Re: manipulating headline level

[Albert Krewinkel <albert+pandoc-9EawChwDxG8hFhg+JK9F0w@public.gmane.org>]

Dear juh,

"'juh' via pandoc-discuss" <pandoc-discuss-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org> writes:

> Many markdown-files hierarchically ordered in folders.
> Every file starts with the top-level headline mark "#".
>
> I want to convert these files like this:
>
> pandoc $(cat outline.txt) -o target.html
>
> where outline.txt contains the relative pathes to all files.
>
>
> 00.md
> 01/00.md
> 01/01/00.md 
> 01/01/01.md 
> 01/01/02.md 
>
> Is there a way to tell pandoc that it should regard the headline levels
> according to the folder structure?

Pandoc can't do this out of the box, but you could use a custom reader
to correct the headings. The below requires the current dev version
(nightly build), but could be rewritten to work with a released version,
too.

The script

 - iterates over all input files,
 - parses their contents as a separate document,
 - shifts headers in each document by the number of subdirs the
   respective file is in, and
 - concatenates the individual per-file documents into a single
   document that is then returned.

Pass the reader file to pandoc via the `--from` parameter.

``` lua
local format = 'markdown+smart'

function Reader (sources, opts)
  local doc = pandoc.Pandoc{}
  for _, source in ipairs(sources) do
    local path_components = #pandoc.path.split(source.name)
    -- last path component is the filename
    local depth = path_components - 1
    doc = doc .. pandoc.read(source, format, opts):walk {
      Header = function (h)
        h.level = h.level + depth
        return h
      end
    }
  end
  return doc
end
```

-- 
Albert Krewinkel
GPG: 8eed e3e2 e8c5 6f18 81fe  e836 388d c0b2 1f63 1124