From basic distance search to a complex multi criteria search
summary
Starts from geodjango distance sorting, moves to elasticsearch + haystack using function_score and explicit decay functions (linear exponential gauss) to combine distance, price, picture_count and an owner_quality_rate (0–1). Implements a custom ElasticsearchSearchQuery / backend and queryset, with examples like picture_count origin 50 offset 40 scale 9 and weighted sums; result is airbnb‑style ranking that runs faster than the original geodjango approach and did not degrade response time; code is on GitHub.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:59AboutCaravanacom
- 1:56Agenda
- 3:27Search result
- 5:08What do guests want
- 7:55haystack elastic search
- 8:53haystack documentation
- 11:31documentation
- 16:41elasticsearch backend
- 22:31results
related talks
41mBeyond the basics with Elasticsearch
41mPainless / Painful? A Custom Ranking System in Elasticsearch
30mBeyond faceted search
44mFinding relevancy: Improving your application search bar!
24mSearch From the Ground Up
28mSuper Search with OpenSearch and Python
21mEvolution of Yelp search to a ranking platform
13mUnconference Session: A Neat Trick for Fuzzy Search