使用 f2py 在 Python 上进行更快的计算
Using f2py for faster calculation on Python
我正在使用 python 进行序列比对项目,python for 循环太慢了。
所以,我决定使用f2py
。我对fortran不太了解,所以我坚持下面的观点。
有两个序列名为'column'和'row',类型为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,我对两个序列(列、行)进行了评分。
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那里得到一些帮助来加快计算速度,所以我用fortran写了一个代码。
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 转换了这个 fortran 代码,并用这个代码在 python 上导入了它。
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...
以下适用于我的情况。
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
编译
$ f2py -c needlemanWunsch.f90 -m needlemanWunsch
正在导入 python 文件 needlemanWunsch.py
。
这就是您的错误来源!
您需要导入已编译的模块,请参见下面的示例。
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]]
我正在使用 python 进行序列比对项目,python for 循环太慢了。
所以,我决定使用f2py
。我对fortran不太了解,所以我坚持下面的观点。
有两个序列名为'column'和'row',类型为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,我对两个序列(列、行)进行了评分。
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那里得到一些帮助来加快计算速度,所以我用fortran写了一个代码。
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 转换了这个 fortran 代码,并用这个代码在 python 上导入了它。
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...
以下适用于我的情况。
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
正在导入 python 文件 needlemanWunsch.py
。
这就是您的错误来源!
您需要导入已编译的模块,请参见下面的示例。
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]]