Readd pykilosort test

[chyumin]

This PR should be merged only after SpikeInterface/spikeinterface#890

[bendichter]

['Warning! The recording is already filtered, but combinato filter is enabled\n', 'Traceback (most recent call last):\n File "/tmp/pytest-of-runner/pytest-0/test_combinato0/in_container_sorter_script.py", line 17, in \n sorting = run_sorter_local(\n File "/home/runner/.local/lib/python3.8/site-packages/spikeinterface/sorters/runsorter.py", line 161, in run_sorter_local\n SorterClass.setup_recording(recording, output_folder, verbose=verbose)\n File "/home/runner/.local/lib/python3.8/site-packages/spikeinterface/sorters/basesorter.py", line 196, in setup_recording\n cls._setup_recording(recording, output_folder, sorter_params, verbose)\n File "/home/runner/.local/lib/python3.8/site-packages/spikeinterface/sorters/combinato/combinato.py", line 138, in setup_recording\n write_to_h5_dataset_format(recording, dataset_path='/data', segment_index=0,\n File "/home/runner/.local/lib/python3.8/site-packages/spikeinterface/core/core_tools.py", line 560, in write_to_h5_dataset_format\n traces = recording.get_traces(segment_index=segment_index,\n File "/home/runner/.local/lib/python3.8/site-packages/spikeinterface/core/baserecording.py", line 144, in get_traces\n raise ValueError('This recording do not support return_scaled=True (need gain_to_uV and offset'\nValueError: This recording do not support return_scaled=True (need gain_to_uV and offset_to_uV properties)\n']

Is it possible to use return_scaled=False?

[chyumin]

@bendichter @alejoe91 This error is happening after this change:
https://github.com/SpikeInterface/spikeinterface/pull/908/files#diff-91a9c642bdff905864e37e0b04e25b5e11b05e0a33169df99863d18d8ac4d703R140

[chyumin]

Converting this PR to draft because pykilosort test is still failing

[chyumin]

This PR should fix pykilosort test error.