add coverage to prototype tests CI

[pmeier]

In the past we abandoned codecov.io since hitting a specific coverage was never a target and the bot comments only added noise to PRs. Still, it would be useful to have some kind of coverage information in the CI. This is especially useful now that we are ramping up testing for the prototype namespaces.

This PR adds coverage as terminal output for the prototype test CI. This only adds a few lines of useful information at the test summary.

[pmeier]

Example output for the transforms tests:

Name                                                              Stmts   Miss  Cover   Missing
-----------------------------------------------------------------------------------------------
torchvision/prototype/transforms/__init__.py                         12      0   100%
torchvision/prototype/transforms/_augment.py                        228     50    78%   38, 112, 119, 133, 144, 174, 188, 229-231, 273-277, 312-331, 334-376
torchvision/prototype/transforms/_auto_augment.py                   189     15    92%   44-45, 48, 50, 144, 217-274, 292-294, 413, 461
torchvision/prototype/transforms/_color.py                          118     32    73%   42, 48-50, 97-102, 105-106, 116-126, 131-151
torchvision/prototype/transforms/_container.py                       62      5    92%   22-25, 33
torchvision/prototype/transforms/_deprecated.py                      46      0   100%
torchvision/prototype/transforms/_geometry.py                       483     18    96%   93, 96, 99, 136-137, 142-143, 187, 208, 446, 451, 561, 590, 664, 811-812, 873, 876
torchvision/prototype/transforms/_meta.py                            41     10    76%   14-17, 20-21, 50, 54, 71-72
torchvision/prototype/transforms/_misc.py                           105     22    79%   16, 29, 32-37, 58, 78, 85, 109, 145-146, 149, 156-157, 160-171, 174
torchvision/prototype/transforms/_presets.py                         42     42     0%   5-69
torchvision/prototype/transforms/_transform.py                       44      9    80%   45-55, 61
torchvision/prototype/transforms/_type_conversion.py                 42      4    90%   16, 34-37
torchvision/prototype/transforms/_utils.py                           67      3    96%   64, 66, 80
torchvision/prototype/transforms/functional/__init__.py               8      0   100%
torchvision/prototype/transforms/functional/_augment.py              18      0   100%
torchvision/prototype/transforms/functional/_color.py                91      6    93%   68, 76-81
torchvision/prototype/transforms/functional/_deprecated.py           30     17    43%   18-27, 31-52, 57-61, 65-69
torchvision/prototype/transforms/functional/_geometry.py            527     31    94%   124, 170, 183, 186, 189, 192, 195, 198, 207, 213, 216, 219, 475, 509-510, 526-527, 634, 661, 685, 777, 801, 887, 981, 1163, 1175-1176, 1195-1196, 1217, 1254
torchvision/prototype/transforms/functional/_meta.py                137     13    91%   33, 37-38, 47-48, 121, 151, 154, 183, 201-202, 205, 218
torchvision/prototype/transforms/functional/_misc.py                 45     12    73%   20, 31, 33, 36, 42, 44, 48, 51, 60-62, 71
torchvision/prototype/transforms/functional/_type_conversion.py      31      7    77%   15-18, 23-24, 34
-----------------------------------------------------------------------------------------------
TOTAL                                                              2366    296    87%

[YosuaMichael]

I think this is a good idea, thanks @pmeier!