Given offsets of GZip headers, this splits on those offsets.

offset_split.py

import sys
input_file = sys.argv[1]
offset_file = sys.argv[2]
output_file = input_file.replace('.gz','')
offsets = map(long, open(offset_file).read().rstrip().split("\n"))

with open(input_file,"rb") as f:
    part = 1
    previous = 0
    for offset in offsets:
        print "Index: " + str(offset)
        if offset == 0:
          continue

        file_size = offset - previous
        byte_data = f.read(file_size)
        print "Offset: " + str(f.tell())
        if byte_data[0:3] == '\x1F\x8B\x08':
          print "valid"

        with open(output_file+'-part-'+str(part)+'-of-'+str(len(offsets))+'.gz', 'wb') as o:
            o.write(byte_data)
        previous = offset
        part += 1
        print ""