matlab runtime error

[lmn369]

I and Mostafa were trying to initialize stemmus-scope on Linux668 using Jupyter notebook, but encountered this error.
Do you have any idea how to solve this problem?

[BSchilperoort]

Do you have any idea how to solve this problem?

It seems like you're running the MATLAB process locally (not in a container). To debug this try to manually start the process from the command line, and pass the config file there. That would allow you to see the full matlab error.

See the docs here; https://ecoextreml.github.io/STEMMUS_SCOPE/run_model/#run-the-model-with-matlab-runtime

[MostafaGomaa93]

Hi Bart, we actually tried to run the model in the terminal using ./STEMMUS_SCOPE/exe/STEMMUS_SCOPE config_file.txt and it runs successfully

[lmn369]

@BSchilperoort Hi bart, we also managed to delete the cache folder, but the error is still there. Do you have any further suggestions?

[BSchilperoort]

Hm. what if you put the code in a .py file, and execute that in the command line? Maybe there's some permission problem.

[lmn369]

@BSchilperoort Hi bart, same error using a .py file

[BSchilperoort]

It seems like it cannot correctly find the cache dir, and you don't have root access so it fails.

Maybe try the following at the top of the python file

import os
os.environ['MCR_CACHE_ROOT'] = "/home/lim1/mcr_cache"

and create that folder in your home dir

[lmn369]

@BSchilperoort I created the mcr_cache folder under /home/lim1/, and added the codes at the top of the python file, still the same error

[BSchilperoort]

Turns out that that environment variable is not passed in the python code;

STEMMUS_SCOPE_Processing/PyStemmusScope/bmi/local_process.py

Lines 103 to 114 in 33adb9e

	env = {
	"LD_LIBRARY_PATH": lib_path,
	"MATLAB_LOG_DIR": str(config["OutputPath"]),
	}
	self.process = subprocess.Popen(
	args,
	stdin=subprocess.PIPE,
	stdout=subprocess.PIPE,
	bufsize=0,
	env=env,
	)

I think something might be going wrong because the subprocess using the wrong shell, which doesn't having the permissions set up, or it's only because of the environment variable.

You could try modifying the pystemmusscope code locally (making sure it's installed in editable mode or reinstalled in your environment):

        env = {
            "LD_LIBRARY_PATH": lib_path,
            "MATLAB_LOG_DIR": str(config["OutputPath"]),
            "MCR_CACHE_ROOT": os.getenv("MCR_CACHE_ROOT")
        }


        self.process = subprocess.Popen(
            args,
            stdin=subprocess.PIPE,
            stdout=subprocess.PIPE,
            bufsize=0,
            env=env,
        )

but I am not sure this will fix it.

[lmn369]

@BSchilperoort After adding this code "MCR_CACHE_ROOT": os.getenv("MCR_CACHE_ROOT"), problem solved! Thanks for your suggestion!

[BSchilperoort]

I'll reopen this issue so we can fix the code on github

[SarahAlidoost]

fixed by #118