Bug#1015169: RE: Bug#1015169: RFS: mir-eval/0.7-1 [ITP] -- Common metrics for common audio/music processing tasks

[Bastian Germann <bage@debian.org>]

Control: tags -1 moreinfo

The 700+ line patch is not acceptable as-is. A general note for patches: Please do not comment out but really remove 
lines that you do not want because commenting results in two patch lines (- and +), which makes it very hard to review a 
patch and get the idea of it.

When I run the tests without your patch, I can see
ERROR: test_input_output.test_load_tempo
----------------------------------------------------------------------
Traceback (most recent call last):
...
FileNotFoundError: [Errno 2] No such file or directory: 'data/tempo/ref01.lab'

======================================================================
ERROR: test_input_output.test_load_tempo_multiline
----------------------------------------------------------------------
...
FileNotFoundError: [Errno 2] No such file or directory: 'data/tempo/bad00.lab'

======================================================================
ERROR: test_input_output.test_load_tempo_badweight
----------------------------------------------------------------------
...
FileNotFoundError: [Errno 2] No such file or directory: 'data/tempo/bad01.lab'

======================================================================
ERROR: test_input_output.test_load_bad_tempi
----------------------------------------------------------------------
...
FileNotFoundError: [Errno 2] No such file or directory: 'data/tempo/bad02.lab'

I think, you can get these tests to run by running the tests with the current working directory set to tests/ in the 
source dir instead of .pybuild/cpython3_3.10_mir-eval/build.
If you cannot get pybuild to change to that dir please use override_dh_auto_test.

In https://github.com/craffel/mir_eval/issues/350, upstream also suggests that the tests should run generally.
The only tests that I expect to fail are the ones which use that old matplotlib feature.

Changing the directory will maybe also fix:

test_melody.test_to_cent_voicing
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/home/bage/mir-eval/.pybuild/cpython3_3.10_mir-eval/build/tests/test_melody.py", line 118, in test_to_cent_voicing
    ref_file = sorted(glob.glob(REF_GLOB))[0]
IndexError: list index out of range

which is the trigger for many other failing tests.