Make a standalone PDF with embedded images?

[richardsprague]

I'm trying to make a PDF book based on some markdown files that use images linked from external URLs. Is there a way to embed the images into the PDF file, similar to how I can create an HTML file with embed-resources: true ?

Here's a simple example of what I mean. Is there a way to turn this into a PDF where the logo is embedded (i.e. visible offline)?

---
title: "Intro"
---

# Sample Page

This is a simple page with the following image:

<figure>
<img src="https://quarto.org/quarto.png" />
<figcaption>An image</figcaption>
</figure>

[cscheid]

Unfortunately, Pandoc will not attempt to resolve external URLs; you need to download them yourself. (And you'll need to use Markdown syntax for the images instead of HTML).

[dragonstyle]

Carlos is right for now, but...

Starting in quarto 1.3 prerelease (including the current prerelease) remote markdown images for PDF should be resolved into local images embedded in the PDF.

[richardsprague]

Thanks! Note that pandoc standalone kinda does what I want: pandoc -f html -t docx -o final.docx docs/index.html will generate a Word file with embedded remote images, and then pandoc -f docx -t pdf -o final.pdf final.docx will convert that to PDF. Is there no way to "chain" these pandoc calls from within Quarto?

Here's my _quarto.yml:

project:
  title: "Sample Book"
  type: book
  output-dir: docs


book:
  title: "This is a book"
  chapters:
    - index.qmd

format:
  html:
    theme: flatly
    embed-resources: true
    standalone: true
  docx:
    fig-align: right
    toc: true
    standalone: true
    extract-media: images
    from: html

I tried to add:

 pdf:
    documentclass: scrbook
    mainfont: "Georgia"
    from: docx

But that gives an error Package babel Error: Bad optionbidi=default'`

[dragonstyle]

If you build your book with Quarto 1.3, it should just work (if you're open to trying our pre-release):

https://quarto.org/docs/download/prerelease

[cscheid]

@dragonstyle TIL, thank you for setting me straight!

[dragonstyle]

It was a slight side project / sneak in for 1.3, but we actually needed to do it because Pandoc actually does the right thing for PDFs when it renders the PDF (resolving remote URLs). Since we take over PDF rendering, however, we weren't getting that feature from Pandoc which wasn't good and you were definitely right for all previous versions of Quarto!

[richardsprague]

I tried with 1.3.249, but it didn't seem to work. Do I need different pdf settings in _quarto.yml.

format:
  html:
    theme: flatly
    embed-resources: true
    standalone: true
  docx:
    fig-align: right
    toc: true
    standalone: true
    extract-media: images
    from: html
  pdf:
    documentclass: scrbook
    mainfont: "Georgia"
    from: docx

I tried from: standalone but that didn't work either.

[dragonstyle]

Can you remove the 'from' under the pdf format? Is there an error when you try this?

[richardsprague]

No errors, either with or without the "from". You can see the full code at https://github.com/richardsprague/quartosample

[dragonstyle]

Looks like you're placing html in your document. Can you try making a markdown figure instead? We aren't seeing that html figure

[richardsprague]

Ah, that works! At least for simple images.

when I tried a more complicated remote URL, it chokes.

I filed a bug PDF with embedded images from remote URL with odd filenames #4747 if you want to track it.