2013-10-17 14:05:27 +00:00
|
|
|
#!/usr/bin/env python
|
2008-04-27 04:27:09 +00:00
|
|
|
import sys
|
2014-06-14 03:00:49 +00:00
|
|
|
import logging
|
2009-11-04 11:28:32 +00:00
|
|
|
try:
|
|
|
|
from cStringIO import StringIO
|
|
|
|
except ImportError:
|
|
|
|
from StringIO import StringIO
|
2008-04-27 04:27:09 +00:00
|
|
|
|
2009-05-17 14:02:57 +00:00
|
|
|
|
2013-03-27 17:05:29 +00:00
|
|
|
class CorruptDataError(Exception):
|
|
|
|
pass
|
|
|
|
|
|
|
|
|
2008-04-27 04:27:09 +00:00
|
|
|
## LZWDecoder
|
|
|
|
##
|
2008-07-09 15:15:32 +00:00
|
|
|
class LZWDecoder(object):
|
2009-01-05 04:40:50 +00:00
|
|
|
|
2009-10-24 04:41:59 +00:00
|
|
|
def __init__(self, fp):
|
|
|
|
self.fp = fp
|
|
|
|
self.buff = 0
|
|
|
|
self.bpos = 8
|
|
|
|
self.nbits = 9
|
|
|
|
self.table = None
|
|
|
|
self.prevbuf = None
|
|
|
|
return
|
|
|
|
|
|
|
|
def readbits(self, bits):
|
|
|
|
v = 0
|
|
|
|
while 1:
|
|
|
|
# the number of remaining bits we can get from the current buffer.
|
|
|
|
r = 8-self.bpos
|
|
|
|
if bits <= r:
|
|
|
|
# |-----8-bits-----|
|
|
|
|
# |-bpos-|-bits-| |
|
|
|
|
# | |----r----|
|
2013-11-07 08:35:04 +00:00
|
|
|
v = (v << bits) | ((self.buff >> (r-bits)) & ((1 << bits)-1))
|
2009-10-24 04:41:59 +00:00
|
|
|
self.bpos += bits
|
|
|
|
break
|
|
|
|
else:
|
|
|
|
# |-----8-bits-----|
|
|
|
|
# |-bpos-|---bits----...
|
|
|
|
# | |----r----|
|
2013-11-07 08:35:04 +00:00
|
|
|
v = (v << r) | (self.buff & ((1 << r)-1))
|
2009-10-24 04:41:59 +00:00
|
|
|
bits -= r
|
|
|
|
x = self.fp.read(1)
|
2013-11-07 08:35:04 +00:00
|
|
|
if not x:
|
|
|
|
raise EOFError
|
2009-10-24 04:41:59 +00:00
|
|
|
self.buff = ord(x)
|
|
|
|
self.bpos = 0
|
|
|
|
return v
|
|
|
|
|
|
|
|
def feed(self, code):
|
2013-10-18 23:19:40 +00:00
|
|
|
x = ''
|
2009-10-24 04:41:59 +00:00
|
|
|
if code == 256:
|
2013-11-07 08:35:04 +00:00
|
|
|
self.table = [chr(c) for c in xrange(256)] # 0-255
|
|
|
|
self.table.append(None) # 256
|
|
|
|
self.table.append(None) # 257
|
2013-10-18 23:19:40 +00:00
|
|
|
self.prevbuf = ''
|
2009-10-24 04:41:59 +00:00
|
|
|
self.nbits = 9
|
|
|
|
elif code == 257:
|
|
|
|
pass
|
|
|
|
elif not self.prevbuf:
|
|
|
|
x = self.prevbuf = self.table[code]
|
|
|
|
else:
|
|
|
|
if code < len(self.table):
|
|
|
|
x = self.table[code]
|
2013-03-27 17:05:29 +00:00
|
|
|
self.table.append(self.prevbuf+x[:1])
|
|
|
|
elif code == len(self.table):
|
|
|
|
self.table.append(self.prevbuf+self.prevbuf[:1])
|
2009-10-24 04:41:59 +00:00
|
|
|
x = self.table[code]
|
2013-03-27 17:05:29 +00:00
|
|
|
else:
|
2013-10-18 23:19:40 +00:00
|
|
|
raise CorruptDataError
|
2009-10-24 04:41:59 +00:00
|
|
|
l = len(self.table)
|
|
|
|
if l == 511:
|
|
|
|
self.nbits = 10
|
|
|
|
elif l == 1023:
|
|
|
|
self.nbits = 11
|
|
|
|
elif l == 2047:
|
|
|
|
self.nbits = 12
|
|
|
|
self.prevbuf = x
|
|
|
|
return x
|
|
|
|
|
|
|
|
def run(self):
|
|
|
|
while 1:
|
|
|
|
try:
|
|
|
|
code = self.readbits(self.nbits)
|
|
|
|
except EOFError:
|
|
|
|
break
|
2013-03-27 17:05:29 +00:00
|
|
|
try:
|
|
|
|
x = self.feed(code)
|
|
|
|
except CorruptDataError:
|
|
|
|
# just ignore corrupt data and stop yielding there
|
|
|
|
break
|
2009-10-24 04:41:59 +00:00
|
|
|
yield x
|
2014-06-14 06:43:10 +00:00
|
|
|
#logging.debug('nbits=%d, code=%d, output=%r, table=%r' %
|
|
|
|
# (self.nbits, code, x, self.table[258:]))
|
2009-10-24 04:41:59 +00:00
|
|
|
return
|
2008-04-27 04:27:09 +00:00
|
|
|
|
2013-11-07 08:35:04 +00:00
|
|
|
|
2009-11-04 11:28:32 +00:00
|
|
|
# lzwdecode
|
|
|
|
def lzwdecode(data):
|
|
|
|
"""
|
|
|
|
>>> lzwdecode('\x80\x0b\x60\x50\x22\x0c\x0c\x85\x01')
|
|
|
|
'\x2d\x2d\x2d\x2d\x2d\x41\x2d\x2d\x2d\x42'
|
|
|
|
"""
|
|
|
|
fp = StringIO(data)
|
|
|
|
return ''.join(LZWDecoder(fp).run())
|
2008-04-27 04:27:09 +00:00
|
|
|
|
2009-11-04 11:28:32 +00:00
|
|
|
if __name__ == '__main__':
|
|
|
|
import doctest
|
|
|
|
doctest.testmod()
|