Open nikohansen opened 5 years ago
Interesting, with the latest code from the development branch, I get this error when trying to postprocess all BFGS algorithms:
ValueError: Data from more than one suites {'bbob-largescale', 'bbob-noisy', 'bbob'} cannot be post-processed together
The same holds true for trying to benchmark single- and multiobjective data sets:
In [6]: cocopp.main('bipop! NSGA!')
Post-processing (2+)
Using:
C:\Users\dimo\.cocopp\data-archives\coco.gforge.inria.fr\data-archive\bbob-noisy\2009\BIPOP-CMA-ES_hansen_noisy.tgz
C:\Users\dimo\.cocopp\data-archives\coco.gforge.inria.fr\data-archive\bbob-biobj\2016\NSGA-II-MATLAB_Auger_bbob-biobj.tgz
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-6-e5a5855e378b> in <module>
----> 1 cocopp.main('bipop! NSGA!')
~\Anaconda3\lib\site-packages\cocopp-2.3.1.31-py3.7.egg\cocopp\rungeneric.py in main(argv)
316 if len(suites) > 1:
317 raise ValueError("Data from more than one suites %s cannot "
--> 318 "be post-processed together" % str(suites))
319
320 if len(args) == 1 or '--include-single' in dict(opts):
ValueError: Data from more than one suites {'bbob-biobj', 'bbob-noisy'} cannot be post-processed together
Currently, the main routine processes the data from different suites with warnings like cocopp/pproc.py:2825: UserWarning: No data for algorithm exdata/random_search_on_bbob_budget0002xD in 2-D. Then only the data from the first suite show up.
This could be desired behavior to be able to process data from the biobjective suites together?
From the user's perspective the warning message should be more expressive.
See also #1900.