Motopy

Motopy is a tool used to translate Matlab/Octave code TO PYthon.

点此查看中文说明.

Introduce

Motopy is a powerful tool used to translate Matlab/Octave code to PYthon. In the process of translation, the python statement generated by mopy will be executed to ensure the translated correctness of subsequent code. For example, the following Matlab/Octave code:

a = ones(1, 3);
b = a';
c = a * b;

will be translated to python code:

import numpy as np
a = np.ones((1, 3))
b = a.T
c = a @ b

The type of the variables a and b are array. So the third statement c = a * b will be translate to: c = a @ b.

install

Please use pip install motopy:

pip install motopy

Quick Start

Motopy is very easy to use. First please prepare your Matlab/Octave files, put the script file and the function files with extetion ".m" in a folder, and ensure that your Matlab/Octave script can be run without exception. And meet Code Preprocessing.Here's a simple example:

• Create a folder named "demo".

• In the "demo" folder, create two ".m" files whose names are "fun.m" and "func_test.m". Input the folowing text:

   % file: func.m
   function s = func(a, b)
       s = sqrt(a.^2 + b.^2);
   end
  

   % file: func_test.m
   a = 3;
   b = 4;
   s = func(a, b);
   disp(s)
  

• In the "demo" folder, create a python script file. Import motopy and use motopy.make() translate your mfile:

  import motopy
  motopy.make(entry_basename='func_test')

  The entryfile parameter specifies the mfile script to be translated(***!!Note that there is no extension!! ***).

  You can also execute the above code directly from the python command line. Ensure that the current directory is the "demo" folder.

Specify the input/output folder

The python script file may not be placed in the same folder as the mfile, and the input mfile and the output python file may be in a different folder. You can use the input_path parameter to specify the location of the input mfile and the output_path parameter to specify the output path of the generated python file.

import motopy

motopy.make(
    entry_basename='<the script filename without extension(*.m)>',
    input_path='<the input path of *.m files>', 
    output_path='<the output path of *.py files>' 
)

Specify replaced function

If you have already translated a function, you can specify an replaced function to that function using the replaced_functions argument in the motopy.make() function.

import motopy
motopy.make(
    entry_basename='func_test', # no extension
    input_path='The path that input ".m" files loaded', 
    output_path='The path that output ".py" files stored',
    replaced_functions={
        'func': ('func', 'func') # 
    }
)

The replaced_functions parameter is a dictionary, the key is the function name that appears in the mfile, and the value is a tuple (module_name, function_name). In the example above, the func function file will not be translated again.

When do you use replaced_functions:

• The .py file generated by motopy has been modified manually, and do not want motopy to regenerate it.

• The .m function that motpy does not support translated yet. You can implement it by yourself.

Code Annotation

The code line started with "%%>" in mfile is a python statment. And will be inserted to generated python file. And the next statment will be skiped. For example, the mfile:

%%> print('this is a code annotation.')
disp('this statment will be skiped.')

will be translated to:

print('this is a code annotation.')

Output Log

By default, motopy generate a log file named "motopy.log" under the output_path folder. You can use the 'logging_file' parameter to specify the output location and name of the log file. Using logging_level set log level: WARN|INFO|DEBUG

import motopy
motopy.make(.., logging_level=motopy.DEBUG, ..)

Indent

By default, the generated .py file uses 4 Spaces for indentation. You can use indent parameter specifies the number of Spaces required for indentation.

Code Preprocessing

The translation will failed if your Matlab/Octave code don't satisfy the folowing requirements:

• Do not use blank spaces to separate elements in arrays and cells. The following code will make failed:

  a = [1 2 3; 4 5 6];
  c = {1 2 'abc'};
  

• The first function name in a function file must be same as the filename.

• Arrays and cells should be defined before used and allocated enough space. The following code will make failed:

  for k=1:5
      A(k) = 2*k; % The variable A is not defined before used.
  end
  

  A = []; % The variable A has not enough space.
  for k=1:5
      A(k) = 2*k; % the size of variable A will grow in iteration.
  end
  

• Do not use "[]" to define empty array. The folowing mcode will translate failed:

  A = [];
  for k=1:5
      B= rand(2,2);
      A = [A;B];
  end
  disp(A)
  

  The expression [A;B] translate failed. because the empty array A with size of 0x0 cannot concatenate with array B with size of 2x2.

  An easy way to resolve this problem is to define the array A as an empty array with size of 0x2:

  A = zeros(0,2);
  for k=1:5
      B= rand(2,2);
      A = [A;B];
  end
  disp(A)
  

Implemented Translation

Creation of Array, Matrix and Cell

Matlab/Octave	Python	Note
a = [1,2,3,4]	a = np.array([1, 2, 3, 4])	The array in matlab will be translated to np.array
a = [1,2;3,4]	a = np.array([[1, 2], [3, 4]])
a = [1;2;3;4]	a = np.array([[1], [2], [3], [4]])
C = {1,2,3;'4',[5,6],{7,8,9}}	C = [[1, 2, 3], ['4', np.array([5, 6]), [7, 8, 9]]]	The cell in matlab will be translated to list
r1 = 1:10;	r1 = arange(1, 11)	low_bound:high_bound in matlab will be translated to arange(low_bound, high_boud + 1)
N = 10; r2 = 1:N;	N = 10 r2 = arange(1, N + 1)
zeros(3)	np.zeros((3, 3))
zeros(2,3)	np.zeros((2, 3))
ones(3)	np.ones((3, 3))
ones((2, 3))	np.ones((2, 3))
C = cell(2,3)	C = [[None for _c in range(3)] for _r in range(2)]

Slice of Array, Matrix and Cell

Matlab/Octave	Python	Note
a(1,1)	a[0, 0]	The value of index will decrease 1, if the index is a number.
a(1,:)	a[0, :]
a(:,1)	a[:, 0]
a(1, 1:2)	a[0, 0:2]
a(1:2, 1)	a[0:2, 0]
a(1,2:end)	a[0, 1:]
m = 1; n = 1; a(m, n*2)	m = 1 n = 1 a[m - 1, n * 2 - 1]	The index will be replaced with index - 1, if index is a variable.

Functions

Matlab/Octave	Python	Note
abs	np.abs
acos	np.arccos
asin	np.arcsin
atan	np.arctan
[y,Fs] = audioread(filename)	Fs, y = wavfile.read(filename)
ceil	np.ceil
cos	np.cos
diag	np.diag
d = dir(name)	d = [{'name':e.name, 'folder':e.path, 'isdir':e.is_dir()} for e in scandir(name)]
disp	print
eye	np.eye
exp	np.exp
fft	np.fft
fix	np.fix
floor	np.floor
fprintf
ifft	np.ifft
inv	linalg.inv
linspace	np.linspace
S = load('data.mat')	S = loadmat('data.mat')	the Variable S is a dict
A = load('data.txt')	A = np.loadtxt('data.txt')	the file "data.txt" is a ASCII data.
load('data.mat')	_mat = loadmat('data.mat'); a = _mat['a']; b = _mat['b']	assume there are two variable a and b in "data.mat"
load('data.txt')	data = np.loadtxt('data.txt')	the file "data.txt" is a ASCII data.
log	np.log
log10	np.log10
log2	np.log2
mod	np.mod
ndims	np.ndim
numel	np.size
pinv	linalg.pinv
rand	random.rand
rank	linalg.matrix_rank
round	np.round
sin	np.sin
sort	np.sort
sprintf('%d%s',a, b)	f'{a}{b}'
sqrt	np.sqrt
s = strcat(s1,...,sN)	s = ''.join([s1,...,sN])
unique	np.unique

Change Log

See changelog.md for more information.