Why can't PIL identify SVG images from stream?

April 10, 2023

I’m working on a system that recieves bytes object containing an image in format which is not known in advance and is supposed to work with it. I am using PIL to do the work for me. The documentation says that PIL.Image.open() can identify the file, even when a file object is passed. I have tried and it works with other file formats. But SVG files seem to be problematic.

Is there a way to make PIL work by itself or do I have to identify the image by some other means and then force PIL.Image.open() to read the image as SVG?

MRE:

#!/usr/bin/env python3

import requests
import io
import PIL.Image

with requests.get("https://github.githubassets.com/favicons/favicon.svg") as r:
    data = r.content

with io.BytesIO(data) as image:
    PIL.Image.open(image)

Traceback (most recent call last):
  File "MRE.py", line 11, in <module>
    PIL.Image.open(image)
  File "/usr/lib/python3.11/site-packages/PIL/Image.py", line 3283, in open
    raise UnidentifiedImageError(msg)
PIL.UnidentifiedImageError: cannot identify image file <_io.BytesIO object at 0x7fe38b2457b0>

>Solution :

Pillow/PIL does not support SVG input.

You need to render to a raster graphic first if you want to process SVG input with PIL. This can be done, e.g., using CairoSVG module.

There is no good support for editing SVG graphics directly in Python. It can be done, however, by editing the underlying XML structure using lxml module.