cocopp does not load data correctly with "*"

[brockho]

When loading bbob data, the "*" does what it is supposed to, namely loading all matching data sets:

import cocopp 
dsl = cocopp.load('bbob/2016/*')
print(dsl)

gives

[DataSet(PSAaLmC-CMA-ES_Nishida on f1 2-D), DataSet(PSAaLmD-CMA-ES_Nishida on f1 2-D), DataSet(PSAaSmC-CMA-ES_Nishida on f1 2-D),
 DataSet(PSAaSmD-CMA-ES_Nishida on f1 2-D), DataSet(PSAaLmC-CMA-ES_Nishida on f2 2-D), DataSet(PSAaLmD-CMA-ES_Nishida on f2 2-D),
 DataSet(PSAaSmC-CMA-ES_Nishida on f2 2-D), DataSet(PSAaSmD-CMA-ES_Nishida on f2 2-D), DataSet(PSAaLmC-CMA-ES_Nishida on f3 2-D), 
[...]

ie. in this case, the data from four algorithms. If I do the same for the bbob-biobj suite, I only get data from a single algorithm (= the first in the matching list):

dsl = cocopp.load('bbob-biobj/2016/*')
print(dsl)

gives

[DataSet(DEMO_Tusar_bbob-biobj on f1 2-D), DataSet(DEMO_Tusar_bbob-biobj on f2 2-D), DataSet(DEMO_Tusar_bbob-biobj on f3 2-D),
 DataSet(DEMO_Tusar_bbob-biobj on f4 2-D), DataSet(DEMO_Tusar_bbob-biobj on f5 2-D), DataSet(DEMO_Tusar_bbob-biobj on f6 2-D),
 DataSet(DEMO_Tusar_bbob-biobj on f7 2-D), DataSet(DEMO_Tusar_bbob-biobj on f8 2-D), DataSet(DEMO_Tusar_bbob-biobj on f9 2-D),
 DataSet(DEMO_Tusar_bbob-biobj on f10 2-D), DataSet(DEMO_Tusar_bbob-biobj on f11 2-D), DataSet(DEMO_Tusar_bbob-biobj on f12 2-D),
 DataSet(DEMO_Tusar_bbob-biobj on f13 2-D), DataSet(DEMO_Tusar_bbob-biobj on f14 2-D), DataSet(DEMO_Tusar_bbob-biobj on f15 2-D), 
[...]

although there are many algorithms matching the string:

dsl = cocopp.load('bbob-biobj/2016/')
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[25], line 1
----> 1 dsl = cocopp.load('bbob-biobj/2016/')

File ~\AppData\Local\anaconda3\envs\coco-web\Lib\site-packages\cocopp\cococommands.py:234, in load(filename)
    222 def load(filename):
  
[...]

ValueError: 'bbob-biobj/2016/' has multiple matches in the data archive:
   bbob-biobj/2016/DEMO_Tusar_bbob-biobj.tgz
   bbob-biobj/2016/HMO-CMA-ES_Loshchilov_bbob-biobj.tgz
   bbob-biobj/2016/MAT-DIRECT_Al-Dujaili_bbob-biobj.tgz
   bbob-biobj/2016/MAT-SMS_Al-Dujaili_bbob-biobj.tgz
   bbob-biobj/2016/MO-DIRECT-HV-Rank_Wong_bbob-biobj.tgz
   bbob-biobj/2016/MO-DIRECT-ND_Wong_bbob-biobj.tgz
   bbob-biobj/2016/MO-DIRECT-Rank_Wong_bbob-biobj.tgz
   bbob-biobj/2016/NSGA-II-MATLAB_Auger_bbob-biobj.tgz
   bbob-biobj/2016/RANDOMSEARCH-100_Auger_bbob-biobj.tgz
   bbob-biobj/2016/RANDOMSEARCH-4_Auger_bbob-biobj.tgz
   bbob-biobj/2016/RANDOMSEARCH-5_Auger_bbob-biobj.tgz
   bbob-biobj/2016/RM-MEDA_Auger_bbob-biobj.tgz
   bbob-biobj/2016/RS-100.tgz
   bbob-biobj/2016/RS-4.tgz
   bbob-biobj/2016/SMS-EMOA-DE_Auger_bbob-biobj.tgz
   bbob-biobj/2016/SMS-EMOA-PM_Auger_bbob-biobj.tgz
   bbob-biobj/2016/UP-MO-CMA-ES_Krause_bbob-biobj.tgz
Either pick a single match, or use the `get_all` or
`get_first` method, or use the ! (first) or * (all)
marker and try again.

Interestingly, it doesn't work either for bbob-largescale data, but for bbob-constrained it works...

[brockho]

Actually, the problem is not only with the "*" operator, also when trying to load two algorithms, only the first is loaded:

print(cocopp.load('K-RVEA MOTPE'))

gives

[DataSet(K-RVEA_Tanabe on f1 2-D), DataSet(K-RVEA_Tanabe on f2 2-D), DataSet(K-RVEA_Tanabe on f3 2-D), 
\DataSet(K-RVEA_Tanabe on f4 2-D), DataSet(K-RVEA_Tanabe on f5 2-D), [...]
]

while

print(cocopp.load('MOTPE K-RVEA'))

gives

[DataSet(MOTPE_Tanabe on f1 2-D), DataSet(MOTPE_Tanabe on f2 2-D), DataSet(MOTPE_Tanabe on f3 2-D), 
DataSet(MOTPE_Tanabe on f4 2-D), DataSet(MOTPE_Tanabe on f5 2-D), DataSet(MOTPE_Tanabe on f6 2-D), 
[...]
DataSet(MOTPE_Tanabe on f53 40-D), DataSet(MOTPE_Tanabe on f54 40-D), DataSet(MOTPE_Tanabe on f55 40-D)]

[brockho]

Probably the last thing for now, before we have to really find the issue: now also bbob-constrained* does not work anymore for me (=loads only the first entry). Maybe the issue is that whenever the data is already downloaded, it behaves differently.

[nikohansen]

This may be related to code introduced handing testbeds that broke the load command, as stated in the documentation of load.

[currently broken when further used within `cocopp`, see `load2`] Create a :py:class:`DataSetList` instance from a file or folder.

Input argument filename can be a single :file:`info` file name, a
single pickle filename or a folder name. In the latter case, the
folder is browsed recursively for :file:`info` or :file:`pickle`
files.

Details: due to newly implemented side effects when data are read in,
the returned data set list may not work anymore when used with plotting
functions of the `cocopp` module, see also `load2`.

cocopp.load2 works as expected in the first case.

As a first quick fix we should issue a warning that load is not working as expected anymore and superseded by load2.