Skip to content

Optimized code using numpy vectorizations #33

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
ajinkya-kulkarni wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Optimized code using numpy vectorizations #33

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 11 additions & 8 deletions Python-module/SpatialDE/base.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,7 +37,7 @@ def get_l_limits(X):

def SE_kernel(X, l):
X = np.array(X)
Xsq = np.sum(np.square(X), 1)
Xsq = np.sum(np.square(X), axis=1)
R2 = -2. * np.dot(X, X.T) + (Xsq[:, None] + Xsq[None, :])
R2 = np.clip(R2, 1e-12, np.inf)
return np.exp(-R2 / (2 * l ** 2))
Expand All @@ -61,13 +61,12 @@ def cosine_kernel(X, p):


def gower_scaling_factor(K):
''' Gower normalization factor for covariance matric K

''' Gower normalization factor for covariance matrix K
Based on https://github.com/PMBio/limix/blob/master/limix/utils/preprocess.py
'''
n = K.shape[0]
P = np.eye(n) - np.ones((n, n)) / n
KP = K - K.mean(0)[:, np.newaxis]
KP = K - K.mean(axis=0, keepdims=True)
trPKP = np.sum(P * KP)

return trPKP / (n - 1)
Expand All @@ -94,8 +93,8 @@ def mu_hat(delta, UTy, UT1, S, n, Yvar=None):
Yvar = np.ones_like(S)

UT1_scaled = UT1 / (S + delta * Yvar)
sum_1 = UT1_scaled.dot(UTy)
sum_2 = UT1_scaled.dot(UT1)
sum_1 = np.dot(UT1_scaled, UTy)
sum_2 = np.dot(UT1_scaled, UT1)

return sum_1 / sum_2

Expand All @@ -107,7 +106,7 @@ def s2_t_hat(delta, UTy, S, n, Yvar=None):
Yvar = np.ones_like(S)

UTy_scaled = UTy / (S + delta * Yvar)
return UTy_scaled.dot(UTy) / n
return np.dot(UTy_scaled, UTy) / n


def LL(delta, UTy, UT1, S, n, Yvar=None):
Expand Down Expand Up @@ -137,7 +136,11 @@ def logdelta_prior_lpdf(log_delta):

def make_objective(UTy, UT1, S, n, Yvar=None):
def LL_obj(log_delta):
return -LL(np.exp(log_delta), UTy, UT1, S, n, Yvar)
delta = np.exp(log_delta)
mu_h = UT1.dot(UTy / (S + delta * Yvar)) / UT1.dot(UT1 / (S + delta * Yvar))
LL_val = -0.5 * (n * np.log(2 * np.pi) + np.sum(np.square((UTy - mu_h * UT1)) / (S + delta * Yvar)) +
np.sum(np.log(S + delta * Yvar)))
return -LL_val

return LL_obj

Expand Down