def generate_ifs_formatted_output(**kwargs):
if 'report_date' in kwargs.keys():
report_date = kwargs['report_date']
else:
report_date = exp.doubledate_shift_bus_days()
output_dir = ts.create_strategy_output_dir(strategy_class='ifs',
report_date=report_date)
ifs_output = ifs.generate_ifs_sheet_4date(date_to=report_date)
intraday_spreads = ifs_output['intraday_spreads']
intraday_spreads.rename(columns={
'ticker_head1': 'tickerHead1',
'ticker_head2': 'tickerHead2',
'ticker_head3': 'tickerHead3'
},
inplace=True)
writer = pd.ExcelWriter(output_dir + '/' + futil.xls_file_names['ifs'] +
'.xlsx',
engine='xlsxwriter')
intraday_spreads.to_excel(writer, sheet_name='all')
def generate_ifs_formatted_output(**kwargs):
if "report_date" in kwargs.keys():
report_date = kwargs["report_date"]
else:
report_date = exp.doubledate_shift_bus_days()
output_dir = ts.create_strategy_output_dir(strategy_class="ifs", report_date=report_date)
ifs_output = ifs.generate_ifs_sheet_4date(date_to=report_date)
intraday_spreads = ifs_output["intraday_spreads"]
intraday_spreads.rename(
columns={"ticker_head1": "tickerHead1", "ticker_head2": "tickerHead2", "ticker_head3": "tickerHead3"},
inplace=True,
)
writer = pd.ExcelWriter(output_dir + "/" + futil.xls_file_names["ifs"] + ".xlsx", engine="xlsxwriter")
intraday_spreads.to_excel(writer, sheet_name="all")
def backtest_ifs2_4date(**kwargs):
report_date = kwargs['report_date']
backtest_date = exp.doubledate_shift_bus_days(double_date=report_date,
shift_in_days=-1)
output_dir = ts.create_strategy_output_dir(strategy_class='ifs',
report_date=report_date)
sheet_output = ifs.generate_ifs_sheet_4date(date_to=report_date)
intraday_spreads = sheet_output['intraday_spreads']
intraday_spreads.sort(['spread_description', 'min_volume'],
ascending=[True, False],
inplace=True)
intraday_spreads.drop_duplicates('spread_description', inplace=True)
intraday_spreads.reset_index(drop=True, inplace=True)
intraday_spreads['pnl'] = np.nan
intraday_spreads['num_trades'] = np.nan
intraday_spreads['mean_holding_period'] = np.nan
signal_name = 'ma40_spread'
for i in range(len(intraday_spreads.index)):
ticker_list = [
intraday_spreads.iloc[i]['contract1'],
intraday_spreads.iloc[i]['contract2'],
intraday_spreads.iloc[i]['contract3']
]
ticker_list = [x for x in ticker_list if x is not None]
ticker_head_list = [
cmi.get_contract_specs(x)['ticker_head'] for x in ticker_list
]
num_contracts = len(ticker_list)
weights_output = sutil.get_spread_weights_4contract_list(
ticker_head_list=ticker_head_list)
contract_multiplier_list = [
cmi.contract_multiplier[x] for x in ticker_head_list
]
spread_weights = weights_output['spread_weights']
intraday_data = iml.get_intraday_data_4spread(
ticker_list=ticker_list,
date_to=backtest_date,
num_days_back=0,
spread_weights=spread_weights)
intraday_data = intraday_data[intraday_data['hour_minute'] > 930]
pnl_list = []
holding_period_list = []
current_position = 0
for j in range(len(intraday_data.index)):
if (current_position
== 0) & (intraday_data[signal_name].iloc[j] <
intraday_spreads['maSpreadLowL'].iloc[i]):
current_position = 1
entry_point = j
entry_price = intraday_data['spread'].iloc[j]
elif (current_position
== 0) & (intraday_data[signal_name].iloc[j] >
intraday_spreads['maSpreadHighL'].iloc[i]):
current_position = -1
entry_point = j
entry_price = intraday_data['spread'].iloc[j]
elif (current_position == 1) & (
(intraday_data[signal_name].iloc[j] > 0) |
(j == len(intraday_data.index) - 1)):
current_position = 0
exit_price = intraday_data['spread'].iloc[j]
pnl_list.append(contract_multiplier_list[0] *
(exit_price - entry_price) /
spread_weights[0] - 2 * num_contracts)
holding_period_list.append(j - entry_point)
elif (current_position == -1) & (
(intraday_data[signal_name].iloc[j] < 0) |
(j == len(intraday_data.index) - 1)):
current_position = 0
exit_price = intraday_data['spread'].iloc[j]
pnl_list.append(contract_multiplier_list[0] *
(entry_price - exit_price) /
spread_weights[0] - 2 * num_contracts)
holding_period_list.append(j - entry_point)
intraday_spreads['pnl'].iloc[i] = sum(pnl_list)
intraday_spreads['num_trades'].iloc[i] = len(pnl_list)
intraday_spreads['mean_holding_period'].iloc[i] = np.mean(
holding_period_list)
return intraday_spreads
def backtest_ifs_4date(**kwargs):
report_date = kwargs['report_date']
output_dir = ts.create_strategy_output_dir(strategy_class='ifs', report_date=report_date)
#if os.path.isfile(output_dir + '/backtest_results.pkl'):
# intraday_spreads = pd.read_pickle(output_dir + '/backtest_results.pkl')
# return intraday_spreads
start_hour = dt.time(9, 0, 0, 0)
end_hour = dt.time(12, 55, 0, 0)
sheet_output = ifs.generate_ifs_sheet_4date(date_to=report_date)
intraday_spreads = sheet_output['intraday_spreads']
intraday_spreads['pnl1'] = 0
intraday_spreads['pnl2'] = 0
intraday_spreads['pnl5'] = 0
intraday_spreads['pnl6'] = 0
intraday_spreads['pnl7'] = 0
intraday_spreads['pnl1_wc'] = 0
intraday_spreads['pnl2_wc'] = 0
intraday_spreads['pnl5_wc'] = 0
intraday_spreads['pnl6_wc'] = 0
intraday_spreads['pnl7_wc'] = 0
intraday_spreads['report_date'] = report_date
intraday_spreads['spread_description'] = intraday_spreads.apply(lambda x: x['ticker_head1']+ '_' +x['ticker_head2'] if x['ticker_head3'] is None else x['ticker_head1']+ '_' +x['ticker_head2'] + '_' + x['ticker_head3'] ,axis=1)
intraday_spreads['min_volume'] = intraday_spreads.apply(lambda x: min(x['volume1'],x['volume2']) if x['ticker_head3'] is None else min(x['volume1'],x['volume2'],x['volume3']) ,axis=1)
intraday_spreads.sort(['spread_description','min_volume'],ascending=[True, False],inplace=True)
intraday_spreads.drop_duplicates('spread_description',inplace=True)
intraday_spreads = intraday_spreads[intraday_spreads['hs']>-1]
intraday_spreads.reset_index(drop=True,inplace=True)
date_list = [exp.doubledate_shift_bus_days(double_date=report_date, shift_in_days=x) for x in [-1,-2]]
for i in range(len(intraday_spreads.index)):
#print(i)
ticker_list = [intraday_spreads.iloc[i]['contract1'],intraday_spreads.iloc[i]['contract2'],intraday_spreads.iloc[i]['contract3']]
ticker_list = [x for x in ticker_list if x is not None]
ticker_head_list = [cmi.get_contract_specs(x)['ticker_head'] for x in ticker_list]
num_contracts = len(ticker_list)
weights_output = sutil.get_spread_weights_4contract_list(ticker_head_list=ticker_head_list)
contract_multiplier_list = [cmi.contract_multiplier[x] for x in ticker_head_list]
spread_weights = weights_output['spread_weights']
intraday_data = opUtil.get_aligned_futures_data_intraday(contract_list=ticker_list,
date_list=date_list)
intraday_data['spread'] = 0
for j in range(num_contracts):
intraday_data['c' + str(j+1), 'mid_p'] = (intraday_data['c' + str(j+1)]['best_bid_p'] +
intraday_data['c' + str(j+1)]['best_ask_p'])/2
intraday_data['spread'] = intraday_data['spread']+intraday_data['c' + str(j+1)]['mid_p']*spread_weights[j]
selection_indx = [x for x in range(len(intraday_data.index)) if
(intraday_data.index[x].to_datetime().time() < end_hour)
and(intraday_data.index[x].to_datetime().time() >= start_hour)]
intraday_data = intraday_data.iloc[selection_indx]
intraday_data['time_stamp'] = [x.to_datetime() for x in intraday_data.index]
intraday_data['settle_date'] = intraday_data['time_stamp'].apply(lambda x: x.date())
unique_settle_dates = intraday_data['settle_date'].unique()
intraday_data['spread1'] = np.nan
if len(unique_settle_dates)<2:
continue
final_spread_price = np.mean(intraday_data['spread'][(intraday_data['settle_date'] == unique_settle_dates[1])])
intraday_data = intraday_data[(intraday_data['settle_date'] == unique_settle_dates[0])]
intraday_data['spread_diff'] = contract_multiplier_list[0]*(final_spread_price-intraday_data['spread'])/spread_weights[0]
mean5 = intraday_spreads.iloc[i]['mean']
std5 = intraday_spreads.iloc[i]['std']
mean1 = intraday_spreads.iloc[i]['mean1']
std1 = intraday_spreads.iloc[i]['std1']
mean2 = intraday_spreads.iloc[i]['mean2']
std2 = intraday_spreads.iloc[i]['std2']
long_qty = -5000/intraday_spreads.iloc[i]['downside']
short_qty = -5000/intraday_spreads.iloc[i]['upside']
intraday_data['z5'] = (intraday_data['spread']-mean5)/std5
intraday_data['z1'] = (intraday_data['spread']-mean1)/std1
intraday_data['z2'] = (intraday_data['spread']-mean2)/std2
intraday_data['z6'] = (intraday_data['spread']-mean1)/std5
intraday_data11 = intraday_data[intraday_data['z1']>1]
intraday_data1_1 = intraday_data[intraday_data['z1']<-1]
if intraday_data1_1.empty:
pnl1_1 = 0
pnl1_1_wc = 0
else:
pnl1_1 = long_qty*intraday_data1_1['spread_diff'].mean()
pnl1_1_wc = pnl1_1 - 2*2*long_qty*num_contracts
if intraday_data11.empty:
pnl11 = 0
pnl11_wc = 0
else:
pnl11 = short_qty*intraday_data11['spread_diff'].mean()
pnl11_wc = pnl11 + 2*2*short_qty*num_contracts
intraday_data21 = intraday_data[intraday_data['z2']>1]
intraday_data2_1 = intraday_data[intraday_data['z2']<-1]
if intraday_data2_1.empty:
pnl2_1 = 0
pnl2_1_wc = 0
else:
pnl2_1 = long_qty*intraday_data2_1['spread_diff'].mean()
pnl2_1_wc = pnl2_1 - 2*2*long_qty*num_contracts
if intraday_data21.empty:
pnl21 = 0
pnl21_wc = 0
else:
pnl21 = short_qty*intraday_data21['spread_diff'].mean()
pnl21_wc = pnl21 + 2*2*short_qty*num_contracts
intraday_data51 = intraday_data[intraday_data['z5']>1]
intraday_data5_1 = intraday_data[intraday_data['z5']<-1]
if intraday_data5_1.empty:
pnl5_1 = 0
pnl5_1_wc = 0
else:
pnl5_1 = long_qty*intraday_data5_1['spread_diff'].mean()
pnl5_1_wc = pnl5_1 - 2*2*long_qty*num_contracts
if intraday_data51.empty:
pnl51 = 0
pnl51_wc = 0
else:
pnl51 = short_qty*intraday_data51['spread_diff'].mean()
pnl51_wc = pnl51 + 2*2*short_qty*num_contracts
intraday_data61 = intraday_data[intraday_data['z6']>0.25]
intraday_data6_1 = intraday_data[intraday_data['z6']<-0.25]
if intraday_data6_1.empty:
pnl6_1 = 0
pnl6_1_wc = 0
else:
pnl6_1 = long_qty*intraday_data6_1['spread_diff'].mean()
pnl6_1_wc = pnl6_1 - 2*2*long_qty*num_contracts
if intraday_data61.empty:
pnl61 = 0
pnl61_wc = 0
else:
pnl61 = short_qty*intraday_data61['spread_diff'].mean()
pnl61_wc = pnl61 + 2*2*short_qty*num_contracts
intraday_data71 = intraday_data[(intraday_data['z6']>0.1)&(intraday_data['z5']>0.5)]
intraday_data7_1 = intraday_data[(intraday_data['z6']<-0.1)&(intraday_data['z5']<0.5)]
if intraday_data7_1.empty:
pnl7_1 = 0
pnl7_1_wc = 0
else:
pnl7_1 = long_qty*intraday_data7_1['spread_diff'].mean()
pnl7_1_wc = pnl7_1 - 2*2*long_qty*num_contracts
if intraday_data71.empty:
pnl71 = 0
pnl71_wc = 0
else:
pnl71 = short_qty*intraday_data71['spread_diff'].mean()
pnl71_wc = pnl71 + 2*2*short_qty*num_contracts
intraday_spreads['pnl1'].iloc[i] = pnl1_1 + pnl11
intraday_spreads['pnl2'].iloc[i] = pnl2_1 + pnl21
intraday_spreads['pnl5'].iloc[i] = pnl5_1 + pnl51
intraday_spreads['pnl6'].iloc[i] = pnl6_1 + pnl61
intraday_spreads['pnl7'].iloc[i] = pnl7_1 + pnl71
intraday_spreads['pnl1_wc'].iloc[i] = pnl1_1_wc + pnl11_wc
intraday_spreads['pnl2_wc'].iloc[i] = pnl2_1_wc + pnl21_wc
intraday_spreads['pnl5_wc'].iloc[i] = pnl5_1_wc + pnl51_wc
intraday_spreads['pnl6_wc'].iloc[i] = pnl6_1_wc + pnl61_wc
intraday_spreads['pnl7_wc'].iloc[i] = pnl7_1_wc + pnl71_wc
intraday_spreads.to_pickle(output_dir + '/backtest_results.pkl')
return intraday_spreads
