Parsing improvements for Arxiv papers

[jseppanen]

Here's a set of parsing improvements implemented when processing tens of thousands of Arxiv papers. We've added the following features:

• parsing TeX primitives (\topskip, ...)
• parsing \cmidrule
• parsing \tikzstyle
• parsing \verb
• better argument handling for \begin

There's also tests included for the new parser features.

Also removes a duplicate test case ("Issue #10").

[clason]

👀