Retrieve JPG pictures from URL (Python)

Question

I'm trying to retrieve JPG images from http ULRs (to display them in a GUI) with the following Python code:

import urllib3
from cStringIO import StringIO
from PIL import Image

conn = urllib3.connection_from_url('http://www.solarspace.co.uk/')
img_file = conn.urlopen('GET', 'http://www.solarspace.co.uk/PlanetPics/Neptune/NeptuneAlt1.jpg')
image = StringIO(img_file.read())
image.seek(0)
resized_image = Image.open(image)

However, this gives me this error message: "IOError: cannot identify image file".

The reason why I'm using urllib3 is because I need a persistent connection (to send multiple requests), which is not available with urllib/urllib2.

Thanks in advance.

Answer 1

Seems to work if you use img_file.data instead of img_file.read(). img_file.read() is meant to be used when you specify preload_content=False on the request. Now that I think about it, it's not very intuitive, perhaps img_file.read() should be aware of the cached preloaded content, or perhaps it should raise an exception if it has already been consumed. The plan was to make preload_content=False default but turns out there are a lot of edge cases which fall under normal usage that are hard to satisfy. I opened a bug to fix this in the future: https://github.com/shazow/urllib3/issues/102

Anyways, using img_file.data should fix your problem. Sorry for the confusion! :)

Also I suggest using conn.request(...) instead of the lower-level conn.urlopen(...), and perhaps use PoolManager if you might be going cross-domain (no reason not to use it, really). Try this:

>>> http = urllib3.PoolManager()
>>> r = http.request('GET', 'http://www.solarspace.co.uk/PlanetPics/Neptune/NeptuneAlt1.jpg')
>>> resized_image = Image.open(StringIO(r.data))

Answer 2

As always, requests to the rescue:

>>> r = requests.get('http://www.solarspace.co.uk/PlanetPics/Neptune/NeptuneAlt1.jpg')
>>> i = Image.open(StringIO.StringIO(r.content))
>>> i.size
(262, 299)

Answer 3

When save it, you can do:

with open('##.jpg','wb') as fout:
     fout.write(r.content)