feat: language aware config step + hana changes

[hntrl]

This adds a build step that has smarter handling for docsets with code fences, and adds a build step to docs.json to more intuitively organize content

• We now automatically expand docsets that have code fenced content (meaning we don't need to hardcode two different references for python + js in docs.json)
• When we encounter a directory that has code fenced content, instead of flagging it as a version we add it as a dropdown to better align with the design enhancements we want to make
• When we encounter a doc that has code fences, we flag the top level directory as needing conditional parsing (instead of hard coding which directories should be parsed)
  • If we were to add code fences to lgp or LS docs today, we would automatically start building js + python docs the same way we do for LG

also removes the version picker since products are the thing we have versions for, not languages (sry brody lol)

[github-actions]

Preview ID generated: preview-hunter-1755235273-2ab5152

[github-actions]

Preview ID generated: preview-hunter-1755235354-1f1ef3d

[github-actions]

Preview ID generated: preview-hunter-1755236608-3c18df7

[github-actions]

Preview ID generated: preview-hunter-1755271106-8297af5