def compress(f, compf, block = None):
"""
Compress a file by bwt-mtf-huff processes.
f: a file-like object, content for compressing
compf: compression result
"""
# craete alphabet for move-to-front
content = f.read()
alphabet_set = set(content)
alphabet = [''] + list(alphabet_set)
# dump alphabet
cPickle.dump(alphabet, compf, 2)
content = None
f.seek(0)
count = 0
while True:
data = f.read(BLOCK_SIZE)
if data == '':
break
if block == None or block == count:
print "block %r:" % count
bwt_encode, I = bwt.encode(data)
mtf_encode = mtf.encode(alphabet, bwt_encode)
# create Huffman tree
freqs = Counter(mtf_encode)
if block:
print freqs
coding, root = huff.generate_coding(freqs)
if block:
print coding
# encoding
huff_encode = ''.join(huff.encode(mtf_encode, coding))
#print "huff_encode:\n%r" % huff_encode
nbits = len(huff_encode)
huff_bytes = tobytes(huff_encode)
compf.write(struct.pack(">2I", nbits, I))
dump_freqs(compf, freqs)
compf.write(huff_bytes)
print "nbits = %r, I = %r, length = %r" % (nbits, I, len(data))
count = count + 1
def test_mix(self):
seq = generator(16*1024)
encode, I = bwt.encode(seq)
decode = bwt.decode(encode, I, True)
self.assertEqual(list(seq), decode)
def test_encode(self):
seq = 'abraca'
encode, I = bwt.encode(seq, indexes = [5, 0, 3, 1, 4, 2])
self.assertEqual(list('caraab'), encode)
self.assertEqual(I, 1)