Pandoc, Weasyprint, and PDF generation

From plain text to PDF


I love writing in plain text. Given the choice between writing a document in Word, Google Docs, or my text editor (I primarily use atom, don’t @ me), I will choose my text editor every time. I prefer my text editor because I don’t want to pay attention to layout / formatting when I’m writing, I just want to write.

But no one wants to read a document in a text editor; they want to read on a tablet, or in the browser, or to print a document. Fortunately, there’s a surprisingly powerful tool, pandoc that I have been using a lot lately and that makes conversion of plain text into readable formats an easy task.

Without spending too much time on the details, pandoc is a command-line tool that converts documents from one format to another. Have a markdown file and want a .docx? Pandoc can do that! Have an ePub and want an HTML? Pandoc to the rescue.

For example, this blog post is, itself, written in markdown. But suppose I wanted a PDF copy of it. Using pandoc, all I need to do is to run:

pandoc content/posts/2020-01-25-pandoc-weasyprint-and-pdfs.md -o static/files/pandoc-test.pdf

If you’re interested, you can see what it looks like.

But let’s say you wanted to get fancy with your PDF, and add some better styles. For example, suppose I wanted to add colors to my links. To do that, I need to pass pandoc a “variable,” specifically: colorlinks: true. At the command line, I run:

pandoc content/posts/2020-01-25-pandoc-weasyprint-and-pdfs.md -V colorlinks:true -o static/files/pandoc-basic-style-test.pdf

Check it out, you can see there are colors now.

Now, let’s say you want to go ahead and really gussy it up. It turns out you can use Cascading Style Sheets when creating a PDF, too. Don’t like the default font for the body text? Just create a styles.css file, add a -t html flag, and run:

pandoc content/posts/2020-01-25-pandoc-weasyprint-and-pdfs.md --css=styles.css -t html -o static/files/pandoc-styled-test.pdf

Another improvement! Though, frustratingly, there seem to be some CSS limitations / elements that just don’t work. For example, I really want my links to have dotted underlines, but it just won’t do it.

So now, for the final touch, if I really really want a beautiful PDF, I can use Weasyprint, which is another tool that specializes in converting HTML to PDF. Here, I use the exact same stylesheet I used before, but pass a -s (standalone) flag and pipe to weasyprint:

pandoc content/posts/2020-01-25-pandoc-weasyprint-and-pdfs.md -s -t html | weasyprint - static/files/pandoc-weasyprint-test.pdf -s styles.css

Now, it’s just like I want it!

There are obviously many more things that are possible options available here, but it’s exciting to realize that plain text can be fun to write, and beautiful to read.