Can I use MATLAB Engine API for Python in a Docker container with an individual license?

I created a Docker container with a R2023b image and installed the MATLAB Engine API for Python. I can use my individual license to start a MATLAB session in the Docker container, but when I attempt to start a MATLAB session from Python I get the following error: "matlab.engine.EngineError: Unable to launch Simple server: Unable to launch /opt/matlab/R2023a/bin/matlab".

 

>>> import matlab.engine
>>> eng=matlab.engine.start_matlab()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python3.8/dist-packages/matlab/engine/__init__.py", line 128, in start_matlab
    eng = future.result()
  File "/usr/local/lib/python3.8/dist-packages/matlab/engine/futureresult.py", line 67, in result
    return self.__future.result(timeout)
  File "/usr/local/lib/python3.8/dist-packages/matlab/engine/matlabfuture.py", line 87, in result
    handle = pythonengine.getMATLAB(self._future)
matlab.engine.EngineError: Unable to launch Simple server: Unable to launch /opt/matlab/R2023a/bin/matlab
because: Timed out reading transport byte from transport

Can I use MATLAB Engine API for Python in a Docker container with my individual license?

Answer 1

[MathWorks Support Team]

With an individual license you need to interactively log in to start a new MATLAB session in a Docker container. For this reason, you cannot start a new MATLAB session from Python using "matlab.engine.start_matlab()" with an individual license. In order to start a new MATLAB session from Python in a Docker container you need a license server, which allows you to start MATLAB non-interactively. However, with an individual license, you can interactively start a new MATLAB session and then connect to the existing MATLAB session from Python with the Engine API. Below is an example workflow for doing so.

 

1) Build and run a container with MATLAB and Python Engine installed (see Dockerfile example, below).

$ docker build -t <image-name> .

2) Start the Dockerfile and interactively log in to start the MATLAB session.

$ docker run -it --rm --shm-size=512M <image-name>

3) In MATLAB, run the following command to make the MATLAB instance a shared engine.

>> matlab.engine.shareEngine

3) Use "docker exec" from a terminal to start a bash shell in the same container.

$ docker exec -it <container-name> /bin/bash

4) Start Python in the bash shell and use "connect_matlab()" to connect to the existing MATLAB session.

>>> import matlab.engine
>>> eng = matlab.engine.connect_matlab()
>>> print(eng.sqrt(4.0))
2.0

Example Dockerfile

 

ARG MATLAB_RELEASE=R2023b

# Build MATLAB image
FROM mathworks/matlab:$MATLAB_RELEASE

# Declare global
ARG MATLAB_RELEASE

# Install MATLAB Engine API for Python
RUN /bin/sh -c 'cd /opt/matlab/$MATLAB_RELEASE/extern/engines/python && sudo python setup.py install'

For more information see:

https://www.mathworks.com/help/cloudcenter/ug/matlab-container-on-docker-hub.html

https://github.com/mathworks-ref-arch/matlab-dockerfile/blob/main/alternates/building-on-matlab-docker-image/Dockerfile