如何在Python中找到base64编码图像的文件扩展名

Question

I have a base64 encoded image that I decode and save into an ImageField in Django. I want to give the file a random name, but I don't know the file extension.

I have "data:image/png;base64," prepended to the string and I know I could do some regex to extract the mimetype, but I'd like to know if there is a best practices way to go from "data:image/png;base64," to ".png" reliably. I don't want to have my handspun function break when someone suddenly wants to upload a strange image filetype that I don't support.

Answer 1

It is best practices to examine the file's contents rather than rely on something external to the file. Many emails attacks, for example, rely on mis-identifying the mime type so that an unsuspecting computer executes a file that it shouldn't. Fortunately, most image file extensions can be determined by looking at the first few bytes (after decoding the base64). Best practices, though, might be to use file magic which can be accessed via a python packages such as this one or this one .

Most image file extensions are obvious from the mimetype. For gif, pxc, png, tiff, and jpeg, the file extension is just whatever follows the 'image/' part of the mime type. To handle the obscure types also, python does provide a standard package:

>>> from mimetypes import guess_extension
>>> guess_extension('image/x-corelphotopaint')
'.cpt'
>>> guess_extension('image/png')
'.png'

Answer 2

It looks like mimetypes stdlib module supports data urls even in Python 2:

>>> from mimetypes import guess_extension, guess_type
>>> guess_extension(guess_type("data:image/png;base64,")[0])
'.png'

Answer 3

Assume having base64 encoded in variable encoded_string , below code works for me:

from base64 import b64decode
import imghdr

encoded_string = 'image base64 encoded'

decoded_string = b64decode(var)
extension = imghdr.what(None, h=decoded_string)

Answer 4

You can use the mimetypes module - http://docs.python.org/2/library/mimetypes.html

Basically mimetypes.guess_extension(mine) should do the job.

Answer 5

I have Written code in Lambda which will find the type of an image and also checks base64 is image or not.

The following code will sure help someone.

import base64
import imghdr
def lambda_handler(event, context):
    image_data = event['img64']    # crate "json event" in lambda 
                                   # Sample JSON Event ========>  { "img64" : BASE64 of an Image }
                                   # Get BASE64 Data of image in image_data variable.
    sample = base64.b64decode(image_data)      # Decode the base64 data and assing to sample.

    for tf in imghdr.tests:
        res = tf(sample, None)
        if res:
            break;
    print("Extension OR Type of the Image =====>",res)
    if(res==None): # if res is None then BASE64 is of not an image.
            return {
            'status': 'False',
           'statusCode': 400,
           'message': 'It is not image, Only images allowed'
          }
    else:
        return 'It is image'  

Note :- The Above code is written Lambda (AWS) in python, You can copy and paste the following code to your local machine and test it as follows.

import base64
import imghdr
image_data = "BASE64 OF AN IMAGE"
sample = base64.b64decode(image_data)      # Decode the base64 data and     assing to sample.

for tf in imghdr.tests:
    res = tf(sample, None)
    if res:
        break;
print("Extension OR Type of the Image =====>",res)
if(res==None):
    print('It is not image, Only images allowed')
else:
    print('It is image')