Convert SVG to PDF (svglib + reportlab not good enough)

Question

I'm creating some SVGs in batches and need to convert those to a PDF document for printing. I've been trying to use svglib and its svg2rlg method but I've just discovered that it's absolutely appalling at preserving the vector graphics in my document. It can barely position text correctly.

My dynamically-generated SVG is well formed and I've tested svglib on the raw input to make sure it's not a problem I'm introducing.

So what are my options past svglib and ReportLab? It either has to be free or very cheap as we're already out of budget on the project this is part of. We can't afford the 1k/year fee for ReportLab Plus.

I'm using Python but at this stage, I'm happy as long as it runs on our Ubuntu server.

Edit: Tested Prince. Better but it's still ignoring half the document.

Answer 1

CairoSVG is the one I am using:

import cairosvg
cairosvg.svg2pdf(url='image.svg', write_to='image.pdf')

Answer 2

I use inkscape for this. In your django view do like:

from subprocess import Popen

x = Popen(['/usr/bin/inkscape', your_svg_input, \
    '--export-pdf=%s' % your_pdf_output])
try:
    waitForResponse(x)
except OSError, e:
    return False

def waitForResponse(x): 
    out, err = x.communicate() 
    if x.returncode < 0: 
        r = "Popen returncode: " + str(x.returncode) 
        raise OSError(r)

You may need to pass as parameters to inkscape all the font files you refer to in your .svg, so keep that in mind if your text does not appear correctly on the .pdf output.

Answer 3

rst2pdf uses reportlab for generating PDFs. It can use inkscape and pdfrw for reading PDFs.

pdfrw itself has some examples that show reading PDFs and using reportlab to output.

Addressing the comment by Martin below (I can edit this answer, but do not have the reputation to comment on a comment on it...):

reportlab knows nothing about SVG files. Some tools, like svg2rlg, attempt to recreate an SVG image into a PDF by drawing them into the reportlab canvas. But you can do this a different way with pdfrw -- if you can use another tool to convert the SVG file into a PDF image, then pdfrw can take that converted PDF, and add it as a form XObject into the PDF that you are generating with reportlab. As far as reportlab is concerned, it is really no different than placing a JPEG image.

Some tools will do terrible things to your SVG files (rasterizing them, for example). In my experience, inkscape usually does a pretty good job, and leaves them in a vector format. You can even do this headless, e.g. "inkscape my.svg -A my.pdf".

The entire reason I wrote pdfrw in the first place was for this exact use-case -- being able to reuse vector images in new PDFs created by reportlab.