Automattic / Automattic/jetpack
Search: Expand content matching
- Dominant language
- PHP
- Stars
- 1.8k
- Forks
- 898
- Avg merge
- 1d 18h
- Merged PRs (30d)
- 774
Description
We have a number of cases where we are not matching content in quite the same way that core WP does.
Examples:
- a partial product name: VCAN1047 and the search is for 1047
- searching for part of a url: greg.wordpress.com and the search is for wordpress.com
- a search that matches a comment on a post
All of these are cases of not having the content indexed.
Additionally, we've found elsewhere that having a single field for "all content" and then performing boosting on all the other fields has worked very well as a default structure for queries. See: https://data.blog/2017/03/15/improving-relevance-and-elasticsearch-query-patterns/
We should create an all_content field, or expand the mlt_content field that already exists. Additionally this field should probably have edge ngrams so we can start building search as you type. Adding it to mlt_content would be my preference, but I am unsure how that will impact related posts performance.
This is also related to issues such as #8662 Any content from fields such as that should also get included in this field.
Contributor guide
Assessment
This issue has not been assessed yet.