Adding encoding alias to python

Question

Is there a way that I can add alias to python for encoding. There are sites on the web that are using the encoding \'windows-1251\' but have their charset set to win-1251, s

Answer 1

Encoding aliases can be added by editing aliases.py file.

# euc_jp codec
'eucjp'              : 'euc_jp',
'ujis'               : 'euc_jp',
'u_jis'              : 'euc_jp',
'euc_jp_linux'       : 'euc_jp',
'euc-jp-linux'       : 'euc_jp',

Above I have added two aliases euc_jp_linux and euc-jp-linux to the encoding euc_jp.

For a 64 bit linux system aliases.py file is generally located under /usr/lib64/python2.6/encodings/

Answer 2

The encodings module is not well documented so I'd instead use codecs, which is:

import codecs

def encalias(oldname, newname):
  old = codecs.lookup(oldname)
  new = codecs.CodecInfo(old.encode, old.decode, 
                         streamreader=old.streamreader,
                         streamwriter=old.streamwriter,
                         incrementalencoder=old.incrementalencoder,
                         incrementaldecoder=old.incrementaldecoder,
                         name=newname)
  def searcher(aname):
    if aname == newname:
      return new
    else:
      return None
  codecs.register(searcher)

This is Python 2.6 -- the interface is different in earlier versions.

If you don't mind relying on a specific version's undocumented internals, @Lennart's aliasing approach is OK, too, of course - and indeed simpler than this;-). But I suspect (as he appears to) that this one is more maintainable.

Answer 3

>>> import encodings
>>> encodings.aliases.aliases['win_1251'] = 'cp1251'
>>> print '\xcc\xce\xd1K\xc2\xc0'.decode('win-1251')
MOCKBA

Although I personally would consider this monkey-patching, and use my own conversion table. But I can't give any good arguments for that position. :)