Disable paths to be indexed in search.json

[benabel]

Description

Hi all quarto team.
I am using quarto for my university courses and I include two version of a same chapter: html and revealjs. When students search in the website they get two results from the same chapter(html and reveal) but I would prefer to include only the html version.
At the moment I found a workaround with a post-render script that delete all occurrences of revealjs content of search.json but it would handy to have some configuration in _quarto.yml file similar to includes and navbar contents based on regexp.
Is there any other option to the post-render script?

[mcanouil]

Use custom index with Algolia for example. More generally I suggest you read the whole documentation on search: https://quarto.org/docs/websites/website-search.html

[benabel]

Thank you for your answer. Algolia is not an option for my use case. I've read carefully the documentation but it was unclear to me if search: false disable the document indexing or the search ui. I have found no way to disable the document indexing with metadata.

[mcanouil]

Do you have a concrete example to share? (as a Git Repository)

[benabel]

Thank you for your interest @mcanouil. No it is not open-sourced because it contains solutions to exercises. I'm investigating the search: false metadata setting on my revealjs pages to see if it prevent my content from being indexed. If not i will stay with my post-render bash script.

jq 'map(select(.href | test("diapo\\.html") | not))' quarto/_site/prod/search.json > /tmp/search.json && mv /tmp/search.json quarto/_site/prod/

[mcanouil]

I mean, could you make a reproducible example using maybe quarto create project?

[benabel]

Ok I'll do that in the week, but I can't do that today. I'll come back on the discussion when I'll get a window time.

[benabel]

Ok @mcanouil, I've created the minimal config at https://github.com/benabel/quarto-issue. And playing with search config clearly shows that the search keyword has a different meaning in formats section and website section. This makes sense.

project:
  type: website
website:
  navbar:
    search: false # enable/disable UI
  search: 
    location: navbar
    type: overlay
format: 
  html:
    search: true # enable or disable indexing
  revealjs: 
    output-file: document-revealjs.html
    search: true # enable or disable indexing
format-links:
  - html
  - format: revealjs
    icon: file-earmark-slides
    text: Diaporama

I think there's could be an enhancement in the docs https://quarto.org/docs/websites/website-search.html with the words:

• enable/disable search UI in https://quarto.org/docs/websites/website-search.html#search-appearance
• forcing/disabling indexing in https://quarto.org/docs/websites/website-search.html#disablingforcing-search.

Thank you for your replies, the search features are OK for me now. I'm closing the discussion. I'll open an issue if you find it pertinent for a documentation enhancement.