Error "cannot find context for 'fork'"

[mlrennerfr]

Hi,
I'm trying to run the example codes of the jupiter notebooks but I obtain the error in the title of this post when I ask to run the cell with the line
responses = twostep_protocol.run(cell_model=simple_cell, param_values=default_params, sim=nrn)
Is it a problem of the version of Python? I'm running Anaconda with Python 3.9 and NEURON 8.0. I've already managed to run the codes without any problem some time ago in other computer (Python 2.7) so this error confuses me.
Thank you in advance for your help!

[DrTaDa]

Hello,
It seems to be a windows related error.
I have tried to solve the issue in the branch https://github.com/BlueBrain/BluePyOpt/tree/win32_compatibility
However, I do not have a windows machine at hand to check if it works.
Could you let me know if it solves your issue ?

[mlrennerfr]

Hello,

Thank you for your quick answer! unfortunately it did not work, unless I did something wrong. I've just open the jupiter notebook for the "simple cell" from this new folder of BluePyOpt... or should I install BluePyOpt again in python?

Anyway, here is the full error message:

---

ValueError Traceback (most recent call last)
~\AppData\Local\Temp\ipykernel_5632\804922019.py in
1 default_params = {'gnabar_hh': 0.1, 'gkbar_hh': 0.03}
----> 2 responses = twostep_protocol.run(cell_model=simple_cell, param_values=default_params, sim=nrn)

~\AppData\Roaming\Python\Python39\site-packages\bluepyopt\ephys\protocols.py in run(self, cell_model, param_values, sim, isolate, timeout)
86 # Try/except added for backward compatibility
87 try:
---> 88 response = protocol.run(
89 cell_model=cell_model,
90 param_values=param_values,

~\AppData\Roaming\Python\Python39\site-packages\bluepyopt\ephys\protocols.py in inner(self, cell_model, param_values, **kwargs)
186 mech.deterministic = True
187
--> 188 responses = func(self, cell_model, param_values, **kwargs)
189
190 if self.deterministic:

~\AppData\Roaming\Python\Python39\site-packages\bluepyopt\ephys\protocols.py in run(self, cell_model, param_values, sim, isolate, timeout)
266 # Default context for python>=3.8 on macos is spawn.
267 # Spwan context would reset NEURON properties, such as dt.
--> 268 multiprocessing_context = multiprocessing.get_context('fork')
269
270 if timeout is not None:

C:\tools\Anaconda3\lib\multiprocessing\context.py in get_context(self, method)
237 return self._actual_context
238 else:
--> 239 return super().get_context(method)
240
241 def set_start_method(self, method, force=False):

C:\tools\Anaconda3\lib\multiprocessing\context.py in get_context(self, method)
191 ctx = _concrete_contexts[method]
192 except KeyError:
--> 193 raise ValueError('cannot find context for %r' % method) from None
194 ctx._check_available()
195 return ctx

ValueError: cannot find context for 'fork'

Thank you again!

[mlrennerfr]

Solved! :)
I did not realized that I was using a computer in which Anaconda was installed for a single user!
Installing it for all users solved this issue.

[DrTaDa]

Great !

[DrTaDa]

Did it with or without the fix ?

[DrTaDa]

Great ! Since it works for you, I will assume that the branch win32_compatibility works as well.

[MatheusHofner]

I'm having this issue now, still in the same tutorial:

`FileNotFoundError Traceback (most recent call last)
Cell In[10], line 5
1 import pickle
2 # pickle.dump(results, open('results.pkl', 'w'))
3 # results = pickle.load(open('results.pkl')
----> 5 cp = pickle.load(open('checkpoints/checkpoint.pkl'))
6 results = (cp['population'],
7 cp['halloffame'],
8 cp['logbook'],
9 cp['history'])
12 pop, hall_of_fame, logs, hist = results

File C:\ProgramData\anaconda3\lib\site-packages\IPython\core\interactiveshell.py:284, in _modified_open(file, *args, **kwargs)
277 if file in {0, 1, 2}:
278 raise ValueError(
279 f"IPython won't let you open fd={file} by default "
280 "as it is likely to crash IPython. If you know what you are doing, "
281 "you can use builtins' open."
282 )
--> 284 return io_open(file, *args, **kwargs)

FileNotFoundError: [Errno 2] No such file or directory: 'checkpoints/checkpoint.pkl'`

[DrTaDa]

If you wish to generate a pickle checkpoint file for the run, you have to provide a filepath for it, for example, in the jupyter notebook, you can replace the line

final_pop, hall_of_fame, logs, hist = optimisation.run(max_ngen=5)

by

final_pop, hall_of_fame, logs, hist = optimisation.run(max_ngen=5, cp_filename="checkpoints/checkpoint.pkl")

[MatheusHofner]

So the line on tutorial that creates the filepath was comment. I uncomment that, and now I got:

FileNotFoundError Traceback (most recent call last)
Cell In[15], line 6
1 optimisation = bpop.optimisations.DEAPOptimisation(
2 evaluator=cell_evaluator,
3 offspring_size = 100,
4 seed=1)
----> 6 results = optimisation.run(max_ngen=10, cp_filename='checkpoints/checkpoint.pkl')

File C:\ProgramData\anaconda3\lib\site-packages\bluepyopt\deapext\optimisations.py:309, in DEAPOptimisation.run(self, max_ngen, offspring_size, continue_cp, cp_filename, cp_frequency, cp_period, parent_population, terminator)
306 if hasattr(self.evaluator, "param_names"):
307 param_names = self.evaluator.param_names
--> 309 pop, hof, log, history = algorithms.eaAlphaMuPlusLambdaCheckpoint(
310 pop,
311 self.toolbox,
312 offspring_size,
313 self.cxpb,
314 self.mutpb,
315 max_ngen,
316 stats=stats,
317 halloffame=self.hof,
318 cp_frequency=cp_frequency,
319 cp_period=None,
320 continue_cp=continue_cp,
321 cp_filename=cp_filename,
322 terminator=terminator,
323 param_names=param_names)
325 # Update hall of fame
326 self.hof = hof

File C:\ProgramData\anaconda3\lib\site-packages\bluepyopt\deapext\algorithms.py:194, in eaAlphaMuPlusLambdaCheckpoint(population, toolbox, mu, cxpb, mutpb, ngen, stats, halloffame, cp_frequency, cp_period, cp_filename, continue_cp, terminator, param_names)
183 if (cp_filename and cp_frequency and
184 gen % cp_frequency == 0 and
185 (cp_period is None or time.time() - time_last_save > cp_period)):
186 cp = dict(population=population,
187 generation=gen,
188 parents=parents,
(...)
192 rndstate=random.getstate(),
193 param_names=param_names)
--> 194 pickle.dump(cp, open(cp_filename_tmp, "wb"))
195 if os.path.isfile(cp_filename_tmp):
196 shutil.copy(cp_filename_tmp, cp_filename)

FileNotFoundError: [Errno 2] No such file or directory: 'checkpoints/checkpoint.pkl.tmp'

[DrTaDa]

Was the file checkpoints/checkpoint.pkl.tmp generated ?
If not, it can be a permission issue.