Index for HTML-Files

[benjaminschlegel]

I really like to be able to create an index in Quarto book with \index{word}. At the moment it just works and displays in PDF-Outputs. It would be nice if it would also work for html documents and be added as an extra document at the end of the quarto website.

[baptiste]

I've been thinking about this too. My first idea was to switch from \index{concept} to a language agnostic syntax such as [concept]{.index} and use the latex-environment to do the conversion. I then thought it might be more convenient to have a Lua filter detect automatically specific words in the manuscript. At that point, I naively concluded that the idea of an index in HTML format was of little use, since one could simply use the website's built-in search function.

However, I've since realised how wrong I was when reading this blogpost: a brute-force listing of all appearances of a keyword is not really an index, but more of a concordance. An index necessarily involves some hard work by the author, to identify a helpful structure for the reader. In that sense, it's as useful in HTML as it is in hard copy format.

All this to say: I do not know of a HTML solution to your question, but I think it'd be a great addition to Quarto! (especially if an output-agnostic convention is developed, that works across formats).

As far as a HTML output is concerned, I think something like this would look quite nice: https://codepen.io/gcyrillus/pen/kNjOjw

[benjaminschlegel]

I don't know how Quarto works in the background, but maybe something like this could work: replace during rendering for HTML all \index{BlaBla} by an HTML-anker and save the word with the file_name and the anker in some sort of list. At the end sort the list and output it as linkes in a new document at the end. Somehow multiple occurence of a word should be grouped.

[baptiste]

The span notation [concept]{.index} would take care of the first part, and to build the HTML index one could use this SO answer, but I haven't been successful in running the filter in practice (I'm guessing Pandoc has changed a bit since then).

[jmgirard]

I wrote an extension that does this, using the [concept]{.index} span syntax suggested above. The same marks build a linked index in HTML and EPUB and a page-numbered index in PDF and Typst. A Quarto book gets one index across all chapters.

quarto add jmgirard/quarto-index

Details and examples: https://jmgirard.github.io/quarto-index/ (announcement: https://github.com/orgs/quarto-dev/discussions/15195)