tmfarrell.github.io

Built with Jekyll and hosted on GitHub Pages.

Quick Start

# Install dependencies
bundle install

# Run locally (http://localhost:4000)
bundle exec jekyll serve

# Or just build
bundle exec jekyll build

Deployment

Changes pushed to main auto-deploy via GitHub Pages.

Search / AI Mode

The site includes a semantic search feature powered by Pinecone vector DB.

Rebuild Search Index

After adding new posts or updating portfolio items:

cd indexer
npm install
node index.js

Requires PINECONE_API_KEY and PINECONE_INDEX in .env.

Frontend Development

To iterate on the search UI without hitting the API:

1. Start the local server: bundle exec jekyll serve
2. Visit: http://localhost:4000/ai-mode/?mock=true

This loads mock results from assets/js/search.js (the getMockResults function). Edit this function to test different result formats, tags, styling, etc.

Content

• Posts: _posts/ (with categories: reading or categories: writing)
• Portfolio: _data/portfolio.yaml
• Pages: writing.html, reading.html, building.html, ai-mode.html

Running jekyll build generates _site/content.json with all searchable content.