Search

Full-text client-side search backed by a Jekyll-generated JSON index.

How it works

1. Index — site/search.json is a Liquid template rendered at build time into a JSON array of all indexable content.
2. UI — site/search/index.html fetches the index and filters in JavaScript as the user types.

What is indexed

Content	type value
Jekyll pages	page
site/_events/	event
site/_news/	news
site/_directory/people/	person
site/_directory/institutions/	institution
site/_directory/projects/	project

Index fields (per item)

{
  "title": "...",
  "url": "/events/2026-03-12-meshup07/",
  "date": "2026-03-12",
  "categories": ["MeshUp", "WP7"],
  "tags": ["Sports Media", "AI Production"],
  "description": "...",
  "content": "first 100 words of body text",
  "type": "event"
}

Filtering

• Full-text across title, description, content, tags, categories
• Type filter chips: Event, News, Page, Person, Institution, Project
• Category and tag filters

Badge colours

Type badges use classes in site/assets/css/custom.css:

Type	Class	Colour
Event	.result-type-event	teal
News	.result-type-news	blue-grey
Person	.result-type-person	terracotta
Institution	.result-type-institution	steel blue
Project	.result-type-project	amber

Tags and search

Person tags and search_keywords are indexed. Keep tags normalised with scripts/merge_tags.py so search and filters stay consistent — see Tags and Keywords.