❄️Festive Industry Pulse
Edge:
πŸ“…
🧠 Self-improving β€” relevance scoring learns from what gets read 0 signals recorded

βš™οΈ Strategy-Driven Intelligence System

Festive Industry Pulse scans 9 Edge Zones β€” each defined by a competitive edge in Festive's business. Sources are mapped to the zones they serve. Articles are relevance-scored (0-100) before appearing β€” junk below threshold is dropped.

πŸ“‘ Method 1: RSS Feeds

Structured XML feeds from publishers. Most reliable β€” headline, URL, summary, date. Sources matched to zones that need them.

πŸ—ΊοΈ Method 2: XML Sitemaps

Sitemaps scanned for recently modified URLs (last 7 days). Catches content not in RSS.

🌐 Method 3: HTML Index

Source homepage parsed for article-like links. Fallback for sites without RSS.

🎯 Relevance Scoring

Each article scored 0-100 against CORE_TERMS (+8), BOOST_TERMS (+3), PENALTY_TERMS (-15). Below threshold = dropped. Scores improve over time as you read articles β€” clicked articles boost their keywords.

🧠 Learning Over Time

When you click an article, the keywords in its headline and summary get a small boost in the scoring system. Over time, the feed learns which topics you care about most. Article reads are stored in localStorage and used to nudge relevance scores up for similar content.

πŸ” SHA256 Dedup

Every article hashed: SHA256(source_url:headline). Same headline from same source via different methods = one article only.

⏰ Schedule

Cron daily at 7:00am NZST. RSS β†’ Sitemap β†’ HTML per source. Full scan ~2 min. Data hot-reloads without container restart.