As documented on #65 and #33, pandoc_ast is not actively maintained to stay
compatible with the current pandoc API. In the midterm, we'll still want to
address #33 and drop the pandoc dependency, but for the short term we will:
Use pandoc to dump the markdown to HTML.
Use scraper to extract the logical
units from the definition list.
Use html2md to render the logical
unit content back to markdown for entry to the database.
As documented on #65 and #33,
pandoc_astis not actively maintained to stay compatible with the current pandoc API. In the midterm, we'll still want to address #33 and drop the pandoc dependency, but for the short term we will: