2024-10-28 16:32:25 +01:00
|
|
|
import sys
|
|
|
|
|
2024-11-01 15:29:13 +01:00
|
|
|
# '19xx' literally would be accepted as well but we won't tell anyone
|
2024-10-28 16:32:25 +01:00
|
|
|
def solve(lines):
|
2024-11-01 15:29:13 +01:00
|
|
|
def detect_date(s):
|
|
|
|
date_f = "19XX r."
|
|
|
|
date_l = len(date_f)
|
|
|
|
|
|
|
|
j = 0
|
|
|
|
for i in range(len(s)):
|
|
|
|
c = s[i]
|
|
|
|
|
|
|
|
if j == date_l:
|
|
|
|
return True
|
|
|
|
elif c == date_f[j] \
|
|
|
|
or (j in (2, 3) and c.isdigit()):
|
|
|
|
j += 1
|
|
|
|
continue
|
|
|
|
|
|
|
|
j = 0
|
|
|
|
|
|
|
|
return False
|
|
|
|
|
|
|
|
return filter(detect_date, lines)
|
|
|
|
|
2024-10-28 16:32:25 +01:00
|
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
|
lines = []
|
|
|
|
|
|
|
|
fp = sys.argv[1]
|
2024-11-01 15:29:13 +01:00
|
|
|
with open(fp, encoding="utf-8") as f: # following symlinks might not work on Windows
|
2024-10-28 16:32:25 +01:00
|
|
|
lines = f.readlines()
|
|
|
|
sol = solve(lines)
|
|
|
|
with open("./polish_wiki_excerpt.out", 'w', encoding="utf-8") as f:
|
|
|
|
f.writelines(sol)
|