Let latexify decide on cdot

[baggepinnen]

cdot --> false leads to very confusing output, like the one below, is there a variable called $adelay$ or is it $a \cdot delay$?

The example comes from https://mtk.sciml.ai/dev/basics/Composition/

In this case, the correct answer is $a \cdot delay$, but if there was actually a variable called $adelay$ present in the system, the printed output would be incorrect.

[codecov-commenter]

Codecov Report

Merging #794 (fc3216d) into master (3bad882) will decrease coverage by 73.36%.
The diff coverage is n/a.

📣 This organization is not using Codecov’s GitHub App Integration. We recommend you install it so Codecov can continue to function properly for your repositories. Learn more

@@            Coverage Diff             @@
##           master    #794       +/-   ##
==========================================
- Coverage   76.41%   3.05%   -73.36%     
==========================================
  Files          26      26               
  Lines        3235    3142       -93     
==========================================
- Hits         2472      96     -2376     
- Misses        763    3046     +2283     

Impacted Files	Coverage Δ
src/latexify_recipes.jl	0.00% <ø> (-52.78%)	⬇️

... and 24 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[isaacsas]

Using cdot in printing a symbolic expression seems unusual for CAS systems. Maybe this should be controlled in ModelingToolkit via a system field that the corresponding recipe there can look at? I’ve found the opposite experience in reaction networks with many species, where the presence of lots of cdots can also make the output hard to read, so being able to turn this on and off for systems would be nice.

[baggepinnen]

The problem is that our current latexification is not smart enough to distinguish

julia> @variables d x dx;

julia> latexify(dx + d*x)
L"\begin{equation}
dx + d x
\end{equation}
"

$dx + d x$

which is a quite severe correctness issue in the printing. Unless we manage to sort out and prevent this, we should really go for the correct, albeit slightly uglier, option as default.

Using cdot in printing a symbolic expression seems unusual for CAS systems.

SymPy does the correct thing here and prints the dot

julia> @vars d x dx
(d, x, dx)

julia> dx + d*x
d⋅x + dx

[isaacsas]

I'd suggest inserting a space via \, instead. That is what, for example, Mathematica does., and much more in line with what someone would do when writing a paper in Latex.

[isaacsas]

Alternatively I guess one could use {\cdot} which should avoid the extra spacing introduced by \cdot itself. That looks consistent with the spacing being introduced in SymPy in the terminal display.

[baggepinnen]

I find both of these suggestions okay, as long as one can distinguish a variable name with multiple characters from multiplication