SFScon19 - Alex Lanz - Tips and Tricks for Building a Full-Featured Full-Text Search using Elasticsearch
1 like108 views
The document discusses building a full-featured full-text search using Elasticsearch at the 2019 Free Software Conference. It highlights key features of Elasticsearch, such as its distributed architecture, schema-free JSON document handling, and support for synonyms, compound words, boosting, and aggregations for enhanced search functionality. Additionally, it provides tips for effectively using Elasticsearch alongside SQL databases.
SFScon19 - Alex Lanz - Tips and Tricks for Building a Full-Featured Full-Text Search using Elasticsearch
- 1. Building a Full-Featured Full-Text Search using Elasticsearch SFSCon - Free Software Conference 2019 1
- 2. About me Alex Lanz 2https://aboutbits.it ● Co-Founder @ AboutBits ● Software Developer @about_bits @alex__lanz
- 3. Shay Banon - Creator of Elasticsearch 3 Search is something that any application should have. https://thedudeabides.com/articles/the_future_of_compass
- 7. But how can I develop a search functionality? 7
- 8. Elasticsearch Search Engine ● Based on the Lucene library ● HTTP web interface (REST) ● Schema-free JSON documents ● Distributed architecture ● High scalability ● Open Source 8 https://www.elastic.co/products/elasticsearch
- 9. Check out the docs! https://www.elastic.co/guide/en/elasticsearch/reference/current/index.html 9
- 10. Storytelling! 10
- 12. 12
- 13. 13
- 14. Different words Solution Synonyms ● Replaces or adds additional words (tokens) ● During text analysis schlagobers => sahne 14 https://www.elastic.co/guide/en/elasticsearch/reference/7.4/analysis-synonym-tokenfilter.html
- 15. 15
- 17. 17
- 18. 18
- 19. Compound words Solution ngrams ● Contiguous sequence of n items from a given word (term) ● Useful for languages that don’t use spaces or that have long compound words Apfelsaft: apf, pfe, fel, els, lsa, saf, aft Apfel: apf, pfe, fel 19 https://www.elastic.co/guide/en/elasticsearch/reference/7.4/analysis-ngram-tokenizer.html
- 20. 20
- 22. 22
- 23. 23
- 24. Not enough words Solution Boosting ● Control the relevance of a search ● Increase (or decrease) the relative weight of a clause Boost of 2 for products within the category “Obst”. 24 https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-term-query.html
- 25. 25
- 26. 26
- 29. Enriched results Solution Aggregations ● Provide aggregated data based on a search query ● Analytic information over a set of documents ● All with one query 29 https://www.elastic.co/guide/en/elasticsearch/reference/7.4/search-aggregations.html
- 30. 30
- 32. Elasticsearch Pro Tipps ● It’s an enhancement not a replacement for a SQL database ● Data must no be persisted forever, rebuild the index if necessary ● Denormalize your data 32
- 33. Questions? 33