feat: add NumberedHeadingsPreprocessor

[sapristi]

Fixes #1586

Add pre-processor that adds numbering to markdown headers:

transforms title in markdown cells, such that e.g.

# title 
## subtitle
content 
# other title

turns into

# 1 title 
## 1.1 subtitle
content 
# 2 other title

[krassowski]

Thank you @sapristi for sending the PR. Was this requested in an issue before? Just trying to understand if there is sufficient community need for this to live in nbconvert given the maintenance costs for adding a new pre-processor.

[sapristi]

Hello, I just checked, it seems that some people were interested in this as well, e.g. #1586

IMO this is a very essential feature for working with a document - it's also a feature provided in the Jupyter web UI, so I think I'm not the only one.

[krassowski]

What if I we have a code in markdown, e.g.:

This is how you create a heading in Markdown
```markdown
# heading
```

It feels like a more robust solution would be to use mistune API for parsing, see https://mistune.lepture.com/en/latest/api.html (markdown_mistune is already in use here, but it should fail nicely).

It is also possible to instruct pandoc to run with --number-sections (see https://www.youtube.com/watch?v=gDV63UMJR_U and https://pandoc.org/chunkedhtml-demo/8.3-headings.html and https://pandoc.org/MANUAL.html) but pandoc is neither the only nor even default markdown parser.