-
-
Save s417-lama/84bf66de1096c4587e8187092fb41684 to your computer and use it in GitHub Desktop.
| #!/bin/bash | |
| # | |
| # Convert an SVG file to a PDF file by using headless Chrome. | |
| # | |
| if [ $# -ne 2 ]; then | |
| echo "Usage: ./svg2pdf.bash input.svg output.pdf" 1>&2 | |
| exit 1 | |
| fi | |
| INPUT=$1 | |
| OUTPUT=$2 | |
| HTML=" | |
| <html> | |
| <head> | |
| <style> | |
| body { | |
| margin: 0; | |
| } | |
| </style> | |
| <script> | |
| function init() { | |
| const element = document.getElementById('targetsvg'); | |
| const positionInfo = element.getBoundingClientRect(); | |
| const height = positionInfo.height; | |
| const width = positionInfo.width; | |
| const style = document.createElement('style'); | |
| style.innerHTML = \`@page {margin: 0; size: \${width}px \${height}px}\`; | |
| document.head.appendChild(style); | |
| } | |
| window.onload = init; | |
| </script> | |
| </head> | |
| <body> | |
| <img id=\"targetsvg\" src=\"${INPUT}\"> | |
| </body> | |
| </html> | |
| " | |
| tmpfile=$(mktemp XXXXXX.html) | |
| trap "rm -f $tmpfile" EXIT | |
| echo $HTML > $tmpfile | |
| google-chrome --headless --disable-gpu --print-to-pdf=$OUTPUT $tmpfile |
I added Mac support in my fork. https://gist.github.com/asmeurer/b9851296578743db7c0e985939a462d9. I think it will also support Chromium (if I understand correctly Chromium uses chrome instead of google-chrome, but I didn't verify this).
In case it helps anyone, I also noticed that in some cases, this produces a PDF with an extra blank page. This answer suggests adding line-height: unset; to the CSS, but I couldn't get that to work. I was able to fix it by editing the SVG file and making the width and height a little tighter. I don't know if the process can be automated.
Can I ask, is it OK to use this script in a BSD licensed project (I need something like this for sympy)?
@asmeurer Thank you for your interest. As I released this script under public domain, you can use it freely without any acknowledgment.
Applied license:
https://gist.github.com/s417-lama/9ff410fdb3e2adead81f7ca035192f43
As a note, I have switched from this script to rsvg-convert command (librsvg2) to convert svg files to pdf files, because Chromium also has some issues regarding svg conversion.
For example, pattern fill in svg is converted as a raster image rather than vectors in Chromium.
rsvg-convert can handle pattern fill in an appropriate way and is stable (at least for my usage), so I'm currently using it.
Example usage:
rsvg-convert -f pdf foo.svg -o foo.pdf
@asmeurer Thank you for your interest. As I released this script under public domain, you can use it freely without any acknowledgment.
Thanks.
As a note, I have switched from this script to rsvg-convert command (librsvg2) to convert svg files to pdf files, because Chromium also has some issues regarding svg conversion.
That's good to know. The svgs I need to convert for SymPy get messed up in some way or other with every tool I've tried (Inkscape, rsvg-convert, convert, cairosvg) except for Chrome.
For everyone looking for a solution to the annoying blank page being appended: when sizing is not super critical just add 1px to the height in line 29.
Looking forward to dominant-baseline support in librsvg2 ... but in the meantime this works nicely.
I had a problem with a UTF-8 char.
That's why I propose to update the <head> with <meta charset="UTF-8"> inside.
Anyone knows what's the backend is used to implement conversion from SVG to PS in chromium? Thanks in advance!
Nice. It works really well Thanks.
I fork it to add support for files with spaces after going throw the same problem as this user: https://stackoverflow.com/questions/47790308/chrome-headless-in-ubuntu.