How do I copy a remote image in python?

Question

I need to copy a remote image (for example http://example.com/image.jpg) to my server. Is this possible?

How do you verify that this is indeed an image?

Answer 1

To download:

import urllib2
img = urllib2.urlopen("http://example.com/image.jpg").read()

To verify can use PIL

import StringIO
from PIL import Image
try:
    im = Image.open(StringIO.StringIO(img))
    im.verify()
except Exception, e:
    # The image is not valid

If you just want to verify this is an image even if the image data is not valid: You can use imghdr

import imghdr
imghdr.what('ignore', img)

The method checks the headers and determines the image type. It will return None if the image was not identifiable.

Answer 2

Downloading stuff

import urllib
url = "http://example.com/image.jpg"
fname = "image.jpg"
urllib.urlretrieve( url, fname )

Verifying that it is a image can be done in many ways. The hardest check is opening the file with the Python Image Library and see if it throws an error.

If you want to check the file type before downloading, look at the mime-type the remote server gives.

import urllib
url = "http://example.com/image.jpg"
fname = "image.jpg"
opener = urllib.urlopen( url )
if opener.headers.maintype == 'image':
    # you get the idea
    open( fname, 'wb').write( opener.read() )

Answer 3

Same thing using httplib2...

from PIL import Image
from StringIO import StringIO
from httplib2 import Http

# retrieve image
http = Http()
request, content = http.request('http://www.server.com/path/to/image.jpg')
im = Image.open(StringIO(content))

# is it valid?
try:
    im.verify()
except Exception:
    pass  # not valid

Answer 4

For the portion of the question with respect to copying a remote image, here's an answer inspired by this answer:

import urllib2
import shutil

url = 'http://dummyimage.com/100' # returns a dynamically generated PNG
local_file_name = 'dummy100x100.png'

remote_file = urllib2.urlopen(url)
with open(local_file_name, 'wb') as local_file:
    shutil.copyfileobj(remote_file, local_file)

Note that this approach will work for copying a remote file of any binary media type.