converting html to pdf - no equations

[wrobell]

Trying to convert a page to PDF with

$ pandoc --pdf-engine=xelatex -f html -t latex https://mlg.eng.cam.ac.uk/blog/2021/03/31/what-keeps-a-bayesian-awake-at-night-part-1.html -o what-keeps.pdf

The equations in PDF file are

$ pandoc --version
pandoc 3.1.11.1
Features: +server +lua
Scripting engine: Lua 5.4
User data directory: /home/wrobell/.local/share/pandoc
Copyright (C) 2006-2023 John MacFarlane. Web: https://pandoc.org
This is free software; see the source for copying conditions. There is no
warranty, not even for merchantability or fitness for a particular purpose.

Seems to be related to #2758

[jgm]

The HTML contains things like

<p>\begin{equation}
    p(D) = \int p(\theta) \prod_{n=1}^N p(D_n \cond \theta) \,\mathrm{d}\theta.
\end{equation}</p>

So you need to enable the raw_tex extension: try with -f html+raw_tex+tex_math_dollars.

[wrobell]

@jgm thanks for the suggestions, i have tried:

greymatter$ pandoc --pdf-engine=xelatex -f html+raw_tex+tex_math_dollars -t latex https://mlg.eng.cam.ac.uk/blog/2021/03/31/what-keeps-a-bayesian-awake-at-night-part-1.html -o what-keeps.pdf
Error producing PDF.
! Undefined control sequence.
l.183     p( X \cond

greymatter$ pandoc --pdf-engine=xelatex -f html+tex_math_dollars -t latex https://mlg.eng.cam.ac.uk/blog/2021/03/31/what-keeps-a-bayesian-awake-at-night-part-1.html -o what-keeps.pdf
Error producing PDF.
! Undefined control sequence.
l.190 variables (\(\hat X_{\text{MAP}} = \argmax

greymatter$ pandoc --pdf-engine=xelatex -f html+raw_tex -t latex https://mlg.eng.cam.ac.uk/blog/2021/03/31/what-keeps-a-bayesian-awake-at-night-part-1.html -o what-keeps.pdf
Error producing PDF.
! Undefined control sequence.
l.183     p( X \cond

[jgm]

Well, these aren't pandoc parsing errors; they are errors produced by pdflatex, presumably because it expects some packages to be loaded that our default template doesn't load. You need to figure out what these are and load them (e.g. by adding \usepackage{PACKAGENAME} in header-includes).
See https://tug.ctan.org/info/symbols/comprehensive/symbols-a4.pdf

[wrobell]

Looking at the source of the page, the following is being setup by for MathJax

    <script type="text/javascript">
    window.MathJax = {
        tex: {
            inlineMath: [['$', '$'], ['$$', '$$'], ['\\(', '\\)']],
            displayMath: [['$$', '$$'], ['\\[', '\\]']],
            packages: {'[+]': ['ams', 'newcommand']},
            tags: 'all',
            macros: {
                Normal: '\\mathcal{N}',
                R: '\\mathbb{R}',
                E: '\\mathbb{E}',
                C: '\\mathbb{C}',
                sd: '\\text{d}',
                pd: '\\partial',
                cond: '\\,|\\,',
                ll: '\\left\\langle',
                rr: '\\right\\rangle',
                fp: '\\operatorname{fp}',
                argmax: '\\operatorname{arg\\,max}',
                argmin: '\\operatorname{arg\\,min}',
                e: '\\varepsilon',
                code: ['{\\small \\texttt{#1}}', 1],
                parens: ['\\left( #1 \\right)', 1]
            }
        }
    };
</script>

This seems documented (see https://docs.mathjax.org/en/latest/web/configuration.html), so I am wondering if it would make sense for Pandoc to support it?

[jgm]

Pandoc doesn't attempt to parse the javascript and figure out mathjax settings, sorry.

[jgm]

It will handle macros, but only if you define them with the standard latex commands.

[wrobell]

@jgm Could that be a bug?

Create packages.tex

\newcommand{\Normal}{\mathcal{N}}
\newcommand{\R}{\mathbb{R}}
\newcommand{\E}{\mathbb{E}}
\newcommand{\C}{\mathbb{C}}
\newcommand{\sd}{\text{d}}
\newcommand{\pd}{\partial}
\newcommand{\cond}{\,|\,}
\newcommand{\rr}{\right\rangle}
\newcommand{\fp}{\operatorname{fp}}
\newcommand{\argmax}{\operatorname{arg\,max}}
\newcommand{\argmin}{\operatorname{arg\,min}}
\newcommand{\e}{\varepsilon}
\newcommand{\parens}[1]{\left( #1 \right)}

Run

$ pandoc -H packages.tex --pdf-engine=xelatex -f html+raw_tex+tex_math_dollars -t latex -o what-keeps.tex https://mlg.eng.cam.ac.uk/blog/2021/03/31/what-keeps-a-bayesian-awake-at-night-part-1.html

Then what-keeps.tex contains the HTML document, but there is no packages.tex content in the file.

Downloading the HTML file first, then converting to LaTeX, does seem to include packages.tex properly.