Is there a Python library function which attempts to guess the character-encoding of some bytes?

Question

I'm writing some mail-processing software in Python that is encountering strange bytes in header fields. I suspect this is just malformed mail; the message itself claims to be us-ascii, so I don't think there is a true encoding, but I'd like to get out a unicode string approximating the original one without throwing a UnicodeDecodeError.

So, I'm looking for a function that takes a str and optionally some hints and does its darndest to give me back a unicode. I could write one of course, but if such a function exists its author has probably thought a bit deeper about the best way to go about this.

I also know that Python's design prefers explicit to implicit and that the standard library is designed to avoid implicit magic in decoding text. I just want to explicitly say "go ahead and guess".

Answer 1

+1 for the chardet module.

It is not in the standard library, but you can easily install it with the following command:

$ pip install chardet

Example:

>>> import urllib.request
>>> rawdata = urllib.request.urlopen('http://yahoo.co.jp/').read()
>>> import chardet
>>> chardet.detect(rawdata)
{'encoding': 'EUC-JP', 'confidence': 0.99}

See Installing Pip if you don't have one.

Answer 2

As far as I can tell, the standard library doesn't have a function, though it's not too difficult to write one as suggested above. I think the real thing I was looking for was a way to decode a string and guarantee that it wouldn't throw an exception. The errors parameter to string.decode does that.

def decode(s, encodings=('ascii', 'utf8', 'latin1')):
    for encoding in encodings:
        try:
            return s.decode(encoding)
        except UnicodeDecodeError:
            pass
    return s.decode('ascii', 'ignore')

Answer 3

The best way to do this that I've found is to iteratively try decoding a prospective with each of the most common encodings inside of a try except block.