make minor modifications to the pdfs generated?

[zenny]

Hi @jgm,

Great package that replaces most of other file format conversion tools! +1

I am using pandoc both from command line, and also using ox-pandoc emacs package (I have raised two issues, reportedly assuming that it has to do with eihther with emacs configurations (jkitchin/scimax#509) or with the ox-pandoc package ( emacsorphanage/ox-pandoc#43) which does not seem to be the case once I have started exploring a bit further as suggested by the developers in the related threads.

I am trying to achieve four things with pandoc because they are bottlenecks to the workflow of any scribe (discussed in pprevos/emacs-writing-studio#24) :

1. Change the font from the default pandoc template while exporting from the orgfile or any other input files pandoc supports.
2. Insert a table of contents on a separate page from the title page and the rest of the contents ( tried appending #+pandoc-options: or latex-header: inside the org file, but did not work).
3. Specify LaTeX template (like kaobook (https://github.com/fmarotta/kaobook) ) either on orgmode file header or on command line to make the output file look as specified when exported to pdf.
4. Using wkhtmltopdf does a fantastic job to wrap the tables which is nearly impossible in orgmode without complex LaTeX configurationy, but generates a redundant horizontal scroll bar as if it is an webpage, not in pdfs.

I appreciate if there are ways to overcome the bottlenecks as above. Thanking you in anticipation,

Cheers,
/zenny

[jgm]

1. Changing the font (for PDF output) just requires setting the mainfont as a variable or in metadata (assuming you're using --pdf-engine=xelatex which is recommended). See the manual under "Variables for LaTeX".
2. You could probably achieve this by beginning your document with \newpage.
3. kaobook is a class not a template. You could try setting documentclass and classoption variables. Not sure how well it would work; it depends on the class.
4. Surprised to hear this. wkhtmltopdf is deprecated anyway, try weasyprint or pagejs-cli. You could also try fiddling with --pdf-engine-opt to pass options to wkhtmltopdf (consult its manual).

[zenny]

Hi again @jgm:

Thanks for taking time to reply promptly.

1. I tried specifying the specific font in the org file #+LaTeX_Header: \usepackage{junicode}, but pandoc export to pdf did not comply.
2. I tried with \newpage in org document, but had no effect. I have tried by inserting some LaTeX options as of below, but it did not generate TOC at all!

#+OPTIONS: timestamp:nil toc:nil
#+BEGIN_EXPORT latex
   \clearpage \setcounter{tocdepth}{5} \tableofcontents \clearpage
#+END_EXPORT

3. Thanks for throwing light on the class stuff. I shall try to learn a bit more about it. Actually, I want to do everything from orgfile without touching the TeX file to make the workflow simple, fyi.
4. Glad that you explained about the pdf-rendering engines. I tried with weasyprint, but the rendering remained the same. If I rendered to pdf using from -f org to -t html5 with in --standalone mode, some tables are rendered alright, but some tables are rendered (the table at the bottom in the screenshot below) with last column contents cropped (the one at the top in the screenshot below), meaning rendering is not consistent, as seen in the screenshot below:
   

I appreciate if pandoc veterans here could hint me to overcome the issues as I have explained above. Have a nice weekend.

Cheers,
/zenny

[zenny]

An update:

1 (font issue) and 4 (table crop issue) above got resolved by appending the following:

#+PANDOC_OPTIONS: standalone:t table-of-contents:true toc-depth:5 pdf-engine:weasyprint number-sections:nil
#+PANDOC_VARIABLES: classoption:10pt,conference mainfont:Junicode sansfont:Monserrat monofont:FiraCodeMono mathfont:TeXGyreDejaVuMath

With help from:

1. https://github.com/kawabata/ox-pandoc/blob/aa37dc7e94213d4ebedb85c384c1ba35007da18e/README.org?plain=1#L135
2. https://superuser.com/questions/1349187/pandoc-how-to-get-pagebreak-between-title-block-and-the-table-of-contents
3. https://tex.stackexchange.com/questions/234786/how-to-set-a-font-family-with-pandoc
4. https://pandoc.org/org.html

Yet, I could not separate cover page and table of contents from the remaining body of text in separate pages even with:

#+PANDOC_METADATA: include-before: - '`\newpage{}`{=latex}' 

as advised in https://superuser.com/a/1349432. Any input appreciated.

[iandol]

For precise control, edit the pandoc templates: https://pandoc.org/MANUAL.html#templates -- you can see them all online here: https://github.com/jgm/pandoc-templates/blob/master/default.latex

Though looking at the default.latex template where include-before is it seems in the right place and should be working: https://github.com/jgm/pandoc-templates/blob/master/default.latex#L65 -- are you sure the latex looks ok (i.e. check the latex output before it becomes a PDF, maybe org mode is causing escapes in the latex characters or soemthing?)

Also PrinceXML is by far the best HTML to PDF engine, though it is not free for commercial use. For PDF overall you should seriously consider Typst. It now does much of what LaTeX can do, all in one tiny 10MB download with orders of magnitude less cruft and complexity...

[iandol]

You probably have some problem with how include-before is converting the string from org mode into metadata; at least with YAML metadata something like:

---
title: A Test
author: Jane Doe
include-before: '`\newpage{}`{=latex}'
---

# Sun

Some *text*.

# Moon

Some *text*.

The \newpage command is in the right place and causes a pagebreak before TOC, i.e. it works. But it could be that going from org mode to YAML or soemthing causes the \n of newpage to get escaped or something?