DNU - Dylan's New York Updates

DNU is a focused NYC news dashboard that pulls fresh local headlines from major New York sources, filters them to the most recent 12-hour window, and presents them in a fast, readable interface.

The project is intentionally small and operational: a Flask backend collects RSS feeds in parallel, exposes a clean JSON endpoint, and serves a responsive front end with local read-state tracking.

What It Does

• Aggregates recent NYC headlines from NYT, THE CITY, Gothamist, NY Post, and NY Daily News.
• Filters results to the latest 12 hours and limits each source to the five newest articles.
• Keeps source failures isolated so one broken feed does not take down the dashboard.
• Provides a browser-friendly /api/feeds endpoint for reuse in other tools.
• Stores read/unread state locally in the browser with no account system required.

Stack

Layer	Technology
Backend	Python, Flask
Data	RSS feeds via feedparser and requests
UI	HTML, CSS, vanilla JavaScript
Runtime	Replit-compatible Python app

Architecture

RSS sources
    -> parallel fetch workers
    -> 12-hour freshness filter
    -> normalized article JSON
    -> Flask API + responsive dashboard

The backend keeps source configuration in main.py, fetches each feed with browser-like headers, normalizes article titles and URLs, and returns a single timestamped payload for the front end.

API

GET /api/feeds

Example response shape:

{
  "fetched_at": "2026-06-01T00:00:00+00:00",
  "sources": [
    {
      "id": "gothamist",
      "name": "Gothamist",
      "site": "https://www.gothamist.com",
      "items": [
        {
          "title": "Example headline",
          "url": "https://example.com/story",
          "published": "2026-06-01T00:00:00+00:00"
        }
      ]
    }
  ]
}

Run Locally

pip install -r requirements.txt
python main.py

Open http://localhost:8080.

Health check:

curl http://localhost:8080/healthz

Run on Replit

1. Create a new Repl and import this repository from GitHub.
2. Use the Python template if prompted.
3. Click Run. The app listens on port 8080; Replit maps it to the public URL.
4. Deploy from Replit when you want a persistent public endpoint.

Repository Map

Path	Purpose
main.py	Flask app, RSS fetching, /api/feeds, /healthz
templates/index.html	Dashboard markup and article templates
static/style.css	Responsive blue circuit-board visual system
static/app.js	Fetching, rendering, refresh, local read state
requirements.txt	Python dependencies
.replit, replit.nix	Replit runtime configuration

Roadmap

• Add source freshness badges and per-source latency.
• Add deploy screenshots for the README.
• Add configurable source groups for borough, politics, transit, and culture feeds.
• Add basic tests around feed normalization and cutoff filtering.

Notes

• Read state is stored per browser with localStorage.
• Feeds that return no recent items display an empty-source state.
• Feed errors are returned per source so the rest of the dashboard stays useful.