gistfile1.py

import re

# Holy crap this library is so unoptimized ...
import spans

class longrange(spans.intrange):
        type = long

class longrangeset(spans.intrangeset):
        type = longrange

successes = set()
regex = re.compile('\[repair #([^\]]+)\] session completed successfully')
with open('successes') as f:
        for l in f.readlines():
                m = regex.match(l)
                if not m:
                        continue
                g = m.groups()
                successes.add(g[0])

regex = re.compile(r'\[repair #([^\]]+)\].*on range \((.*)] for ([^.]+).')
ranges = longrangeset([])
with open('repairs') as f:
        for l in f.readlines():
                m = regex.match(l)
                if not m:
                    continue
                g = m.groups()
                if g[0] not in successes:
                        continue
                (lower, upper) = [long(x) for x in g[1].split(',')]
                if lower < upper:
                        ranges.add(longrange(lower, upper))
                else:   
                        ranges.add(longrange(-2**63, upper))
                        ranges.add(longrange(lower, 2**63-1))

m = longrangeset([longrange(-2**63, 2**63-1)])
failed = m.difference(ranges)
for pair in failed._list:
        print "nodetool repair -dcpar -st {} -et {}".format(pair.lower, pair.upper)