[python] aggregate m,...,n to m-n

README.md

The program reads a sequence of integers (sorted), and outputs x-y ranges for consecutive integers, outputting it verbatim otherwise.

cat | python ranges.py
1
2
3
7
8
9
^D
1-3
7-9

ranges is actually a DFA/FSM!

Use case: for a NTFS partition, show which files are in bad sectors marked by ddrescue.

ddrescuelog outputs individual blocks (sectors). On the other hand, the -s option to ntfscluster can accept x-y ranges. By "factoring" the list of blocks into ranges, we can reduce the number of (slow) disk accesses.

ddrescuelog --domain-logfile=sdN.domain --list-blocks=- rescue.log \
| python ranges.py \
| xargs -n1 -i \
  sudo ntfscluster -s \{} 2>/dev/null

(redirected stderr for ntfscluster as it tends to output a lot of "Error reading inode NNN" lines)

ranges.py

def ranges(ints):
  start = None
  end = None

  for x in iter(ints):
    shallreset = True
    if end is None:
      pass
    elif end + 1 == x:
      end = x
      shallreset = False
    else:
      yield (start, end)
      
    if shallreset:
      start = x
      end = x

  # in case the last int consumed was part of a range
  if end > start:
    yield (start,end)

def read_ints():
  import fileinput

  for line in fileinput.input():
    yield int(line)


if __name__ == '__main__':
  ints = read_ints()
  r = ranges(ints)
  for rr in r:
    print "%d-%d" % rr