Numpy 逐块减少操作

Numpy blockwise reduce operations

我认为自己是一位经验丰富的 numpy 用户,但我无法找到以下问题的解决方案。假设有以下数组:

# sorted array of times
t = numpy.cumsum(numpy.random.random(size = 100))
#  some values associated with the times
x = numpy.random.random(size=100)
# some indices into the time/data array
indices = numpy.cumsum(numpy.random.randint(low = 1, high=10,size = 20)) 
indices = indices[indices <90] # respect size of 100
if len(indices) % 2: # make number of indices even
    indices = indices[:-1]

# select some starting and end indices
istart = indices[0::2]
iend   = indices[1::2]

我现在想要的是在给定 istartiend 表示的间隔的情况下减少值数组 x。即

# e.g. use max reduce, I'll probably also need mean and stdv
what_i_want = numpy.array([numpy.max(x[is:ie]) for is,ie in zip(istart,iend)])

我已经用谷歌搜索了很多,但我所能找到的只是通过 stride_tricks 进行的分块操作,它只允许常规块。如果不执行 python 循环,我无法找到解决方案:-( 在我的实际应用程序中,数组要大得多,性能也很重要,所以我暂时使用 numba.jit

我是否缺少任何能够做到这一点的 numpy 函数?

你可以使用numpy.r_
像这样:

what_i_want = np.array([np.max(x[np.r_[ib:ie]]) for ib,ie in zip(istart,iend)])

你看过ufunc.reduceat了吗?使用 np.maximum,您可以执行以下操作:

>>> np.maximum.reduceat(x, indices)

沿着切片 x[indices[i]:indices[i+1]] 产生最大值。要得到你想要的(x[indices[2i]:indices[2i+1]),你可以做

>>> np.maximum.reduceat(x, indices)[::2]

如果您不介意 x[inidices[2i-1]:indices[2i]] 的额外计算。这会产生以下结果:

>>> numpy.array([numpy.max(x[ib:ie]) for ib,ie in zip(istart,iend)])
array([ 0.60265618,  0.97866485,  0.78869449,  0.79371198,  0.15463711,
        0.72413702,  0.97669218,  0.86605981])

>>> np.maximum.reduceat(x, indices)[::2]
array([ 0.60265618,  0.97866485,  0.78869449,  0.79371198,  0.15463711,
        0.72413702,  0.97669218,  0.86605981])

(非numpy解决方案,使用astropy)

import numpy as np
from astropy.nddata.utils import block_reduce
data = np.arange(16).reshape(4, 4)
block_reduce(data, 2)    

将转换:

array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11],
       [12, 13, 14, 15]])

至:

array([[10, 18],
       [42, 50]])

有关更多示例,请参阅 this