pass inputs from python to matlab and return output back to python

I have a python flask restful API to accept web requests in JSON format.

{
"imageurl":"http://cf067b.medialib.glogster.com/media/ea/ea717410f781940999f24dadaeb0b547dc9aa1289dc17b98ee4d134cec74e73c/t-shirt-in-modal-stretch-png.png"
  "itemlist": [item1, item2, item3]
}

I want to pass the contents of imageurl and itemlist as inputs to a Matlab function. I used subprocess to pass parameters to Matlab function. Here is how far I got:

@app.route('/testMatlab',methods = ['POST'])
def testMatlab():
url=request.json['imageurl']
taglist=request.json.get('taglist',[])
solve_it(url,taglist)
return jsonify(out)

def solve_it(arg1,arg2):
# Modify this code to run your matlab code
proc=subprocess.call(["matlab -nosplash -nodisplay -wait -r \"myFunction(\'%s\',\'%s\')\"" % (arg1, arg2)],shell=True, stdin=subprocess.PIPE, stout=subprocess.PIPE);
# Read result from file
out, err = proc.communicate()
print(out)
Right now myFunction.m simply returns taglist itself.

How can I return the output of myFunction as a string in JSON. Suppose the results of myFunction is out, I want to pass this results in JSON back to web request, e.g.

return jsonify(out)

Answer 1

[Bo Li]

In R2014b, the MATLAB Engine for Python provides a Python package for Python to call MATLAB functions:

http://www.mathworks.com/help/matlab/matlab-engine-for-python.html

Using Python Engine, you still need to convert the result into JSON types.