Decode python base64 String

Question

I have extracted base64 string of forecolor, texture and edgemap values of images, I have a list with following structure:

forecolor=AgCEAg4DUQQCBQQGARMBFQE1AmUB
edge=AfCAFg5iIATCPwTAEIiBFggBDw
forecolor=AgAsAQ0CJAMcDRgOGg8DHQYeBzYBPQ4-DU0ETgNtBm4CfQI

I am trying to decode these values, but I am getting Incorrect Padding error, here is the exact error:

Traceback (most recent call last):
  File "ImageVectorData.py", line 44, in <module>
    print "Decoded String: " + decoded.decode('base64', 'strict')
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/encodings/base64_codec.py", line 42, in base64_decode
    output = base64.decodestring(input)
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/base64.py", line 321, in decodestring
    return binascii.a2b_base64(s)
binascii.Error: Incorrect padding

Here is my code:

for item in value:
    print "String before Split: " + item
    if item.split("=")[0] == "forecolor":
        decoded = (item.split("=")[1])
        print "String to be decoded: " + decoded
        print "Decoded String: " + decoded.decode('base64', 'strict')

I also saw an interesting out put when the first forecolor base64 string got decoded: Here is the out put of that:

String before Split: forecolor=AgCEAg4DUQQCBQQGARMBFQE1AmUB
String to be decoded: AgCEAg4DUQQCBQQGARMBFQE1AmUB
Decoded String: ?Q5e

I am not really sure what I am doing wrong here. I looked at following python document and tried that but that didn't work either: http://docs.python.org/library/base64.html

Solution

You are trying to decode a Base64 String which does not have padding. Although many flavors of Base64 do not have padding, Python requires padding for standard base64 decoding. This StackOverflow question has a more in-depth explanation: Python: Ignore 'Incorrect padding' error when base64 decoding

For your code I would make modifications similar to below:

for item in value:
    print "String before Split: " + item
    if item.split("=")[0] == "forecolor":
        decoded = (item.split("=")[1])
        print "String to be decoded: " + decoded
        # Add Padding if needed
        decoded += "===" # Add extra padding if needed
        print "Decoded String: " + decoded.decode('base64', 'strict')

Based on your comment it seemed that you also need the byte array returned from the base64 decoding turned into a list of integers. I made an assumption that the integers are little endian short ints.

import struct
x = "AgAsAQ0CJAMcDRgOGg8DHQYeBzYBPQ4-DU0ETgNtBm4CfQI"
x += "==="
y = x.decode('base64', 'strict')
intList = [struct.unpack('<h', y[i] + y[i+1]) for i in xrange(0, len(y), 2)]
print intList

The result was:

[(2,), (300,), (525,), (804,), (3356,), (3608,), (3866,), (7427,), (7686,), (13831,), (15617,), (782,), (16723,), (-32749,), (16859,), (-32613,), (16543,)]

Comments

[add-semi-colons]

I still get the same error, I used the exact line you had put.

[Jesse Harris]

I found another technique in another answer that works for your data. I have updated the answer. It is simply adding enough padding to account for all scenarios.

[add-semi-colons]

Thanks but the problem is decoded value should be a vector like 2,3,4,5 instead its giving me string like d#CSs padding issue is resolve the data is completely wrong. These encoded values are generated by a C++ program do you that has something to do with this?

[Jesse Harris]

I have appended my answer to convert from the byte array to a list of ints. If the base64 piece is working, I would recommend closing this question and opening another with the byte array issue.