Hi Bhaskar,
I went through the Tomd python tool as mentioned by you, I saw that it is quite similar to Pandoc with limitation of converting simple HTML( generated from markdown) back to Markdown. If we use to convert the nested TAG and CLASS used in HT, it will note produce the expected result (i.e. still leaving a lot of manual work). Thus a custom convector( if possible) should be made so manual conversion can be avoided ;). I would be glad to listen to your opinions about :).
P.S. Meanwhile I went through other software and custom scripts written by me, and also able to produce some result (available on
http://ankitdhankhar.me/FreeType/)