Call image.verify() on downloads to check they aren't corrupt. about comics HOT 5 CLOSED

Two alternatives, pasted by @adamcik the other day:

import pil

def option1(file):
    parser = pil.ImageFileParser()
    try:
        while True:
            data = file.read(1024)
            if not data:
                break
            parser.feed(data)
        image = parser.close()
    except IOError as e:
        print 'Bad image: %s' % e

def option2(path):
    image = pil.Image.open(path)
    try:
        image.load()
    except IOError as e:
        print 'Bad image: %s' % e

from comics.

Seems image.verify() will not catch half downloaded files it seems, while Image Magick's identify will. However, using image.load() will cause IOError: image file is truncated (232 bytes not processed) so that would probably be worth a try.

from comics.

Other possible checks could be:

• Ensure that file has some minimum number of colors
• Ensure that content-size from http header matches size of downloaded file
• Ensure that file size matches file size that header indicates

from comics.

Also im.getcolors(maxcolors) returns None once maxcolors is passed, so one could use im.getcolors(3) is None as at test to ensure that image has at least three colors. Not sure if this would add any value over just .verify() - but nice to know about at least.

http://www.pythonware.com/library/pil/handbook/image.htm#Image.getcolors

from comics.

option1 in the listing above is exactly the same as Django do in django/core/files/images.py to get the height and width from the image, so I would think option would be redundant in addition to using Django's ImageField.

option2 has been implemented in ac358a0.

from comics.