Lunr integration - how to avoid indexing code in code fences


I’m using hugo for a little more than a month and I’ve almost added search with lunr. I more or less followed this blog post. I’ve got my index.json, my gulp task that generate lunr own index file. That went well

My only remaining problem is that I have a lot of code (in code fences) in my markdown files. I would like to exclude it from the search. I think the easiest way would be to adapt the way .Plain work to avoid generating all the code, is it possible ?

Thanks in advance.