How carve out Text Translator and Builder

[oglop zelgo]

Hello,

I'm trying to carve out Sphinx Text translator to be able to use it for my Pelican project and other things.

This is my repository: https://github.com/1oglop1/rst2text/blob/master/src/rst_2_plain/translators.py#L20

I kinda understand how docutils.writers._html_base.HTMLTranslator work and that it (somehow) stores results of states inside self.body and self.fragment

After looking for some inspiration (Flask RSTPages) I came up with this https://github.com/1oglop1/rst2text/blob/master/src/rst_2_plain/document.py#L30
The idea behind this was to compare behaviour of my TextTranslator with docutils original HTMLTranslator to be able to understand on what is happning there.

But as you can see my object txtrst lacks the actual processed content and I do not know where the problem is at the moment.
I assume it might be something with the method depart_document() but it can be something completely different.

I'd appreciate all help I can get.

Thank you!