使用f2py在Python上更快地进行计算 [英] Using f2py for faster calculation on Python

问题描述

我正在使用python进行序列比对项目，而python for循环太慢了.

I'm working on a sequence alignment project with python, and python for loop is too slow.

因此，我决定使用 f2py .我对fortran不太了解，所以我坚持下面的观点.

So, I decided to use f2py. I don't know much about fortran, so I'm stuck to the point below.

有两个名为"column"的序列和"row"，其类型为 np.array

There are two sequence named 'column', and 'row' whose type is np.array

例如:

column = ['A', 'T', 'G', 'C']
row = ['A', 'A', 'C', 'C'] 

我为 Needleman-Wunsch 算法创建了一个矩阵，并且给两个序列打分(列，行).

I created a matrix for the Needleman-Wunsch algorithm, and I scored two sequences (column, row).

    import numpy as np
    column = np.array(list('ATGC'))
    row = np.array(list('AACC'))
    matrix = np.zeros((len(column) + 1, len(row) + 1), dtype='int')

    for i in range(1, len(column)+1):
        self.matrix[i][0] = -1 * i

    for j in range(1, len(row)+1):
        self.matrix[0][j] = -1 * j

    matchCheck = 0

    for i in range(1, len(column) + 1):
        for j in range(1, len(row) + 1):
            if column[i-1] == row[j-1]:
                matchCheck = 1 
            else:
                matchCheck = -1 
            top = matrix[i-1][j] + -1
            left = matrix[i][j-1] + -1
            top_left = matrix[i-1][j-1] + matchCheck
            matrix[i][j] = max(top, left, top_left)

我想从fortran获得一些帮助以加快计算速度，所以我用fortran编写了代码.

I wanted to get some help from fortran for faster calculation, so I wrote a code with fortran.

subroutine needlemanWunsch(matrix, column, row, cc, rr, new_matrix)
integer, intent(in) :: cc, rr
character, intent(in) :: column(0:cc-1), row(0:rr-1)
integer, intent(in) :: matrix(0:cc, 0:rr)
integer, intent(out) :: new_matrix(0:cc, 0:rr)
integer :: matchcheck, top, left, top_left

do i = 1, cc
    new_matrix(i, 0) = -1 * i
end do

do j = 1, rr
    new_matrix(i, 0) = -1 * j
end do

do k = 1, cc
    do l = 1, rr
        if (column(i-1).EQ.row(j-1)) then
            matchcheck = 1
        else
            matchcheck = -1 
        
        top = matrix(i-1, j) + inDel
        left = matrix(i, j-1) + inDel
        top_left = matrix(i-1, j-1) + matchCheck
        new_matrix(i, j) = max(top, left, top_left)
        end if 
    end do
end do 
return
end subroutine

然后，我用f2py转换了此fortran代码，并使用此代码将其导入了python.

Then, I converted this fortran code with f2py, and imported it on python with this code.

    import numpy as np
    column = np.array(list('ATGC'))
    row = np.array(list('AACC'))
    matrix = np.zeros((len(column) + 1, len(row) + 1), dtype='int')
    
    # import my fortran code 
    matrix = algorithm.needlemanwunsch(matrix, column, row, cc, rr)

每当我尝试导入fortran代码时
它崩溃了...

whenever I tried to import the fortran code
it crasheds...

推荐答案

以下情况适用于我.

文件 neeldemanWunsch.f90

subroutine needlemanWunsch(matrix, column, row, cc, rr, new_matrix)
  integer, intent(in) :: cc, rr
  character, intent(in) :: column(0:cc-1), row(0:rr-1)
  integer, intent(in) :: matrix(0:cc, 0:rr)
  integer, intent(out) :: new_matrix(0:cc, 0:rr)
  integer :: matchcheck, top, left, top_left

  do i = 1, cc
      new_matrix(i, 0) = -1 * i
  end do

  do j = 1, rr
      new_matrix(i, 0) = -1 * j
  end do

  do k = 1, cc
      do l = 1, rr
          if (column(i-1).EQ.row(j-1)) then
              matchcheck = 1
          else
              matchcheck = -1

          top = matrix(i-1, j) + inDel
          left = matrix(i, j-1) + inDel
          top_left = matrix(i-1, j-1) + matchCheck
          new_matrix(i, j) = max(top, left, top_left)
          end if
      end do
  end do
  return
end subroutine

通过 f2py

$ f2py -c needlemanWunsch.f90 -m needlemanWunsch

导入python文件 needlemanWunsch.py​​ .这是您的错误出处！您需要导入已编译的模块，请参见下面的示例.

Importing into python file needlemanWunsch.py. This is where your error comes from! You need to import the compiled module see example below.

import needlemanWunsch         # THIS IS MISSING IN YOUR CODE!!
import numpy as np

# create matrix
_column = ['A', 'T', 'G', 'C']
_row = ['A', 'A', 'C', 'C']
column = np.array(list(_column))
row = np.array(list(_row))
cc = len(column)
rr = len(column)
matrix = np.zeros((len(column) + 1, len(row) + 1), dtype='int')

# import my fortran code
matrix = needlemanWunsch.needlemanwunsch(matrix, column, row, cc, rr)

print(matrix)

输出为

$ python needlemanWunsch.py 
[[ 0 -4  0  0  0]
 [-1  0  0  0  0]
 [-2  0  0  0  0]
 [-3  0  0  0  0]
 [-4  0  0  0  0]]

这篇关于使用f2py在Python上更快地进行计算的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！