fix(news): one-hour World News API retention, refunded reads, pins, disputed areas, counts

Keep World News API results an hour at most and say caching needs its
permission; refund a read that never reached it; match Local by any pin;
assign a disputed area by a point inside it; untitled GDELT stories are
minor; hotspot counts leave out minor reports.

Co-Authored-By: Claude Fable 5.2 <[email protected]>

DATA_SOURCES.md

@@ -43,7 +43,7 @@ How to read this:
 | **News server** (`GEV_NEWS_ORIGIN`, optional; none by default) | News layer, both sections: the last 24 h of reports from many publishers. Set, the local server forwards the layer's `/api/news/hotspots`, `/feed` and `/outline` GET requests to it and Core reads none of the three sources below. Which providers stand behind its feed is the news server's to state | Each report stays its publisher's: headlines and summaries link to the original article, and pictures are hotlinked from the publisher's site (no referrer), never stored | Per report: publisher name and article link on every card; picture cards name the picture's publisher. The Data attribution popover states that places, categories, severity and English headlines are derived automatically and can be wrong |
 | **GDELT 2.0 events and GKG** (`data.gdeltproject.org/gdeltv2/`; no key; used only without a news server) | News layer, Conflicts: the last 24 h of GDELT's 15-minute event exports (`*.export.CSV.zip`, about 100 KB each), read server-side as they appear. Physical violence only (CAMEO roots 18–20); one report per story, copies of one agency story merged; placed at GDELT's location for it (a town the article's words name beats a bare country). Titled by the article's page title from the GKG export of the same quarter hour (`*.gkg.csv.zip`, about 6 MB each) for the newest 3 h, else by the words of the article's address. GDELT's actor coding is never shown. Categories, relevance and severity by bundled rules | [GDELT Terms of Use](https://www.gdeltproject.org/about.html#termsofuse): unrestricted academic, commercial and governmental use, with citation and a link to the GDELT Project. Coded events, page titles and article links only; linked articles keep their publishers' terms | "GDELT Project" with a link to gdeltproject.org in the Data attribution popover; "via GDELT Project", the publisher's domain and the article link on every card |
 | **Liveuamap** (`a.liveuamap.com/api`, `LIVEUAMAP_API_KEY`, optional, paid; used only without a news server) | News layer, Conflicts, beside GDELT: Liveuamap's worldwide mapped items of the last 24 h (`a=mpts`), read server-side every 5 minutes; categories and severity are read from the headline by bundled rules | Liveuamap API terms of the operator's own subscription. Headlines, places and links only; item pictures and clips are hotlinked from Liveuamap, never stored | "Liveuamap" with a link in the Data attribution popover once its items show; "Liveuamap" and the item's Liveuamap link on every card |
-| **World News API** (`api.worldnewsapi.com/search-news`, `WORLDNEWSAPI_KEY`, optional; used only without a news server) | News layer, Local: the last 24 h of news by the source country at the view centre (country from the bundled Natural Earth admin-0 pack; none in a disputed area), read server-side once per country every 30 minutes, about 1.5 points a read; at most `WORLDNEWSAPI_DAILY_READS` reads a day (default 30), after which Local says the daily limit is reached. Each story keeps World News API's own category (politics, business, else other) with no severity of its own; sport, entertainment, lifestyle and travel left out | [World News API terms](https://worldnewsapi.com/terms/) of the operator's own plan; the free plan (50 points a day) requires a visible backlink to worldnewsapi.com in the app, paid plans do not. Titles and links only; article text and summaries are not kept or shown; pictures are hotlinked from the publisher | The "World News API" link to worldnewsapi.com, always in the Data attribution popover (the free plan's required backlink); "via World News API", the publisher's domain and the article link on every card |
+| **World News API** (`api.worldnewsapi.com/search-news`, `WORLDNEWSAPI_KEY`, optional; used only without a news server) | News layer, Local: the last 24 h of news by the source country at the view centre (country from the bundled Natural Earth admin-0 pack; none in a disputed area), read server-side once per country every 30 minutes, about 1.5 points a read, kept in memory for at most 1 hour and then deleted (a read that fails is served from that copy only within the hour); at most `WORLDNEWSAPI_DAILY_READS` reads a day (default 30), after which Local says the daily limit is reached. Each story keeps World News API's own category (politics, business, else other) with no severity of its own; sport, entertainment, lifestyle and travel left out | [World News API terms](https://worldnewsapi.com/terms/) of the operator's own plan. The terms allow caching only with World News API's prior written permission, and then for at most 1 hour: getting that permission before setting the key is the operator's responsibility. Titles and links only; article text and summaries are not kept or shown; pictures are hotlinked from the publisher | The "World News API" link to worldnewsapi.com, always in the Data attribution popover (a credit; the terms require a backlink only of invited free plans); "via World News API", the publisher's domain and the article link on every card |
 | **InciWeb** (inciweb.wildfire.gov)                                    | Per-incident "InciWeb ↗" card links to official incident pages (narratives, evacuation/closure notices) | U.S. government public incident information. Links are matched best-effort by incident name and state, with incident page origin and update times checked on selection. Same-origin `/api/fire-perimeters/inciweb` routes cache the catalog for 1 hour and publication metadata for 30 minutes; linked pages remain InciWeb content | "Incident information: InciWeb (inciweb.wildfire.gov)" |
 | **FAA Special Use Airspace** (`services6.arcgis.com` FeatureServer `Special_Use_Airspace`) | Prohibited & Restricted Airspace layer: US prohibited (P) and restricted (R) special-use airspace polygons (not the UAS Facility Map or UAS flight restrictions), altitude limits from the `*_VAL`/`*_UOM`/`*_CODE` fields, times of use, and controlling agency | U.S. public domain (FAA aeronautical data published through the [FAA AIS Open Data portal](https://adds-faa.opendata.arcgis.com/)); fetched keyless through `/api/drone-airspace` (1-hour server cache). Informational only — not a substitute for official UAS flight-planning tools | "Prohibited & restricted airspace: FAA Special Use Airspace" |
 | **iNaturalist API** (`api.inaturalist.org/v2/observations`, `/v1/grid/{z}/{x}/{y}.grid.json`) | Wildlife Sightings layer: verifiable wildlife observed within the last 48 h (whales 7 d), by observation time, in the current view, with taxon, observation time, quality grade and eligible photo | Per-record licences; observations default CC BY-NC. Photos CC0/BY/BY-SA/BY-NC/BY-NC-SA shown; ND and all-rights-reserved photos link-only. Keyless via `/api/inaturalist/observations` (5-min cache; global sample 10 min; stale ≤ 1 h; 60 requests/min and 8,000/day upstream caps). **NonCommercial content:** commercial redistributors must filter to CC0/BY/BY-SA or remove the layer. No AI training on this content | "Wildlife sightings: iNaturalist contributors (licences per record; photos credited on each card)" |