Naming of datasets

[nikohansen]

import cocopp
cocopp.archives.all.find(".*bob/.*BFGS")

['bbob/2009/BFGS_ros_noiseless.tgz',
 'bbob/2012/DE-BFGS_voglis_noiseless.tgz',
 'bbob/2012/PSO-BFGS_voglis_noiseless.tgz',
 'bbob/2014-others/BFGS-scipy-Baudis.tgz',
 'bbob/2014-others/L-BFGS-B-scipy-Baudis.tgz',
 'bbob/2018/BFGS-M-17.tgz',
 'bbob/2018/BFGS-P-09.tgz',
 'bbob/2018/BFGS-P-Instances.tgz',
 'bbob/2018/BFGS-P-range.tgz',
 'bbob/2018/BFGS-P-StPt.tgz']

I think we should get back to use the original naming convention which gives algorithm name and author of the data set.

[brockho]

I updated the names of the zip files on the remote server as well as the corresponding archive definition files. After a

import cocopp
cocopp.archives.update_all()

I get now with the above command (cocopp.archives.all.find(".*bob/.*BFGS")):

['bbob/2009/BFGS_ros_noiseless.tgz',
 'bbob/2012/DE-BFGS_voglis_noiseless.tgz',
 'bbob/2012/PSO-BFGS_voglis_noiseless.tgz',
 'bbob/2014-others/BFGS-scipy_Baudis.tgz',
 'bbob/2014-others/L-BFGS-B-scipy_Baudis.tgz',
 'bbob/2018/BFGS-M-17_Blelly.tgz',
 'bbob/2018/BFGS-P-09_Blelly.tgz',
 'bbob/2018/BFGS-P-Instances_Blelly.tgz',
 'bbob/2018/BFGS-P-StPt_Blelly.tgz',
 'bbob/2018/BFGS-P-range_Blelly.tgz',
 'bbob/2019/BFGS-scipy-2019_Varelas.tgz',
 'bbob/2019/L-BFGS-B-scipy-2019_Varelas.tgz']

As you can see, I left the _noiseless in the files that had them (same for noisy). Otherwise, the default naming convention is now

algorithm-name-with-potential-dashes_First-author-name-with-potential-dashes.tgz

Biobjective and largescale data sets have the suite name or largescale in the name as last part, such as 'DEMO_Tusar_bbob-biobj.tgz' or CMA_Varelas_largescale.tgz.