News Aggregator - Project Context

Overview

A React-based news aggregator with global intelligence features, built with Vite and deployed on Netlify.

Tech Stack

• Frontend: React 18, TypeScript, TailwindCSS, React Query
• Backend: Netlify Functions (serverless)
• APIs: GDELT (news data), MyMemory (translation), Pinecone (semantic search)
• Routing: React Router v6

Project Structure

src/
├── api/           # API hooks (React Query)
├── components/
│   ├── articles/  # Article display components
│   ├── filters/   # Topic and date filters
│   ├── global/    # Global intelligence widgets
│   ├── layout/    # Header, Sidebar, MainLayout
│   └── search/    # Search modal
├── config/        # Country configurations
├── hooks/         # Custom React hooks
├── pages/         # Route pages (HomePage, GlobalDashboard, TopicPage)
├── types/         # TypeScript type definitions
└── utils/         # Utility functions

netlify/functions/  # Serverless functions (.cjs files)

Key Features

Global News Tab (/global)

• Components: GlobalDashboard.tsx, TrendingByCountry.tsx, RegulatoryRadar.tsx, EuropeMap.tsx
• API: global-intel.cjs serverless function
• Actions:
  • trending - Fetches trending topics for a country from GDELT
  • regulatory - Fetches regulatory news for a country/topic
  • trending-batch - Returns pre-computed heat data for Europe map (optimized for speed)

Language & Translation

• Language selector supports: EN, FR, DE, ES, IT, PT, NL, PL
• Each language has domain-specific news sources (e.g., lemonde.fr for French)
• Translation toggle uses MyMemory API to translate titles to English
• Original titles preserved in article.originalTitle

Topic Categories

Valid TrendingCategory values: technology, business, politics, science, culture, sports

Netlify Functions

Function	Purpose
global-intel.cjs	Global trending, regulatory data, batch map data
rss-proxy.cjs	RSS feed proxy
gdelt-proxy.cjs	GDELT API proxy
pinecone-search.cjs	Semantic search via Pinecone
ft-scraper.cjs	Financial Times scraping
finance-scraper.cjs	Finance data scraping

Environment Variables

Required in .env.local:

• PERPLEXITY_API_KEY
• PINECONE_API_KEY
• PINECONE_INDEX_HOST
• FT_MYFT_RSS_URL
• FT_SESSION_COOKIE

Development Commands

npm run dev          # Start Vite dev server (port 5173)
npx netlify dev      # Start with Netlify functions (port 8888)
npm run build        # Production build

Common Patterns

GDELT Filtering

Articles are filtered for language using:

1. URL path patterns (e.g., /afrique/, /mundo/)
2. Character detection (non-Latin scripts)
3. Common word matching (language-specific)

Category Priority

Articles are assigned ONE category based on priority order: politics > technology > business > science > sports > culture

Known Considerations

• GDELT has a 250 record limit per request
• Netlify functions have 30-second timeout
• Europe map uses pre-computed heat values (not real-time) for performance
• Translation limited to 5 articles per topic to avoid rate limits