Export nm_solver to external computing sources

[aliyyanurr]

Hello,

I have been using PHOEBE 2.3 and have no problem exporting solvers and samplers for massive computations, using 2.3 code

b.export_solver(kind='nelder_mead', maxiter=100000, solution='nm_sol', script_fname='nm_TIC93316328.py'

However, when I migrated to 2.4, I encountered difficulties exporting the solver using the newest tutorial.

Here is my error message when using crimpl

[54]: b.run_compute(use_server='mylocalserver', detach=True)
# crimpl: ls -d /home2/s30323301/crimpl/crimpl-job-*
# crimpl: ls /home2/s30323301/crimpl/crimpl-job-2024.09.04-15.49.59/*
# crimpl: mkdir -p /home2/s30323301/crimpl
# crimpl: cp exportpath.sh /home2/s30323301/crimpl/
# crimpl: source /home2/s30323301/crimpl/exportpath.sh; conda -V
# crimpl: source /home2/s30323301/crimpl/exportpath.sh; conda info --envs
# crimpl: ls -d /home2/s30323301/crimpl/crimpl-envs/*
# crimpl: ls -d /home2/s30323301/crimpl/crimpl-job-2024.09.04-15.49.59/crimpl-envs/*
# crimpl: source /home2/s30323301/crimpl/exportpath.sh; mkdir -p /home2/s30323301/crimpl/crimpl-job-2024.09.04-15.49.59
# crimpl: cp crimpl_submit_script.sh /home2/s30323301/_zaxZIaFeVdqzKrovpXtuOqghOjwpqX.py /home2/s30323301/crimpl/crimpl-job-2024.09.04-15.49.59/
# crimpl: source /home2/s30323301/crimpl/exportpath.sh; echo '_zaxZIaFeVdqzKrovpXtuOqghOjwpqX.py' >> /home2/s30323301/crimpl/crimpl-job-2024.09.04-15.49.59/crimpl-input-files.list
# crimpl: source /home2/s30323301/crimpl/exportpath.sh; echo 'default' > /home2/s30323301/crimpl/crimpl-job-2024.09.04-15.49.59/crimpl-conda-environment
# crimpl (detached): source /home2/s30323301/crimpl/exportpath.sh; cd /home2/s30323301/crimpl/crimpl-job-2024.09.04-15.49.59; chmod +x ./crimpl_submit_script.sh; nohup bash ./crimpl_submit_script.sh 2> ./crimpl_submit_script.sh.err & echo $! > crimpl-nohup.pid
/bin/sh: 1: source: not found

[54]: <Parameter: detached_job=unknown | keys: description, value, job_name, uniqueid, readonly, advanced, latexfmt>
-------------------------------------------------------
Primary job  terminated normally, but 1 process returned
a non-zero exit code.. Per user-direction, the job has been aborted.
-------------------------------------------------------
--------------------------------------------------------------------------
mpirun detected that one or more processes exited with non-zero status, thus causing
the job to be terminated. The first process to do so was:

  Process name: [[42364,1],3]
  Exit code:    1
--------------------------------------------------------------------------

And here is my error message when using export script and submit manually

[66]: script_fname, results_fname = b.export_solver('nm_TIC93316328.py', use_server='none')

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
/tmp/ipykernel_21171/3986932395.py in ?()
----> 1 script_fname, results_fname = b.export_compute('export_compute_to_server.py',
      2                                                use_server='none')

~/anaconda3/envs/phoebe24/lib/python3.12/site-packages/phoebe/frontend/bundle.py in ?(self, script_fname, out_fname, compute, model, dataset, pause, log_level, import_from_older, **kwargs)
  11527 
  11528         """
  11529         use_server = kwargs.get('use_server', kwargs.get('server', None))
  11530 
> 11531         model, computes, datasets, do_create_fig_params, changed_params, overwrite_ps, kwargs = self._prepare_compute(compute, model, dataset, from_export=True, **kwargs)
  11532 
  11533         if use_server is None:
  11534             for compute in computes:

~/anaconda3/envs/phoebe24/lib/python3.12/site-packages/phoebe/frontend/bundle.py in ?(self, compute, model, dataset, from_export, **kwargs)
  11202 
  11203         if model in self.models and kwargs.get('overwrite', model=='latest'):
  11204             # NOTE: default (instead of detached_job=) is correct here
  11205             if self.get_value(qualifier='detached_job', model=model, context='model', default='loaded') not in ['loaded', 'error', 'killed']:
> 11206                 raise ValueError("model '{}' cannot be overwritten until it is complete and loaded (or killed).".format(model))
  11207             if model=='latest':
  11208                 logger.warning("overwriting model: {}".format(model))
  11209             else:

ValueError: model 'latest' cannot be overwritten until it is complete and loaded (or killed).

Is there anything I can do to tackle these errors?

For additional information: I run PHOEBE in my institution's computing facility (external server), not in my local computer (I don't know if there is another set up that I have to do for this condition)

[kecnry]

This can sometimes happen if the job gets interrupted. This is a weird case though since its complaining about the model when you're exporting a solver 🤔 . If you delete the "latest" model (b.remove_model('latest')), does the export_solver command then work?

[aliyyanurr]

Now it works, thank you for your suggestion! I don't know what happened before, seems like I made an error when changing constraints, parameters, or creating models before exporting the NM solver...