def update_pts(self, lst, rec, r): # 更新得分类数据
pt = lst[1]
ix = 17
if lst[0] not in self.qtr_stats[
lst[2]] or lst[0] not in self.plyrs_oncourt[
lst[2]]: # 若球员不在场，则此时刻所有得分同时作用于下场和上场球员（下场球员的遗留影响、上场球员已在场上）
# print(self.gm, rec, r)
if lst[0] not in self.plyrs_mp:
self.plyrs_mp[lst[0]] = MPTime(rec['T'])
try:
assert rec['T'] == r['T']
except:
print(self.gm, rec, r)
raise KeyError
if lst[0] not in self.qtr_stats[lst[2]]:
self.qtr_stats[lst[2]][lst[0]] = [
np.zeros((self.num_items, )),
[
np.zeros((self.num_items, )),
np.zeros((self.num_items, ))
]
]
self.qtr_stats[lst[2]][lst[0]][1][0][ix] += pt
self.qtr_stats[lst[2]][lst[0]][0][ix] += pt
self.qtr_stats[lst[2]]['team'][ix] += pt
for tm in range(2):
for pm in self.plyrs_oncourt[tm]:
self.qtr_stats[tm][pm][1][0 if tm == lst[2] else 1][ix] += pt
def singleplayer(self, pm):
player = Player(pm, self.RoP)
if not player.exists or isinstance(player.data, list):
return '', ''
pres = {} # 球员结果
atds = 0 # 总人数
for ss, yy in player.yieldSeasons():
yy = yy.replace('-', '_' + yy[:2] if yy[2:4] != '99' else '_20')
try:
sf = self.seasonfile(yy)
except:
continue
sgn = sf.shape[0]
ig = 0
for game in player.yieldGames(ss):
gm, wol, diff = game['Date' if self.RoP == 'regular' else 'Playoffs'],\
game['WoL'][0], int(game['WoL'][3:-1])
while sf['gm'][ig] != gm and ig < sgn - 1:
ig += 1
gb = sf.loc[ig]
refs, atd, tog = gb['Referees'].split(', '),\
gb['Attendance'] if gb['Attendance'] != -1 else 0, gb['Time of Game']
if atd:
atds += atd # 球馆观众人数
tog = MPTime(tog, reverse=False) if tog else MPTime(
'0:00.0', reverse=False)
if refs[0]:
for r in refs: # 执法裁判
if r not in pres:
pres[r] = [
0, 0, 0, 0,
MPTime('0:00.0', reverse=False)
] # 胜率、胜场数、总场数、分差、比赛时长
if wol == 'W':
pres[r][1] += 1
pres[r][2] += 1
pres[r][3] += diff
pres[r][4] += tog
for i in pres:
pres[i][0] = '%.1f%%' % (pres[i][1] / pres[i][2] * 100)
pres[i][4] = pres[i][4].average(pres[i][2])
pres[i][3] = '%.1f' % (pres[i][3] / pres[i][2])
return sorted(pres.items(), key=lambda x: x[1][1],
reverse=True), atds / player.games
def new_player(pm, ss):
if pm not in plyrs: # 新建球员统计
plyrs[pm] = [[
0,
MPTime('0:00.0'), 0, {
'TOV': [0, 0]
}, 0, [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], {}
],
[
0,
MPTime('0:00.0'), 0, {
'TOV': [0, 0]
}, 0, [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], {}
]]
if ss not in plyrs[pm][i][-1]: # 新建球员赛季统计
plyrs[pm][i][-1][ss] = [
0,
MPTime('0:00.0'), 0, {
'TOV': [0, 0]
}, 0, [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0]
]
def swt(
self,
new,
t=0
): # 节内发生换人，初始化新上场球员数据，更新场上阵容 t = 0 节中换人 1 节中换人且最后一条记录未转换球权（202012220BRK {'Q': 2, 'T': '28:03.0', 'MS': ['jordade01', 1, 1], 'D': [2, 2], 'M': '', 'BP': 1, 'S': [49, 77]}）
ix = 19
for tm in range(2):
on_pms = set(new['R'][tm]) - set(self.plyrs_oncourt[tm])
off_pms = set(self.plyrs_oncourt[tm]) - set(new['R'][tm])
# 上场球员：记录上场时刻，初始化数据统计格式
for on_pm in on_pms:
if on_pm not in self.plyrs_mp:
self.plyrs_mp[on_pm] = MPTime(new['T'])
if on_pm not in self.qtr_stats[tm]:
self.qtr_stats[tm][on_pm] = [
np.zeros((self.num_items, )),
[
np.zeros((self.num_items, )),
np.zeros((self.num_items, ))
]
]
# if t: # 节中换上场球员pace修正
self.qtr_stats[tm][on_pm][0][18] += 0.5
self.qtr_stats[tm][on_pm][1][0][18] += 0.5
self.qtr_stats[tm][on_pm][1][1][18] += 0.5
# 下场球员：累计上场时间
for off_pm in off_pms:
assert off_pm in self.plyrs_mp
self.qtr_stats[tm][off_pm][0][ix] += (
MPTime(new['T']) - self.plyrs_mp[off_pm]).secs()
self.plyrs_mp.pop(off_pm)
# if off_pm == 'wiggian01':
# print(self.qtr_stats[0]['wiggian01'])
if t: # 节中换下场球员pace修正
self.qtr_stats[tm][off_pm][0][18] -= 0.5
self.qtr_stats[tm][off_pm][1][0][18] -= 0.5
self.qtr_stats[tm][off_pm][1][1][18] -= 0.5
# print(off_pm, self.qtr_stats[tm][off_pm][0][18])
self.plyrs_oncourt = new['R']
def splitLineups(lines5):
print(list(lines5[0])[0], '...', list(lines5[0])[-1])
print(list(lines5[1])[0], '...', list(lines5[1])[-1])
lines_all = [[{}, {}], [{}, {}], [{}, {}],
[{}, {}]] # 0一人组合1二人组合2三人组合3四人组合 0常规赛1季后赛
for season in range(2002, 2021):
ss = '%d_%d' % (season, season + 1)
print(ss)
for i in range(2):
lines1234 = [{}, {}, {}, {}] # 单赛季按球队 0一人组合1二人组合2三人组合3四人组合
for tm in lines5[i][ss]:
# 球队lineups初始化
for ix in range(4):
if tm not in lines1234[ix]:
lines1234[ix][tm] = {}
lines_all[ix][i][ss] = {}
line5_tm = lines5[i][ss][tm] # 单队5人lineups
# print(tm, len(line5_tm))
for L in line5_tm:
pms = L.split(' ') # 遍历每组5人组合
# print(pms)
pms1234 = [[], [], [], []] # 0一人组合1二人组合2三人组合3四人组合
for c in range(1, 5): # 按照每组5人组合，使用（排列）组合方法穷尽所有1-4人的搭配
for iter in itertools.combinations(pms, c):
pms1234[c - 1].append(list(iter))
for ix, combs_i in enumerate(
pms1234): # 遍历1-4人的所有搭配，累加时间和正负值
for combs in combs_i:
combstr = ' '.join(sorted(combs))
if combstr not in lines1234[ix][tm]: # 每种组合初始化
lines1234[ix][tm][combstr] = [
MPTime('0:00.0'), 0
]
lines1234[ix][tm][combstr][0] += line5_tm[L][0]
lines1234[ix][tm][combstr][1] += line5_tm[L][1]
for ix in range(4):
for tm in lines1234[ix]:
lines_sorted = {}
for k in sorted(lines1234[ix][tm],
key=lines1234[ix][tm].__getitem__,
reverse=True): # 字典按值排序
lines_sorted[k] = lines1234[ix][tm][k]
lines_all[ix][i][ss][tm] = lines_sorted
for ix in range(4):
writeToPickle(
'D:/sunyiwu/stat/data/Lineups/anaSeason%dLineups.pickle' %
(ix + 1), lines_all[ix])
def detectgame(self, gm, playeres, season=0):
game = Game(gm[:-7], 'playoff' if self.RoP else 'regular')
for qtr in range(3, game.quarters):
for ply in game.yieldPlay(qtr):
play = Play(ply, qtr)
if play.time() <= MPTime(self.lm): # 在规定时间内
rec, ind = play.record()
s = play.score(ind=ind)
if s and play.diffbeforescore(s) <= self.dp:
try:
p = self.pm2pn[rec.split(' ')[0]]
except KeyError:
p = rec.split(' ')[0]
if season:
p += ' %d_%d' % (season, season + 1)
if p not in playeres:
playeres[p] = np.zeros((1, len(self.columns)))
if 'makes' in rec: # 得分增加命中数
playeres[p][0, 0] += s
playeres[p][0, s * 3 - 1] += 1
if 'assist' in rec: # 受助攻
playeres[p][0, s + 11] += 1
playeres[p][0, 15] += 1
try:
astp = self.pm2pn[rec.split(' ')[-1][:-1]]
except:
astp = rec.split(' ')[-1][:-1]
if season:
astp += ' %d_%d' % (season, season + 1)
if astp not in playeres:
playeres[astp] = np.zeros(
(1, len(self.columns)))
playeres[astp][0, s + 14] += 1
playeres = self.onlycountonce(
astp, gm[:-7], playeres)
playeres[p][0, s * 3] += 1 # 得分或投失增加出手数
playeres = self.onlycountonce(p, gm[:-7], playeres)
# if 'LeBron James' in p and season == 2019:
# print(s, play.play, gm[:-7], self.plyrcgs['LeBron James 2019_2020'])
return playeres
(ss, regularOrPlayoffs[i]))
for gm in tqdm(gms):
game = Game(gm, regularOrPlayoffs[i])
record = LoadPickle(gameMarkToDir(gm, regularOrPlayoffs[i], tp=3))
rot = game.rotation(record)
bxs, rot = game.replayer(record, rot)
et = '%d:00.0' % (48 + 5 * (game.quarters - 4))
for tmix, tm in enumerate(list(game.bxscr[0])): # 分别回溯两支球队
if tm not in lines:
lines[tm] = {}
RoH = list(game.bxscr[0]).index(tm)
for ix, r in enumerate(rot):
line = inline(r)
if ' '.join(line) not in lines[tm]:
lines[tm][' '.join(line)] = [
MPTime('0:00.0'), 0,
[np.zeros((21, )),
np.zeros((21, ))]
] # 0时间 1+/- 2详细数据 0本队数据1对手数据
if ix == 0:
tt = '0:00.0'
stats = np.zeros((21, ))
s = [0, 0]
else:
try:
assert r['T'] != tt or (ix > 0 and
r['Q'] != rot[ix - 1]['Q'])
except:
print(gm, tt, r)
print(rot[ix - 1])
if set(r['R'][RoH]) != set(rot[ix - 1]['R'][RoH]):
print('防守篮板/总篮板占比', lg_drbp)
print('联盟运动战单打能力', lg_factor)
print('bbr复杂pace估计', lg_pace)
print('传统pace估计',
(lg_stats[1] - lg_stats[9] + lg_stats[15] + 0.44 * lg_stats[7]) /
lg_games / 2)
lgsum, lgames, lgmax, lgmin = 0, 0, [0, []], [30, []]
for pm in tqdm(pms):
if ss in enforce['Turnover'][pm][i][-1]:
pmdir = playerMarkToDir(pm, RoF[i])
if pmdir:
plyr = Player(pm, RoF[i])
ssgms = plyr.getSeason(ss)
num_games = ssgms.shape[0]
if num_games >= games_td[i]:
mpsum, scoresum = MPTime('0:00'), 0
for g in range(num_games):
sgm = ssgms[g:g + 1]
mp = MPTime(sgm['MP'].values[0])
if mp.mf() != 0:
# ================ 球员单场数据准备 ================
mpsum += mp
pts = float(sgm['PTS'])
fta, ft, fga, fg, tpa, tp = float(
sgm['FTA']), float(sgm['FT']), float(
sgm['FGA']), float(sgm['FG']), float(
sgm['3PA']), float(sgm['3P'])
drb, orb, trb = float(sgm['DRB']), float(
sgm['ORB']), float(sgm['TRB'])
ast, stl, blk, tov, pf = float(
sgm['AST']), float(sgm['STL']), float(
tar_item = 1 # 0抢断1失误
sentence = "'TOV' in rec and rec['plyr'] != 'Team' and rec['plyr'] != ''" if tar_item else "'TOV' in rec and 'STL' in rec"
def new_player(pm, ss):
if pm not in plyrs: # 新建球员统计
plyrs[pm] = [[0, MPTime('0:00.0'), 0, {'TOV': [0, 0]}, 0, [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], {}],
[0, MPTime('0:00.0'), 0, {'TOV': [0, 0]}, 0, [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], {}]]
if ss not in plyrs[pm][i][-1]: # 新建球员赛季统计
plyrs[pm][i][-1][ss] = [0, MPTime('0:00.0'), 0, {'TOV': [0, 0]}, 0, [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0]]
for season in range(2000, 2001):
count_games = [0, 0] # 0reg1plf
count_item = np.zeros((2, 3, 9)) # 0reg1plf 0客场球队1主场球队2总 0总1第一节2第二节3第三节4第四节5f加时6s加时7t加时8f加时
count_time = [[[MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0')],
[MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0')],
[MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0')]],
[[MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0')],
[MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0')],
[MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0'), MPTime('0:00.0')]]]
average_time = [[['', '', '', '', '', '', '', '', ''], ['', '', '', '', '', '', '', '', ''], ['', '', '', '', '', '', '', '', '']],
[['', '', '', '', '', '', '', '', ''], ['', '', '', '', '', '', '', '', ''], ['', '', '', '', '', '', '', '', '']]]
count_score = np.zeros((2, 3, 9))
ss = '%d_%d' % (season, season + 1)
# print(ss)
for i in range(2): # 分别统计常规赛、季后赛
gms = os.listdir('D:/sunyiwu/stat/data/seasons/%s/%s/' % (ss, regularOrPlayoffs[i]))
for gm in tqdm(gms):
count_games[i] += 1
# print('\t\t\t' + gm)
# 0命中率1投中2出手3罚球命中率4罚球投中5罚球出手6两分命中率7两分投中8两分出手9三分命中率10三分投中11三分出手
# 12助攻两分13助攻三分14盖帽15总得分16负责得分17eFG%18TS%
for season in tqdm(range(1996, 2020)):
ss = '%d_%d' % (season, season + 1)
# print('starting to analysing season %s:' % ss)
gms = os.listdir('D:/sunyiwu/stat/data/seasons/%s/%s/' %
(ss, regularOrPlayoffs[i]))
for gm in gms:
count_games += 1
game = Game(gm[:-7], regularOrPlayoffs[i])
record = LoadPickle(
gameMarkToDir(gm[:-7], regularOrPlayoffs[i], tp=3))
# _, _, _, record = game.game_scanner(gm[:-7])
lastqtr = record[-1]['Q'] # 最后一节节次
assert lastqtr > 2
lastsec = MPTime('%d:00.0' % (48 + 5 * (lastqtr - 3))) # 本场最后时刻
ix = -1 # 从后往前遍历比赛过程
for qtr in range(lastqtr -
2): # 从最后时刻开始统计（若有加时则依次从后往前统计每节次最后时刻至第四节为止）
if qtr > 0:
lastsec = lastsec - MPTime('5:00.0') # 本节最后时刻
timetd = lastsec - MPTime(lastSecs) # 统计时间范围
while record[ix]['Q'] > lastqtr - qtr:
ix -= 1
ft = [-1, -1, -1]
ftt = [-1, -1]
ftplus = [-1, -1, -1, -1,
''] # 第4个元素用于判断罚球方与运动战投篮出手方是否一致，第5个元素用于记录罚球人
while MPTime(record[ix]['T']
) >= timetd and record[ix]['Q'] == lastqtr - qtr:
tmp = record[ix]
def default_mp(self, now):
for tm in range(2):
for pm in self.plyrs_oncourt[tm]:
self.plyrs_mp[pm] = MPTime(now)
def qtr_end(self, qtr, new, end=0): # 一节结束，整理本节数据（更新所有上过场球员的数据，包括基础和进阶）
# 更新打至节末的球员的上场时间
endt = MPTime(new['T']) if qtr < 4 else MPTime(
'%d:00.0' % (48 if qtr == 4 else 48 + 5 * (qtr - 4)))
for tm in range(2):
for off_pm in self.plyrs_oncourt[tm]:
# print(qtr, off_pm)
# print(new)
assert off_pm in self.plyrs_mp
self.qtr_stats[tm][off_pm][0][19] += (
endt - self.plyrs_mp[off_pm]).secs()
# 更新本节出场球员其它数据
periods = [0, qtr]
if qtr < 5:
periods.append(9 if qtr < 3 else 10)
for n in periods:
self.qtr_end_pm(n, self.tdbxs, self.qtr_stats)
# 待完成：更新单节进阶数据
if not end:
# 更新新的一节上场球员及上场时间点
self.plyrs_oncourt = new['R']
self.qtr_stats = [{}, {}]
self.default_qtr()
self.plyrs_mp = {}
self.default_mp(new['T'])
if end:
# 计算各项命中率
for tm in range(2):
for qtr in range(11):
for pm in self.tdbxs[tm][qtr]:
if pm == 'team':
self.cal_perc(self.tdbxs[tm][qtr][pm])
else:
self.cal_perc(self.tdbxs[tm][qtr][pm][0])
self.cal_perc(self.tdbxs[tm][qtr][pm][1][0])
self.cal_perc(self.tdbxs[tm][qtr][pm][1][1])
if end and 0:
print('客队')
for i in self.tdbxs[0][0]:
if i == 'team':
print(i, list(self.tdbxs[0][0][i]))
else:
print(i, list(self.tdbxs[0][0][i][0]))
print('主队')
for i in self.tdbxs[1][0]:
if i == 'team':
print(i, list(self.tdbxs[1][0][i]))
else:
print(i, list(self.tdbxs[1][0][i][0]))
print(self.tdbxs[0][0]['team'])
print(self.tdbxs[1][0]['team'])
# 0FG 1FGA 2FG% 33P 43PA 53P% 6FT 7FTA 8FT% 9ORB 10DRB 11TRB 12AST 13STL 14BLK 15TOV 16PF 17PTS 18BP 19MP 20+/-
for pm in self.tdbxs[0][0]:
if pm != 'team':
pmself = self.tdbxs[0][0][pm][0] # 球员自身数据
pmtm = self.tdbxs[0][0][pm][1][0] # 球员在场期间本队数据
pmop = self.tdbxs[0][0][pm][1][1] # 球员在场期间敌队数据
astp = pmself[12] / (
pmtm[0] - pmself[0]) if pmtm[0] - pmself[0] else float(
'nan') # 助攻率
ast2to = pmself[12] / pmself[15] if pmself[15] else float(
'nan') # 助攻失误比
drbp = pmself[10] / (
pmtm[10] + pmop[9]) if pmtm[10] + pmop[9] else float(
'nan') # 防守篮板率
orbp = pmself[9] / (
pmtm[9] + pmop[10]) if pmtm[9] + pmop[10] else float(
'nan') # 进攻篮板率
trbp = (pmself[9] + pmself[10]) / (
pmtm[10] + pmop[9] + pmtm[9] + pmop[10]
) if pmtm[10] + pmop[9] + pmtm[9] + pmop[10] else float(
'nan') # 总篮板率
print(pm, astp, ast2to, orbp, drbp, trbp)
print('联盟进攻效率（分/回合）', lg_vop)
print('防守篮板/总篮板占比', lg_drbp)
print('联盟运动战单打能力', lg_factor)
lgsum, lgames, lgmax, lgmin = 0, 0, [0, []], [30, []]
for pm in pms:
pmdir = playerMarkToDir(pm, RoF[i], tp=1)
if pmdir:
cr = LoadPickle(pmdir)
if ss in cr['Season'].values and ss in plyr_ss[pm]:
cr_a = cr[cr['Season'] == ss][:1]
cr_s = cr[cr['Season'] == ss][1:]
num_games = cr_a['G']
num_games = int(num_games.values[0][:-3])
if num_games >= games_td[i] or int(
cr_s['MP'].values[0].split(':')[0]) > 1600:
mp = MPTime(cr_a['MP'].values[0])
pts = float(cr_a['PTS'])
fta = float(cr_a['FTA'])
ft = float(cr_a['FT'])
fga = float(cr_a['FGA'])
fg = float(cr_a['FG'])
tpa = float(cr_a['3PA'])
tp = float(cr_a['3P'])
bga = fga - tpa
bg = fg - tp
fmiss = fta - ft
tmiss = tpa - tp
bmiss = bga - bg
drb = float(cr_a['DRB'])
orb = float(cr_a['ORB'])
trb = float(cr_a['TRB'])
def ave_and_sum(self, tmp, type=2): # type=0:计算总和 1:计算均值 2:计算均值和总和
if isinstance(tmp, list):
tmp = pd.DataFrame(tmp,
columns=regular_items_en.keys()
if not self.RoP else playoff_items_en.keys())
if type:
ave = []
sumn = []
for k in tmp.columns:
if k == 'G': # 首列
if type:
ave.append('%d场平均' % tmp.shape[0])
sumn.append('%d场总和' % tmp.shape[0])
elif k in ['Playoffs', 'Date', 'Age', 'Tm', 'Opp', 'Series', 'G#']:
if type:
ave.append('/')
sumn.append('/')
elif k == 'RoH': # 统计主客场数量
try:
at = np.sum(tmp[k] == '@')
except:
at = 0
if type:
ave.append('%dR/%dH' % (at, tmp.shape[0] - at))
sumn.append('%dR/%dH' % (at, tmp.shape[0] - at))
elif k == 'WoL': # 统计几胜几负
origin = WinLoseCounter(False)
for i in tmp[k]:
origin += WinLoseCounter(True, strwl=i)
if type:
ave.append(origin.average())
sumn.append(origin)
elif k == 'GS': # 统计首发次数
cs = tmp[k].value_counts()
if cs.empty:
if type:
ave.append('')
sumn.append('')
else:
try:
s = np.sum(tmp[k] == '1')
except:
s = 0
if type:
ave.append('%d/%d' % (s, tmp.shape[0]))
sumn.append('%d/%d' % (s, tmp.shape[0]))
elif k == 'MP': # 时间加和与平均
sum_time = MPTime('0:00.0', reverse=False)
for i in tmp[k]:
if isinstance(i, str):
sum_time += MPTime(i, reverse=False)
if type:
ave.append(sum_time.average(tmp.shape[0]))
sumn.append(sum_time.strtime[:-2])
elif '%' in k: # 命中率单独计算
goal = tmp[[k[:-1], k[:-1] + 'A']]
goal = goal.dropna(axis=0, how='any')
goal = goal.astype('float').sum(axis=0)
p = goal[k[:-1]] / goal[k[:-1] +
'A'] if goal[k[:-1] +
'A'] else float('nan')
if not math.isnan(p):
p = '%.3f' % p
if p != '1.000':
p = p[1:]
else:
p = ''
if type:
ave.append(p)
sumn.append(p)
else:
tmp_sg = tmp[k][tmp[k].notna()]
# tmp_sg = tmp_sg[tmp_sg != '']
if not tmp_sg.empty:
if type:
# print(tmp_sg)
try:
a = '%.1f' % tmp_sg.astype('float').mean()
except:
tmp_sg = tmp_sg[tmp_sg != '']
a = '%.1f' % tmp_sg.astype('float').mean()
# 除比赛评分以外其他求和结果均为整数
s = '%.1f' % tmp_sg.astype('float').sum(
) if k == 'GmSc' else int(tmp_sg.astype('int').sum())
if k == '+/-': # 正负值加+号
if type and s != 0 and a[0] != '-':
a = '+' + a
if s > 0:
s = '+%d' % s
if s == 'nan':
if type:
a = ''
s = ''
else:
a, s = '', ''
if type:
ave.append(a)
sumn.append(s)
if type == 0:
return [sumn]
elif type == 1:
return [ave]
else:
return [ave, sumn]
def time(self): # 当前时间（本节倒计时）
return MPTime(self.play[0])
if ss not in plyrs[pm][i][-1]: # 新建球员赛季统计
plyrs[pm][i][-1][ss] = [
0,
MPTime('0:00.0'), 0, {
'TOV': [0, 0]
}, 0, [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0]
]
for season in range(1996, 2020):
count_games = [0, 0] # 0reg1plf
count_item = np.zeros(
(2, 3,
9)) # 0reg1plf 0客场球队1主场球队2总 0总1第一节2第二节3第三节4第四节5f加时6s加时7t加时8f加时
count_time = [[[
MPTime('0:00.0'),
MPTime('0:00.0'),
MPTime('0:00.0'),
MPTime('0:00.0'),
MPTime('0:00.0'),
MPTime('0:00.0'),
MPTime('0:00.0'),
MPTime('0:00.0'),
MPTime('0:00.0')
],
[
MPTime('0:00.0'),
MPTime('0:00.0'),
MPTime('0:00.0'),
MPTime('0:00.0'),
MPTime('0:00.0'),