python - 使用 oct2py 进行矩阵乘法

Question

我正在使用 oct2py 通过 python 运行 Octave 函数。它正在工作，但当我尝试乘以 2 个矩阵时出现错误。我该怎么做才能解决这个问题？

这是一个示例 matlab 函数

%% MATLAB
function lol = jk2(arg1,arg2)
    arg1 = arg1;
    arg2 = arg2;
    lol = arg1*arg2;

end

这是调用函数的代码

import oct2py
from oct2py import octave
a=3
b=4
octave.call("/MATLAB/jk2.m",a,b) # this call works
a=np.array([[1,2],[3,4]])
b=np.array([[5,6],[1,2]])
octave.call("/MATLAB/jk2.m",a,b) # this call report an errors 

这是错误消息

>>> octave.call("/home/donbeo/Documents/MATLAB/jk2.m",a,b)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python2.7/dist-packages/oct2py-1.2.0-py2.7.egg/oct2py/session.py", line 210, in call
    resp = self._eval(cmd, verbose=verbose)
  File "/usr/local/lib/python2.7/dist-packages/oct2py-1.2.0-py2.7.egg/oct2py/session.py", line 350, in _eval
    return self._session.evaluate(cmds, verbose, log, self.logger)
  File "/usr/local/lib/python2.7/dist-packages/oct2py-1.2.0-py2.7.egg/oct2py/session.py", line 523, in evaluate
    raise Oct2PyError(msg)
oct2py.utils.Oct2PyError: Oct2Py tried to run:
"""
[a__] = jk2(A__, B__)
"""
Octave returned:
binary operator '*' not implemented for 'int64 matrix' by 'int64 matrix' operations
>>> 

Answer 1

这是一个 Python 和 Octave 之间的界限变得模糊的示例。 Numpy 将数组解释为整数类型(因为没有显式 float )，但 Octave 会将数组视为 double 。如果您在数组定义中的任何位置添加句点，它都会起作用。

修复(测试)示例:

from oct2py import octave
import numpy as np
a = np.array([[1, 2], [3, 4.]])  # notice the addition of the period
b = np.array([[5, 6], [1, 2], dtype=float])  # another way to specify floating point type
octave.call("/MATLAB/jk2.m", a, b)  # this call works just fine