support HTML definition lists (<dl>, <dt>, and <dd>)

[chrispy-snps]

Fixes #172.

New convert_dt() and convert_dd() functions are added that follow the PHP Markdown Extra syntax:

https://michelf.ca/projects/php-markdown/extra/#def-list

If additional definition list dialects are requested in the future, a configuration option can be added to select the format.

No convert_dl() function is added; the child-tag conversion functions do all the work.

The regression tests are updated to test various structures. I also used Pandoc to confirm that all Markdownify results are converted back to the expected HTML source.

Note: This pull request requires that #171 be merged first; otherwise the test_dl unit test will fail.

Limitations

There are two limitations in this support, both related to the fact that blank lines are added outside the convert_dt() and convert_dd() function scopes.

Limitation 1 - multiple terms sharing the same definition are not handled properly (the term lines are separated by a blank line instead of kept directly adjacent):

<dl>
  <dt>term 1a</dt>
  <dt>term 1b</dt>
  <dd>definition</dd>
</dl>

Limitation 2 - a blank line is always inserted before definitions, causing them to signify paragraph-based definitions even when they were not:

<dl>
  <dt>term 1</dt>
  <dd>bare definition</dd>
  <dt>term 2</dt>
  <dd><p>definition in paragraph</p></dd>
</dl>

[ninsbl]

I just tested the feature locally as I am interested in the functionality for markdown conversion. It is great to see this becoming available, I noticed that the result is one big text paragraph. I think an additional newline would be needed after the <dd> elements (for example).

[chrispy-snps]

@ninsbl - can you share a small testcase here for me to reproduce it?

[ninsbl]

Sure. We are currently looking into moving documentation for GRASS GIS from HTML to Markdown.

One manual page with <dl>, <dt> and <dd> is for example: https://grass.osgeo.org/grass84/manuals/grass.html

Here is the code used for translation to Markdown (with python-markdownify installed for the branch)

import bs4
import requests

from markdownify import markdownify, MarkdownConverter

resp = requests.get("https://grass.osgeo.org/grass84/manuals/grass.html")
soup = bs4.BeautifulSoup(resp.text, 'html5lib')
MarkdownConverter(
    **{
        "heading_style": "atx",
        "escape_misc": True,
        "code_language": "shell",
        "newline_style": "backslash",
        "wrap_width": 79
    }
).convert_soup(soup)

You can see resulting markdown here: https://github.com/ninsbl/grass/edit/md_test/md/grass.md. Please look at the "FLAGS" section...

Note the line break before the : (which becomes a white space) and only one linebreak after the <dd> element becoming space as well...

Also, element content seems only parially wrapped...