可以参考官方的说明文档：

MATLAB Engine API的使用文档：

原材料：

1、MATLAB 2015a  32位的

2、Python 2.7.13    32位的

安装：

1、运行cmd，切换到matlab的目录下：

C:\Program Files (x86)\MATLAB\MATLAB Production Server\R2015a\extern\engines\python

由于这个文件夹是在C盘的，安装时有可能会遇到写权限的问题。如果出现write permission，可以通过右击该文件夹，选择 属性->安全->编辑，给当前用户赋予完全控制的权限。

> python setup.py install

即可完成安装。

2、python中调用matlab的API示例：

#coding=utf-8

import matlab.engine

if __name__ == ‘__main__‘:

eng = matlab.engine.start_matlab(‘MATLAB_R2015a‘)

a = eng.sqrt(4.0)

print type(a),a

eng.quit()

pass

在Python中创建MATLAB数组

1、创建1XN数组

import matlab.engine

A = matlab.int8([1,2,3,4,5])

print type(A),A.size,A

#输出：

(1, 5) [[1,2,3,4,5]]

Attribute or MethodPurpose

size

Size of array returned as a tuple

reshape(size)

Reshape array as specified by sequence size

2、创建多维数组

import matlab.engine

A = matlab.double([[1,2,3,4,5], [6,7,8,9,10]])

print(A)

3、在Python中索引MATLAB数组

这里的MATLAB数组索引跟在MATLAB的IDE里面不一样，MATLAB是从1开始，而在Python中是从0开始索引

import matlab.engine

A = matlab.int8([1,2,3,4,5])

print(A[0])

#输出：

[1,2,3,4,5]

由于A是1X5的矩阵，A[0]就是[1,2,3,4,5]，如果要在A里面索引出4，则需要输入：

print A[0][3]

4、在Python中对MATLAB数组切片

这里语法跟Python中没多少差别，直接使用即可

import matlab.engine

A = matlab.int8([1,2,3,4,5])

print(A[0][1:4])

#输出：

[2,3,4]

切片赋值，也可以从一个MATLAB数组赋值到另一个MATLAB数组：

A = matlab.double([[1,2,3,4],[5,6,7,8]]);

A[0] = [10,20,30,40]

print(A)

#输出：

[[10.0,20.0,30.0,40.0],[5.0,6.0,7.0,8.0]]

A = matlab.int8([1,2,3,4,5,6,7,8]);

A[0][2:4] = [30,40]

A[0][6:8] = [70,80]

print(A)

#输出：

[[1,2,30,40,5,6,70,80]]

注意：

Note: Slicing MATLAB arrays behaves differently from slicing a Python list. Slicing a MATLAB array returns a view instead of a shallow copy.

Given a MATLAB array and a Python list with the same values, assigning a slice results in different results as shown by the following code.

A = matlab.int32([[1,2],[3,4],[5,6]])

L = [[1,2],[3,4],[5,6]]

A[0] = A[0][::-1]

L[0] = L[0][::-1]

print(A)

[[2,2],[3,4],[5,6]]

print(L)

[[2, 1], [3, 4], [5, 6]]

数组reshape

import matlab.engine

A = matlab.int8([1,2,3,4,5,6,7,8,9])

A.reshape((3,3))

print(A)

[[1,4,7],[2,5,8],[3,6,9]]

Python中MATLAB支持的数据类型：

matlab ClassConstructor Call in Python

matlab.double

matlab.double(initializer=None, size=None, is_complex=False)

matlab.single

matlab.single(initializer=None, size=None, is_complex=False)

matlab.int8

matlab.int8(initializer=None, size=None, is_complex=False)

matlab.int16

matlab.int16(initializer=None, size=None, is_complex=False)

matlab.int32

matlab.int32(initializer=None, size=None, is_complex=False)

matlab.int64[a]

matlab.int64(initializer=None, size=None, is_complex=False)

matlab.uint8

matlab.uint8(initializer=None, size=None, is_complex=False)

matlab.uint16

matlab.uint16(initializer=None, size=None, is_complex=False)

matlab.uint32

matlab.uint32(initializer=None, size=None, is_complex=False)

matlab.uint64[b]

matlab.uint64(initializer=None, size=None, is_complex=False)

matlab.logical

matlab.logical(initializer=None, size=None)[c]

matlab.object

No constructor. When a function returns a handle to a MATLAB object, the engine returns a matlab.object to Python.

[a] In Python 2.7 on Windows, matlab.int64 is converted to int32 in MATLAB. Also, MATLAB cannot return an int64 array to Python.

[b] In Python 2.7 on Windows, matlab.uint64 is converted to uint32 in MATLAB. Also, MATLAB cannot return a uint64 array to Python.

[c] Logicals cannot be made into an array of complex numbers.

奇异值分解示例：

#coding=utf-8

import matlab.engine

from numpy import *

if __name__ == ‘__main__‘:

eng = matlab.engine.start_matlab(‘MATLAB_R2015a‘)

A = matlab.double([[1,2],[5,6]])

print type(A),A.size,A

print eng.eig(A)

eng.quit()

pass

Examples and How To

To start the MATLAB engine within a Python session, you first must install the engine API as a Python package.

By default, the installer builds the engine API for Python in the matlabroot\extern\engines\python folder. The installer installs the engine in the default Python folder.

Options for starting the MATLAB Engine for Python.

How to connect the MATLAB Engine for Python to a shared MATLAB session that is already running on your local machine.

How to return an output argument from a MATLAB function. How to read multiple outputs from a function. What to do when the MATLAB function does not return an output argument.

This example shows how to call the MATLAB sqrt function asynchronously from Python and retrieve the square root later.

This example shows how to call a MATLAB script to compute the area of a triangle from Python.

This example shows how to redirect standard output and standard error from a MATLAB function to Python StringIO objects.

This example shows how to add variables to the MATLAB engine workspace in Python.

This example shows how to create an object from a MATLAB handle class and call its methods in Python.

This example shows how to create a MATLAB array in Python and pass it as the input argument to the MATLAB sqrt function.

This example shows how to sort data about patients into lists of smokers and nonsmokers in Python and plot blood pressure readings for the patients with MATLAB.

From Python, you can access supporting documentation for all MATLAB functions.

Concepts

The MATLAB Engine API for Python provides a Python package named matlab that enables you to call MATLAB functions from Python.

What you need to write and build MATLAB engine applications.

When you pass Python data as input arguments to MATLAB functions, the MATLAB Engine for Python converts the data into equivalent MATLAB data types.

When MATLAB functions return output arguments, the MATLAB Engine API for Python converts the data into equivalent Python data types.

The matlab Python package provides array classes to represent arrays of MATLAB numeric types as Python variables so that MATLAB arrays can be passed between Python and MATLAB.

MATLAB stores all numeric values as double-precision floating point numbers by default.

Troubleshooting

The engine cannot start or connect to MATLAB on a remote machine.

When a MATLAB function raises an error, the MATLAB Engine for Python stops the function and catches the exception raised by MATLAB.