-#!/usr/bin/env python
+#!/usr/bin/env python3
# Copyright (C) 2013 Intra2net AG
#
'''
i = 0
pos = 0
- buf = ""
+ buf = b""
sep_len = len(separator)
if sep_len == 0:
raise Exception("empty separator")
if new_file_func is None:
- new_file_func = lambda prefix, i: open(prefix + str(i), 'w')
+ new_file_func = lambda prefix, i: open(prefix + str(i), 'wb')
output = new_file_func(prefix, i)
# buffered search. we try not to have the while input file in memory, as
# it's not needed
- with open(input_file, 'r') as f:
+ with open(input_file, 'rb') as f:
while True:
buf += f.read(BUFSIZE)
if len(buf) == 0:
# else: continue writing to the current output and iterate
output.write(buf)
- buf = ""
+ buf = b""
output.close()
+def chunk_file(input_file, output_file, from_pos, to_pos):
+ ifl = open(input_file, 'rb')
+ ofl = open(output_file, 'wb')
+
+ ifl.seek(from_pos)
+ ofl.write(ifl.read(to_pos-from_pos))
+ ifl.close()
+ ofl.close()
+
if __name__ == "__main__":
parser = argparse.ArgumentParser()
- parser.add_argument("-s", "--separator", required=True,
- help="string for the separator")
- parser.add_argument("-p", "--prefix", required=True,
- help="prefix for split files")
+ parser.add_argument("-s", "--separator", help="string for the separator")
+ parser.add_argument("-p", "--prefix", help="prefix for split files")
parser.add_argument("input_file", help="input file")
+ parser.add_argument("-f", "--from-pos", type=int, default=-1)
+ parser.add_argument("-t", "--to-pos", type=int, default=-1)
+ parser.add_argument("-o", "--output", default=None)
args = parser.parse_args()
- split_file(separator=args.separator, prefix=args.prefix, input_file=args.input_file)
+ if args.from_pos > -1 and args.to_pos > -1:
+ chunk_file(input_file=args.input_file, output_file=args.output,
+ from_pos=args.from_pos, to_pos=args.to_pos)
+ else:
+ split_file(separator=args.separator.encode('UTF-8', errors="surrogateescape"),
+ prefix=args.prefix, input_file=args.input_file)