Problem with fitness optimizer.

[mariuszslabicki]

When i try to run code from examples (http://esa.github.io/pygmo/tutorials/analysis_unconstrained_problem.html)

prob = problem.kur(4)
inspector = util.analysis(prob, 500, method='sobol', output_to_file=False)
inspector.f_distribution()
inspector.f_linearity_convexity()
inspector.f_regression(degree=[1,1,2,2,3,3],interaction=[False,True,False,True,False,True])
inspector.f_sensitivity()
inspector.levelset(threshold=45, k_tune=2)
inspector.local_search(clusters_to_show=3, scatter_plot_dimensions=[1,2], algo=algorithm.gsl_nm(), decomposition_method='weighted', weights='uniform')
inspector.f_correlation()

I've got an error:

Traceback (most recent call last):
File "/home/mariusz/Dropbox/SON_research/simulations/comnet_local/main.py", line 309, in
inspector = util.analysis(prob, 500, method='sobol', output_to_file=False)
File "/usr/local/lib/python3.4/dist-packages/PyGMO/util/_analysis.py", line 60, in init
if isinstance(input_object, core._core.population):
NameError: name 'core' is not defined

[mariuszslabicki]

Yup, i found a solution.

At the top of the file: _analysis.py i had to add (below the line from __future...)

from PyGMO import *

Maybe it can be done nicer, but it works.

[darioizzo]

as I said its a namespace problem ... import the whole PyGMO its a bit of an overkill .. maybe something like from PyGMO import core would also work?

Thanks for spotting this, I will post a definitive solution for the next release.

[mariuszslabicki]

Probably you are right, i did not check it.