Add math-mode support for horizontal Parsons problems

[geoff-cox]

This patch extends the Runestone hparsons component to support math-mode horizontal Parsons problems.

[geoff-cox]

Here is one of the test cases I used in Pretext:

<exercise label="horizontal-parson-math-simple" language="natural"><title>Parsons Problem, Math</title>
    <statement>
        <p>
            Problem Statement
        </p>
    </statement>
    <blocks layout="horizontal" reuse="no" randomize="no">
        <block order="4"><m>\alpha</m></block>
        <block order="2"><m>\beta</m></block>
        <block order="1"><m>\displaystyle\sum_{i=1}^{\infty}</m></block>
        <block order="3"><m>x^2</m></block>
        <block order="5" correct="no"><m>\displaystyle\int x^2</m></block>
    </blocks>
</exercise>

For quick reference, here is the correct order

[bnmnetp]

Thanks for this!

[bnmnetp]

Thanks for this addition!

You should have a look at parsons.js and others for how we use the this.queueMathJax(specific item).

MathJax can create a big set of timing problems and this allows us to take control. Also trying to re-typeset the whole page can be a performance issue.

[bnmnetp]

Maybe some comments about what the simulateSolution function is all about? Its not clear to me.. Is this for scoring? I would have guessed that block scoring would be fine? Maybe not for some expression.

[geoff-cox]

@bnmnetp
The Micro-parsons code was formatting the math blocks with unpredictable Ascii characters, so I simulated the correct solution in the answer area and asked Micro-parsons to output what a correct solution would look like from their end. I needed this because a correct answer is determined by comparing the text on the blocks. I was hoping there would have been an ID on the blocks, but I didn't see any, so this was my workaround. I am sure there is a better way. Let me know if you have any ideas.

I can add more details to the PR tomorrow.

[bnmnetp]

Thanks for clarifying! I assume all of the \ ing of characters was the cause of the weird text.

I does not look like the html representation lends itself easily to adding ids.... But it seems like it should. maybe worth investigating with Rob, and looking at a better html (or json) representation to use in the javascript.

I'm ok with your solution for now, but thinking long term.

[geoff-cox]

I think adding IDs could streamline future improvements and make grading in blockGrader.js more straightforward.

I designed this PR thinking that it would be a good way to construct formulas and equations term-by-term. Unfortunately, it doesn't currently support multiple correct orderings, which would be essential for commutative operations like addition and multiplication. So getting Rob (@rbeezer) involved in coordinating how multiple orderings can be communicated to Runestone would be good.

[geoff-cox]

Also, display math tags (<me>) are not currently supported with this PR, but <m>\displaystyle * </m> is good to go.