Python calls MATLAB built-in function griddedInterpolant(), how do I get the value of the query point?

In fact, I need to use the 4D spline interpolation function in the experiment, but python has no API to call.The following is a simple one-dimensional example, I want to know how to get the value of the query point xq. Thanks a lot!

#---------------------------------------python code----------------------------------------#

import numpy as np

import matlab.engine

eng = matlab.engine.start_matlab()

x = np.array([1, 2, 3, 4, 5])

v = np.array([1.2, 0.6, 0.9, 4.5, 7.9])

x_new = matlab.double(x.tolist())

v_new = matlab.double(v.tolist())

F = eng.griddedInterpolant(x_new, v_new)

xq = np.array([2.5, 2.7, 3.6])

xq = matlab.double(xq.tolist())

vq = F(xq)

print(vq)

eng.quit() # quit Matlab engine

#--------------------------------------error message --------------------------------#

Traceback (most recent call last):

File "D:/Data/GitHub/CAE_CSI_mutillayer/test.py", line 84, in <module>

vq = F(xq)

TypeError: 'matlab.object' object is not callable

Answer 1

[Madheswaran]

Hi @xaofeng he

The problem you are facing is because MATLAB objects (like griddedInterpolant) cannot be called directly in Python. To solve this, you can use the following approach:

import numpy as np
import matlab.engine

eng = matlab.engine.start_matlab()

x = np.array([1, 2, 3, 4, 5])
v = np.array([1.2, 0.6, 0.9, 4.5, 7.9])

x_new = matlab.double(x.tolist())
v_new = matlab.double(v.tolist())

xq = np.array([2.5, 2.7, 3.6])
xq_new = matlab.double(xq.tolist())

eng.workspace['x'] = x_new
eng.workspace['v'] = v_new
eng.workspace['xq'] = xq_new

eng.eval('F = griddedInterpolant(x, v);', nargout=0)
vq = eng.eval('F(xq)', nargout=1)
print(vq)

eng.quit()

This code creates the interpolant and evaluates it by placing variables in the MATLAB workspace and using MATLAB's eval function to execute the commands.

You could also create a dedicated MATLAB function if you would like to avoid the use of 'eval' function. Create a file named 'interpolate_data.m' in your working directory,

% interpolate_data.m
function vq = interpolate_data(x, v, xq)
    % Create interpolant
    F = griddedInterpolant(x, v);
    % Evaluate at query points
    vq = F(xq);
end

and in python,

import numpy as np
import matlab.engine

eng = matlab.engine.start_matlab()

x = np.array([1, 2, 3, 4, 5])
v = np.array([1.2, 0.6, 0.9, 4.5, 7.9])

x_new = matlab.double(x.tolist())
v_new = matlab.double(v.tolist())
xq_new = matlab.double(np.array([2.5, 2.7, 3.6]).tolist())

vq = eng.interpolate_data(x_new, v_new, xq_new)
print(vq)

eng.quit()

This approach creates a dedicated MATLAB function that handles the interpolation, making the cleaer code.

For more information on using MATLAB with Python, refer to the following documentation: https://www.mathworks.com/help/matlab/matlab-engine-for-python.html

I hope this helps!