How do I control the way the PDF index is generated?

For example:



The above index is, as far as I can tell, generated from <H1><H2><H3><...> tags.

The issue is that we have content such as:

<h1>Title</h1><h2>Chapter Foo</h2>{Bar}[...]

Where {Bar} is user generated HTML that can contain arbitrary <H> tags.

I need a way to pick and choose.

Ideas?

1. pdf_index.png‎ 68.5 kB ‎
   Bad Index

The bookmark properties can control this. For example, you could disable bookmarks for user content:

div.user { bookmark-level: none }