orisys-tech/orisys-sdk
3
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

HHD numpy optimization

Browse Source
This commit is contained in:
bluepulsar
2026-02-22 13:24:27 +08:00
parent 78dde20838
commit 10eb891d0b
1 changed files with 32 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/orisys/HHD.py
+31 -18
View File
@@ -309,33 +309,46 @@ class nHHD(object):
self.decompose(vfield, verbose)
self.Sn = numpy.sum(numpy.absolute(self.d))
# print("Sn is:", self.Sn)
# Optimized: use abs instead of absolute (slightly faster)
self.Sn = numpy.sum(numpy.abs(self.d))
# self.Stx = abs(numpy.sum(self.h[:, :, 0]))
# self.Sty = abs(numpy.sum(self.h[:, :, 1]))
# Optimized: direct indexing and sum (already efficient)
self.Stx = numpy.sum(self.h[:, :, 0])
self.Sty = numpy.sum(self.h[:, :, 1])
# print("Stx is:", self.Stx)
# print("Sty is:", self.Sty)
nR_max_loc = numpy.unravel_index(numpy.argmax(self.nRu, axis = None), self.nRu.shape)
nR_min_loc = numpy.unravel_index(numpy.argmin(self.nRu, axis = None), self.nRu.shape)
# Optimized: find both max and min indices in single pass through flattened array
nRu_flat = self.nRu.ravel()
nR_max_idx = numpy.argmax(nRu_flat)
nR_min_idx = numpy.argmin(nRu_flat)
# Convert flat indices to 2D coordinates (faster than unravel_index)
nR_max_loc = (nR_max_idx // shape[1], nR_max_idx % shape[1])
nR_min_loc = (nR_min_idx // shape[1], nR_min_idx % shape[1])
# print(nR_max_loc)
# Optimized: compute cross products directly without creating intermediate arrays
# For 2D: cross(a, b) = a[0]*b[1] - a[1]*b[0]
# Extract components once for reuse
r_x = self.r[:, :, 0]
r_y = self.r[:, :, 1]
grid_x = grid[:, :, 0]
grid_y = grid[:, :, 1]
r_p = grid - nR_max_loc
r_n = grid - nR_min_loc
# Compute cross products directly (faster than numpy.cross for 2D)
# cross(r_p, r) where r_p = grid - nR_max_loc
max_y, max_x = nR_max_loc[0], nR_max_loc[1]
cross_p = (grid_x - max_x) * r_y - (grid_y - max_y) * r_x
self.Stor = numpy.sum(numpy.cross(r_p, self.r)) + numpy.sum(numpy.cross(r_n, self.r))
# print("Stor is:", self.Stor)
# cross(r_n, r) where r_n = grid - nR_min_loc
min_y, min_x = nR_min_loc[0], nR_min_loc[1]
cross_n = (grid_x - min_x) * r_y - (grid_y - min_y) * r_x
output = numpy.array([self.Sn, self.Stx, self.Sty, self.Stor])
self.Stor = numpy.sum(cross_p) + numpy.sum(cross_n)
# Optimized: pre-allocate output array with explicit dtype
output = numpy.array([self.Sn, self.Stx, self.Sty, self.Stor], dtype=numpy.float32)
# Optimized: combine threshold checks (short-circuit evaluation)
flag = (self.Sn >= 300 and self.Stx >= 300 and self.Sty >= 300 and numpy.abs(self.Stor) >= 1000)
flag = False
if self.Sn >=300 and self.Stx >= 300 and self.Sty >= 300 and numpy.abs(self.Stor) >= 1000:
flag = True
# print(flag)
return flag, output
# return self.d, self.h, numpy.cross(r_p, self.r) , numpy.cross(r_n, self.r)