👌 Improve parsing of nested amsmath

[chrisjsewell]

the previous logic was problematic for amsmath blocks nested in other blocs (such as blockquotes)

the parsing code now principally follows the logic in markdown_it/rules_block/fence.py
(see also https://spec.commonmark.org/0.30/#fenced-code-blocks),
except that:

1. it allows for a closing tag on the same line as the opening tag, and
2. it does not allow for an opening tag without closing tag (i.e. no auto-closing)

fixes #117

[chrisjsewell]

cc @tovrstra

[codecov]

Codecov Report

Attention: Patch coverage is 90.47619% with 2 lines in your changes missing coverage. Please review.

Project coverage is 92.74%. Comparing base (d11bdaf) to head (8f5cf97).
Report is 2 commits behind head on master.

Files with missing lines	Patch %	Lines
mdit_py_plugins/amsmath/__init__.py	90.47%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #119      +/-   ##
==========================================
- Coverage   92.80%   92.74%   -0.07%     
==========================================
  Files          31       31              
  Lines        1835     1833       -2     
==========================================
- Hits         1703     1700       -3     
- Misses        132      133       +1     

Flag	Coverage Δ
pytests	92.74% <90.47%> (-0.07%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[tovrstra]

If compatibility with Jupyter is of importance, then the HTML should be different here:

<blockquote>
<div class="math amsmath">
\begin{matrix}
&gt; -0.707 &amp;  0.408 &amp;  0.577 \\
&gt; -0.707 &amp; -0.408 &amp; -0.577 \\
&gt; -0.    &amp; -0.816 &amp;  0.577
&gt; \end{matrix}
</div>
</blockquote>

The markdown of this example is rendered in Jupyter Lab as:

Or should this be treated as a Jupyter issue? In Jupyter, one would write the Markdown as:

> \begin{matrix}
-0.707 &  0.408 &  0.577 \\
-0.707 & -0.408 & -0.577 \\
-0.    & -0.816 &  0.577
\end{matrix}

Given that Jupyter is so widespread, it seems prudent to also include this example in the tests.

[chrisjsewell]

Or should this be treated as a Jupyter issue?

very much so, that is definitely very wrong 😅

I'm interested in "correct parsing", not mimicking bugs of other programs 😬

If compatibility with Jupyter is of importance

As you note, what you are actually talking about here is only the jupyter lab implementation,
which relies on things like markedjs and mathjax,
and is known to have a bunch of issues with math rendering (jupyterlab/jupyterlab#8645, jupyterlab/jupyterlab#14570, etc)

by contrast, try this in https://code.visualstudio.com/docs/datascience/jupyter-notebooks and you will get perfectly correct math 😄

[chrisjsewell]

it seems prudent to also include this example in the tests

I've added it for completeness, but to be clear, this is just a case of block quote lazy continuation: https://spec.commonmark.org/0.30/#example-233

[tovrstra]

I've opened a Jupyter issue: jupyterlab/jupyterlab#16755