sgf timeout statistics

import os
import re
import pandas as pd

rex_move = re.compile(r';[BW]\[[a-z]{2}\]')
rex_res = re.compile(r'RE\[[BW]\+T\]')
rex_tm = re.compile(r'OT\[([^ ]{0,100}) ?([^]]{2,100})\]')

data_timeout = []
data_tmode = []
total = 0

for fn in filter(lambda x: x.endswith('.sgf'), os.listdir('.')):
    with open(fn, 'r') as f:
        content = f.read()
    has_timeout = rex_res.search(content) is not None
    tm = rex_tm.search(content)
    tm_mode = None
    if tm is not None:
        tm_mode = tm.group(2)
    if len(rex_move.findall(content)) > 5:
        data_timeout += [has_timeout]
        data_tmode += [tm_mode]
    total += 1

df = pd.DataFrame({
    'timeout': data_timeout,
    'tmode': data_tmode,
    'count':1})

sum_all = df.filter(items=['tmode','count']).groupby(by='tmode').sum()

df2 = df[df.timeout == True]
sum_to = df2.filter(items=['tmode','count']).groupby(by='tmode').sum()

df5 = sum_all.merge(sum_to, on='tmode', suffixes=['_all', '_timeout'])
df5.loc['TOTAL'] = df5.sum()
df5['%timeout'] = df5['count_timeout'] / df5['count_all'] * 100
df5.at['TOTAL','%timeout'] += 1000
df5 = df5.sort_values(by='%timeout')
df5.at['TOTAL','%timeout'] -= 1000

print('ignored', total - len(data_timeout) + 1, 'immediately aborted games')
print()

print(df5)
print()

df5 = df5.drop(labels='TOTAL')

import matplotlib.pyplot as plt
fig,ax = plt.subplots()
ax.pie(df5['%timeout'], labels=df5.index, autopct='%.1f%%')
plt.show()