Currently, the search index is duplicated in searchindex.js and searchindex.json. It seems the reason for the duplication is that the .json allows for async loading via fetch() and the .js provides a fallback in case CORS causes issues with the fetch().
The problem with having the duplicated index is that for large books, the index becomes quite large (2.5MB for my company's book!). It would be nice if the index didn't have to be duplicated.
This PR removes searchindex.json from the site output. It replicates the async functionality by reading in the searchindex.js file and parsing out the JSON within. The fallback remains unchanged.
SichangHe
commented
1 week ago
Currently, the search index is duplicated in
searchindex.jsandsearchindex.json. It seems the reason for the duplication is that the.jsonallows for async loading viafetch()and the.jsprovides a fallback in case CORS causes issues with thefetch().The problem with having the duplicated index is that for large books, the index becomes quite large (2.5MB for my company's book!). It would be nice if the index didn't have to be duplicated.
This PR removes
searchindex.jsonfrom the site output. It replicates the async functionality by reading in thesearchindex.jsfile and parsing out the JSON within. The fallback remains unchanged.