diff options
Diffstat (limited to 'python/analytics/tranche_functions.py')
| -rw-r--r-- | python/analytics/tranche_functions.py | 18 |
1 files changed, 9 insertions, 9 deletions
diff --git a/python/analytics/tranche_functions.py b/python/analytics/tranche_functions.py index 21c7b328..24f06484 100644 --- a/python/analytics/tranche_functions.py +++ b/python/analytics/tranche_functions.py @@ -2,8 +2,8 @@ import numpy as np from ctypes import * from numpy.ctypeslib import ndpointer from quantlib.time.schedule import Schedule, CDS2015 -from quantlib.time.api import Actual360, Period, WeekendsOnly, ModifiedFollowing, Unadjusted, today -from quantlib.util.converter import qldate_to_pydate, pydate_to_qldate +from quantlib.time.api import Actual360, Period, WeekendsOnly, ModifiedFollowing, Unadjusted +from quantlib.util.converter import pydate_to_qldate import pandas as pd from scipy.special import h_roots import os @@ -293,7 +293,7 @@ def tranche_cl(L, R, cs, K1, K2, scaled=False): else: support = np.linspace(0, 1, L.shape[0]) size = K2 - K1 - np.dot(trancheloss(support, K1, K2), L) - \ - np.dot(trancherecov(support, K1, K2), R) + np.dot(trancherecov(support, K1, K2), R) sizeadj = 0.5 * (size + np.hstack((K2-K1, size[:-1]))) if scaled: return 1 / (K2-K1) * np.dot(sizeadj * cs["coupons"], cs["df"]) @@ -343,7 +343,8 @@ def cds_accrued(tradedate, coupon): start_protection = tradedate + 1 DC = Actual360() cal = WeekendsOnly() - sched = Schedule(tradedate, end, Period('3M'), cal, date_generation_rule=CDS2015) + sched = Schedule.from_rule(tradedate, end, Period('3M'), cal, + date_generation_rule=CDS2015) prevpaydate = sched.previous_date(start_protection) return DC.year_fraction(prevpaydate, start_protection) * coupon @@ -377,15 +378,14 @@ def dist_transform2(q): return distDR def compute_pv(q, strike): - """ compute E(1_{R^\bar \leq strike} * D)""" - D = 0 + r""" compute E(1_{R^\bar \leq strike} * D)""" for i in range(q.shape): - val += sum(v for k, v in q[i].items() if k < strike) * 1/Ngrid + val += sum(v for k, v in q[i].items() if k < strike) * 1 / Ngrid return val def average_recov(p, R, Ngrid): - q = np.zeros((p.shape[0]+1, Ngrid)) - q[0,0] = 1 + q = np.zeros((p.shape[0] + 1, Ngrid)) + q[0, 0] = 1 lu = 1 / (Ngrid-1) weights = np.empty(Ngrid) index = np.empty(Ngrid) |