DominikPeters

Indeed I have scripts for compiling a given tikz figure using latex (in particular dvisvgm so I get an SVG instead of a PDF) as well as my js-based renderer. I apply that script to various corpuses, mostly particular pages from the tikz manual (see https://tikz.dev), but there are also a few books about TikZ that have downloadable zips of all the examples they use. I then inspect the correspondence between the two renderers by eye and give Codex a list of which figures are wrong and why, and it then goes and fixes the underlying issues.

You'd think that finding discrepancies between the renderers could be done automatically, but it hasn't worked well in my experience. The models are multimodal but still kinda blind; they think two pictures are the same even if they are very much not the same. But once you tell them whats wrong, they're then pretty good at iterating until it is fixed. (One could also try to do a pixel diff of rasterized images, but that's super noisy, and text rendering isn't going to be pixel perfect anyway.)

  \begin{tikzpicture}
    \draw (0,2) rectangle (1,1);
    \draw (1.5,2) rectangle (2.5,1);
  \end{tikzpicture}

Elsewhere in this discussion, we talked about positioning like "right of", and some good suggestions were made (https://news.ycombinator.com/item?id=48647683).