4

I'm extracting files from MIME messages in a python milter and am running across issues with files named as such:

=?ISO-8859-1?Q?Certificado=5FZonificaci=F3n=5F2010=2Epdf?=

I can't seem to decode this name into UTF. In order to solve a prior ISO-8859-1 issue, I started passing all filenames to this function:

def unicodeConvert(self, fname):
    normalized = False

    while normalized == False:
        try:
            fname  = unicodedata.normalize('NFKD', unicode(fname, 'utf-8')).encode('ascii', 'ignore')
            normalized = True
        except UnicodeDecodeError:
            fname = fname.decode('iso-8859-1')#.encode('utf-8')
            normalized = True
        except UnicodeError:
            fname = unicode(fname.content.strip(codecs.BOM_UTF8), 'utf-8')
            normalized = True
        except TypeError:
            fname = fname.encode('utf-8')

    return fname

which was working until I got to this filename.

Ideas are appreciated as always.

Larry G. Wapnitsky
  • 1,216
  • 2
  • 16
  • 35
  • 2
    This is an RFC 2047 `encoded-word`. It's completely incorrect for that to appear in a `Content-Disposition` parameter, but Outlook does so anyway, due to the technical matter of it sucking so much. – bobince Jul 25 '12 at 22:17

1 Answers1

8

Your string is encoded using the Quoted-printable format for MIME headers. The email.header module handles this for you:

>>> from email.header import decode_header
>>> try:
...     string_type = unicode  # Python 2
... except NameError:
...     string_type = str      # Python 3
...
>>> for part in decode_header('=?ISO-8859-1?Q?Certificado=5FZonificaci=F3n=5F2010=2Epdf?='):
...     decoded = string_type(*part)
...     print(decoded)
...
Certificado_Zonificación_2010.pdf
Martijn Pieters
  • 1,048,767
  • 296
  • 4,058
  • 3,343