NoiseHandler not work in parallel_mode=True

[Landau1908]

Hi,
The code and outputs are as follows.
Regards

import numpy as np
import cma

def fun(x,a):
    return sum(x)

if __name__ == "__main__":    
    nprocesses = 1
    args = (0,)
    parallel = Parallel(fitness_function=fun, num_processes=nprocesses)
    es = cma.CMAEvolutionStrategy(np.ones(10), 1, options={'verb_disp_overwrite':1, 'maxfevals': 5e6})  #doctest: +ELLIPSIS
    nh = cma.NoiseHandler(es.N, maxevals=[1, 1, 200])

    with cma.optimization_tools.EvalParallel2(fitness_function=fun, number_of_processes=nprocesses) as p_objective:
          X, fit_vals = es.ask_and_eval(fun, evaluations=nh.evaluations, args=args, parallel_mode=True)
          es.tell(X, fit_vals)  # prepare for next iteration
          es.sigma *= nh(X, fit_vals, fun, es.ask, args=args)  # see method __call__
          es.countevals += nh.evaluations_just_done  # this is a hack, not important though
          es.logger.add(more_data = [nh.evaluations, nh.noiseS])  # add a data point
          es.disp()

(5_w,10)-aCMA-ES (mu_w=3.2,w_1=45%) in dimension 10 (seed=1037176, Mon Nov 18 17:29:57 2024)
Traceback (most recent call last):

  File E:\ProgramData\Lib\site-packages\spyder_kernels\py3compat.py:356 in compat_exec
    exec(code, globals, locals)

  File f:\git\test_cma_noisehandler_parallel.py:50
    X, fit_vals = es.ask_and_eval(fun, evaluations=nh.evaluations, args=args, parallel_mode=True)

  File E:\ProgramData\Lib\site-packages\cma\evolution_strategy.py:2526 in ask_and_eval
    x, f = X_first.pop(0), fit_first.pop(0)

AttributeError: 'numpy.ndarray' object has no attribute 'pop'

[nikohansen]

I am not 100% sure this is the reason for the bail, but p_objective is defined but never used, hence nothing is parallelized and, I think, fit_vals is computed as the scalar sum over all sample variables.

AFAICS, the fun argument to es.ask_and_eval must be p_objective.