mirror of
https://github.com/koala73/worldmonitor.git
synced 2026-04-25 17:14:57 +02:00
* Add Security Advisories panel with government travel alerts (#460) * feat: add Security Advisories panel with government travel advisory feeds Adds a new panel aggregating travel/security advisories from official government foreign affairs agencies (US State Dept, AU DFAT Smartraveller, UK FCDO, NZ MFAT). Advisories are categorized by severity level (Do Not Travel, Reconsider, Caution, Normal) with filter tabs by source country. Includes summary counts, auto-refresh, and persistent caching via the existing data-freshness system. * chore: update package-lock.json * fix: event delegation, localization, and cleanup for SecurityAdvisories panel P1 fixes: - Use event delegation on this.content (bound once in constructor) instead of direct addEventListener after each innerHTML replacement — prevents memory leaks and stale listener issues on re-render - Use setContent() consistently instead of mixing with this.content.innerHTML - Add securityAdvisories translations to all 16 non-English locale files (panels name, component strings, common.all key) - Revert unrelated package-lock.json version bump P2 fixes: - Deduplicate loadSecurityAdvisories — loadIntelligenceData now calls the shared method instead of inlining duplicate fetch+set logic - Add Accept header to fetch calls for better content negotiation * feat(advisories): add US embassy alerts, CDC, ECDC, and WHO health feeds Adds 21 new advisory RSS feeds: - 13 US Embassy per-country security alerts (TH, AE, DE, UA, MX, IN, PK, CO, PL, BD, IT, DO, MM) - CDC Travel Notices - 5 ECDC feeds (epidemiological, threats, risk assessments, avian flu, publications) - 2 WHO feeds (global news, Africa emergencies) Panel gains a Health filter tab for CDC/ECDC/WHO sources. All new domains added to RSS proxy allowlist. i18n "health" key added across all 17 locales. * feat(cache): add negative-result caching to cachedFetchJson (#466) When upstream APIs return errors (HTTP 403, 429, timeout), fetchers return null. Previously null results were not cached, causing repeated request storms against broken APIs every refresh cycle. Now caches a sentinel value ('__WM_NEG__') with a short 2-minute TTL on null results. Subsequent requests within that window get null immediately without hitting upstream. Thrown errors (transient) skip sentinel caching and retry immediately. Also filters sentinels from getCachedJsonBatch pipeline reads and fixes theater posture coalescing test (expected 2 OpenSky fetches for 2 theater query regions, not 1). * feat: convert 52 API endpoints from POST to GET for edge caching (#468) * feat: convert 52 API endpoints from POST to GET for edge caching Convert all cacheable sebuf RPC endpoints to HTTP GET with query/path parameters, enabling CDN edge caching to reduce costs. Flatten nested request types (TimeRange, PaginationRequest, BoundingBox) into scalar query params. Add path params for resource lookups (GetFredSeries, GetHumanitarianSummary, GetCountryStockIndex, GetCountryIntelBrief, GetAircraftDetails). Rewrite router with hybrid static/dynamic matching for path param support. Kept as POST: SummarizeArticle, ClassifyEvent, RecordBaselineSnapshot, GetAircraftDetailsBatch, RegisterInterest. Generated with sebuf v0.9.0 (protoc-gen-ts-client, protoc-gen-ts-server). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: add rate_limited field to market response protos The rateLimited field was hand-patched into generated files on main but never declared in the proto definitions. Regenerating wiped it out, breaking the build. Now properly defined in both ListEtfFlowsResponse and ListMarketQuotesResponse protos. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: remove accidentally committed .planning files Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * feat: add Cloudflare edge caching infrastructure for api.worldmonitor.app (#471) Route web production RPC traffic through api.worldmonitor.app via fetch interceptor (installWebApiRedirect). Add default Cache-Control headers (s-maxage=300, stale-while-revalidate=60) on GET 200 responses, with no-store override for real-time endpoints (vessel snapshot). Update CORS to allow GET method. Skip Vercel bot middleware for API subdomain using hostname check (non-spoofable, replacing CF-Ray header approach). Update desktop cloud fallback to route through api.worldmonitor.app. * fix(beta): eagerly load T5-small model when beta mode is enabled BETA_MODE now couples the badge AND model loading — the summarization-beta model starts loading on startup instead of waiting for the first summarization call. * fix: move 5 path-param endpoints to query params for Vercel routing (#472) Vercel's `api/[domain]/v1/[rpc].ts` captures one dynamic segment. Path params like `/get-humanitarian-summary/SA` add an extra segment that has no matching route file, causing 404 on both OPTIONS preflight and direct requests. These endpoints were broken in production. Changes: - Remove `{param}` from 5 service.proto HTTP paths - Add `(sebuf.http.query)` annotations to request message fields - Update generated client/server code to use URLSearchParams - Update OpenAPI specs (YAML + JSON) to declare query params - Add early-return guards in 4 handlers for missing required params - Add happy.worldmonitor.app to runtime.ts redirect hosts Affected endpoints: - GET /api/conflict/v1/get-humanitarian-summary?country_code=SA - GET /api/economic/v1/get-fred-series?series_id=T10Y2Y&limit=120 - GET /api/market/v1/get-country-stock-index?country_code=US - GET /api/intelligence/v1/get-country-intel-brief?country_code=US - GET /api/military/v1/get-aircraft-details?icao24=a12345 * fix(security-advisories): route feeds through RSS proxy to avoid CORS blocks (#473) - Advisory feeds were fetched directly from the browser, hitting CORS on all 21 feeds (US State Dept, AU Smartraveller, US Embassies, ECDC, CDC, WHO). Route through /api/rss-proxy on web, keep proxyUrl for desktop. - Fix double slash in ECDC Avian Influenza URL (323//feed → 323/feed) - Add feeds.news24.com to RSS proxy allowlist (was returning 403) * feat(cache): tiered edge Cache-Control aligned to upstream TTLs (#474) * fix: move 5 path-param endpoints to query params for Vercel routing Vercel's `api/[domain]/v1/[rpc].ts` captures one dynamic segment. Path params like `/get-humanitarian-summary/SA` add an extra segment that has no matching route file, causing 404 on both OPTIONS preflight and direct requests. These endpoints were broken in production. Changes: - Remove `{param}` from 5 service.proto HTTP paths - Add `(sebuf.http.query)` annotations to request message fields - Update generated client/server code to use URLSearchParams - Update OpenAPI specs (YAML + JSON) to declare query params - Add early-return guards in 4 handlers for missing required params - Add happy.worldmonitor.app to runtime.ts redirect hosts Affected endpoints: - GET /api/conflict/v1/get-humanitarian-summary?country_code=SA - GET /api/economic/v1/get-fred-series?series_id=T10Y2Y&limit=120 - GET /api/market/v1/get-country-stock-index?country_code=US - GET /api/intelligence/v1/get-country-intel-brief?country_code=US - GET /api/military/v1/get-aircraft-details?icao24=a12345 * feat(cache): add tiered edge Cache-Control aligned to upstream TTLs Replace flat s-maxage=300 with 5 tiers (fast/medium/slow/static/no-store) mapped per-endpoint to respect upstream Redis TTLs. Adds stale-if-error resilience headers and X-No-Cache plumbing for future degraded responses. X-Cache-Tier debug header gated behind ?_debug query param. * fix(tech): use rss() for CISA feed, drop build from pre-push hook (#475) - CISA Advisories used dead rss.worldmonitor.app domain (404), switch to rss() helper - Remove Vite build from pre-push hook (tsc already catches errors) * fix(desktop): enable click-to-play YouTube embeds + CISA feed fixes (#476) * fix(tech): use rss() for CISA feed, drop build from pre-push hook - CISA Advisories used dead rss.worldmonitor.app domain (404), switch to rss() helper - Remove Vite build from pre-push hook (tsc already catches errors) * fix(desktop): enable click-to-play for YouTube embeds in WKWebView WKWebView blocks programmatic autoplay in cross-origin iframes regardless of allow attributes, Permissions-Policy, mute-first retries, or secure context. Documented all 10 approaches tested in docs/internal/. Changes: - Switch sidecar embed origin from 127.0.0.1 to localhost (secure context) - Add MutationObserver + retry chain as best-effort autoplay attempts - Use postMessage('*') to fix tauri://localhost cross-origin messaging - Make sidecar play overlay non-interactive (pointer-events:none) - Fix .webcam-iframe pointer-events:none blocking clicks in grid view - Add expand button to grid cells for switching to single view on desktop - Add http://localhost:* to CSP frame-src in index.html and tauri.conf.json * fix(gateway): convert stale POST requests to GET for backwards compat (#477) Stale cached client bundles still send POST to endpoints converted to GET in PR #468, causing 404s. The gateway now parses the POST JSON body into query params and retries the match as GET. * feat(proxy): add Cloudflare edge caching for proxy.worldmonitor.app (#478) Add CDN-Cache-Control headers to all proxy endpoints so Cloudflare can cache responses at the edge independently of browser Cache-Control: - RSS: 600s edge + stale-while-revalidate=300 (browser: 300s) - UCDP: 3600s edge (matches browser) - OpenSky: 15s edge (browser: 30s) for fresher flight data - WorldBank: 1800s/86400s edge (matches browser) - Polymarket: 120s edge (matches browser) - Telegram: 10s edge (matches browser) - AIS snapshot: 2s edge (matches browser) Also fixes: - Vary header merging: sendCompressed/sendPreGzipped now merge existing Vary: Origin instead of overwriting, preventing cross-origin cache poisoning at the edge - Stale fallback responses (OpenSky, WorldBank, Polymarket, RSS) now set Cache-Control: no-store + CDN-Cache-Control: no-store to prevent edge caching of degraded responses - All no-cache branches get CDN-Cache-Control: no-store - /opensky-reset gets no-store (state-changing endpoint) * fix(sentry): add noise filters for 4 unresolved issues (#479) - Tighten AbortError filter to match "AbortError: The operation was aborted" - Filter "The user aborted a request" (normal navigation cancellation) - Filter UltraViewer service worker injection errors (/uv/service/) - Filter Huawei WebView __isInQueue__ injection * feat: configurable VITE_WS_API_URL + harden POST→GET shim (#480) * fix(gateway): harden POST→GET shim with scalar guard and size limit - Only convert string/number/boolean values to query params (skip objects, nested arrays, __proto__ etc.) to prevent prototype pollution vectors - Skip body parsing for Content-Length > 1MB to avoid memory pressure * feat: make API base URL configurable via VITE_WS_API_URL Replace hardcoded api.worldmonitor.app with VITE_WS_API_URL env var. When empty, installWebApiRedirect() is skipped entirely — relative /api/* calls stay on the same domain (local installs). When set, browser fetch is redirected to that URL. Also adds VITE_WS_API_URL and VITE_WS_RELAY_URL hostnames to APP_HOSTS allowlist dynamically. * fix(analytics): use greedy regex in PostHog ingest rewrites (#481) Vercel's :path* wildcard doesn't match trailing slashes that PostHog SDK appends (e.g. /ingest/s/?compression=...), causing 404s. Switch to :path(.*) which matches all path segments including trailing slashes. Ref: PostHog/posthog#17596 * perf(proxy): increase AIS snapshot edge TTL from 2s to 10s (#482) With 20k requests/30min (60% of proxy traffic) and per-PoP caching, a 2s edge TTL expires before the next request from the same PoP arrives, resulting in near-zero cache hits. 10s allows same-PoP dedup while keeping browser TTL at 2s for fresh vessel positions. * fix(markets): commodities panel showing stocks instead of commodities (#483) The shared circuit breaker (cacheTtlMs: 0) cached the stocks response, then the stale-while-revalidate path returned that cached stocks data for the subsequent commodities fetch. Skip SWR when caching is disabled. * feat(gateway): complete edge cache tier coverage + degraded-response policy (#484) - Add 11 missing GET routes to RPC_CACHE_TIER map (8 slow, 3 medium) - Add response-headers side-channel (WeakMap) so handlers can signal X-No-Cache without codegen changes; wire into military-flights and positive-geo-events handlers on upstream failure - Add env-controlled per-endpoint tier override (CACHE_TIER_OVERRIDE_*) for incident response rollback - Add VITE_WS_API_URL hostname allowlist (*.worldmonitor.app + localhost) - Fix fetch.bind(globalThis) in positive-events-geo.ts (deferred lambda) - Add CI test asserting every generated GET route has an explicit cache tier entry (prevents silent default-tier drift) * chore: bump version to 2.5.20 + changelog Covers PRs #452–#484: Cloudflare edge caching, commodities SWR fix, security advisories panel, settings redesign, 52 POST→GET migrations. * fix(rss): remove stale indianewsnetwork.com from proxy allowlist (#486) Feed has no <pubDate> fields and latest content is from April 2022. Not referenced in any feed config — only in the proxy domain allowlist. * feat(i18n): add Korean (한국어) localization (#487) - Add ko.json with all 1606 translation keys matching en.json structure - Register 'ko' in SUPPORTED_LANGUAGES, LANGUAGES display array, and locale map - Korean appears as 🇰🇷 한국어 in the language dropdown * feat: add Polish tv livestreams (#488) * feat(rss): add Axios (api.axios.com/feed) as US news source (#494) Add api.axios.com to proxy allowlist and CSP connect-src, register Axios feed under US category as Tier 2 mainstream source. * perf: bootstrap endpoint + polling optimization (#495) * perf: bootstrap endpoint + polling optimization (phases 3-4) Replace 15+ individual RPC calls on startup with a single /api/bootstrap batch call that fetches pre-cached data from Redis. Consolidate 6 panel setInterval timers into the central RefreshScheduler for hidden-tab awareness (10x multiplier) and adaptive backoff (up to 4x for unchanged data). Convert IntelligenceGapBadge from 10s polling to event-driven updates with 60s safety fallback. * fix(bootstrap): inline Redis + cache keys in edge function Vercel Edge Functions cannot resolve cross-directory TypeScript imports from server/_shared/. Inline getCachedJsonBatch and BOOTSTRAP_CACHE_KEYS directly in api/bootstrap.js. Add sync test to ensure inlined keys stay in sync with the canonical server/_shared/cache-keys.ts registry. * test: add Edge Function module isolation guard for all api/*.js files Prevents any Edge Function from importing from ../server/ or ../src/ which breaks Vercel builds. Scans all 12 non-helper Edge Functions. * fix(bootstrap): read unprefixed cache keys on all environments Preview deploys set VERCEL_ENV=preview which caused getKeyPrefix() to prefix Redis keys with preview:<sha>:, but handlers only write to unprefixed keys on production. Bootstrap is a read-only consumer of production cache — always read unprefixed keys. * fix(bootstrap): wire sectors hydration + add coverage guard - Wire getHydratedData('sectors') in data-loader to skip Yahoo Finance fetch when bootstrap provides sector data - Add test ensuring every bootstrap key has a getHydratedData consumer — prevents adding keys without wiring them * fix(server): resolve 25 TypeScript errors + add server typecheck to CI - _shared.ts: remove unused `delay` variable - list-etf-flows.ts: add missing `rateLimited` field to 3 return literals - list-market-quotes.ts: add missing `rateLimited` field to 4 return literals - get-cable-health.ts: add non-null assertions for regex groups and array access - list-positive-geo-events.ts: add non-null assertion for array index - get-chokepoint-status.ts: add required fields to request objects - CI: run `typecheck:api` (tsconfig.api.json) alongside `typecheck` to catch server/ TS errors before merge * feat(military): server-side military bases 125K + rate limiting (#496) * feat(military): server-side military bases with 125K entries + rate limiting (#485) Migrate military bases from 224 static client-side entries to 125,380 server-side entries stored in Redis GEO sorted sets, served via bbox-filtered GEOSEARCH endpoint with server-side clustering. Data pipeline: - Pizzint/Polyglobe: 79,156 entries (Supabase extraction) - OpenStreetMap: 45,185 entries - MIRTA: 821 entries - Curated strategic: 218 entries - 277 proximity duplicates removed Server: - ListMilitaryBases RPC with GEOSEARCH + HMGET + tier/filter/clustering - Antimeridian handling (split bbox queries) - Blue-green Redis deployment with atomic version pointer switch - geoSearchByBox() + getHashFieldsBatch() helpers in redis.ts Security: - @upstash/ratelimit: 60 req/min sliding window per IP - IP spoofing fix: prioritize x-real-ip (Vercel-injected) over x-forwarded-for - Require API key for non-browser requests (blocks unauthenticated curl/scripts) - Input validation: allowlisted types/kinds, regex country, clamped bbox/zoom Frontend: - Viewport-driven loading with bbox quantization + debounce - Server-side grid clustering at low zoom levels - Enriched popup with kind, category badges (airforce/naval/nuclear/space) - Static 224 bases kept as search fallback + initial render * fix(military): fallback to production Redis keys in preview deployments Preview deployments prefix Redis keys with `preview:{sha}:` but military bases data is seeded to unprefixed (production) keys. When the prefixed `military:bases:active` key is missing, fall back to the unprefixed key and use raw (unprefixed) keys for geo/meta lookups. * fix: remove unused 'remaining' destructure in rate-limit (TS6133) * ci: add typecheck:api to pre-push hook to catch server-side TS errors * debug(military): add X-Bases-Debug response header for preview diagnostics * fix(bases): trigger initial server fetch on map load fetchServerBases() was only called on moveend — if the user never panned/zoomed, the API was never called and only the 224 static fallback bases showed. * perf(military): debounce base fetches + upgrade edge cache to static tier (#497) - Add 300ms debounce on moveend to prevent rapid pan flooding - Fixes stale-bbox bug where pendingFetch returns old viewport data - Upgrade edge cache tier from medium (5min) to static (1hr) — bases are static infrastructure, aligned with server-side cachedFetchJson TTL - Keep error logging in catch blocks for production diagnostics * fix(cyber): make GeoIP centroid fallback jitter deterministic (#498) Replace Math.random() jitter with DJB2 hash seeded by the threat indicator (IP/URL), so the same threat always maps to the same coordinates across requests while different threats from the same country still spread out. Closes #203 Co-authored-by: Chris Chen <fuleinist@users.noreply.github.com> * fix: use cross-env for Windows-compatible npm scripts (#499) Replace direct `VAR=value command` syntax with cross-env/cross-env-shell so dev, build, test, and desktop scripts work on Windows PowerShell/CMD. Co-authored-by: facusturla <facusturla@users.noreply.github.com> * feat(live-news): add CBC News to optional North America channels (#502) YouTube handle @CBCNews with fallback video ID 5vfaDsMhCF4. * fix(bootstrap): harden hydration cache + polling review fixes (#504) - Filter null/undefined values before storing in hydration cache to prevent future consumers using !== undefined from misinterpreting null as valid data - Debounce wm:intelligence-updated event handler via requestAnimationFrame to coalesce rapid alert generation into a single render pass - Include alert IDs in StrategicRiskPanel change fingerprint so content changes are detected even when alert count stays the same - Replace JSON.stringify change detection in ServiceStatusPanel with lightweight name:status fingerprint - Document max effective refresh interval (40x base) in scheduler * fix(geo): tokenization-based keyword matching to prevent false positives (#503) * fix(geo): tokenization-based keyword matching to prevent false positives Replace String.includes() with tokenization-based Set.has() matching across the geo-tagging pipeline. Prevents false positives like "assad" matching inside "ambassador" and "hts" matching inside "rights". - Add src/utils/keyword-match.ts as single source of truth - Decompose possessives/hyphens ("Assad's" → includes "assad") - Support multi-word phrase matching ("white house" as contiguous) - Remove false-positive-prone DC keywords ('house', 'us ') - Update 9 consumer files across geo-hub, map, CII, and asset systems - Add 44 tests covering false positives, true positives, edge cases Co-authored-by: karim <mirakijka@gmail.com> Fixes #324 * fix(geo): add inflection suffix matching + fix test imports Address code review feedback: P1a: Add suffix-aware matching for plurals and demonyms so existing keyword lists don't regress (houthi→houthis, ukraine→ukrainian, iran→iranian, israel→israeli, russia→russian, taiwan→taiwanese). Uses curated suffix list + e-dropping rule to avoid false positives. P1b: Expand conflictTopics arrays in DeckGLMap and Map with demonym forms so "Iranian senate..." correctly registers as conflict topic. P2: Replace inline test functions with real module import via tsx. Tests now exercise the production keyword-match.ts directly. * fix: wire geo-keyword tests into test:data command The .mts test file wasn't covered by `node --test tests/*.test.mjs`. Add `npx tsx --test tests/*.test.mts` so test:data runs both suites. * fix: cross-platform test:data + pin tsx in devDependencies - Use tsx as test runner for both .mjs and .mts (single invocation) - Removes ; separator which breaks on Windows cmd.exe - Add tsx to devDependencies so it works in offline/CI environments * fix(geo): multi-word demonym matching + short-keyword suffix guard - Add wordMatches() for suffix-aware phrase matching so "South Korean" matches keyword "south korea" and "North Korean" matches "north korea" - Add MIN_SUFFIX_KEYWORD_LEN=4 guard so short keywords like "ai", "us", "hts" only do exact-match (prevents "ais"→"ai", "uses"→"us" false positives) - Add 5 new tests covering both fixes (58 total, all passing) * fix(geo): support plural demonyms in keyword matching Add compound suffixes (ians, eans, ans, ns, is) to handle plural demonym forms like "Iranians"→"iran", "Ukrainians"→"ukraine", "Russians"→"russia", "Israelis"→"israel". Adds 5 new tests (63 total). --------- Co-authored-by: karim <mirakijka@gmail.com> * chore: strip 61 debug console.log calls from 20 service files (#501) * chore: strip 61 debug console.log calls from services Remove development/tracing console.log statements from 20 files. These add noise to production browser consoles and increase bundle size. Preserved: all console.error (error handling) and console.warn (warnings). Preserved: debug-gated logs in runtime.ts (controlled by verbose flag). Removed: debugInjectTestEvents() from geo-convergence.ts (test-only code). Removed: logSummary()/logReport() methods that were pure console.log wrappers. * fix: remove orphaned stubs and remaining debug logs from stripped services - Remove empty logReport() method and unused startTime variable (parallel-analysis.ts) - Remove orphaned console.group/console.groupEnd pair (parallel-analysis.ts) - Remove empty logSignalSummary() export (signal-aggregator.ts) - Remove logSignalSummary import/call and 3 remaining console.logs (InsightsPanel.ts) - Remove no-op logDirectFetchBlockedOnce() and dead infrastructure (prediction/index.ts) * fix: generalize Vercel preview origin regex + include filters in bases cache key (#506) - api/_api-key.js: preview URL pattern was user-specific (-elie-), rejecting other collaborators' Vercel preview deployments. Generalized to match any worldmonitor-*.vercel.app origin. - military-bases.ts: client cache key only checked bbox/zoom, ignoring type/kind/country filters. Switching filters without panning returned stale results. Unified into single cacheKey string. * fix(prediction): filter stale/expired markets from Polymarket panel (#507) Prediction panel was showing expired markets (e.g. "Will US strike Iran on Feb 9" at 0%). Root causes: no active/archived API filters, no end_date_min param, no client-side expiry guard, and sub-market selection picking highest volume before filtering expired ones. - Add active=true, archived=false, end_date_min API params to all 3 Gamma API call sites (events, markets, probe) - Pre-filter sub-markets by closed/expired BEFORE volume selection in both fetchPredictions() and fetchCountryMarkets() - Add defense-in-depth isExpired() client-side filter on final results - Propagate endDate through all market object paths including sebuf fallback - Show expiry date in PredictionPanel UI with new .prediction-meta layout - Add "closes" i18n key to all 18 locale files - Add endDate to server handler GammaMarket/GammaEvent interfaces and map to proto closesAt field * fix(relay): guard proxy handlers against ERR_HTTP_HEADERS_SENT crash (#509) Polymarket and World Bank proxy handlers had unguarded res.writeHead() calls in error/timeout callbacks that race with the response callback. When upstream partially responds then times out, both paths write headers → process crash. Replace 5 raw writeHead+end calls with safeEnd() which checks res.headersSent before writing. * feat(breaking-news): add active alert banner with audio for critical/high RSS items (#508) RSS items classified as critical/high threat now trigger a full-width breaking news banner with audio alert, auto-dismiss (60s/30s by severity), visibility-aware timer pause, dedup, and a toggle in the Intelligence Findings dropdown. * fix(sentry): filter Android OEM WebView bridge injection errors (#510) Add ignoreErrors pattern for LIDNotifyId, onWebViewAppeared, and onGetWiFiBSSID — native bridge functions injected by Lenovo/Huawei device SDKs into Chrome Mobile WebView. No stack frames in our code. * chore: add validated telegram channels list (global + ME + Iran + cyber) (#249) * feat(conflict): add Iran Attacks map layer + strip debug logs (#511) * chore: strip 61 debug console.log calls from services Remove development/tracing console.log statements from 20 files. These add noise to production browser consoles and increase bundle size. Preserved: all console.error (error handling) and console.warn (warnings). Preserved: debug-gated logs in runtime.ts (controlled by verbose flag). Removed: debugInjectTestEvents() from geo-convergence.ts (test-only code). Removed: logSummary()/logReport() methods that were pure console.log wrappers. * fix: remove orphaned stubs and remaining debug logs from stripped services - Remove empty logReport() method and unused startTime variable (parallel-analysis.ts) - Remove orphaned console.group/console.groupEnd pair (parallel-analysis.ts) - Remove empty logSignalSummary() export (signal-aggregator.ts) - Remove logSignalSummary import/call and 3 remaining console.logs (InsightsPanel.ts) - Remove no-op logDirectFetchBlockedOnce() and dead infrastructure (prediction/index.ts) * feat(conflict): add Iran Attacks map layer Adds a new Iran-focused conflict events layer that aggregates real-time events, geocodes via 40-city lookup table, caches 15min in Redis, and renders as a toggleable DeckGL ScatterplotLayer with severity coloring. - New proto + codegen for ListIranEvents RPC - Server handler with HTML parsing, city geocoding, category mapping - Frontend service with circuit breaker - DeckGL ScatterplotLayer with severity-based color/size - MapPopup with sanitized source links - iranAttacks toggle across all variants, harnesses, and URL state * fix: resolve bootstrap 401 and 429 rate limiting on page init (#512) Same-origin browser requests don't send Origin header (per CORS spec), causing validateApiKey to reject them. Extract origin from Referer as fallback. Increase rate limit from 60 to 200 req/min to accommodate the ~50 requests fired during page initialization. * fix(relay): prevent Polymarket OOM via request deduplication (#513) Concurrent Polymarket requests for the same cache key each fired independent https.get() calls. With 12 categories × multiple clients, 740 requests piled up in 10s, all buffering response bodies → 4.1GB heap → OOM crash on Railway. Fix: in-flight promise map deduplicates concurrent requests to the same cache key. 429/error responses are negative-cached for 30s to prevent retry storms. * fix(threat-classifier): add military/conflict keyword gaps and news-to-conflict bridge (#514) Breaking news headlines like "Israel's strike on Iran" were classified as info level because the keyword classifier lacked standalone conflict phrases. Additionally, the conflict instability score depended solely on ACLED data (1-7 day lag) with no bridge from real-time breaking news. - Add 3 critical + 18 high contextual military/conflict keywords - Preserve threat classification on semantically merged clusters - Add news-derived conflict floor when ACLED/HAPI report zero signal - Upsert news events by cluster ID to prevent duplicates - Extract newsEventIndex to module-level Map for serialization safety * fix(breaking-news): let critical alerts bypass global cooldown and replace HIGH alerts (#516) Global cooldown (60s) was blocking critical alerts when a less important HIGH alert fired from an earlier RSS batch. Added priority-aware cooldown so critical alerts always break through. Banner now auto-dismisses HIGH alerts when a CRITICAL arrives. Added Iran/strikes keywords to classifier. * fix(rate-limit): increase sliding window to 300 req/min (#515) App init fires many concurrent classify-event, summarize-article, and record-baseline-snapshot calls, exhausting the 200/min limit and causing 429s. Bump to 300 as a temporary measure while client-side batching is implemented. * fix(breaking-news): fix fake pubDate fallback and filter noisy think-tank alerts (#517) Two bugs causing stale CrisisWatch article to fire as breaking alert: 1. Non-standard pubDate format ("Friday, February 27, 2026 - 12:38") failed to parse → fallback was `new Date()` (NOW) → day-old articles appeared as "just now" and passed recency gate on every fetch 2. Tier 3+ sources (think tanks) firing alerts on keyword-only matches like "War" in policy analysis titles — too noisy for breaking alerts Fix: parsePubDate() handles non-standard formats and falls back to epoch (not now). Tier 3+ sources require LLM classification to fire. * fix: make iran-events handler read-only from Redis (#518) Remove server-side LiveUAMap scraper (blocked by Cloudflare 403 on Vercel IPs). Handler now reads pre-populated Redis cache pushed from local browser scraping. Change cache tier from slow to fast to prevent CDN from serving stale empty responses for 30+ minutes. * fix(relay): Polymarket circuit breaker + concurrency limiter (OOM fix) (#519) * fix(rate-limit): increase sliding window to 300 req/min App init fires many concurrent classify-event, summarize-article, and record-baseline-snapshot calls, exhausting the 200/min limit and causing 429s. Bump to 300 as a temporary measure while client-side batching is implemented. * fix(relay): add Polymarket circuit breaker + concurrency limiter to prevent OOM Railway relay OOM crash: 280 Polymarket 429 errors in 8s, heap hit 3.7GB. Multiple unique cache keys bypassed per-key dedup, flooding upstream. - Circuit breaker: trips after 5 consecutive failures, 60s cooldown - Concurrent upstream limiter: max 3 simultaneous requests - Negative cache TTL: 30s → 60s to reduce retry frequency - Upstream slot freed on response.on('end'), not headers, preventing body buffer accumulation past the concurrency cap * fix(relay): guard against double-finalization on Polymarket timeout request.destroy() in timeout handler also fires request.on('error'), causing double decrement of polymarketActiveUpstream (counter goes negative, disabling concurrency cap) and double circuit breaker trip. Add finalized guard so decrement + failure accounting happens exactly once per request regardless of which error path fires first. * fix(threat-classifier): stagger AI classification requests to avoid Groq 429 (#520) flushBatch() fired up to 20 classifyEvent RPCs simultaneously via Promise.all, instantly hitting Groq's ~30 req/min rate limit. - Sequential execution with 2s min-gap between requests (~28 req/min) - waitForGap() enforces hard floor + jitter across batch boundaries - batchInFlight guard prevents concurrent flush loops - 429/5xx: requeue failed job (with retry cap) + remaining untouched jobs - Queue cap at 100 items with warn on overflow * fix(relay): regenerate package-lock.json with telegram dependency The lockfile was missing resolved entries for the telegram package, causing Railway to skip installation despite it being in package.json. * chore: trigger deploy to flush CDN cache for iran-events endpoint * Revert "fix(relay): regenerate package-lock.json with telegram dependency" This reverts commita8d5e1dbbd. * fix(relay): add POLYMARKET_ENABLED env flag kill switch (#523) Set POLYMARKET_ENABLED=false on Railway to disable all Polymarket upstream requests. Returns 503 immediately, preventing OOM crashes. * fix(breaking-news): fill keyword gaps missing real Iran attack headlines (#521) * fix(breaking-news): fill keyword gaps that miss real Iran attack headlines Three root causes for zero alerts during the Iran war: 1. Keyword gaps — top Iran headlines failed classification: - "US and Israel attack Iran" → info (no "attack iran" keyword) - "attacked Iran" → info (only "attacks iran" existed, plural) - "Explosions heard in Tehran" → info (no "explosions" keyword) Added: attack iran, attacked iran, attack on iran, attack against iran, bombing/bombed iran, war against iran (CRITICAL); explosions, launched/launches attacks, retaliatory/preemptive/preventive attack (HIGH) 2. 5-item RSS limit — Al Jazeera's CRITICAL "major combat operations" headline was item #7 and never reached the classifier. Increased per-feed limit from 5 to 10. 3. False positive — "OpenAI strikes deal with Pentagon" matched HIGH keyword "strikes". Added "strikes deal/agreement/partnership" to exclusions. * fix(threat-classifier): prevent Iran compound keyword false positives "attack iran" as plain substring matched "Iran-backed" and "Iranian" in headlines about proxy groups, not direct attacks on Iran. Added TRAILING_BOUNDARY_KEYWORDS set with negative lookahead (?![\w-]) for all Iran compound keywords. This rejects "Iran-backed militias" and "Iranian targets" while still matching "attack Iran:" and "attack Iran" at end of string. Addresses Codex review comment on PR #521. * fix(relay): regenerate package-lock.json with telegram dependency (#522) The lockfile was missing resolved entries for the telegram package, causing Railway to skip installation despite it being in package.json. * fix(iran): bypass stale CDN cache for iran-events endpoint (#524) The CDN cached empty {events:[],scrapedAt:0} from the pre-Redis deployment and Vercel deploy didn't purge all edge nodes. Add ?_v=2 query param to force cache miss until CDN naturally expires the stale slow-tier entry. * fix(focal-points): attribute theater military activity to target nations (#525) The signal aggregator attributed military flights/vessels to the country they're physically over (point-in-polygon). Aircraft attacking Iran from the Persian Gulf got attributed to XX/IQ/SA, not IR — so Iran showed ELEVATED in Focal Points despite being under active attack (CRIT in Strategic Posture). Feed theater-level posture data back into the signal aggregator for target nations (Iran, Taiwan, North Korea, Gaza, Yemen) so they get credited for military activity in their theater bounding box. Includes double-count guard to skip if the nation already has signals. Also fixes stale "sebuf" comment in threat-classifier. * fix(relay): block rsshub.app requests with 410 Gone (#526) Stale clients still send RSS requests to rsshub.app (NHK, MOFCOM, MIIT). These feeds were migrated to Google News RSS but cached PWA clients keep hitting the relay, which forwards to rsshub.app and gets 403. - Add explicit blocklist returning 410 Gone before allowlist check - Remove rsshub.app from all allowlists (relay, edge proxy, vite) - Remove dead AP News dev proxy target * feat(map): prioritize Iran Attacks layer (#527) * feat(map): move Iran Attacks layer to first position and enable by default Move iranAttacks to the top of the layer toggle list in the full (geopolitical) variant so it appears first. Enable it by default on both desktop and mobile during the active conflict. * feat(map): add Iran Attacks layer support to SVG/mobile map - Implement setIranEvents() in SVG Map (was no-op) - Render severity-colored circle markers matching DeckGL layer - Add iranAttacks to mobile layer toggles (first position) - Forward setIranEvents to SVG map in MapContainer - Add IranEventPopupData to PopupData union for click popups - Add .iran-event-marker CSS with pulse animation - Add data-layer-hidden-iranAttacks CSS toggle * fix(geo): expand geo hub index with 60+ missing world locations (#528) The geo hub index only had ~30 entries, missing all Gulf states (UAE, Qatar, Bahrain, Kuwait, Oman), Iraq cities, and many world capitals. News mentioning Abu Dhabi, Dubai, Baghdad, etc. had no lat/lon assigned so they never appeared on the map. Added: Gulf capitals (Abu Dhabi, Dubai, Doha, Manama, Kuwait, Muscat), Iraq (Baghdad, Erbil, Basra), Jordan, Istanbul, Haifa, Dimona, Isfahan, Kabul, Mumbai, Shanghai, Hong Kong, Singapore, Manila, Jakarta, Bangkok, Hanoi, Canberra, all major European capitals (Rome, Madrid, Warsaw, Bucharest, Helsinki, Stockholm, Oslo, Baltics, Athens, Belgrade, Minsk, Tbilisi, Chisinau, Yerevan, Baku), Americas (Ottawa, Mexico City, Brasilia, Buenos Aires, Caracas, Bogota, Havana), Africa (Nairobi, Pretoria, Lagos, Kinshasa, Mogadishu, Tripoli, Tunis, Algiers, Rabat), conflict zones (Iraq, Kashmir, Golan), chokepoints (Malacca, Panama, Gibraltar), and US military bases (Ramstein, Incirlik, Diego Garcia, Guam, Okinawa). * fix(iran): bust CDN cache to serve updated Gulf-geocoded events (#532) CDN edge cache was still serving stale 93-event response without Gulf state coordinates (UAE, Bahrain, Qatar, Kuwait). Bump cache key from ?_v=2 to ?_v=3 so browsers fetch fresh 100-event data. Also gitignore internal/ for private tooling scripts. * fix(alerts): remove SESSION_START gate that blocked pre-existing breaking news (#533) The isRecent() function used Math.max(now - 15min, SESSION_START) as recency cutoff. Since SESSION_START = Date.now() at module load, items published before page load could never trigger alerts — they failed the SESSION_START gate in the first 15 min, then aged past the 15-min window. Now uses only the 15-minute recency window. Spam prevention remains via per-event dedup (30 min), global cooldown (60s), and source tier filter. * fix(relay): Telegram + OOM + memory cleanup (#531) * fix(relay): resolve Telegram missing package, OOM crashes, and memory cleanup - Add `telegram` and `ws` to root dependencies so Railway's `npm install` installs them - Log V8 heap limit at startup to confirm NODE_OPTIONS is active - Make MAX_VESSELS/MAX_VESSEL_HISTORY env-configurable (default 20k, down from 50k) - Add permanent latch to skip Telegram import retries when package is missing - Raise memory cleanup threshold from 450MB to 2GB (env-configurable) - Clear all caches (RSS, Polymarket, WorldBank) during emergency cleanup * fix(relay): treat blank env vars as unset in safeInt Number('') === 0 passes isFinite, silently clamping caps to 1000 instead of using the 20000 default. Guard empty/null before parsing. * fix(live-news): replace 7 stale YouTube fallback video IDs (#535) Validated all 23 YouTube fallbackVideoIds via oEmbed API and all 9 HLS URLs. Found 5 broken IDs (403 embed-restricted or 404 deleted) plus 2 previously identified stale IDs: - Fox News: QaftgYkG-ek → ZvdiJUYGBis - Sky News Arabia: MN50dHFHMKE → U--OjmpjF5o - RTVE 24H: 7_srED6k0bE → -7GEFgUKilA - CNN Brasil: 1kWRw-DA6Ns → 6ZkOlaGfxq4 - C5N: NdQSOItOQ5Y → SF06Qy1Ct6Y - TBS NEWS DIG: ohI356mwBp8 → Anr15FA9OCI - TRT World: CV5Fooi8WDI → ABfFhWzWs0s All 9 HLS URLs validated OK. 16 remaining YouTube IDs validated OK. * fix(relay): fix telegram ESM import path and broaden latch regex - `import('telegram/sessions')` fails with "Directory import is not supported resolving ES modules" — use explicit `telegram/sessions/index.js` - Broaden permanent-disable latch to also catch "Directory import" errors * fix(ui): move download banner to bottom-right (#536) * fix(alerts): remove SESSION_START gate that blocked pre-existing breaking news The isRecent() function used Math.max(now - 15min, SESSION_START) as recency cutoff. Since SESSION_START = Date.now() at module load, items published before page load could never trigger alerts — they failed the SESSION_START gate in the first 15 min, then aged past the 15-min window. Now uses only the 15-minute recency window. Spam prevention remains via per-event dedup (30 min), global cooldown (60s), and source tier filter. * fix(ui): move download banner to bottom-right of screen Repositioned from top-right (overlapping content) to bottom-right. Dismissal already persists via localStorage. Added TODO for header download link. * Revert "fix(relay): fix telegram ESM import path and broaden latch regex" This reverts commit1f2f0175ab. * Revert "Revert "fix(relay): fix telegram ESM import path and broaden latch regex"" (#537) This reverts commit ad41a2e2d245850fd4b699af2adbe53acca80325. * feat: add day/night solar terminator overlay to map (#529) * Trigger redeploy with preview env vars * Trigger deployment * chore: trigger redeploy for PR #41 * chore: trigger Vercel redeploy (edge function transient failure) * chore: retrigger Vercel deploy * feat: add Nigeria feeds and Greek locale feeds (#271) - Add 5 Nigeria news sources to Africa section (Premium Times, Vanguard, Channels TV, Daily Trust, ThisDay) - Add 5 Greek feeds with lang: 'el' for locale-aware filtering (Kathimerini, Naftemporiki, in.gr, iefimerida, Proto Thema) - Add source tiers for all new outlets - Allowlist 8 new domains in RSS proxy * fix: enforce military bbox filtering and add behavioral cache tests (#284) * fix: add request coalescing to Redis cache layer Concurrent cache misses for the same key now share a single upstream fetch instead of each triggering redundant API calls. This eliminates duplicate work within Edge Function invocations under burst traffic. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: reduce AIS polling frequency from 10s to 30s Vessel positions do not change meaningfully in 10 seconds at sea. Reduces Railway relay requests by 66% with negligible UX impact. Stale threshold bumped to 45s to match the new interval. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: quantize military flights bbox cache keys to 1-degree grid Precise bounding box coordinates caused near-zero cache hit rate since every map pan/zoom produced a unique key. Snapping to a 1-degree grid lets nearby viewports share cache entries, dramatically reducing redundant OpenSky API calls. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: parallelize ETF chart fetches instead of sequential await loop The loop awaited each ETF chart fetch individually, blocking on every Yahoo gate delay. Using Promise.allSettled lets all 10 fetches queue concurrently through the Yahoo gate, cutting wall time from ~12s to ~6s. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: add Redis pipeline batch GET to reduce round-trips Add getCachedJsonBatch() using the Upstash pipeline API to fetch multiple keys in a single HTTP call. Refactor aircraft details batch handler from 20 sequential GETs to 1 pipelined request. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * test: add structural tests for Redis caching optimizations 18 tests covering: cachedFetchJson request coalescing (in-flight dedup, cache-before-fetch ordering, cleanup), getCachedJsonBatch pipeline API, aircraft batch handler pipeline usage, bbox grid quantization (1-degree step, expanded fetch bbox), and ETF parallel fetch. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: enforce military bbox contract and add behavioral cache tests --------- Co-authored-by: Elias El Khoury <efk@anghami.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * fix: add User-Agent and Cloudflare 403 detection to all secret validation probes (#296) Sidecar validation probes were missing User-Agent headers, causing Cloudflare-fronted APIs (e.g. Wingbits) to return 403 which was incorrectly treated as an auth rejection. Added CHROME_UA to all 13 probes and isCloudflare403() helper to soft-pass CDN blocks. * fix: open external links in system browser on Tauri desktop (#297) Tauri WKWebView/WebView2 traps target="_blank" navigation, so news links and other external URLs silently fail to open. Added a global capture-phase click interceptor that routes cross-origin links through the existing open_url Tauri command, falling back to window.open. * fix: add Greek flag mapping to language selector (#307) * fix: add missing country brief i18n keys and export PDF option (#308) - Add levels, trends, fallback keys to top-level countryBrief in en/el/th/vi locales (fixes raw key display in intelligence brief and header badge) - Add Export PDF option to country brief dropdown using scoped print dialog - Add exportPdf i18n key to all 17 locale files * feat: add day/night solar terminator overlay to map Add a real-time day/night overlay layer using deck.gl PolygonLayer that renders the solar terminator (boundary between day and night zones). The overlay uses astronomical formulas (Meeus) to compute the subsolar point and trace the terminator line at 1° resolution. - New toggleable "Day/Night" layer in all 3 variants (full/tech/finance) - Theme-aware styling (lighter fill on light theme, darker on dark) - Auto-refresh every 5 minutes with conditional timer (only runs when layer is enabled, pauses when render is paused) - Cached polygon computation to avoid recomputing on every render - i18n translations for all 17 locales - Updated documentation with new layer entry Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: address review feedback — equinox terminator + locale indentation - Replace safeTanDecl epsilon clamp with proper equinox handling: when |tanDecl| < 1e-6, draw terminator as vertical great circle through the poles (subsolar meridian ±90°) instead of clamping - Fix JSON indentation in all 17 locale files: dayNight and tradeRoutes keys were left-aligned instead of matching 8-space indentation of surrounding keys --------- Co-authored-by: Elie Habib <elie.habib@gmail.com> Co-authored-by: Elias El Khoury <efk@anghami.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * fix(relay): auto-reconnect on Telegram AUTH_KEY_DUPLICATED and fix IranIntl handle (#539) - On AUTH_KEY_DUPLICATED (406), disconnect client and set to null so next poll cycle reconnects fresh — self-heals after competing client dies - Fix IranIntl → iranintltv (correct Telegram channel handle) * fix(live-news): add fallback video ID for LiveNOW from FOX channel (#538) The livenow-fox channel had no fallbackVideoId, relying solely on YouTube handle lookup which fails intermittently. Added ZvdiJUYGBis (confirmed live stream) as fallback. * fix(iran): bump CDN cache-bust to v4 for fresh event data (#544) 100 new events pushed to Redis covering active Iran-Israel-US conflict theater including Gulf states (UAE, Bahrain, Qatar, Kuwait, Jordan). Bump ?_v=3 to ?_v=4 to bypass stale CDN. * fix(telegram): fix ESM import path in session-auth script (#542) telegram/sessions → telegram/sessions/index.js (same fix as relay) * fix(telegram): latch AUTH_KEY_DUPLICATED to stop retry spam (#543) AUTH_KEY_DUPLICATED is permanent — the session string is invalidated and no amount of retrying will fix it. Previously the relay retried every 60s, spamming logs. Now it logs a clear error message with instructions to regenerate the session and stops retrying. Renamed telegramImportFailed → telegramPermanentlyDisabled to cover both import failures and auth failures under one latch. * fix(live-news): fix broken Europe channel handles + add fallback video IDs (#541) * fix(live-news): fix broken Europe channel handles + add fallback video IDs - Fix France 24 English handle: @FRANCE24English (404) → @France24_en - Fix WELT handle: @WELTNachrichtensender (hijacked to "Movie Glow") → @WELTVideoTV - Add fallbackVideoId for BBC News, France 24 EN, TRT Haber, NTV Turkey, CNN TURK, TVP Info, Telewizja Republika (verified via Playwright) - Update stale fallback IDs for Fox News, RTVE, CNN Brasil, C5N, TBS News, Sky News Arabia, TRT World * fix(live-news): update CBS News fallback video ID * fix(live-news): update Newsmax fallback video ID * fix(live-news): add NBC News fallback video ID * fix(live-news): full channel audit — fix 10 broken handles + update 8 stale fallbacks Broken handles fixed: - Bloomberg: @Bloomberg (404) → @markets - WION: @WIONews (wrong channel "Write It Out") → @WION - CTI News: @CtiTv (404) → @中天新聞CtiNews - VTC NOW: @VTCNOW (404) → @VTCNowOfficial - Record News: @recordnewsoficial (404) → @RecordNews - T13: @T13 (404) → @Teletrece - Channels TV: @channelstv (404) → @ChannelsTelevision - KTN News: @KTNNewsKE (404) → @ktnnews_kenya - eNCA: @enewschannel (404) → @eNCA - SABC News: @SABCNews (404) → @SABCDigitalNews Stale fallback video IDs refreshed: - Sky News, NASA, CBC News, CNN Brasil, C5N, TBS NEWS DIG, Sky News Arabia, TRT World * feat(oref): add OREF sirens panel with Hebrew-to-English translation (#545) Add real-time Israel Home Front Command (OREF) siren alerts panel: - Edge Function proxy at api/oref-alerts.js - OrefSirensPanel component with live/history views - oref-alerts service with 10s polling and update callbacks - Hebrew→English translation via existing translateText() LLM chain with 3-layer caching (in-memory Map → server Redis → circuit breaker) - i18n strings for all 23 locales - Panel registration, data-loader integration, CSS styles * fix(relay): use execFileSync for OREF curl to avoid shell injection (#546) Proxy credentials with special characters (semicolons, dollar signs) were interpolated into a shell command via execSync. Switch to execFileSync which passes args directly without shell parsing. * gave the user freedom to resize panels "fixes issue #426" (#489) * gave the user freedom to resize panles * feat(panels): add horizontal resize with col-span persistence * feat(cii): integrate Iran strike events into CII scoring, country brief & timeline (#547) Iran had ~100 geolocated strike events but the CII was unaware of them: conflict score stuck at 70 (ACLED only), no strike chip in Active Signals, timeline conflict lane empty, intelligence brief silent on strikes. Changes: - Add strikes[] to CountryData and ingestStrikesForCII() with geo-lookup fallback (bounding boxes when GeoJSON not yet loaded) - Boost CII conflict score with 7-day freshness window (min(50, count*3 + highSev*5)) - Cache iranEvents in IntelligenceCache, preserve across refresh cycles - Wire data loading: always load Iran events (not gated by map layer), ingest into CII, trigger panel refresh - Add activeStrikes to CountryBriefSignals with geo-lookup counting - Render strike chip in Active Signals and include in fallback brief - Feed strike events into 7-day timeline (conflict lane) - Add structured strikeCount/highSeverityStrikeCount fields to GeoSignal (replaces fragile regex parsing in focal-point-detector) - Add active_strike signal type to InsightsPanel focal points - Add bounding-box fallback to signal aggregator for conflict events - Add i18n keys for activeStrikes * fix(alerts): add compound escalation for military action + geopolitical target (#548) Keyword matching was too rigid — "attacks on iran" didn't match CRITICAL because only "attack on iran" (singular) existed. Headlines like "strikes by US and Israel on Iran" also missed because words weren't adjacent. Added compound escalation: if a HIGH military/conflict keyword matches AND the headline mentions a critical geopolitical target (Iran, Russia, China, Taiwan, NATO, US forces), escalate to CRITICAL. Also added missing Iran keyword variants (plural forms, "Iran retaliates/strikes"). * feat(conflict): enhance Iran events popup with severity badge and related events (#549) Rewrite the Iran events popup to follow the established popup pattern (conflict/protest) with severity-colored header, badge, close button, stat rows, and source link using CSS classes instead of inline styles. - Add normalizeSeverity helper (clamps unknown values to 'low') - Show related events from same location (normalized matching, max 5) - Add IranEventPopupData to PopupData union (removes unsafe double cast) - Add iranEvent header CSS with severity border-left colors - Add i18n keys for en/ar/fr * feat(telegram): add Telegram Intel panel (#550) * feat(telegram): add Telegram Intel panel consuming relay feed - Service layer: fetchTelegramFeed() with 30s cache, types matching relay shape - Panel component: topic filter tabs, safe DOM rendering via h()+replaceChildren() - DataLoader + RefreshScheduler pattern (60s interval, hidden-tab aware) - Handles enabled=false and empty states from relay - CSS following existing gdelt-intel pattern - Panel title localized across all 18 locales * fix(i18n): add components.telegramIntel translations to 10 remaining locales * feat(live-news): add residential proxy + gzip decompression for YouTube detection (#551) YouTube blocks Vercel datacenter IPs — returns HTML without videoDetails/isLive data. Switch from edge runtime to Node.js serverless to enable HTTP CONNECT tunnel proxy via YOUTUBE_PROXY_URL env var. Add zlib decompression for gzip responses (YouTube returns empty body without Accept-Encoding header). Also adds missing fallback video IDs for WELT, KTN News, CNA NewsAsia, and updates TBS NEWS DIG fallback. * debug(live-news): add debug param to diagnose proxy env var on Vercel * fix(live-news): set explicit runtime: 'nodejs' for proxy support Vercel defaults to edge runtime when not specified. node:http/https/zlib imports are unavailable in edge — causing FUNCTION_INVOCATION_FAILED. Remove debug param added in previous commit. * fix(live-news): lazy-load node modules + proxy fallback to direct fetch Top-level import of node:http/https/zlib crashes if Vercel bundles for edge despite runtime: 'nodejs' config. Use dynamic import() to lazy-load at call time. Also add try/catch around proxy so it falls back to direct fetch if proxy connection fails. * feat(aviation): integrate AviationStack API for non-US airport delays (#552) Replace 100% simulated delay data for international airports with real flight data from AviationStack API. Add 28 Middle East/conflict-zone airports (Iran, Iraq, Lebanon, Syria, Yemen, Pakistan, Libya, Sudan). Key changes: - AviationStack integration with bounded concurrency (5 parallel), rotating batch (20 airports/cycle), and 20s deadline - Redis SETNX lock prevents cross-isolate cache stampede on expiry - Split FAA/intl caches (both 30min TTL) with isolated error handling - Fix severity colors (was checking 'GS'/'GDP', now minor/moderate/major/severe) - Fix tooltip (was obj.airport, now obj.name + obj.iata) - Add FLIGHT_DELAY_TYPE_CLOSURE for airport/airspace closures - Add closure i18n key across all 18 locales - Graceful fallback: no API key → simulation; API failure → simulation * feat(live-news): move YouTube proxy scraping to Railway relay Vercel serverless cannot use node:http/https for HTTP CONNECT proxy tunnels. Move the residential proxy YouTube scraping to the Railway relay (ais-relay.cjs) which has full Node.js access. - Add /youtube-live route to relay with proxy + direct fetch fallback - Add 5-min in-memory cache for channel lookups, 1hr for oembed - Revert Vercel api/youtube/live.js to edge runtime — now proxies to Railway first, falls back to direct scrape * feat(settings): add AVIATIONSTACK_API to desktop settings page (#553) Register the key in Rust keychain (SUPPORTED_SECRET_KEYS), frontend RuntimeSecretKey type, feature toggle, and settings UI under "Tracking & Sensing" category. * fix(live-news): use correct relay auth header for YouTube proxy (#554) Edge function was sending X-Relay-Auth header with RELAY_AUTH_TOKEN env var, but the Railway relay expects x-relay-key header validated against RELAY_SHARED_SECRET. This mismatch caused the relay to reject requests from Vercel, falling back to direct YouTube scrape (which fails from datacenter IPs for many channels). * fix(live-news): align YouTube edge function with relay auth pattern (#555) Use same getRelayBaseUrl/getRelayHeaders as other edge functions: - WS_RELAY_URL env var instead of VITE_WS_API_URL - RELAY_SHARED_SECRET + RELAY_AUTH_HEADER for flexible auth - Dual x-relay-key + Authorization headers * fix(i18n): rename OREF Sirens panel to Israel Sirens (#556) Remove internal implementation references (OREF, proxy relay, oref.org.il) from all user-facing strings across 18 locales and panel config. * fix(live-news): annotate empty catches and sanitize error output (#560) - Add context comments to empty catch blocks for debuggability - Replace error.message leak with generic client-safe message * fix(sentry): add noise filters and fix beforeSend null-filename leak (#561) - Add 8 new ignoreErrors patterns: signal timeout, premium gate, hybridExecute/mag/webkit bridge injections, postMessage null, NotSupportedError, appendChild injection, luma assignment - Fix LIDNotify regex to match both LIDNotify and LIDNotifyId - Fix beforeSend: strip null/anonymous filename frames so deck.gl TypeErrors (28 events, 8 users) are properly suppressed * feat(cii): wire OREF sirens into CII score & country brief (#559) * feat(cii): wire OREF sirens into CII score and country brief Active OREF sirens now boost Israel's CII score through two channels: - Conflict component: +25 base + min(25, alertCount*5) for active sirens - Blended score: +15 for active sirens, +5/+10 for 24h history thresholds Country brief for Israel shows a siren signal chip when alerts are active. * refactor(cii): extract getOrefBlendBoost helper to DRY scoring paths * fix(relay): add graceful shutdown + poll concurrency guard for Telegram (#562) - SIGTERM/SIGINT handler disconnects Telegram client before container dies - telegramPollInFlight guard prevents overlapping poll cycles - Mid-poll AUTH_KEY_DUPLICATED now permanently disables (was reconnect loop) * fix(aviation): query all airports instead of rotating batch (#557) * fix(aviation): query all airports instead of rotating batch of 20 The rotating batch (20 airports/cycle) caused major airports like DXB (52% cancellations) to be missed entirely for multiple cache cycles. With a paid AviationStack plan, query all ~90 non-US airports per refresh with concurrency 10 and 50s deadline (~9 chunks × 5s = 45s). * feat(cii): feed airport disruptions into CII and country brief Major/severe airport delays and closures now boost the CII security score and appear as signal chips in country briefs. Only major+ severity alerts are ingested to avoid noise from minor delays. - Add aviationDisruptions to CountryData and ingestAviationForCII() - Boost security score: closure +20, severe +15, major +10, moderate +5 - Store flight delays in intelligenceCache for country brief access - Add aviation disruptions chip in country brief signals grid * fix(relay): replace smart quotes crashing relay on startup (#563) * fix(relay): replace Unicode smart quotes crashing Node.js CJS parser * fix(relay): await Telegram disconnect + guard startup poll * fix(cii): resolve Gulf country strike misattribution via multi-match bbox disambiguation (#564) Dubai/Doha/Bahrain/Kuwait coordinates matched Iran's bounding box first due to iteration order. Now collects ALL matching bboxes, disambiguates via isCoordinateInCountry() geometry, and falls back to smallest-area bbox. - Add BH, QA, KW, JO, OM to bounds tables (previously missing entirely) - Extract ME_STRIKE_BOUNDS + resolveCountryFromBounds() into country-geometry.ts - All 4 consumer files use shared constant (single source of truth) - Bump CDN cache-bust param for iran-events endpoint * fix(relay): upstreamWs → upstreamSocket in graceful shutdown (#565) * fix(relay): install curl in Railway container for OREF polling (#567) * fix(relay): increase Polymarket cache TTL to 10 minutes (#568) * fix(relay): increase Polymarket cache TTL to 10 minutes All requests were MISS with 2-min TTL under concurrent load. Bump to 10-min cache and 5-min negative cache to reduce upstream pressure. * fix(relay): normalize Polymarket cache key from canonical params Raw url.search as cache key meant ?tag=fed&endpoint=events and ?endpoint=events&tag_slug=fed produced different keys for the same upstream request — defeating both cache and inflight dedup, causing 121 MISS entries in 3 seconds. Build cache key from parsed canonical params (endpoint + sorted query string) so all equivalent requests share one cache entry. * feat(webcams): add Iran tab to live webcams panel (#569) Add dedicated Iran region tab as the first/default tab with 4 feeds: Tehran, Middle East overview, Tehran (alt angle), and Jerusalem. * fix(relay): replace nixpacks.toml with railpack.json for curl (#571) Railway uses Railpack (not Nixpacks). nixpacks.toml in scripts/ was silently skipped. Use railpack.json at repo root with deploy.aptPackages to install curl at runtime for OREF polling. * fix(webcams): replace duplicate Tehran feed with Tel Aviv, rename Iran tab (#572) - Remove duplicate iran-tehran2 feed (same channel/video as iran-tehran) - Remove iran-mideast feed - Add Tel Aviv feed (-VLcYT5QBrY) to Iran Attacks tab - Rename tab label from "IRAN" to "IRAN ATTACKS" across all 18 locales * feat(scripts): add Iran events seed script and latest data (#575) Add seed-iran-events.mjs for importing Iran conflict events into Redis (conflict:iran-events:v1). Includes geocoding by location keywords and category-to-severity mapping. Data file contains 100 events from 2026-02-28. * fix(relay): add timeouts and logging to Telegram poll loop (#578) GramJS getEntity/getMessages have no built-in timeout. When the first channel hangs (FLOOD_WAIT, MTProto stall), telegramPollInFlight stays true forever, blocking all future polls — zero messages collected, zero errors logged, frontend shows "No messages available". - Add 15s per-channel timeout on getEntity + getMessages calls - Add 3-min overall poll cycle timeout - Force-clear stuck in-flight flag after 3.5 minutes - Detect FLOOD_WAIT errors and break loop early - Log per-cycle summary: channels polled, new msgs, errors, duration - Track media-only messages separately (no text → not a bug) - Expose lastError, pollInFlight, pollInFlightSince on /status endpoint * feat(cii): hook security advisories into CII scoring & country briefs (#579) Travel advisories (Do Not Travel, Reconsider, Caution) from US, AU, UK, NZ now act as a floor and boost on CII scores. Do Not Travel guarantees a minimum score of 60 (elevated), Reconsider floors at 50. Multi-source corroboration (3+ govts) adds +5 bonus. Advisory chips appear in country brief signal grid with level-appropriate styling, and advisory context is passed to AI brief generation. - Extract target country from advisory titles via embassy feed tags and country name matching - Add advisoryMaxLevel/advisoryCount/advisorySources to CII CountryData - Wire ingestAdvisoriesForCII into data loader pipeline - Add travelAdvisories/travelAdvisoryMaxLevel to CountryBriefSignals - Render advisory signal chips in CountryBriefPage * fix(sentry): guard setView against invalid preset + filter translateNotifyError (#580) - DeckGLMap.setView(): early-return if VIEW_PRESETS[view] is undefined, preventing TypeError on 'longitude' when select value is invalid - Add ignoreErrors pattern for Google Translate widget crash * feat(relay): bootstrap OREF 24h history on startup (#582) * fix(relay): improve OREF curl error logging with stderr capture -s flag silenced curl errors. Add -S to show errors, capture stderr via stdio pipes, and log curl's actual error message instead of generic "Command failed" from execFileSync. * feat(relay): bootstrap OREF 24h history on startup and add missing headers - Fetch AlertsHistory.json once on startup to populate orefState.history immediately instead of starting empty - Add X-Requested-With: XMLHttpRequest header required by Akamai WAF - Add IST→UTC date converter handling DST ambiguity - Redact proxy credentials from error logs and client responses - Fix historyCount24h to count individual alert records, not snapshots - Guard historyCount24h reducer for both array and string data shapes - Add input validation to orefDateToUTC for malformed dates * feat(aviation): add comprehensive logging to flight delay pipeline (#581) Adds structured [Aviation] logging throughout the handler chain: - Handler: timing, FAA/intl alert counts - fetchIntlWithLock: cache hit/miss, lock status, fallback triggers - fetchAviationStackDelays: airport count, deadline hits, success/fail stats - fetchSingleAirport: per-airport flight stats, API errors, severity - Annotate empty catches with error context * feat: add GPS/GNSS jamming map layer + CII integration (#570) * feat: add GPS/GNSS jamming data ingestion from gpsjam.org - scripts/fetch-gpsjam.mjs: standalone fetcher that downloads daily H3 hex data, filters medium/high interference, converts to lat/lon via h3-js, and writes JSON. Can be run on cron. - api/gpsjam.js: Vercel Edge Function that proxies gpsjam.org data with 1hr cache, returns medium/high hexes for frontend consumption. - src/services/gps-interference.ts: frontend service that fetches from the Edge API, converts H3→lat/lon, and classifies by conflict region. - h3-js added as dependency for hex→coordinate conversion. * feat: add GPS jamming map layer, CII integration, and country brief signals Wire gpsjam.org data into map visualization, instability scoring, and country intelligence. ScatterplotLayer renders high (red) and medium (orange) interference hexes. CII security score incorporates jamming counts per country via h3→country geocoding with cache. Country briefs show jamming zone chip. Full i18n across 18 locales including popup labels. Data loads with intelligence signals cycle (15min), gated by 1hr client-side cache. * feat(aviation): add NOTAM closure detection via ICAO API (#583) * feat(aviation): add NOTAM closure detection via ICAO API Adds international airport closure detection via ICAO NOTAMs: - New fetchNotamClosures() queries ICAO realtime-notams endpoint - Detects closures via Q-codes (FA/AH/AL/AW/AC/AM) and text patterns - Batches airports in groups of 20 per API call - 4-hour cache TTL via cachedFetchJson (stampede-safe) - NOTAM closures override existing AviationStack alerts for same airport - Graceful: no ICAO_API_KEY env var = silently skipped To activate: set ICAO_API_KEY env var (register at dataservices.icao.int) * feat(settings): add ICAO_API_KEY to desktop app settings Adds ICAO NOTAM API key to the desktop settings UI: - Rust: SUPPORTED_SECRET_KEYS [23→24] - TypeScript: RuntimeSecretKey + RuntimeFeatureId unions - Feature definition: 'icaoNotams' in Tracking & Sensing category - Settings UI: label, signup URL, analytics name * feat(aviation): limit NOTAM queries to MENA airports only Per user request, ICAO NOTAM closure detection is scoped to Middle East airports only (region='mena', ~35 airports). This reduces API calls (2 batches vs 5) and focuses on the region where closures are most relevant. * fix(aviation): align NOTAM cache TTL to 30 min (matching FAA/intl) * feat(risk): wire theater posture + breaking news into strategic risk score (#584) The composite score ignored theater military buildup and breaking news alerts, causing misleadingly low scores during active military events. - Add theater boost from raw asset counts + strike capability (capped +25) - Add breaking news severity boost (critical=15, high=8, capped +15) - Listen for wm:breaking-news events with 30-min TTL and auto-expiry - Read cached theater postures via getCachedPosture() with stale discount - Lower trend threshold from ±5 to ±3 for faster escalation detection - Cleanup listeners and timers in destroy() * fix(relay): delay Telegram connect 60s on startup to prevent AUTH_KEY_DUPLICATED (#587) Railway zero-downtime deploys start the new container before the old one receives SIGTERM. Both containers connect with the same session string simultaneously, triggering Telegram's AUTH_KEY_DUPLICATED which permanently invalidates the session. A 60s startup delay gives the old container time to disconnect gracefully. Configurable via TELEGRAM_STARTUP_DELAY_MS env. * feat(feeds): add RT (Russia Today) RSS feeds (#585) - Add RT main feed (rt.com/rss/) and RT Russia desk (rt.com/rss/russia/) - Add to SOURCE_TIERS (tier 3), SOURCE_TYPES (wire), SOURCE_CREDIBILITY - Add rt.com to rss-proxy ALLOWED_DOMAINS * feat(live-news): add RT channel via HLS + enable HLS on web (#586) - Add RT (rt.com) as optional channel in Europe region - HLS stream: rt-glb.rttv.com/dvr/rtnews/playlist.m3u8 (CORS: *) - Enable native <video> HLS playback on web (was desktop-only) - Channels in DIRECT_HLS_MAP with CORS headers now work everywhere * fix(telegram): add missing relay auth headers to telegram-feed edge function (#590) The edge function was the only relay proxy missing RELAY_SHARED_SECRET auth headers. The relay returns 401 for all non-public routes, so the panel always received an error response → "No messages available". * fix(aviation): replace broken lock mechanism with direct cache, add cancellation tiers (#591) The Redis SETNX lock for AviationStack had a 3s wait but the API takes ~25s for 50 airports. Every lock-loser fell back to simulation — meaning AviationStack data was never served despite the API key being configured. Changes: - Remove fetchIntlWithLock/tryAcquireLock, use getCachedJson/setCachedJson with conditional TTL (30min real data, 5min simulation fallback) - Add cancellation severity tiers: ≥20% moderate, ≥10% minor (DXB at 32% cancelled was previously dropped as null) - Bump cache key to v2 to invalidate stale simulation data - Add HTML content-type detection for NOTAM API (ICAO CloudFront returns HTML challenge pages from Vercel datacenter IPs) * fix(relay): stop Polymarket cache stampede from concurrent limit + CDN bypass (#592) Three issues caused continuous MISS every 5 seconds: 1. Concurrent limit rejection poisoned cache: 11 tags fire via Promise.all but POLYMARKET_MAX_CONCURRENT=3, so 8 tags got negative-cached with empty [] (5 min TTL). Those 8 tags NEVER got positive cache because they were always throttled. Fix: replace reject-with-negative-cache with a proper queue — excess requests wait for a slot instead of being silently rejected. 2. Cache key fragmentation: fetchPredictions(limit=20) and fetchCountryMarkets(limit=30) created separate cache entries for the same tag. Fix: normalize to canonical limit=50 upstream, cache key is shared regardless of caller's requested limit. 3. CDN bypass: end_date_min timestamp in query string made every URL unique, preventing Vercel CDN caching entirely. Fix: strip end_date_min, active, archived from proxy params (relay ignores them anyway). * fix(polymarket): add queue backpressure and response limit slicing (#593) - Add POLYMARKET_MAX_QUEUED=20 cap to prevent unbounded queue growth under sustained load (rejects with negative cache when full) - Use requestedLimit to slice cached responses — callers requesting limit=20 now get 20 items instead of the full 50-item upstream payload - Hoist PROXY_STRIP_KEYS Set to module level (avoids per-call allocation) * fix(webcams): add 4th Iran Attacks feed to fill 2x2 grid (#601) Add Middle East multi-cam (4E-iFtUM2kk) as 4th Iran region feed. Previously only 3 feeds for a 4-cell grid, leaving one cell black. * fix(aviation): route NOTAM through relay + improve intl logging (#599) Root cause: ICAO NOTAM API times out from Vercel edge (>10s). AviationStack alerts indistinguishable from simulation in logs. Changes: - Add /notam proxy endpoint to Railway relay (25s timeout, 30min cache) - Route fetchNotamClosures through relay when WS_RELAY_URL is set - Fall back to direct ICAO call (20s timeout) when no relay - Log cache hits with real vs simulated alert counts - Send all MENA airports in single NOTAM request (was batched by 20) Requires: ICAO_API_KEY env var on Railway relay * chore(telegram): update channel list — remove nexta_live, air_alert_ua; add wfwitness (#600) * fix(sentry): guard YT player methods + filter GM/InvalidState noise (#602) - Guard loadVideoById/cueVideoById with typeof check in LiveNewsPanel (race condition: YT.Player object exists before onReady fires) - Add ignoreErrors for GM_getValue (Greasemonkey extension) and InvalidStateError (IndexedDB/DOM state errors from browser internals) * fix(aviation): always show all monitored airports on flight delays map (#603) * fix(aviation): always show MENA airports on map regardless of delay status MENA airports only appeared when they had active delays/closures. GCC airports like DOH, AUH, RUH were invisible during normal operations. Now fills in "normal operations" entries for all MENA airports without alerts so they always render as gray dots on the flight delays layer. * fix(aviation): show all monitored airports globally, not just MENA Extend normal-operations fill to all 128 monitored airports worldwide, not just the 35 MENA airports. Any airport without active delays now appears as a gray dot on the flight delays map layer. * fix(webcams): fix broken live news channels — eNCA handle, remove VTC NOW, fix CTI News (#604) - eNCA: fix YouTube handle from @eNCA to @encanews - VTC NOW: remove — VTC shut down Jan 2025 (Vietnam govt restructuring) - CTI News: remove stale fallbackVideoId and useFallbackOnly, let auto-detect work * docs(readme): comprehensive update for Telegram, OREF, GPS jamming, airports, and more (#606) Reflects 30+ recent commits with updates across the README: - Update data layer counts (36 → 40+), webcam counts (19 → 22), data sources (16 → 28+) - Add 6 new How It Works sections: Telegram OSINT, OREF Rocket Alerts, GPS/GNSS Interference, Security Advisories, Airport Delays/NOTAMs, Strategic Risk Score - Expand Railway relay section with service table (AIS, OpenSky, Telegram, OREF, Polymarket, NOTAM) and update architecture ASCII diagram - Add Iran/Attacks webcam tab, HLS native streaming details, RT coverage - Add Algorithmic Design Decisions section (log vs linear scoring, Welford's, H3 hex grids, cosine-lat correction, negative caching) - Add 2 new architecture principles (graceful degradation, multi-source corroboration) - Add 8 new roadmap items for recently shipped features - Update Tech Stack table with Telegram, OREF, gpsjam.org in Geopolitical APIs * chore: bump version to 2.5.21 (#605) * fix(aviation): invalidate stale IndexedDB cache + reduce CDN TTL (#607) Bump circuit breaker name from 'FAA Flight Delays' to 'Flight Delays v2' to force all clients to discard stale IndexedDB entries that predate PR #603 (which added normal-ops airport fill). Also downgrade CDN cache tier from 'slow' (15 min) to 'medium' (5 min) since airport status changes more frequently than other slow-tier endpoints. * fix(relay): increase OREF curl maxBuffer to prevent ENOBUFS (#609) * fix(relay): increase OREF curl maxBuffer to 10MB to prevent ENOBUFS AlertsHistory.json response exceeds execFileSync default 1MB buffer, causing spawnSync ENOBUFS on Railway container at startup. * fix(relay): use curl -o tmpfile for OREF history instead of stdout buffer Large AlertsHistory.json overflows execFileSync stdout buffer (ENOBUFS). Now writes to temp file via curl -o, reads with fs.readFileSync, cleans up. Live alerts (tiny payload) still use stdout path. * fix: RT channel HLS-only recovery, test shim, and LiveNOW fallback (#610) - Remove useFallbackOnly from RT channel — RT is HLS-only (banned from YouTube), so the flag was causing undefined videoId on HLS failure instead of graceful offline state - Add response-headers shim to redis-caching test so military flights bbox tests can import list-military-flights.ts - Restore LiveNOW from FOX fallbackVideoId (removed in channel audit) * fix(docs): add blank lines after CHANGELOG headings for markdownlint (#608) * Expand country brief and CII signal coverage (#611) * perf(rss): raise news refresh interval to 10min and cache TTL to 20min (#612) Reduces upstream RSS polling frequency and extends client-side cache lifetime to lower API load and bandwidth usage. * fix(api): harden cache-control headers for polymarket and rss-proxy (#613) * docs(changelog): add v2.5.21 entry covering 86 merged PRs (#616) Comprehensive changelog for 2026-03-01 release including Iran Attacks layer, Telegram Intel panel, OREF sirens, GPS jamming, AviationStack, breaking news alerts, and strategic risk score. * fix(aviation): increase cache TTL from 30min to 2h to reduce API quota usage (#617) * feat(oref): show history waves timeline with translation and NaN fix (#618) - Fetch and display alert history waves in OrefSirensPanel (cap 50 most recent) - Last-hour waves highlighted with amber border and RECENT badge - Translate Hebrew history alerts via existing translateAlerts pipeline - Guard formatAlertTime/formatWaveTime against NaN from unparseable OREF dates - Cap relay history bootstrap to 500 records - Add 3-minute TTL to prevent re-fetching history on every 10s poll - Remove dead .oref-footer/.oref-history CSS; add i18n key for history summary * feat(map): native mobile map experience with location detection and full feature parity (#619) - Fix URL restore: lat/lon now override view center when explicitly provided - Fix touch scroll: 8px threshold before drag activation, preventDefault once active - Add location bootstrap: timezone-first detection, optional geolocation upgrade - Enable DeckGL on mobile with deviceMemory capability guard - Add DeckGL state sync on moveend/zoomend for URL param updates - Fix breakpoint off-by-one: JS now uses <= to match CSS max-width: 768px - Add country-click on SVG fallback with CSS transform inversion - Add fitCountry() to both map engines (DeckGL uses fitBounds, SVG uses projection) - Add SVG inertial touch animation with exponential velocity decay - Add mobile map e2e tests for timezone, URL restore, touch, and breakpoint * fix(sentry): guard pauseVideo optional chaining + add 4 noise filters (#624) - Fix: `this.player.pauseVideo()` → `this.player.pauseVideo?.()` at LiveNewsPanel line 1301 (6 Sentry issues, 33 events) - Noise: Chrome extension "Could not establish connection" - Noise: Safari "webkitCurrentPlaybackTargetIsWireless" internal - Noise: Sentry SDK crash on iOS "a.theme" - Noise: Broaden Can't find variable to include EmptyRanges - Noise: Catch stale cached builds with /this.player.\w+ is not a function/ * fix(aviation): prevent AviationStack API quota blowout (#623) - Replace manual getCachedJson/setCachedJson with cachedFetchJson for intl delays — prevents thundering herd (concurrent misses each firing 93 API calls independently) - Bump Redis cache TTL from 30min to 2h - Bump frontend polling from 10min to 2h to match server cache - Bump circuit breaker browser cache from 5min to 2h - Bump CDN edge tier from medium (5min) to static (1h) - Bump cache key to v3 to force fresh entry with new TTL * feat(news): server-side feed aggregation to reduce edge invocations by ~95% (#622) Phase 1: Force CDN caching on rss-proxy (s-maxage=300 for 2xx, short TTL for errors) — fixes bug where upstream no-cache headers were passed through verbatim, defeating Vercel CDN. Phase 2: Add ListFeedDigest RPC that aggregates all feeds server-side into a single Redis-cached response. Client makes 1 request instead of ~90 per cycle. Includes circuit breaker with persistent cache fallback, per-feed AI reclassification, and headline ingestion parity. Phase 3: Increase polling interval from 5min to 7min to offset CDN cache alignment. New files: - proto/worldmonitor/news/v1/list_feed_digest.proto - server/worldmonitor/news/v1/{_feeds,_classifier,list-feed-digest}.ts - src/services/ai-classify-queue.ts (extracted from rss.ts) * feat(rss): add conditional GET (ETag/If-Modified-Since) to Railway relay (#625) When RSS cache expires, send If-None-Match/If-Modified-Since headers on revalidation. Upstream 304 responses refresh the cache timestamp and serve cached body with zero bandwidth, cutting egress ~80-95% for feeds that support conditional GET. * fix(map): sync layer toggles to URL for shareable links (#576) (#621) Layer toggles were not updating the browser URL, so shared links would not carry the user's current layer state. Extracted the debounced URL sync to a reusable class method and call it on every layer change. * chore(cache): bump all sub-5min cache TTLs and polling intervals (#626) Audit and raise every cache/poll under 5 min to reduce upstream API pressure and Vercel function invocations across 31 files: Frontend polling: markets 4→8min, crypto 4→8min, predictions 5→10min, natural 5→1hr, cableHealth 5→2hr, oref 10s→2min, maritime 30s→5min, IntelligenceGapBadge 1→3min. Circuit breakers: PizzINT 2→30min, aviation 5→20min, seismology/weather/ outages/statuses/wildfires 5→30min, military-vessels/flights/GDACS/ maritime/polymarket/GDELT 5→10min, chokepoint 5→20min. Server Redis: market-quotes 2→8min, stablecoins 2→8min/5→10min, vessel-snapshot 10s→5min, earthquakes 5→30min, sector 5→10min, predictions/crypto/commodities 5→10min, outages/statuses 5→30min, macro-signals/chokepoints 5→15min, aviation-sim 5→15min. CDN edge: market RPCs fast→medium, infra/seismology fast→slow. * fix(military): narrow ICAO hex ranges to stop civilian false positives (#627) * fix(military): narrow hex ranges and callsign regex to stop civilian false positives (#462) MILITARY_HEX_RANGES used entire country ICAO allocations instead of military-specific sub-ranges (sourced from tar1090-db/ranges.json). This flagged ALL commercial aircraft from Italy, Spain, Japan, India, South Korea, etc. as military activity. Key changes: - Remove A00000-A3FFFF (US civilian N-numbers) — military starts at ADF7C8 - Italy 300000-33FFFF → 33FF00-33FFFF (top 256 codes only) - Spain 340000-37FFFF → 350000-37FFFF (upper 3/4 confirmed military) - Japan 840000-87FFFF removed (no confirmed JASDF sub-range) - France narrowed to 3AA000-3AFFFF + 3B7000-3BFFFF - Germany narrowed to 3EA000-3EBFFF + 3F4000-3FBFFF - India 800000-83FFFF → 800200-8002FF (256 codes) - Canada C00000-C0FFFF → C20000-C3FFFF (upper half) - Remove unconfirmed: South Korea, Sweden, Singapore, Pakistan - Add confirmed: Austria, Belgium, Switzerland, Brazil - Drop overly broad /^[A-Z]{4,}\d{1,3}$/ callsign regex from server * fix(military): constrain short prefixes + add classification tests Move ambiguous 2-letter prefixes (AE, RF, TF, PAT, SAM, OPS, CTF, IRG, TAF) to SHORT_MILITARY_PREFIXES — these now only match when followed by a digit (e.g. AE1234=military, AEE123=Aegean Airlines). Add 97-case test suite covering: - Military callsign detection (19 known patterns) - Short prefix digit gating (6 cases) - Civilian airline non-detection (26 airlines) - Short prefix letter rejection (6 cases) - Military hex range boundaries (7 confirmed ranges) - Civilian hex non-detection (19 codes) - Boundary precision (ADF7C8 start, 33FF00 start, etc.) - No-full-allocation guard (10 countries) * fix: use charAt() instead of bracket indexing for strict TS * fix(sidecar): add AVIATIONSTACK_API and ICAO_API_KEY to env allowlist (#632) Both keys were added to Rust SUPPORTED_SECRET_KEYS and runtime-config.ts but the sidecar's own ALLOWED_ENV_KEYS was never updated. This caused "key not in allowlist" 403 when saving/verifying these keys from the desktop settings UI. Also adds AviationStack API validation in validateSecretAgainstProvider. * fix(linux): sanitize env for xdg-open in AppImage (#631) * fix(desktop): backoff on errors to stop CPU abuse + shrink settings window (#633) Three bugs combine to burn 130% CPU when sidecar auth fails: 1. RefreshScheduler resets backoff multiplier to 1 (fastest) on error, causing failed endpoints to poll at base interval instead of backing off. Fix: exponential backoff on errors, same as unchanged-data path. 2. classify-event batch system ignores 401 (auth failure) — only pauses on 429/5xx. Hundreds of classify calls fire every 2s, each wasted. Fix: pause 120s on 401, matching the 429/5xx pattern. 3. Fetch patch retries every 401 (refresh token + retry), doubling all requests to the sidecar even when token refresh consistently fails. Fix: 60s cooldown after a retry-401 still returns 401. Also shrinks settings window from 760→600px (min 620→480) to reduce the empty whitespace below content on all tabs. * feat: implement deduct situation feature (#636) * feat: Add new intelligence panels for situation deduction, strategic posture, and tech events with associated backend services and E2E tests. * fix(deduction): desktop-only gating, XSS fix, and code review fixes - Gate DeductionPanel creation behind isDesktopApp check - Gate deduce buttons in StrategicPosturePanel and TechEventsPanel behind isDesktopRuntime() - Remove deduction from FULL_PANELS (runtime-only on desktop) - Add DOMPurify to sanitize marked.parse() output before innerHTML - Revert shared _shared.ts: restore UPSTREAM_TIMEOUT_MS=30_000, hardcoded GROQ_API_URL/MODEL - Define local DEDUCT_TIMEOUT_MS=120_000, DEDUCT_API_URL, DEDUCT_MODEL in handler - Fix dead API key guard: remove 'ollama-local' fallback, skip if no key - Strip all console.log from deduct-situation handler (keep console.error) - Bump max_tokens 500→1500 for meaningful analysis - Fix cache key: deduct:sebuf→deduct:situation - Add 5s client-side rate limit cooldown on submit button - Move @types/marked to devDependencies, add dompurify + @types/dompurify - Add DeductContextDetail shared interface in types/index.ts - Extract buildNewsContext helper to src/utils/news-context.ts - Pass getLatestNews to StrategicPosturePanel constructor - Fix E2E test: .deduction-markdown-content→.deduction-result - Revert emoji escapes in commands.ts to unicode format - Revert variant-switcher whitespace in panel-layout.ts * fix: resolve IranEvent timestamp type mismatches (string vs number) Generated IranEvent.timestamp is string but consumers expected number. - Coerce timestamps via Number() in data-loader before passing to aggregators - Accept string|number in MapPopup IranEventPopupData interface - Fix scrapedAt fallback type in conflict/index.ts - Coerce timestamp in country-intel.ts and DeckGLMap.ts sort * fix: restore generated code from main, keep intelligence deductSituation RPC Reverts all src/generated/ files to origin/main except the intelligence service client/server which retains the new deductSituation RPC added by this PR. Fixes scrapedAt type to match generated number type. * fix: add missing listGulfQuotes handler from main The Gulf Economies panel (PR #667) added this RPC after the feature branch was forked. Bring handler and registration from main. * fix: address code review findings (P1+P2+P3) - P1: Fix event listener leak — store handler ref, remove in destroy() - P1: Prevent cooldown bypass via requestSubmit when button disabled - P2: Read process.env inside function body for runtime secret updates - P3: Use \u{1F9E0} unicode escape in StrategicPosturePanel (consistency) --------- Co-authored-by: Elie Habib <elie.habib@gmail.com> Co-authored-by: Sebastien Melki <sebastien@anghami.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: Konstanty Szumigaj <kostekszumigaj@gmail.com> Co-authored-by: Chris Chen <fuleinist@users.noreply.github.com> Co-authored-by: facusturla <facusturla@users.noreply.github.com> Co-authored-by: karim <mirakijka@gmail.com> Co-authored-by: maxime.io <maxime.de.visscher@gmail.com> Co-authored-by: Elias El Khoury <efk@anghami.com> Co-authored-by: Fayez Bast <96446332+FayezBast@users.noreply.github.com>
107 lines
5.7 KiB
JSON
107 lines
5.7 KiB
JSON
{
|
|
"name": "world-monitor",
|
|
"private": true,
|
|
"version": "2.5.21",
|
|
"license": "AGPL-3.0-only",
|
|
"type": "module",
|
|
"scripts": {
|
|
"lint:md": "markdownlint-cli2 '**/*.md' '!.agent/**' '!.agents/**' '!.claude/**' '!.factory/**' '!.windsurf/**' '!skills/**' '!docs/internal/**' '!docs/Docs_To_Review/**'",
|
|
"version:sync": "node scripts/sync-desktop-version.mjs",
|
|
"version:check": "node scripts/sync-desktop-version.mjs --check",
|
|
"dev": "vite",
|
|
"dev:tech": "cross-env VITE_VARIANT=tech vite",
|
|
"dev:finance": "cross-env VITE_VARIANT=finance vite",
|
|
"dev:happy": "cross-env VITE_VARIANT=happy vite",
|
|
"build": "tsc && vite build",
|
|
"build:sidecar-sebuf": "node scripts/build-sidecar-sebuf.mjs",
|
|
"build:desktop": "node scripts/build-sidecar-sebuf.mjs && tsc && vite build",
|
|
"build:full": "cross-env-shell VITE_VARIANT=full \"tsc && vite build\"",
|
|
"build:tech": "cross-env-shell VITE_VARIANT=tech \"tsc && vite build\"",
|
|
"build:finance": "cross-env-shell VITE_VARIANT=finance \"tsc && vite build\"",
|
|
"build:happy": "cross-env-shell VITE_VARIANT=happy \"tsc && vite build\"",
|
|
"typecheck": "tsc --noEmit",
|
|
"typecheck:api": "tsc --noEmit -p tsconfig.api.json",
|
|
"typecheck:all": "tsc --noEmit && tsc --noEmit -p tsconfig.api.json",
|
|
"tauri": "tauri",
|
|
"preview": "vite preview",
|
|
"test:e2e:full": "cross-env VITE_VARIANT=full playwright test",
|
|
"test:e2e:tech": "cross-env VITE_VARIANT=tech playwright test",
|
|
"test:e2e:finance": "cross-env VITE_VARIANT=finance playwright test",
|
|
"test:e2e:runtime": "cross-env VITE_VARIANT=full playwright test e2e/runtime-fetch.spec.ts",
|
|
"test:e2e": "npm run test:e2e:runtime && npm run test:e2e:full && npm run test:e2e:tech && npm run test:e2e:finance",
|
|
"test:data": "tsx --test tests/*.test.mjs tests/*.test.mts",
|
|
"test:feeds": "node scripts/validate-rss-feeds.mjs",
|
|
"test:sidecar": "node --test src-tauri/sidecar/local-api-server.test.mjs api/_cors.test.mjs api/youtube/embed.test.mjs api/cyber-threats.test.mjs api/usni-fleet.test.mjs scripts/ais-relay-rss.test.cjs api/loaders-xml-wms-regression.test.mjs",
|
|
"test:e2e:visual:full": "cross-env VITE_VARIANT=full playwright test -g \"matches golden screenshots per layer and zoom\"",
|
|
"test:e2e:visual:tech": "cross-env VITE_VARIANT=tech playwright test -g \"matches golden screenshots per layer and zoom\"",
|
|
"test:e2e:visual": "npm run test:e2e:visual:full && npm run test:e2e:visual:tech",
|
|
"test:e2e:visual:update:full": "cross-env VITE_VARIANT=full playwright test -g \"matches golden screenshots per layer and zoom\" --update-snapshots",
|
|
"test:e2e:visual:update:tech": "cross-env VITE_VARIANT=tech playwright test -g \"matches golden screenshots per layer and zoom\" --update-snapshots",
|
|
"test:e2e:visual:update": "npm run test:e2e:visual:update:full && npm run test:e2e:visual:update:tech",
|
|
"desktop:dev": "npm run version:sync && cross-env VITE_DESKTOP_RUNTIME=1 tauri dev -f devtools",
|
|
"desktop:build:full": "npm run version:sync && cross-env VITE_VARIANT=full VITE_DESKTOP_RUNTIME=1 tauri build",
|
|
"desktop:build:tech": "npm run version:sync && cross-env VITE_VARIANT=tech VITE_DESKTOP_RUNTIME=1 tauri build --config src-tauri/tauri.tech.conf.json",
|
|
"desktop:build:finance": "npm run version:sync && cross-env VITE_VARIANT=finance VITE_DESKTOP_RUNTIME=1 tauri build --config src-tauri/tauri.finance.conf.json",
|
|
"desktop:package:macos:full": "node scripts/desktop-package.mjs --os macos --variant full",
|
|
"desktop:package:macos:tech": "node scripts/desktop-package.mjs --os macos --variant tech",
|
|
"desktop:package:windows:full": "node scripts/desktop-package.mjs --os windows --variant full",
|
|
"desktop:package:windows:tech": "node scripts/desktop-package.mjs --os windows --variant tech",
|
|
"desktop:package:macos:full:sign": "node scripts/desktop-package.mjs --os macos --variant full --sign",
|
|
"desktop:package:macos:tech:sign": "node scripts/desktop-package.mjs --os macos --variant tech --sign",
|
|
"desktop:package:windows:full:sign": "node scripts/desktop-package.mjs --os windows --variant full --sign",
|
|
"desktop:package:windows:tech:sign": "node scripts/desktop-package.mjs --os windows --variant tech --sign",
|
|
"desktop:package": "node scripts/desktop-package.mjs"
|
|
},
|
|
"devDependencies": {
|
|
"@playwright/test": "^1.52.0",
|
|
"@tauri-apps/cli": "^2.10.0",
|
|
"@types/canvas-confetti": "^1.9.0",
|
|
"@types/d3": "^7.4.3",
|
|
"@types/dompurify": "^3.0.5",
|
|
"@types/marked": "^5.0.2",
|
|
"@types/maplibre-gl": "^1.13.2",
|
|
"@types/papaparse": "^5.5.2",
|
|
"@types/topojson-client": "^3.1.5",
|
|
"@types/topojson-specification": "^1.0.5",
|
|
"cross-env": "^10.1.0",
|
|
"esbuild": "^0.27.3",
|
|
"h3-js": "^4.4.0",
|
|
"markdownlint-cli2": "^0.20.0",
|
|
"tsx": "^4.21.0",
|
|
"typescript": "^5.7.2",
|
|
"vite": "^6.0.7",
|
|
"vite-plugin-pwa": "^1.2.0"
|
|
},
|
|
"dependencies": {
|
|
"@deck.gl/aggregation-layers": "^9.2.6",
|
|
"@deck.gl/core": "^9.2.6",
|
|
"@deck.gl/geo-layers": "^9.2.6",
|
|
"@deck.gl/layers": "^9.2.6",
|
|
"@deck.gl/mapbox": "^9.2.6",
|
|
"@sentry/browser": "^10.39.0",
|
|
"@upstash/ratelimit": "^2.0.8",
|
|
"@upstash/redis": "^1.36.1",
|
|
"@vercel/analytics": "^1.6.1",
|
|
"@xenova/transformers": "^2.17.2",
|
|
"canvas-confetti": "^1.9.4",
|
|
"convex": "^1.32.0",
|
|
"d3": "^7.9.0",
|
|
"deck.gl": "^9.2.6",
|
|
"dompurify": "^3.1.7",
|
|
"fast-xml-parser": "^5.3.7",
|
|
"i18next": "^25.8.10",
|
|
"i18next-browser-languagedetector": "^8.2.1",
|
|
"maplibre-gl": "^5.16.0",
|
|
"marked": "^17.0.3",
|
|
"onnxruntime-web": "^1.23.2",
|
|
"papaparse": "^5.5.3",
|
|
"telegram": "^2.26.22",
|
|
"topojson-client": "^3.1.0",
|
|
"ws": "^8.19.0",
|
|
"youtubei.js": "^16.0.1"
|
|
},
|
|
"overrides": {
|
|
"fast-xml-parser": "^5.3.7"
|
|
}
|
|
}
|