Simple LaTeX filter before PDF is produced?

[MHellmund]

PDF output: Is it possible to run the LaTeX file produced by quarto through a simple (e.g., shell or python) filter before quarto runs xelatex?
It is just a small question.
I can live with LaTeX output and a post-render script.

[mcanouil]

Could you elaborate on your real need and what you really want to do/achieve?

[MHellmund]

Formatting of code output cells.

• Pandoc wraps them in a verbatim environment.
• I want to use Verbatim from the fancyvrb package which has a lot of options, e.g. \begin{Verbatim}[xleftmargin=2em] for better alignment with the input code.
• \renewenvironment{verbatim}{\begin{Verbatim}{\end{Verbatim}} in the header makes strange TeX errors.
• pandoc has a listings option (and quarto passes it through), but it is used for both code and output and stops syntax highlighting.

Furthermore, there is a small ugliness with Julia. Example:

---
format:
  pdf:
    keep-tex: true
    keep-md: true
    keep-ipynb: true

jupyter: julia-1.8
---

```{julia}
x = 3
println(x)
x = 4
```

The generated ipynb file shows that Jupyter/IJulia produces 2 output cells of slightly different types, since writing to stdout and displaying the result of the last expression are handled differently. The .tex looks like

\begin{verbatim}
3
\end{verbatim}

\begin{verbatim}
4
\end{verbatim}

and this generates a large gap between the two lines. My script merges consecutive verbatim environments.

I guess all this can be done with a pandoc lua filter, but for me some lines of python or shell are much faster written and tested.

[mcanouil]

It would probably be easier to tweak the LaTeX partials to fit your needs instead of adding new LaTeX via the include-* YAML keys, see https://quarto.org/docs/journals/templates.html#latex-partials.

[MHellmund]

This is helpful. Thank you for pointing it out!