def book_greek_scen_report(value_date, book_name, \
req_greeks = ['pv', 'theta', 'cmdelta', 'cmgamma', 'cmvega_atm'], \
scens = ['COMFwd', 'SGXIRO', [- 5.0, -3.0, -1.0, 0.0, 1.0, 3.0, 5.0], cmq_inst.CurveShiftType.Abs]):
book = cmq_book.get_book_from_db(book_name, status=[2])
base_mkt = cmq_market_data.load_market_data(book.mkt_deps,
value_date=value_date)
output = {}
for shift in scens[2]:
mkt_data = cmq_inst_risk.generate_scen(base_mkt,
scens[0],
scens[1],
curve_tenor='ALL',
shift_size=shift,
shift_type=scens[3])
re = cmq_risk_engine.CMQRiskEngine(book, mkt_data, req_greeks)
re.run_risk()
output[shift] = copy.deepcopy(re.book_risks)
for greek in req_greeks:
if greek in [
'cmdelta', 'cmvega_atm', 'cmgamma', 'cmvega_v90',
'cmvega_v75', 'cmvega_v25', 'cmvega_v10'
]:
output[shift][greek] = sum(output[shift][greek].values())
df = pd.DataFrame.from_dict(output, orient='index').round(2)
df.index = numpy.round(scens[2], 2)
return df
def greeks_ladder_report(value_date, book_name, \
req_greeks = ['cmdeltas', 'cmgammas', 'cmvegas_atm'], \
scens = ['COMFwd', 'SGXIRO', [-5.0, -4.0, -3.0, -2.0, -1.0, 0.0, 1.0, 2.0, 3.0, 4.0, 5.0], cmq_inst.CurveShiftType.Abs]):
book = cmq_book.get_book_from_db(book_name, status=[2])
base_mkt = cmq_market_data.load_market_data(book.mkt_deps,
value_date=value_date)
output = dict([(greek, {}) for greek in req_greeks])
for shift in scens[2]:
mkt_data = cmq_inst_risk.generate_scen(base_mkt,
scens[0],
scens[1],
curve_tenor='ALL',
shift_size=shift,
shift_type=scens[3])
re = cmq_risk_engine.CMQRiskEngine(book, mkt_data, req_greeks)
re.run_risk()
for greek in req_greeks:
for idx in re.book_risks[greek]:
if idx not in output[greek]:
output[greek][idx] = {}
output[greek][idx][shift] = copy.deepcopy(
re.book_risks[greek][idx])
res = dict([(greek, {}) for greek in req_greeks])
for greek in req_greeks:
for idx in output[greek]:
if idx not in res[greek]:
res[greek][idx] = {}
res[greek][idx] = pd.DataFrame.from_dict(output[greek][idx],
orient='columns')
return res
def book_greek_scen_report(value_date, book, \
req_greeks = ['pv', 'theta', 'cmdelta', 'cmgamma', 'cmvega_atm'], \
scens = ['COMFwd', 'SGXIRO', [- 5.0, -3.0, -1.0, 0.0, 1.0, 3.0, 5.0], 0], \
base_mkt = {}, use_pool = False):
if len(base_mkt) == 0:
base_mkt = cmq_market_data.load_market_data(book.mkt_deps,
value_date=value_date)
output = {}
if use_pool:
num_cpus = mp.cpu_count() - 1
pool = mp.Pool(num_cpus)
else:
pool = None
re = {}
for shift in scens[2]:
mkt_data = cmq_inst_risk.generate_scen(base_mkt,
scens[0],
scens[1],
curve_tenor='ALL',
shift_size=shift,
shift_type=scens[3])
re[shift] = cmq_risk_engine.CMQRiskEngine(book, mkt_data, req_greeks,
pool)
re[shift].run_scenarios()
for shift in scens[2]:
re[shift].summerize_risks()
output[shift] = copy.deepcopy(re[shift].book_risks)
for greek in req_greeks:
if greek in [
'cmdelta', 'cmvega_atm', 'cmgamma', 'cmvega_v90',
'cmvega_v75', 'cmvega_v25', 'cmvega_v10'
]:
output[shift][greek] = sum(output[shift][greek].values())
df = pd.DataFrame.from_dict(output, orient='index')
return df
def scenario_2d_report(value_date, book_name, greeks, \
scens = [['COMFwd', 'SGXIRO', [-3.0, -2.0, -1.0, 0.0, 1.0, 2.0, 3.0], cmq_inst.CurveShiftType.Abs], \
['COMVolATM', 'SGXIRO', [-0.02, -0.01, 0.0, 0.01, 0.02], cmq_inst.CurveShiftType.Abs]]):
book = cmq_book.get_book_from_db(book_name, status=[2])
base_mkt = cmq_market_data.load_market_data(book.mkt_deps,
value_date=value_date)
res = {}
for greek in greeks:
output = {}
for shift_x in scens[0][2]:
mkt_data_x = cmq_inst_risk.generate_scen(base_mkt, scens[0][0], scens[0][1], 'ALL', \
shift_size=shift_x, shift_type=scens[0][3])
x_key = scens[0][0] + '_' + str(shift_x)
output[x_key] = {}
for shift_y in scens[1][2]:
mkt_data_y = cmq_inst_risk.generate_scen(mkt_data_x, scens[1][0], scens[1][1], 'ALL', \
shift_size = shift_y, shift_type = scens[1][3])
re = cmq_risk_engine.CMQRiskEngine(book, mkt_data_y, [greek])
re.run_risk()
y_key = scens[1][0] + '_' + str(shift_y)
output[x_key][y_key] = re.book_risks[greek]
if greek in ['cmdelta', 'cmvega_atm', 'cmvega_v90', 'cmvega_v75', \
'cmvega_v25', 'cmvega_v10', 'cmgamma']:
output[x_key][y_key] = sum(output[x_key][y_key].values())
res[greek] = pd.DataFrame.from_dict(output, orient='index')
return res
def run_book_report(value_date, book, fwd_idx = 'SGXIRO', req_greeks = ['pv','cmdeltas', 'theta'], \
base_mkt = {}):
if len(base_mkt) == 0:
mkt_data = cmq_market_data.load_market_data(book.mkt_deps,
value_date=value_date)
else:
mkt_data = base_mkt
re = cmq_risk_engine.CMQRiskEngine(book, mkt_data, req_greeks)
re.run_risk()
# assume only one instrument for one deal
deal_results = copy.deepcopy(re.deal_risks)
for deal in book.deal_list:
deal_id = deal.id
deal_results[deal_id]['strike'] = deal.positions[0][0].strike
deal_results[deal_id]['otype'] = deal.positions[0][0].otype
deal_results[deal_id]['end'] = deal.positions[0][0].end
deal_results[deal_id]['volume'] = deal.positions[0][1]
for greek in req_greeks:
if greek in [
'cmdelta', 'cmvega_atm', 'cmvega_v90', 'cmvega_v75',
'cmvega_v25', 'cmvega_v10', 'cmgamma'
]:
deal_results[deal_id][greek] = deal_results[deal_id][greek][
fwd_idx]
df = pd.DataFrame.from_dict(deal_results, orient='index')
return re, df
def run_book(value_date, book, req_greeks = ['cmdelta', 'cmgamma', 'cmvega_atm', 'theta'], \
base_mkt = {}, region = 'EOD'):
book_obj = get_book(book)
if len(base_mkt) == 0:
mkt_data = cmq_market_data.load_market_data(book_obj.mkt_deps,
value_date=value_date,
region=region)
else:
mkt_data = base_mkt
re = cmq_risk_engine.CMQRiskEngine(book_obj, mkt_data, req_greeks)
re.run_risk()
return re, book_obj, mkt_data
def greek_ladder_report(value_date, book, \
req_greeks = ['cmdeltas', 'cmgammas', 'cmvegas_atm'], \
scens = ['COMFwd', [-0.05, -0.03, -0.02, -0.01, 0.0, 0.01, 0.02, 0.03, 0.05], 1],\
base_mkt = {}, region = 'EOD', use_pool = False):
if len(base_mkt) == 0:
base_mkt = cmq_market_data.load_market_data(book.mkt_deps,
value_date=value_date,
region=region)
output = dict([(greek, {}) for greek in req_greeks])
if use_pool:
num_cpus = mp.cpu_count() - 1
pool = mp.Pool(num_cpus)
else:
pool = None
re = {}
fwd_list = book.mkt_deps[scens[0]].keys()
for fwd_idx in fwd_list:
for shift in scens[1]:
key = (fwd_idx, shift)
mkt_data = cmq_inst_risk.generate_scen(base_mkt,
scens[0],
fwd_idx,
curve_tenor='ALL',
shift_size=shift,
shift_type=scens[2])
re[key] = cmq_risk_engine.CMQRiskEngine(book, mkt_data, req_greeks,
pool)
re[key].run_scenarios()
for fwd_idx in fwd_list:
output[greek][fwd_idx] = {}
for shift in scens[1]:
key = (fwd_idx, shift)
re[key].summerize_risks()
if fwd_idx in re[key].book_risks[greek]:
for greek in req_greeks:
output[greek][fwd_idx][shift] = copy.deepcopy(
re[key].book_risks[greek][fwd_idx])
res = dict([(greek, pd.DataFrame()) for greek in req_greeks])
for greek in req_greeks:
for fwd_idx in fwd_list:
df = pd.DataFrame.from_dict(output[greek][fwd_idx],
orient='columns')
df.loc['Total', :] = df.sum()
df = df.reset_index()
df.rename(columns={'index': 'tenor'}, inplace=True)
cols = list(df.columns)
df['fwd_idx'] = fwd_idx
df = df[['fwd_idx'] + cols]
res[greek] = res[greek].append(df)
res[greek] = res[greek].set_index(['fwd_idx', 'tenor'])
return res
def run_book_report(value_date, book, req_greeks = ['cmdelta', 'cmgamma', 'cmvega_atm', 'theta'], \
base_mkt = {}, region = 'EOD'):
if len(base_mkt) == 0:
mkt_data = cmq_market_data.load_market_data(book.mkt_deps,
value_date=value_date,
region=region)
else:
mkt_data = base_mkt
re = cmq_risk_engine.CMQRiskEngine(book, mkt_data, req_greeks)
re.run_risk()
# assume only one instrument for one deal
deal_results = {}
for deal in book.deal_list:
if deal.positions[0][0].inst_type in ["ComEuroOption", 'ComMthAsian']:
deal_id = deal.id
deal_results[deal_id] = {}
deal_results[deal_id]['internal_id'] = deal.internal_id
deal_results[deal_id]['strike'] = deal.positions[0][0].strike
deal_results[deal_id]['otype'] = deal.positions[0][0].otype
deal_results[deal_id]['end'] = deal.positions[0][0].end
deal_results[deal_id][
'volume'] = deal.positions[0][1] * deal.positions[0][0].volume
deal_results[deal_id]['fwd_idx'] = deal.positions[0][0].fwd_index
for greek in req_greeks:
if greek in [
'cmdelta', 'cmvega_atm', 'cmvega_v90', 'cmvega_v75',
'cmvega_v25', 'cmvega_v10', 'cmgamma'
]:
for fwd_idx in re.deal_risks[deal_id][greek]:
deal_results[deal_id][greek] = re.deal_risks[deal_id][
greek][fwd_idx]
else:
deal_results[deal_id][greek] = re.deal_risks[deal_id][
greek]
df = pd.DataFrame.from_dict(deal_results, orient='index')
volume = pd.pivot_table(df, values='volume', index=['fwd_idx', 'end'], \
columns=['otype', 'strike'], aggfunc=np.sum, margins=True).fillna(0)
delta = pd.pivot_table(df, values = 'cmdelta', index=['fwd_idx', 'end'], \
columns=['otype', 'strike'], aggfunc=np.sum, margins=True).fillna(0)
gamma = pd.pivot_table(df, values='cmgamma', index=['fwd_idx', 'end'], \
columns=['otype', 'strike'], aggfunc=np.sum, margins=True).fillna(0)
vega = pd.pivot_table(df, values='cmvega_atm', index=['fwd_idx', 'end'], \
columns=['otype', 'strike'], aggfunc=np.sum, margins=True).fillna(0)
theta = pd.pivot_table(df, values='theta', index=['end'], \
columns=['otype', 'strike'], aggfunc=np.sum, margins=True).fillna(0)
return volume, delta, gamma, vega, theta, df
def book_greek_scen_report(value_date, book, \
req_greeks = ['pv', 'theta', 'cmdelta', 'cmgamma', 'cmvega_atm'], \
scens = ['COMFwd', [- 5.0, -3.0, -1.0, 0.0, 1.0, 3.0, 5.0], 0], \
base_mkt = {}, region = 'EOD', use_pool = False):
if len(base_mkt) == 0:
base_mkt = cmq_market_data.load_market_data(book.mkt_deps,
value_date=value_date,
region=region)
output = {}
if use_pool:
num_cpus = mp.cpu_count() - 1
pool = mp.Pool(num_cpus)
else:
pool = None
re = {}
fwd_list = book.mkt_deps[scens[0]].keys()
for fwd_idx in fwd_list:
for shift in scens[1]:
mkt_data = cmq_inst_risk.generate_scen(base_mkt,
scens[0],
fwd_idx,
curve_tenor='ALL',
shift_size=shift,
shift_type=scens[2])
key = (fwd_idx, shift)
re[key] = cmq_risk_engine.CMQRiskEngine(book, mkt_data, req_greeks,
pool)
re[key].run_scenarios()
for fwd_idx in fwd_list:
for shift in scens[1]:
key = (fwd_idx, shift)
re[key].summerize_risks()
output[key] = copy.deepcopy(re[key].book_risks)
for greek in req_greeks:
if greek in [
'cmdelta', 'cmvega_atm', 'cmgamma', 'cmvega_v90',
'cmvega_v75', 'cmvega_v25', 'cmvega_v10'
]:
if fwd_idx in output[key][greek]:
output[key][greek] = output[key][greek][fwd_idx]
else:
output[key][greek] = 0.0
df = pd.DataFrame.from_dict(output, orient='index')
return df
def run_book_report(value_date, book_name, fwd_idx='SGXIRO'):
book = cmq_book.get_book_from_db(book_name, status=[2])
mkt_data = cmq_market_data.load_market_data(book.mkt_deps,
value_date=value_date)
req_greeks = ['pv', 'cmdelta', 'cmgamma', 'cmvega_atm', 'theta']
re = cmq_risk_engine.CMQRiskEngine(book, mkt_data, req_greeks)
re.run_risk()
# assume only one instrument for one deal
deal_results = copy.deepcopy(re.deal_risks)
for deal in book.deal_list:
deal_id = deal.id
deal_results[deal_id]['strike'] = deal.positions[0][0].strike
deal_results[deal_id]['otype'] = deal.positions[0][0].otype
deal_results[deal_id]['end'] = deal.positions[0][0].end
deal_results[deal_id]['volume'] = deal.positions[0][1]
for greek in ['cmdelta', 'cmvega_atm', 'cmgamma']:
deal_results[deal_id][greek] = deal_results[deal_id][greek][
fwd_idx]
df = pd.DataFrame.from_dict(deal_results, orient='index').round(2)
vol_tbl = df.pivot_table(columns=['otype', 'strike'],
index=['end'],
values=['volume'],
aggfunc='sum').round(2)
delta_tbl = df.pivot_table(columns=['otype', 'strike'],
index=['end'],
values=['cmdelta'],
aggfunc='sum').round(2)
gamma_tbl = df.pivot_table(columns=['otype', 'strike'],
index=['end'],
values=['cmgamma'],
aggfunc='sum').round(2)
vega_tbl = df.pivot_table(columns=['otype', 'strike'],
index=['end'],
values=['cmvega_atm'],
aggfunc='sum').round(2)
theta_tbl = df.pivot_table(columns=['otype', 'strike'],
index=['end'],
values=['theta'],
aggfunc='sum').round(2)
return re
def greeks_ladder_report(value_date, book, \
req_greeks = ['cmdeltas', 'cmgammas', 'cmvegas_atm'], \
scens = ['COMFwd', 'SGXIRO', [-5.0, -4.0, -3.0, -2.0, -1.0, 0.0, 1.0, 2.0, 3.0, 4.0, 5.0], 0],\
base_mkt = {}, use_pool = False):
if len(base_mkt) == 0:
base_mkt = cmq_market_data.load_market_data(book.mkt_deps,
value_date=value_date)
output = dict([(greek, {}) for greek in req_greeks])
if use_pool:
num_cpus = mp.cpu_count() - 1
pool = mp.Pool(num_cpus)
else:
pool = None
re = {}
for shift in scens[2]:
mkt_data = cmq_inst_risk.generate_scen(base_mkt,
scens[0],
scens[1],
curve_tenor='ALL',
shift_size=shift,
shift_type=scens[3])
re[shift] = cmq_risk_engine.CMQRiskEngine(book, mkt_data, req_greeks,
pool)
re[shift].run_scenarios()
for shift in scens[2]:
re[shift].summerize_risks()
for greek in req_greeks:
for idx in re[shift].book_risks[greek]:
if idx not in output[greek]:
output[greek][idx] = {}
output[greek][idx][shift] = copy.deepcopy(
re[shift].book_risks[greek][idx])
res = dict([(greek, {}) for greek in req_greeks])
for greek in req_greeks:
for idx in output[greek]:
if idx not in res[greek]:
res[greek][idx] = {}
res[greek][idx] = pd.DataFrame.from_dict(output[greek][idx],
orient='columns')
return res
def get_engine(book, mkt, greeks):
return cmq_risk_engine.CMQRiskEngine(book, mkt, greeks)
def test():
book = cmq_book.get_book_from_db('BOF', status=[2])
mkt_data = cmq_market_data.load_market_data(book.mkt_deps, value_date = datetime.date(2017,9,29))
req_greeks = ['pv','cmdelta','cmgamma','cmvega']
re = cmq_risk_engine.CMQRiskEngine(book, mkt_data, req_greeks)
re.run_risk()