Table of contents

I've put code at the links below that constructs a table of contents in HTML form given Asciidoctor text as
input.  I will post something shortly that describes better what it does and what the specification is.  I am using this
now at http://www.scripta.io; I would like evolve it to a state in which it could be useful for others.  To this end, I wold very much appreciate your comments.
​
https://github.com/jxxcarlson/noteshare/blob/master/lib/noteshare/modules/asciidoctor_helpers.rb[Code for table_of_contents] +
https://github.com/jxxcarlson/noteshare/blob/master/spec/noteshare/modules/asciidoctor_helpers_spec.rb[Code for the tests]
​
Here is an article that describes this in somewhat more detail:

http://home.noteshareblog.io/2015/12/09/Experiments-with-Asciidoctor-no-3-a-better-table-of-contents.html

I would like to endow the table of contents code with a feature allowing the user to re-arrange sections of a document by dragging the titles around in a suitable display.  This already works for the "outer nodes" of the table in the demo at http://www.scripta.com.  (scripta is an open source project that will eventually replace notheshare).  You can play around with this buy logging into scripta with user name coram@foo.com, with password hohoho1234 .

I can do this using the AST that the Asciidoctor parser generates, but need to know the name of the method that returns the raw content, jus as #content yields the HTML.                                                                                                                                                                  

Let's call the needed method 'raw_content'.  i see how it could
be written using the blocks an line methods plus a little
recursion, but if the method already exists, so much the better.

One way to get the "raw_content" would be to have converter
that produced Asciidoc text instead of HTML.  Let's call this method
'convert back'.  Applied to a node in the AST, it would produce  Asciidoctor
text that when parsed would yield the give node.

This may have been discussed before -- an "Asciidoc backend".  It should
be very easy to write as an extension -- it would be like  a drastic simplification
of the latex converter: instead of doing something to a node's content, it
just passes it through.

Having said this I realize  in the latex converter, n.content is the html text.
So why am I getting html when I apply #content?  The problem must already
be solved!!??