rules.
- Intraday modal now blocks stale cached detail during refresh, so users do not briefly trade off old city/date data before full detail arrives.
- Terminal city cards now combine settlement observations, DEB hourly consensus, model cluster context, calibrated probability, and market-bucket mapping without blocking the chart on AI text generation.
- Terminal data uses page memory cache, browser `localStorage`, backend short-TTL cache, SSE patch replay, and foreground refresh so returning from another tab restores the latest visible chart state quickly.
- Market bucket matching now uses the full `all_buckets` surface and strict exact / range / or-higher / or-lower direction checks, reducing bad matches to unreasonable tail buckets.
- The card label “model-market difference” means `model probability - market-implied probability`; positive values indicate weather probability above market pricing, while negative values indicate the YES is already priced more fully.
- Calibrated model probability is now the primary probability panel. It shows the active production probability engine (legacy Gaussian or EMOS), while model consensus remains a secondary reference.
- Frontend design system overhauled: unified CSS token system, eliminated `!important` abuse (134→49 in light theme), consolidated breakpoints (18→10), migrated hardcoded colors to CSS variables, added ARIA attributes and focus-visible keyboard navigation. See `docs/frontend-ui-design-review.md` for the full audit trail.
- Public in repo: weather aggregation, core analysis, dashboard, bot baseline, and standard payment flow.
- Not included in this repository: private production data, internal operating thresholds, commercial risk rules, pricing strategy details, and growth tooling.
- Trademark, brand, domain, production databases, and hosted-service operations are **not** granted by the code license.
- Adds official nearby-network and runway-level enhancement layers for China, Japan, Korea (AMOS runway sensors for Seoul/Busan), Hong Kong, Taiwan, and Turkey without replacing airport settlement anchors.
- APAC: Seoul, Busan, Hong Kong, Lau Fau Shan, Taipei, Shanghai, Beijing, Qingdao, Wuhan, Chengdu, Chongqing, Shenzhen, Guangzhou, Singapore, Tokyo, Kuala Lumpur, Jakarta, Manila, Wellington
- Americas: Toronto, New York, Los Angeles, San Francisco, Aurora, Austin, Houston, Chicago, Dallas, Miami, Atlanta, Seattle, Mexico City, Buenos Aires, Sao Paulo, Panama City
- Airport-linked contracts use the METAR / airport primary observing site as the settlement anchor. Wunderground pages are reference/history pages, not stations.
- Taipei and Shenzhen retain their explicitly configured station history pages for reconciliation, but the docs avoid describing Wunderground itself as a physical station.
- Calibrated probability uses legacy Gaussian (default) or EMOS/CRPS when evaluated; model vote counts remain an explanatory consensus line, not the final probability.
Do not run full EMOS retraining on a small VPS. The VPS should collect data and load approved calibration files; training should run on a local/dev machine using a copied production SQLite database:
Promote a generated `default.json` only when `auto_retrain_report.json` has `ready_for_promotion=true`, and prefer `emos_shadow` before enabling `emos_primary`.
Production payment routes are configured by the backend. Polygon remains the default checkout-contract chain, while Ethereum mainnet USDC can be enabled as a direct-transfer route so users who pay on their wallet default network are still confirmed by `intent.chain_id`.
