Pandoc Templates | Hacker News Reader

Pandoc Templates

pandoc-templates.org

441 pointsankitg1259 comments

I used to use pandoc for my bachelors papers, which needed to be submitted as word documents. I never used templates but had a rather large "one-liner" pandoc command to convert my markdown files.

At the time I'd not got round to understanding the yaml front matter etc. I even user Zettlr for a while [0].

I then discovered quarto [1] and this changed everything. Much nicer experience. I used this for my masters papers.

I think the tooling around pandoc is what makes it such a good tool. I remember attempting restructured text and latex and having a right hard time.

[0] https://zettlr.com/ [1] https://quarto.org/

Neat. Now I need Quarto, but with Elixir and Livebook!

I have my crazy notes on Quarto and word documents here, https://github.com/apwheele/Blog_Code/tree/master/Quarto/Rep.... Hopefully useful for others reading these comments.

I don't even know what magic buttons I need to push to get that template to correctly inherit the table format I wanted from pandoc, but it does. I tend to have other scripts though for more complicated tables though. So if I want a table to have a certain row highlighted a different color, I would write a Powershell script to run after the table was generated.

I was never able to figure out how to use LibreOffice to insert the table of contents and then export to PDF (although I can do it via the GUI).

pandoc is my "document converter" go-to tool.

Quarto is my documentation tool.

For me, they are both massively used, but cover different usages.

Pandoc is such an amazing piece of software. I used it to format my novel and made it part of a GitHub action to produce all the formats I required. I wasn't aware of templates, but some look really sleek.

I keep thinking that modern text editors are just flawed and markdown, with all its downsides and limitations, is what 99% is the people need.

For the short, simple documents that most people make, a versioned, wysiwyg word processor is going to beat everything else.

I mean, they don't want to think about building the output, never mind controlling the process.

For most of the short simple documents I create, I don't want to redo the formating for every document. Simply writing it in something simple like Markdown ( possibly a markdown wysiwig editor) and having my software automatically apply appropriate standard formats to it is ideal.

Right, most people don't want to do that, they want the burden of applying styles to the couple headings or whatever.

Unfortunately, most people don't use paragraph styles, but if you do, it's a couple clicks.

Agreed. There is actually a lot better control in openoffice / libreoffice than most people know. You just have to set up your styles and be systematic about (virtually) never using direct formatting, instead always applying a pre-configured style. There is a distinct value in seeing your final product as you work, when the final product is visual.

This is more of a utopia than expecting the average office drone to learn emacs.

And if you define shortcut keys for your styles it's as quick to type as markdown.

This is what Word is to most people. They just use the default styles, or their company's template. No special formatting, styling etc.

Also if you do want to add a table or a figure, for most people Word will be much easier than doing the same in Markdown.

Having handled Word documents a few times in my years, oh, how I wish this was true.

Building my resume in a wysiwyg editor was an exercise in frustration. Formatting was inconsistent, they were only searchable from inside the editor and versioning was useless because diff had no meaning.

My markdown resume has its own problems but having this level of control has been a huge load off my mind.

I started maintaining mine in plain text, no markup at all, 20 years ago and try to use it wherever I can. My favorite employer, a startup, liked it. Regrettably most companies these days lack a sense of humor.

I'm a programmer and even I like writing in a non-programmable environment. Programming in the document system just stimulates the more primitive parts of my brain that love the processing and programming more than the writing itself. So it's distracting in that way.

WYSIWYG pretty consistently leads to visual and structural messes. It's only going to "beat" everything else if you don't care about quality.

Most people don't—and don't have to—care about quality for their short, simple documents, but that is neither good nor inevitable, and it's always worth trying to do better.

You may be overestimating technical abilities of 99% of the people. I tried to convert some to pandoc and failed miserably. Personally I love it, markdown is becoming more and more central to my workflows.

Haskell thriving in the wild!

You should be aware that pandoc markdown has extensive non-standard additions to the language to make it usable.

If you want a pure markup language that is simple, plain text readable and able to produce text more complex than what a type writer could manage in 1920 then restructured text is the way to go.

pandoc has infinitely many devices for including various commands. A lua filter - to take one standard example - can manage basically anything you cook up. The going AIs can write them for you and your triggering syntax at the drop of a hat. Inter alia, you can write your restructured text in markdown and include bits like this.

Here is normal Pandoc Markdown text.

```{=rst} .. note:: This is a native reStructuredText directive! Pandoc will not parse or change this text. It goes straight to the rST output writer. ```

Back to normal Markdown text.

Pandoc is an impressive piece of software but I could never quite get PDF generation working nicely with it.

Table layouts were often broken, with text overlapping into adjacent fields. Unicode font fallback didn't work properly, with characters like "→" being silently dropped because they didn't exist in the main font. Having predictable control of page breaks, to avoid situations where header text didn't stick to the following paragraph and instead had header and paragraph text split over a page boundary, was pretty much impossible.

I ended up concluding that Markdown isn't a sufficiently powerful markup language for page-based documents, and went back to using Word in all its WYSIWYG delight.

That said, maybe there were ways of doing all of the above but I couldn't figure it out and found the whole process of wrestling with with both Markdown and LaTeX templates, and Pandoc configuration, unintuitive and annoying.

I broke out a little HTML/CSS for this with page-break: after, now deprecated but the improved directives look better:

https://stackoverflow.com/questions/1664049/can-i-force-a-pa...

To address that, a long while back I used a Markdown to Ruby Prawn to PDF generator. I could not find a generic mainstream tool supporting this…

Use Pandoc w/ Typst. I've published many books using it.

In fact, just had a friend with a traditionally published book who is now self-pubbing ask me yesterday about my tools. I recommended Pandoc and Typst. He (surprisingly to me) had never heard of Typst, but within the hour replied that (with the use of AI) he had a great-looking template for the book. (Try doing that with LaTeX).

Could you elaborate on your workflow? Which tool do you start in (presumably it's a sequential flow?)? When is it that you go across the tools?

Jupyter to markdown to typst. Pandoc does the last conversation. (Pandoc also creates epubs for me).

I have some custom filters for index entries and code formatting.

I'm curious if Quarto could coordinate all of that for you automatically. It supports both code execution via Jupyter, and output via Typst (including books).

https://quarto.org/docs/computations/python.html

https://quarto.org/docs/output-formats/typst.html

https://quarto.org/docs/books/book-output.html#typst-output

https://quarto.org/docs/output-formats/typst-custom.html#boo...

"→" is standard unicode; pandoc would have put it into the internal AST. If you were writing a pdf e.g. via pdflatex and tex fonts in olden times, a number of things might have gone wrong. Such a thing will pass unproblematically into the things like the going luatex - but similarly if you write the pdf via e.g. typst or one of the other engines.

It is true one ends up familiarizing oneself with peculiarities of the typesetting engine. The LLMs are lately teaching me bits and pieces about typst and typst+pandoc which I'm using for a writing project. They all seem infallible about pandoc curiosa.

Tables I agree are a nightmare in any plain text representation whatsoever - except with editor support which alas Word makes easy.

I am a heavy user of Pandoc. As I write all my text in markdown using Obsidian, but have to create content for the MS Office environment, I use Pandoc to convert my markdown content into ms office formated content.

I would be lost had I have to use the Office tools to edit and format my text.

So thank you to all the maintainers of Pandoc.

I used it to output my doctoral thesis in LaTeX from Markdown 10 years ago, and similarly for going back and forth between my supervisor's Word documents and the main thesis text.

Embarrassingly, a horrible little script for converting Pandoc's Markdown endnotes to inline format remains my most-starred GitHub repo: https://github.com/ltrgoddard/inliner/

Basic familiarity with the paragraph styles in Word is like a 20 minute task.

If you are using markdown, you already understand the conceptual basis for it, so you just need to understand how it's implemented over there.

I'm not arguing that it is something you should do, just rolling my eyes at "I would be lost".

Not lost because it's hard to learn, but because I don't like writing in ms office products. It's not just word, I write formated long emails in outlook as well.

Somehow related is https://www.metanorma.org/ (using Markdown to produce norms-compatible outputs).

Oh wow, I use Pandoc fairly extensively, and have my own templates, and I never knew you could make things as colorful as some of these.

Oh no, inspiration has arrived. Guess I know what I'm wasting my weekend into, hah.

Also this page seems to have existed for a while and I never heard of it! I'm glad I stumbled upon this. A lot of nice ideas here.

https://typst.app/

Isn't this a paid for app?

The upper tier features on the webapp are paid, but the local tools are all free. That includes local editor support. IIRC the compiler is apache 2.

You against paying for things?

That's your takeaway from my question?

Yeah idk about all the typst HN shilling. Latex works great and is as free as it can get.

You can use Typst 100% for free, just like LaTeX. There's a CLI for compilation, there's an LSP, code formatter, etc. Complaining that the web app has a paid tier is like complaining that Overleaf is paid.

As for why people shill Typst over LaTeX, it's just a better overall experience. Things that are annoying in LaTeX are easy in Typst. I've written plenty of LaTeX for academic papers and my Master's dissertation, but I'm now writing my PhD thesis in Typst. It's so much better. The only barrier to using it everywhere is that my colleagues still prefer to use Overleaf for collaboration, which forces me to use LaTeX for papers.

Fellow Typst user for my PhD thesis, can concur.

Beware it does not export to docx/libreoffice format. So not usable in the exact same situations.

https://typst.app/docs/web-app/export-and-preview/

To me, Pandoc is "never have to touch Word again", in other _words_, salvation.

Sadly, the .docx templates aren't enough - they have the styling, but for those forced to produce a .docx output as a necessity for "corporate" consumption - you will find it that it's not quite straightforward to insert a title page, and out of the reference docx file, only the styles are used, not the content. You can define headers and footers in the reference file, but not more than that.

As of 3.2.1, Pandoc supports OpenXML template files for its docx writer as well as pre and post body insertion: --template, --include-before-body and --include-after-body, but if you prefer to define your title page in Word, or haven't gotten your head around doing an OpenXML thing like me, I found this Windows tool:

https://github.com/jamessantiago/DocxMerge

You can define your title page in one docx file and then merge it with Pandoc's output. But since styles from both get merged, it's best to start your title page with a copy of your reference and/or delete as many styles as possible from your title page so you don't have to make changes to both.

There are also some other quirks like fitting tables to contents properly in docx - a workaround is a PowerShell script (shock, horror), that opens the file, iterates over tables, sets autofit and saves the document.

Doing "Werd" right: edit markdown, type "make". -> docx, pdf. The world is a good place sometimes.

I’ve been looking for a template to use for fancy business reports, so I can do my stuff in R/Python/QMD and management can get something colorful to look at without me having to copy paste everything into PowerPoint

You might find Quarto ( https://quarto.org/ ) interesting. It sounds like it does what you want.

I have been relying on pandoc for many years and had no idea I could use templates like this, which I suppose is pathetic but also indicates just how powerful the defaults are on their own.

Eisvogel template is a fantastic way to use Latex without knowing much about it, by transforming your markdown with Pandoc, mind you i used it before LLMs existed, but it was great to turn the notes of a project of a tesis into a super pro formated version without learning Latex.

I have been using Typst exclusively for a year. I've commercially published with it, no issues.

I’ve been building something somewhat adjacent to this. It’s https://sdocs.dev. It’s as 100% private browser based Markdown renderer

Pandoc can be a bit painful to set up, but once it’s in your workflow, it’s hard to replace. Especially when you care about repeatable formatting and multiple outputs.

Pandoc templates need an update for better logical operators. The supported yaml and conditionals are poor for even light use cases.

I am stunned by the beauty of these. The Tufte template is particularly unexpected and lovely

I've always wanted to make a GUI client for pandoc

Looks very nice but still takes some effort to undertsand how to use it. Isn't it simpler to ask the AI to create the PDFs for you?

typst templates would be a great addition to these.

2 hours, 56 points and not a single comment?

Perhaps people are quasi-bookmarking it

Reply on news.ycombinator.com