covariance_matrix_plot runcard fails when executing procs_chi2_table

[andreab1997]

Running the covariance_matrix_plot runcard fails when entering in the function groups_chi2_table

nnpdf/validphys2/src/validphys/results.py

Line 666 in 9840fc6

def groups_chi2_table(groups_data, pdf, groups_chi2, groups_each_dataset_chi2):

This could be linked to the wrong shape of groups_each_dataset_chi2 as constructed by

nnpdf/validphys2/src/validphys/results.py

Line 663 in 9840fc6

groups_each_dataset_chi2 = collect("each_dataset_chi2", ("group_dataset_inputs_by_metadata",))

[scarlehoff]

Note that interestingly enough, groups_data is grouped correctly instead:

nnpdf/validphys2/src/validphys/results.py

Line 803 in 9840fc6

groups_chi2 = collect(

despise using exactly the same collect syntax! I guess we need @Zaharid for this one.

[andreab1997]

Runcard to reproduce the error:
covariance_matrix_plots_7pt_reduced.txt

[Zaharid]

I tried reproducing this and got the error

Traceback (most recent call last):
  File "/home/zk261/data2/miniconda3/envs/nnpdfdev/bin/validphys", line 33, in <module>
    sys.exit(load_entry_point('validphys', 'console_scripts', 'validphys')())
  File "/mhome/damtp/s/zk261/nnpdf/validphys2/src/validphys/scripts/main.py", line 10, in main
    vp.main()
  File "/mhome/damtp/s/zk261/nnpdf/validphys2/src/validphys/app.py", line 161, in main
    a.main()
  File "/home/zk261/data2/miniconda3/envs/nnpdfdev/lib/python3.9/site-packages/reportengine/app.py", line 395, in main
    self.run()
  File "/mhome/damtp/s/zk261/nnpdf/validphys2/src/validphys/app.py", line 156, in run
    super().run()
  File "/home/zk261/data2/miniconda3/envs/nnpdfdev/lib/python3.9/site-packages/reportengine/app.py", line 380, in run
    rb.execute_sequential()
  File "/home/zk261/data2/miniconda3/envs/nnpdfdev/lib/python3.9/site-packages/reportengine/resourcebuilder.py", line 166, in execute_sequential
    result = self.get_result(callspec.function,
  File "/home/zk261/data2/miniconda3/envs/nnpdfdev/lib/python3.9/site-packages/reportengine/resourcebuilder.py", line 175, in get_result
    fres =  function(**kwdict)
  File "/mhome/damtp/s/zk261/nnpdf/validphys2/src/validphys/results.py", line 685, in procs_chi2_table
    return groups_chi2_table(procs_data, pdf, procs_chi2, each_dataset_chi2)
  File "/mhome/damtp/s/zk261/nnpdf/validphys2/src/validphys/results.py", line 677, in groups_chi2_table
    stats = chi2_stats(dsres)
  File "/mhome/damtp/s/zk261/nnpdf/validphys2/src/validphys/results.py", line 785, in chi2_stats
    rep_data, central_result, npoints = abs_chi2_data
TypeError: cannot unpack non-iterable MCStats object

Is that it?

[scarlehoff]

Yes. If you inspect the error you will see that the reason is that the grouping of the input arguments seems to be different (despite having similar collect)

[Zaharid]

There might be a problem related to the fact that collect always returns a list, so when you use the results you might need to take the first element.

[scarlehoff]

The problem is that the list contains all dataset in the runcard, with no grouping, so if you do that the dataset you use has no connection with the ones you get on the other two arguments.

If you do that the code "works" because zips stop when one of the lists finishes and by construction the subgroups will always be smaller than the full dataset list.

[Zaharid]

I see that groups_chi2_table takes groups_each_dataset_chi2 while procs_chi2_table takes each_dataset_chi2, which seems wrong.

[Zaharid]

After I fix that one as

cf4e3eb

there are still problems in various other places, e.g.

  File "/mhome/damtp/s/zk261/nnpdf/validphys2/src/validphys/theorycovariance/output.py", line 32, in procs_chi2_table_theory
    return groups_chi2_table(
  File "/mhome/damtp/s/zk261/nnpdf/validphys2/src/validphys/results.py", line 675, in groups_chi2_table
    stats = chi2_stats(dsres)
  File "/mhome/damtp/s/zk261/nnpdf/validphys2/src/validphys/results.py", line 781, in chi2_stats
    rep_data, central_result, npoints = abs_chi2_data
TypeError: cannot unpack non-iterable MCStats object

[Zaharid]

FWIW what broke these things was
79604bd