import json,re,glob,sys
a=json.load(open('part1.json'))
tz=re.compile(r'(Africa|America|Asia|Atlantic|Australia|Europe|Indian|Pacific|Antarctica|Arctic|Etc|US|Brazil|Canada|Chile|Mexico)/')
d={}
for f in sorted(glob.glob('part1.ru.chunk*.txt')):
    for l in open(f).read().splitlines():
        if ' => ' in l:
            k,v=l.split(' => ',1);d[k]=v
for f in glob.glob('keep*.txt'):
    for l in open(f).read().splitlines():
        if l:d[l]=l
lo,hi=int(sys.argv[1]),int(sys.argv[2])
for i,s in enumerate(a[lo:hi],lo):
    if s not in d and not tz.match(s):print('MISS',i,json.dumps(s,ensure_ascii=False))
bad=set(d)-set(a)
if bad:print('EXTRA',list(bad)[:20])