Convert svg images to png for pdf documentation

[ale5000-git]

Details

• Read the Docs project URL: http://pycompatlayer.readthedocs.io/en/latest/intro.html

Expected Result

Hi,
I have some svg dynamically generated server-side by external services (badges).

Is it possible to automatically convert them to png in the pdf documentation?

Actual Result

Currently they doesn't appear at all in the pdf.

[stsewd]

This is a sphinx issue sphinx-doc/sphinx#4720. There are some solutions there.

[ale5000-git]

Hi,
I got a reply in the sphinx issue:

It seems readthedocs.org does not provide imagemagick. So sphinx.ext.imgconverter will not work.
I don't know how to enable it on readthedocs. Please ask to the administrator.

How can it be fixed?

[stsewd]

I'm not sure how the imagemagick plugin works, I'll try to search about that later

[stsewd]

Ok, reading the docs http://www.sphinx-doc.org/en/master/ext/imgconverter.html RTD needs to install imagemagick in its docker images https://github.com/rtfd/readthedocs-docker-images.

This would be a feature request.

[mgeier]

Alternatively, the extension sphinxcontrib-svg2pdfconverter could be used to convert SVG to PDF (which doesn't turn the images into bitmaps!). See #4797 (and #905).

[Moini]

Tbh. I don't care very much about the 'how', but I would definitely like to see vector graphics used at readthedocs in exported pdfs. I'm trying to build an Inkscape Manual here....

[humitos]

Alternatively, the extension sphinxcontrib-svg2pdfconverter could be used to convert SVG to PDF (which doesn't turn the images into bitmaps!). See #4797 (and #905).

I suppose that this extension needs readthedocs/readthedocs-docker-images#79 (librsvg2-bin package), right?

By default, without any kind of extension, Sphinx will automatically use imagemagick, right?

If I'm right in both, I would say that we want to add imagemagick to the docker image also.

[humitos]

I read your suggestion from readthedocs/readthedocs-docker-images#70 (comment) and I think that I'm right and you have a point with that extension. Although, I think we should support both possibilities.

[mgeier]

@humitos

Yes, it needs the librsvg2-bin package.

By default, without any kind of extension, Sphinx will automatically use imagemagick, right?

You still need to explicitly enable the built-in extension sphinx.ext.imgconverter.

Although, I think we should support both possibilities.

I agree, having both would be good.

Even though most of the times SVG -> PDF looks better, there are also cases where SVG -> PNG is preferable (e.g. when the SVG is overly complex).

[humitos]

@mgeier thanks!

I'm closing this issue in favor of readthedocs/readthedocs-docker-images#78 and readthedocs/readthedocs-docker-images#79 and readthedocs/readthedocs-docker-images#80