Python:元素明智除法运算符错误

Question

我想知道是否有更好的方法在 python 中执行元素除法运算符。下面的代码假设用 B1 行执行除法 A1，用 B2 行执行除法 A2，因此我的预期输出只有两行。然而，划分部分是A1与B1、A1与B2、A2与B1、A2与B2。谁能帮我吗？

二进制文件使用 1000,0100,0010,0001 表示 A、C、G、T。 除法文件有四列，每列 A、C、G、T，因此获得的值 之前必须相应地划分。

代码

import numpy as np
from numpy import genfromtxt
import csv
csvfile = open('output.csv', 'wb')
writer = csv.writer(csvfile)

#open csv file into arrays
with open('binary.csv') as actg:
    actg=actg.readlines()
    with open('single.csv') as single:
        single=single.readlines()
        with open('division.csv') as division:
            division=division.readlines()

            # Converting binary line and single line into 3 rows and 4 columns 
            # binary values using reshape
            for line in actg:
                myarray = np.fromstring(line, dtype=float, sep=',')                
                myarray = myarray.reshape((-1, 3, 4))
                for line2 in single:                    
                    single1 = np.fromstring(line2, dtype=float, sep=',')
                    single1 = single1.reshape((-1, 4))
                    # This division is in 2 rows and 4 column: first column 
                    # represents 1000, 2nd-0100, 3rd-0010, 4th-0001 in the
                    # binary.csv. Therefore the division part where 1000's
                    # value should be divided by 1st column, 0010 should be
                    # divided by 3rd column value
                    for line1 in division:
                        division1 = np.fromstring(line1, dtype=float, sep=',')
                        m=np.asmatrix(division1)
                        m=np.array(m)
                        res2 = (single1[np.newaxis,:,:] / m[:,np.newaxis,:] * myarray).sum(axis=-1)                        
                        print(res2)
                        writer.writerow(res2)


csvfile.close()

二进制.csv

0,1,0,0,1,0,0,0,0,0,0,1
0,0,1,0,1,0,0,0,1,0,0,0

single.csv:

0.28,0.22,0.23,0.27,0.12,0.29,0.34,0.21,0.44,0.56,0.51,0.65

division.csv

0.4,0.5,0.7,0.1
0.2,0.8,0.9,0.3

预期输出

 0.44,0.3,6.5
 0.26,0.6,2.2

实际输出

0.44,0.3,6.5
0.275,0.6,2.16666667
0.32857143,0.3,1.1       
0.25555556,0.6,2.2       

错误说明

让分割文件如下:

A,B,C,D
E,F,G,H

令单次和二进制计算结果如下:

1,3,4
2,2,1

让数字1,2,3,4分配给位置A,B,C,D和下一行E,F,G,H

1/A,3/C,4/D
2/F,2/F,1/E

其中 1 除以 A，3 除以 C，依此类推。基本上这就是代码可以做的事情。不幸的是，除法部分恰好像前面描述的那样。 221 与 BBC 一起运行，134 与 EGH 一起运行，因此输出有 4 行，这不是我想要的。

Answer 1

我不知道这是否是您正在寻找的东西，但这里有一个简短的方法来获得(我认为)您想要的东西:

import numpy as np

binary = np.genfromtxt('binary.csv', delimiter = ',').reshape((2, 3, 4))
single = np.genfromtxt('single.csv', delimiter = ',').reshape((1, 3, 4))
divisi = np.genfromtxt('division.csv', delimiter = ',').reshape((2, 1, 4))

print(np.sum(single / divisi * binary, axis = -1))

输出:

[[ 0.44        0.3         6.5       ]
 [ 0.25555556  0.6         2.2       ]]