Fuzzy Search in the Browser Over Half a Million Lines: Building an Index on a CDN
summary
Describes building a client-side fuzzy search over ~500,000 lines by constructing a searchable index hosted on a CDN, using NLP algorithms for typo-tolerant matching. Explains how HTTP/2 enabled serving the index from CDN, and shows code examples and demos of the indexing and browser-side search implementation.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 1:282015 banking search
- 8:32Server-side search
- 11:26Static site interactivity
- 13:40SQLite in browser
- 19:42Gatsby plugin
- 20:27Movie search demo
- 23:21Search algorithm
- 26:16Stemming and lemmatization
- 29:55Pre-indexing
- 33:06Fuzzy search and Levenshtein
- 35:12N-grams
- 36:36Indexing at build
- 38:41Data updates and pagination
- 40:26Q&A
related talks
57mOptimizing Frontend Performance of a High-Load Search
46mFull-Text Search Internals with Elasticsearch
47mImplement Al Semantic Search in a Nuxt Blog w/ Transformer.js
31mIntelligent Japanese Search With OpenSearch
08mImplement Al Semantic Search in a Nuxt Blog w/ Transformer.js
21mBring AI-Based Search to Your Web App
50mTechnical SEO for JS developers
37mSearch and Recommendations: 3 Sides of the Same Coin #bbuzz