Parsing the content of math environments?

[jonathan-laurent]

Currently, it seems to me that TexSoup treats the content of a math environment ($...$, (...) or [...]) like a single token.
Is there any way to get a structured representation of this content too? I was thinking of getting a string representation of the inside of $...$ and then calling TexSoup on it recursively but I am not sure this would work.

To be even more concrete, here is an example of a task I am interested in performing:

• Find all instances of \infer{A}{B} within a math environment and replace it by \infer{B}{A} where A and B could be arbitrary Latex code.

Could one use TexSoup to do this?

[alvinwan]

@jonathan-laurent Thanks for creating this issue! I've just made a change that will allow this. The linked PR will have a passing test for this math env replacement, reproduced below:

    soup = TexSoup(r'$$\infer{A}{B}\infer{C}{D}$$')
    assert soup.infer is not None, repr(soup.expr)
    for infer in soup.find_all('infer'):
        infer.args = infer.args[::-1]
    assert str(soup) == r'$$\infer{B}{A}\infer{D}{C}$$'

[qwenger]

Wow, that's exactly the issue I was searching for.

Thank you very much @alvinwan for this really handy library! Would you consider pushing a release to PyPI for this new math parsing feature?

[jonathan-laurent]

I just tried it and it works great. Thanks!

[alvinwan]

@qwenger yup, just pushed! https://pypi.org/project/TexSoup/0.3.1/