updates to the news feed

First, let's handle the news feed like we do the people, publications, and software pages (via spreadsheets).

Second, I'd like to add a bluesky feed (to the news page) for @contextlab.bsky.social that automatically pulls from bluesky to show a themed, scrollable list of the latest posts from the lab account.

also add new (brief) news articles (sort so most recent appears at the top):

January 10, 2025:
Title: Why are we so preoccupied by the past? Content: Check out our reflection on our recent paper on prediction and retrodiction in Scientific American; thumbnail:
August 1, 2025:
Title: Claudia Gonciulea joins the lab! Content: We've managed to recruit superstar-scientist Claudia Gonciulea to our incoming PhD student cohort at Dartmouth!
Link below: [READ CLAUDIA'S BIO]
Thumbnail:
July 1, 2024:
Title: Jeremy gets tenure! Content: Jeremy managed to sneak past the tenure review committee-- they never even noticed he was actually three kittens in a lab coat! He even picked up some awards along the way.
Thumbnail:
January 25, 2022:
Title: We CAREER'd! Content: The CDL was awarded an NSF CAREER Award to map and enhance how students learn, using natural language processing models. We're excited to do some awesome science with our new-found riches!

Thumbnail:

Implementation Plan for Issue #9: News Feed Updates

Overview

This issue has three main components:

Convert news page to spreadsheet-based generation (like people, publications, software)
Add Bluesky feed integration for @contextlab.bsky.social
Add 4 new news articles from the comment

Phase 1: Spreadsheet-Based News Feed

1.1 Create Data Structure

Files to create:

data/news.xlsx - Single sheet with columns:
- date (YYYY-MM-DD format for sorting)
- title - News headline
- title_url - Optional link for the headline
- content - HTML-safe content (can include links)
- links_html - Additional links (e.g., "[READ BIO]")
- image - Thumbnail filename

Migrate existing news items from news.html to spreadsheet

1.2 Create Build Script

File: scripts/build_news.py

Follow pattern of build_people.py and build_publications.py
Load data from data/news.xlsx
Sort by date (most recent first)
Generate HTML for each news item
Inject into template

Functions:

load_news(xlsx_path) - Load and sort news items
generate_news_item(item) - Generate HTML for single item
generate_news_content(items) - Generate full news list HTML
build_news(data_path, template_path, output_path) - Main build function

1.3 Create Template

File: templates/news.html

Copy current news.html structure
Replace news items with  marker

1.4 Update Build Infrastructure

Files to modify:

scripts/build.py - Add news build to master script
scripts/validate_data.py - Add news validation
.github/workflows/build-content.yml - Add news.html to commit list

Phase 2: Bluesky Feed Integration

2.1 Architecture Decision

Client-side approach (recommended for static site):

JavaScript fetches Bluesky posts at page load
No server-side infrastructure needed
Uses public Bluesky API (app.bsky.feed.getAuthorFeed)

API Endpoint:

https://public.api.bsky.app/xrpc/app.bsky.feed.getAuthorFeed?actor=contextlab.bsky.social&limit=10

2.2 Implementation

Files to create/modify:

js/bluesky-feed.js - New module for Bluesky integration
- fetchBlueskyPosts(handle, limit) - Fetch posts from API
- renderBlueskyPost(post) - Generate HTML for single post
- initBlueskyFeed() - Initialize feed on page load

CSS additions to css/style.css:

.bluesky-feed - Container styling
.bluesky-post - Individual post styling
.bluesky-post-header - Avatar + handle + timestamp
.bluesky-post-content - Post text
.bluesky-post-embed - Images/links

2.3 Template Integration

Modify: templates/news.html

Add Bluesky section with container div
Include js/bluesky-feed.js script

Phase 3: Add New News Articles

3.1 Download and Process Thumbnails

The comment provides 4 articles with GitHub-hosted images that need to be:

Downloaded to images/news/
Renamed appropriately
Possibly resized/bordered using existing scripts

New articles:

"Why are we so preoccupied by the past?" (Jan 10, 2025)
"Claudia Gonciulea joins the lab!" (Aug 1, 2025)
"Jeremy gets tenure!" (Jul 1, 2024)
"We CAREER'd!" (Jan 25, 2022)

3.2 Add to Spreadsheet

Add all 4 new articles to data/news.xlsx with proper dates for sorting.

Execution Plan (Parallel Agent Assignment)

Batch 1 (Can run in parallel):

Agent A: Create data/news.xlsx spreadsheet with all existing + new news items
Agent B: Create scripts/build_news.py following existing patterns
Agent C: Create templates/news.html template file

Batch 2 (After Batch 1):

Agent D: Implement Bluesky feed JavaScript (js/bluesky-feed.js)
Agent E: Add Bluesky CSS styles to css/style.css

Batch 3 (After Batch 2):

Agent F: Update scripts/build.py, scripts/validate_data.py, and GitHub Actions workflow
Agent G: Download and process new thumbnail images

Batch 4 (Final integration):

Agent H: Run full build and comprehensive testing

Testing Plan

Unit Tests (NO MOCKS - Real Files)

File: tests/test_build_news.py

test_loads_news_from_real_xlsx() - Load actual news.xlsx
test_news_sorted_by_date() - Verify reverse chronological order
test_generates_valid_html() - Check HTML structure
test_handles_empty_fields() - Graceful empty handling
test_handles_special_characters() - HTML entities preserved

Integration Tests (Real Build)

test_build_produces_valid_news_html() - Full build creates valid file
test_news_count_matches_spreadsheet() - All items rendered

Bluesky Integration Tests (Real API)

File: tests/test_bluesky_feed.py

Test fetching real posts from @contextlab.bsky.social
Test HTML rendering of actual post data
Test error handling for network failures
Test rate limiting behavior

Visual Verification

Open local news.html in browser
Verify news items display correctly
Verify Bluesky feed loads and displays
Test on mobile viewport

Files to Create/Modify Summary

New Files:

data/news.xlsx
scripts/build_news.py
templates/news.html
js/bluesky-feed.js
tests/test_build_news.py
tests/test_bluesky_feed.py

Modified Files:

scripts/build.py - Add news build
scripts/validate_data.py - Add news validation
css/style.css - Add Bluesky styles
.github/workflows/build-content.yml - Add news.html to commit
news.html - Will be auto-generated

Notes Folder:

notes/issue_9_news_feed.md - Implementation notes and progress tracking

Cleanup Plan

Before signing off:

Remove any temporary test files
Clean up notes folder (keep only final summary)
Remove any .pyc or __pycache__ artifacts
Verify no debug code left in production files

Pre-Push Checklist

Run python scripts/validate_data.py - All validations pass
Run python scripts/build.py - All builds succeed
Run python -m pytest tests/ -v - All tests pass
Open news.html locally - Visual verification
Check Bluesky feed loads correctly
Run linters if configured
Push to GitHub
Verify GitHub Actions pass
Check live site after deployment

Skills/Tools to Leverage

Existing:

utils.py functions for spreadsheet loading
citation_utils.py for link resolution
build_*.py patterns for template injection

Potentially Create:

Bluesky API helper skill for future use
Image download/processing automation

Ready for review. Please approve or suggest modifications before implementation begins.

Add spreadsheet-based news feed with Bluesky integration (Issue #9)

Completed in PR #11

News Feed Implementation

Created data/news.xlsx spreadsheet for managing news entries with columns: date, type (news/bluesky), title, description, image, link, bluesky_url
Built scripts/build_news.py to generate news.html from the spreadsheet
Added templates/news.html template with proper structure

Bluesky Integration

Implemented js/bluesky-feed.js for embedding Bluesky posts from @contextlab.bsky.social
Posts are fetched client-side and displayed in a themed, scrollable feed
Supports rich content including images, link cards, and quoted posts
Added CSS styling in css/style.css to match the site's green theme

How to Use

Manual news entries: Add rows to data/news.xlsx with type="news"
Bluesky embeds: Add rows with type="bluesky" and the post URL in bluesky_url column
Run python scripts/build_news.py to regenerate news.html

The Bluesky feed automatically pulls the latest posts and displays them alongside manual news entries.

updates to the news feed #9

Description

Activity

jeremymanning commented on Dec 18, 2025