Re: HTML attributes not being stripped off

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

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

> The idea there, is to be able to convert both html (generated by a rich
> text editor)  and markdown (or other similar markup language) file
> through a similar pipeline to a pdf with similar style. Using a
> different templating engine somewhere in the pipeline mean more
> complexity, so i'm considering the idea of using pandoc templating if
> the html result is okay.

OK, I see. How about the following approach then: use a custom reader
that passes the input through as raw HTML if any of the files have an
`.html` extension, but otherwise treats the input as Markdown.

``` lua
function Reader (sources, opts)
  local raw_html = false
  for _, source in ipairs(sources) do
    if source.name:match '%.htm[l]$' then
      raw_html = true
    end
  end
  if raw_html then
    local blocks = sources:map(
      function (source)
        return pandoc.RawBlock('html', tostring(source))
      end
    )
    return pandoc.Pandoc(blocks)
  else
    return pandoc.read(sources, 'markdown', opts)
  end
end
```

See also <https://pandoc.org/custom-readers.html>.

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