From __icontains to search
summary
Shows how full‑text search is built from an inverted index (the sorted posting lists that power ~90% of searches), how analysis (tokenization, lowercasing, stemming, synonyms) happens at index time and must be mirrored for queries, why changing analysis requires reindexing, how phrase/proximity queries use stored positions, how relevance comes from TF‑IDF, how highlighting uses byte offsets, and how Elasticsearch aggregations and Kibana produce fast facet counts and visualizations.
AI-written summary · about
chapters (15)
- 0:00Intro
- 1:23search is interface to data
- 1:57Find documents about Django
- 3:40Bible concordance
- 5:12search(python AND django)
- 6:37Phrase search
- 7:12Search("python web")
- 9:34Building an inverted index
- 12:45Analysis
- 15:46Relevancy!
- 16:43Relevancy, for humans
- 19:15Highlighting
- 20:55Facets & Filtering
- 22:23Facets/Aggregations
- 23:47Visualizations
related talks
55mFull-Text Search Demystified
24mSearch From the Ground Up
46mFull-Text Search Internals with Elasticsearch
31mUnspoken overhead of the Inverted Index
26mOn The Look-Out For Your Data
45mAnything New in Search? What If I Find It?
38mIntroduction to Search
44mFinding relevancy: Improving your application search bar!