def decode(self):
""" algorithm derived from:
http://www.rasip.fer.hr/research/compress/algorithms/fund/lz/lzw.html
and the PDFReference
"""
cW = self.CLEARDICT
baos = ""
while True:
pW = cW
cW = self.nextCode()
if cW == -1:
raise PdfReadError("Missed the stop code in LZWDecode!")
if cW == self.STOP:
break
elif cW == self.CLEARDICT:
self.resetDict()
elif pW == self.CLEARDICT:
baos += self.dict[cW]
else:
if cW < self.dictlen:
baos += self.dict[cW]
p = self.dict[pW] + self.dict[cW][0]
self.dict[self.dictlen] = p
self.dictlen += 1
else:
p = self.dict[pW] + self.dict[pW][0]
baos += p
self.dict[self.dictlen] = p
self.dictlen += 1
if (self.dictlen >= (1 << self.bitspercode) - 1
and self.bitspercode < 12):
self.bitspercode += 1
return baos
def decode(data, decodeParms):
data = decompress(data)
predictor = 1
if decodeParms:
try:
predictor = decodeParms.get("/Predictor", 1)
except AttributeError:
pass # usually an array with a null object was read
# predictor 1 == no predictor
if predictor != 1:
columns = decodeParms["/Columns"]
# PNG prediction:
if predictor >= 10 and predictor <= 15:
output = StringIO()
# PNG prediction can vary from row to row
rowlength = columns + 1
assert len(data) % rowlength == 0
prev_rowdata = (0, ) * rowlength
for row in xrange(len(data) / rowlength):
rowdata = [
ord(x) for x in data[(row * rowlength):((row + 1) *
rowlength)]
]
filterByte = rowdata[0]
if filterByte == 0:
pass
elif filterByte == 1:
for i in range(2, rowlength):
rowdata[i] = (rowdata[i] + rowdata[i - 1]) % 256
elif filterByte == 2:
for i in range(1, rowlength):
rowdata[i] = (rowdata[i] + prev_rowdata[i]) % 256
else:
# unsupported PNG filter
raise PdfReadError("Unsupported PNG filter %r" %
filterByte)
prev_rowdata = rowdata
output.write(''.join([chr(x) for x in rowdata[1:]]))
data = output.getvalue()
else:
# unsupported predictor
raise PdfReadError("Unsupported flatedecode predictor %r" %
predictor)
return data