def main_get(timStr='', nday=2):
#
#1---init.sys
print('\nmain_get,nday:', nday)
tfsys.xnday_down = nday
zsys.web_get001txtFg = True
#
#2---init.tfb
rs0 = '/tfbDat/'
fgid = rs0 + 'gid2017.dat'
xtfb = tft.fb_init(rs0, fgid)
if nday == -1:
tfsys.xnday_down = xtfb.gid_nday + 10
print('nday,', tfsys.xnday_down)
#
#3---update.data
print('\n#3,update.data')
if nday != 0:
tft.fb_gid_get_nday(xtfb, timStr, fgExt=True)
#
#4
tn = zt.timNSec(timStr, xtfb.tim0, '')
print('\n#4,update.data,tim:{0:.2f} s'.format(tn))
def main_ai_bt(timStr='', nday=2):
#
#1---init.sys
print('\nmain_bt,nday:', nday)
tfsys.xnday_down = nday
zsys.web_get001txtFg = True
#2---init.tfb
rs0 = '/tfbDat/'
fgid = rs0 + 'gid2018.dat'
xtfb = tft.fb_init(rs0, fgid)
if nday == -1:
tfsys.xnday_down = xtfb.gid_nday + 10
print('nday,', tfsys.xnday_down)
#
#3---backtest
print('\n#3,backtest')
if nday != 0:
xtfb.funPre = tfsty.sta00_pre #bt_1dayMain
xtfb.funSta = tfsty.sta_ai_log01
xtfb.preVars = []
xtfb.staVars = [99, 99, 99]
#
#
#3.a------ai.init
zai.xmodel = {}
xtfb.ai_mxfFN0 = rs0 + 'mlib/p7y2016_'
xtfb.ai_mx_sgn_lst = ['log']
xtfb.ai_ysgn = 'kwin'
xtfb.ai_xlst = [
'cid', 'pwin0', 'pdraw0', 'plost0', 'pwin9', 'pdraw9', 'plost9'
]
#
zai.ai_f_mxRdlst(xtfb.ai_mxfFN0, xtfb.ai_mx_sgn_lst)
#3.b-------预测
xtfb.kcid = '1' #cn,3=bet365
tfbt.bt_main(xtfb, timStr)
#
#4---main_ret
print('\n#4,result.anz')
tfbt.bt_main_ret(xtfb, True)
print('kcid,', xtfb.kcid, ',nday,', nday)
print('preVar,', xtfb.preVars)
print('staVar,', xtfb.staVars)
#
#5
tn = zt.timNSec('', xtfb.tim0, '')
print('\n#5,backtest,tim:{0:.2f} s'.format(tn))
#
#6---end.main
print('\n#6,end.main')
def main_download_samples():
#1---init.sys
print('\ndownload samples\n')
#
#2---init.tfb
rs0='/tfbDat/'
fgid=rs0+'gid2019_js.dat'
xtfb=tft.fb_init(rs0,fgid)
#---download samples
tft.fb_download_league_data('英冠', '2018-2019', fgSample=True)
def save_odds_to_file_for_xgboost(file_name, num):
cv_sgn = [
'win_cv_0', 'draw_cv_0', 'lost_cv_0', 'win_cv_9', 'draw_cv_9',
'lost_cv_9'
]
#2---init.tfb
rs0 = '/tfbDat/'
fgid = rs0 + 'gid2018-2019(xgboost).dat'
xtfb = tft.fb_init(rs0, fgid)
df = tfsys.gids
p_data = pd.DataFrame()
p_data_tz = pd.DataFrame()
oz_cv = pd.DataFrame()
for i, row in df.iterrows():
if ((i + 1) % 2000 == 0):
print((i + 1) / len(df) * 100, "%")
print('now:', zt.tim_now_str())
if i >= num[0] and i < num[1]:
kend = row['kend']
kwin = row['kwin']
if kend == '0' or kwin == -1: continue
gid = row['gid']
fxdat_tz = tfsys.rxdat + gid + '_tz.dat'
fxdat_oz = tfsys.rxdat + gid + '_oz_1.dat'
if os.path.exists(fxdat_tz) and os.path.exists(fxdat_oz):
odds_tz = pd.read_csv(fxdat_tz,
index_col=False,
dtype=float,
encoding='gb18030')
odds_oz = pd.read_csv(fxdat_oz,
index_col=False,
dtype=float,
encoding='gb18030')
if odds_tz.isnull().values.any(): continue ##如果有Nan值就不保存该数据
win_cv_0, draw_cv_0, lost_cv_0, win_cv_9, draw_cv_9, lost_cv_9 = calculate_odds_cv(
odds_oz)
# cov_win0, cov_lost0, cov_win9, cov_lost9 = calculate_odds_cov(odds_oz)
features = odds_tz
features['win_cv_0'] = win_cv_0
features['draw_cv_0'] = draw_cv_0
features['lost_cv_0'] = lost_cv_0
features['win_cv_9'] = win_cv_9
features['draw_cv_9'] = draw_cv_9
features['lost_cv_9'] = lost_cv_9
p_data = p_data.append(features, ignore_index=True)
p_data.to_csv(file_name, index=False, encoding='gb18030')
def main_bt(timStr='',nday=2):
#
#1---init.sys
print('\nmain_bt,nday:',nday)
tfsys.xnday_down=nday
# zsys.web_get001txtFg= True
#2---init.tfb
rs0='/tfbDat/'
fgid=rs0+'gid2019_js.dat'
xtfb=tft.fb_init(rs0,fgid)
if nday==-1:
tfsys.xnday_down=xtfb.gid_nday+2
print('nday,',tfsys.xnday_down)
#
#3---backtest
print('\n#3,backtest')
if nday!=0:
xtfb.funPre=tfsty.sta_lstm_pre ##预测算法
xtfb.funSta=tfsty.sta00_sta
xtfb.preVars=[]
xtfb.staVars=[1.5]
xtfb.kcid='1' #cn,3=bet365
#
tfbt.bt_main(xtfb,timStr)
print('\n#4, 完成预测')
#4---main_ret 按预测计算预计回报
#暂时不分析预计回报，以后再开发
'''
print('\n#4,result.anz')
tfbt.bt_main_ret(xtfb, True)
print('kcid,',xtfb.kcid,',nday,',nday)
print('preVar,',xtfb.preVars)
print('staVar,',xtfb.staVars)
'''
#
#5
tn=zt.timNSec('',xtfb.tim0,'')
print('\n#5,backtest,tim:{0:.2f} s'.format(tn))
#
#6---end.main
print('\n#6,end.main')
def read_lstm_sample_files(file_name, num):
#2---init.tfb
rs0 = '/tfbDat/'
fgid = rs0 + 'gid2019_js.dat'
xtfb = tft.fb_init(rs0, fgid)
df = tfsys.gids
data = pd.DataFrame()
for i, row in df.iterrows():
if ((i + 1) % 2000 == 0):
print((i + 1) / len(df) * 100, "%")
print('now:', zt.tim_now_str())
if i >= num[0] and i < num[1]:
kend = row['kend']
kwin = row['kwin']
if kend == '0' or kwin == -1: continue
gid = row['gid']
fxdat_ftr = tfsys.rxdat + gid + '_ftr.dat'
if os.path.exists(fxdat_ftr):
features = pd.read_csv(fxdat_ftr,
index_col=False,
encoding='gb18030')
if features.isnull().values.any(): continue ##如果有Nan值就不保存该数据
data = data.append(features, ignore_index=True)
data.loc[:, "FTR"][data.loc[:, "FTR"] == 1] = 0
data.loc[:, "FTR"][data.loc[:, "FTR"] == 3] = 1
data = data.drop([
'vol_prob_h', 'vol_prob_d', 'vol_prob_g', 'loss_idx', 'loss_vol_ratio'
],
axis=1)
data = data.drop([
'avg_win', 'avg_draw', 'avg_lost', 'avg_win_diff', 'avg_draw_diff',
'avg_lost_diff'
],
axis=1)
return data
def save_odds_to_file(file_name, num):
#2---init.tfb
xtfb = tft.fb_init(rs0, fgid)
df = tfsys.gids
p_data = pd.DataFrame()
# p_data_oz = pd.DataFrame()
# p_data_az = pd.DataFrame()
for i, row in df.iterrows():
if ((i+1) % 2000 == 0):
print((i+1)/len(df) * 100, "%")
print('now:',zt.tim_now_str())
if i>=num[0] and i<num[1]:
gid = row['gid']
fxdat_oz = tfsys.rxdat + gid + '_oz.dat'
fxdat_az = tfsys.rxdat + gid + '_az.dat'
fxdat_tz = tfsys.rxdat + gid + '_tz.dat'
if os.path.exists(fxdat_oz) and os.path.exists(fxdat_az) and os.path.exists(fxdat_tz):
odds_oz = pd.read_csv(fxdat_oz, index_col = False, dtype = str, encoding = 'gb18030')
odds_az = pd.read_csv(fxdat_az, index_col = False, dtype = str, encoding = 'gb18030')
odds_tz = pd.read_csv(fxdat_tz, index_col = False, dtype = str, encoding = 'gb18030')
if len(odds_oz) >= tfsys.cidrows \
and odds_oz.loc[0, 'kwin'] != '-1' \
and len(odds_az) >= tfsys.cidrows \
and odds_az.loc[0, 'kwin'] != '-1' \
and len(odds_tz) == 1: #如果数据有CID_ROWS行并且有比赛结果才处理数据
odds_oz = odds_oz[odds_oz['cid'] == '3']
odds_az = odds_az[odds_az['cid'] == '3']
if odds_oz.empty or odds_az.empty:
continue
label = odds_oz['kwin']
odds_oz = odds_oz[tfsys.usedSgn_oz]
odds_az = odds_az[tfsys.usedSgn_az]
odds_tz = odds_tz[tfsys.usedSng_tz]
odds_oz = odds_oz.reset_index(drop=True)
odds_az = odds_az.reset_index(drop=True)
odds_tz = odds_tz.reset_index(drop=True)
label = label.reset_index(drop=True)
# odds_oz = odds_oz.astype(float)
# odds_az = odds_az.astype(float)
flag = is_filter_sample(odds_az)
if flag:
continue
merge_data = pd.concat([odds_az, odds_oz, odds_tz, label], axis=1)
p_data = p_data.append(merge_data, ignore_index=True)
p_data.to_csv(file_name, index=False, encoding='gb18030')
def save_odds_to_file_for_svm(file_name, num):
#2---init.tfb
xtfb = tft.fb_init(rs0, fgid)
df = tfsys.gids
p_data = pd.DataFrame()
p_data_oz = pd.DataFrame()
p_data_az = pd.DataFrame()
oz_cols = [
'pwin0', 'pdraw0', 'plost0', 'pwin9', 'pdraw9', 'plost9', 'vwin0',
'vdraw0', 'vlost0', 'vback0', 'vwin0kali', 'vdraw0kali', 'vlost0kali',
'vwin9', 'vdraw9', 'vlost9', 'vback9', 'vwin9kali', 'vdraw9kali',
'vlost9kali'
]
az_cols = ['mshui0', 'pan0', 'gshui0', 'mshui9', 'pan9', 'gshui9']
kali0 = ['vwin0kali', 'vdraw0kali', 'vlost0kali']
kali9 = ['vwin9kali', 'vdraw9kali', 'vlost9kali']
oz0 = ['pwin0', 'pdraw0', 'plost0']
oz9 = ['pwin9', 'pdraw9', 'plost9']
az0 = ['mshui0', 'pan0', 'gshui0']
az9 = ['mshui9', 'pan9', 'gshui9']
pb0 = ['vwin0', 'vdraw0', 'vlost0']
pb9 = ['vwin9', 'vdraw9', 'vlost9']
vb0 = ['vback0']
vb9 = ['vback9']
oz = ['pwin0', 'pwin9', 'pdraw0', 'pdraw9', 'plost0', 'plost9']
kali_columns = ['vwinkali', 'vdrawkali', 'vlostkali']
oz_columns = ['pwin', 'pdraw', 'plost']
az_columns = ['mshui', 'gshui', 'pan']
for i, row in df.iterrows():
if ((i + 1) % 2000 == 0):
print((i + 1) / len(df) * 100, "%")
print('now:', zt.tim_now_str())
if i >= num[0] and i < num[1]:
gid = row['gid']
fxdat_oz = tfsys.rxdat + gid + '_oz.dat'
fxdat_az = tfsys.rxdat + gid + '_az.dat'
if os.path.exists(fxdat_oz) and os.path.exists(fxdat_az):
odds_oz = pd.read_csv(fxdat_oz,
index_col=False,
dtype=str,
encoding='gb18030')
odds_az = pd.read_csv(fxdat_az,
index_col=False,
dtype=str,
encoding='gb18030')
if len(odds_oz) >= tfsys.cidrows and odds_oz.loc[
0, 'kwin'] != '-1' and len(
odds_az) >= tfsys.cidrows and odds_az.loc[
0,
'kwin'] != '-1': #如果数据有CID_ROWS行并且有比赛结果才处理数据
if float(odds_oz.loc[0, 'pwin0']) <= 1.5 or float(
odds_oz.loc[0, 'plost0']) <= 1.5 or float(
odds_oz.loc[0, 'pwin9']) <= 1.5 or float(
odds_oz.loc[0, 'plost9']) <= 1.5:
continue
cids = ['2', '3', '293']
odds_oz = odds_oz[odds_oz['cid'].isin(cids)]
if len(odds_oz) != 3: continue
# odds_oz = odds_oz[1:tfsys.cidrows+1]
# odds_az = odds_az[0:tfsys.cidrows]
target = odds_oz.pop('kwin')
odds_oz = odds_oz[oz_cols].astype(float)
# odds_az = odds_az[az_cols].astype(float)
#1
odds_oz0_9 = odds_oz[oz]
odds_oz0 = odds_oz[oz0]
odds_oz9 = odds_oz[oz9]
# odds_az0 = odds_az[az0]
# odds_az9 = odds_az[az9]
odds_oz0.columns = oz_columns
odds_oz9.columns = oz_columns
# odds_az0.columns = az_columns
# odds_az9.columns = az_columns
# az_diff = odds_az9 - odds_az0
oz_diff = odds_oz9 / odds_oz0
#2
odds_kali0 = odds_oz[kali0]
odds_kali9 = odds_oz[kali9]
odds_kali0.columns = kali_columns
odds_kali9.columns = kali_columns
kali_diff = odds_kali9 - odds_kali0
#3
vback0 = odds_oz['vback0'] / 100
vback9 = odds_oz['vback0'] / 100
vback0_diff = odds_kali0.sub(vback0, axis=0)
vback9_diff = odds_kali9.sub(vback9, axis=0)
#4
prob0 = odds_oz[pb0]
prob9 = odds_oz[pb9]
target = target.astype(int)
target[target == 3] = 2
target = target.reset_index(drop=True)
# az_diff = az_diff.reset_index(drop=True)
oz_diff = oz_diff.reset_index(drop=True)
kali_diff = kali_diff.reset_index(drop=True)
vback0_diff = vback0_diff.reset_index(drop=True)
vback9_diff = vback9_diff.reset_index(drop=True)
# odds_az0 = odds_az0.reset_index(drop=True)
odds_oz0 = odds_oz0.reset_index(drop=True)
# odds_az9 = odds_az9.reset_index(drop=True)
odds_oz9 = odds_oz9.reset_index(drop=True)
odds_oz0_9 = odds_oz0_9.reset_index(drop=True)
odds_kali0 = odds_kali0.reset_index(drop=True)
odds_kali9 = odds_kali9.reset_index(drop=True)
vback0 = vback0.reset_index(drop=True)
vback9 = vback9.reset_index(drop=True)
prob0 = prob0.reset_index(drop=True)
prob9 = prob9.reset_index(drop=True)
kelly0, kelly9, kelly0_var, kelly9_var, kelly0_diff, kelly9_diff, ky0_var, ky9_var = kelly_variance(
prob0, prob9, odds_oz0, odds_oz9, vback0, vback9)
# merge_data = pd.concat([target, az_diff, oz_diff, kali_diff, vback0_diff, vback9_diff], axis=1)
merge_data = pd.concat(
[target, odds_oz0_9, ky0_var, ky9_var], axis=1)
p_data = p_data.append(merge_data, ignore_index=True)
p_data.to_csv(file_name, index=False, encoding='gb18030')