Comments in Code cells are taken as section and increment numbering

[mcanouil]

Bug description

A comment in a code cell (qmd/ipynb) # a comment lead to the wrong overall header level to be used in the document for Typst and LaTeX formats.

Note that the behaviour is observed for all engines.

Steps to reproduce

Render the following document into Typst or LaTeX PDF.

Input	Output
--- title: XXXX format: typst: default pdf: default number-sections: true jupyter: python3 --- ## Introduction ```{python} # Import libraries for loading the dataset print("Hello World") ``` ### Subsection {{< lipsum 1 >}}

Expected behavior

Without having to double the comment symbol or use shift-heading-level-by: -1.

Input	Output
--- title: XXXX format: typst: default pdf: default number-sections: true jupyter: python3 --- ## Introduction ```{python} ## Import libraries for loading the dataset print("Hello World") ``` ### Subsection {{< lipsum 1 >}}

Your environment

• macOS Sequoia (latest)

Quarto check output

Quarto 99.9.9
[✓] Checking environment information...
      Quarto cache location: /Users/mcanouil/Library/Caches/quarto
[✓] Checking versions of quarto binary dependencies...
      Pandoc version 3.4.0: OK
      Dart Sass version 1.70.0: OK
      Deno version 1.46.3: OK
      Typst version 0.11.0: OK
[✓] Checking versions of quarto dependencies......OK
[✓] Checking Quarto installation......OK
      Version: 99.9.9
      commit: d0d6f87e18e7522386bc8b1ab5b390ef7eca3799
      Path: /Users/mcanouil/Projects/quarto/quarto-cli/package/dist/bin

[✓] Checking tools....................OK
      TinyTeX: v2024.11
      Chromium: (not installed)

[✓] Checking LaTeX....................OK
      Using: TinyTex
      Path: /Users/mcanouil/Library/TinyTeX/bin/universal-darwin
      Version: 2024

[✓] Checking basic markdown render....OK

[✓] Checking Python 3 installation....OK
      Version: 3.13.1
      Path: /Users/mcanouil/Projects/quarto/quarto-playground/.venv/bin/python3
      Jupyter: 5.7.2
      Kernels: julia-1.11, python3

[✓] Checking Jupyter engine render....OK

(|) Checking R installation...........ℹ R version 4.4.2 (2024-10-31)
! Config '~/.Rprofile' was loaded!
[✓] Checking R installation...........OK
      Version: 4.4.2
      Path: /Library/Frameworks/R.framework/Resources
      LibPaths:
        - /Users/mcanouil/Projects/quarto/quarto-playground/renv/library/macos/R-4.4/aarch64-apple-darwin20
        - /Users/mcanouil/Library/Caches/org.R-project.R/R/renv/sandbox/macos/R-4.4/aarch64-apple-darwin20/f7156815
      knitr: 1.48
      rmarkdown: 2.29

[✓] Checking Knitr engine render......OK

[cscheid]

This bug almost certainly from the code that detects the highest level of headings. I bet we're using regexps there.

Another point for the "AST everywhere" project I keep talking about...

[cderv]

It does not happen for HTML it seems, so I was curious.

This is indeed LaTeX and Typst specific

quarto-cli/src/format/pdf/format-pdf.ts

Lines 256 to 273 in 52a6a5b

	// Don't shift the headings if we see any H1s (we can't shift up any longer)
	const hasLevelOneHeadings = !!markdown.match(/\n^#\s.*$/gm);
	// pdfs with no other heading level oriented options get their heading level shifted by -1
	if (
	!hasLevelOneHeadings &&
	autoShiftHeadings &&
	(flags?.[kNumberSections] === true ||
	format.pandoc[kNumberSections] === true) &&
	flags?.[kTopLevelDivision] === undefined &&
	format.pandoc?.[kTopLevelDivision] === undefined &&
	flags?.[kShiftHeadingLevelBy] === undefined &&
	format.pandoc?.[kShiftHeadingLevelBy] === undefined
	) {
	extras.pandoc = {
	[kShiftHeadingLevelBy]: -1,
	};
	}

quarto-cli/src/format/typst/format-typst.ts

Lines 69 to 78 in 52a6a5b

	// unless otherwise specified, pdfs with only level 2 or greater headings get their
	// heading level shifted by -1.
	const hasLevelOneHeadings = !!markdown.match(/\n^#\s.*$/gm);
	if (
	!hasLevelOneHeadings &&
	flags?.[kShiftHeadingLevelBy] === undefined &&
	format.pandoc?.[kShiftHeadingLevelBy] === undefined
	) {
	pandoc[kShiftHeadingLevelBy] = -1;
	}

So this is a processing specific to Quarto and we do indeed use regex in both place to search for headings.

quarto-cli/src/format/typst/format-typst.ts

Lines 70 to 71 in 52a6a5b

	// heading level shifted by -1.
	const hasLevelOneHeadings = !!markdown.match(/\n^#\s.*$/gm);

quarto-cli/src/format/pdf/format-pdf.ts

Lines 256 to 257 in 52a6a5b

	// Don't shift the headings if we see any H1s (we can't shift up any longer)
	const hasLevelOneHeadings = !!markdown.match(/\n^#\s.*$/gm);

On way to fix this would be a Pandoc call with a Lua filter giving us information if any H1.
Or a conversion with pandoc to json ast format to parse. Something like that, that does provided us with AST. In R Markdown ecosystem, I think we needed something like this, and we used a basic --to html conversion to then query DOM for any expected node.

In any case, this is another call to pandoc if we go that road.

[mcanouil]

It does not happen for HTML it seems, so I was curious.

This is indeed LaTeX and Typst specific

Yes, sorry I only implied it.
I did not test docx though.

[cscheid]

The truly correct fix here would be for us to operate on the AST (either via Pandoc or some other parser).

A slightly better way to this would be to use the MappedString infrastructure and breakQuartoMd, but I think it will require adding a few functions to the mapped-text.ts API, and I don't know if that's worth the trouble. But, you could:

• create a version of searchAll in src/core/lib/mapped-text.ts that takes a mappedstring returns an array of mappedstring results.
• use breakQuartoMd function to find code blocks, and then check that the returned mappedstring values are not inside any code blocks.

Edit: it's even easier than that, because formatExtras uses a string instead of a MappedString. I'll make a PR.