add transpose method to _Audio

[artpelling]

This PR adds a transpose method to all Signals, very similar to np.ndarray.transpose.

[f-brinkmann]

Thanks for the PR. Only a few comments on the docstring and some gneral points to consider/discuss

• If we specify axes for audio objects we have the concept of ignoring the last axis holding the time frequency bins. This means, that axis -1 in an audio object refers to axis -2 of the underlying numpy array _data. You can account for this by axes = [a-1 if a<0 else a for a in axes] and add a description for this in the docstring.
• I noted that the test are failing. Might be because you branched main. Can you branch develop instead?
• Once the code is approved we should add tests to tests/test_audio_signal.py, tests/test_audio_time_data.py, and tests/test_audio_frequency_data.py

[artpelling]

I've implemented your suggestions. Out of lazyness, I removed all assertions and simply transpose a mock array with identical ndim before deep copying instead of reverse engineering all rules. This will throw numpy's exceptions. Let me know if you like it or not.

• Once the code is approved we should add tests to tests/test_audio_signal.py, tests/test_audio_time_data.py, and tests/test_audio_frequency_data.py

Will do that once the code is agreed on.

[ahms5]

Thanks for the PR, very elegant, looks fine to me. for me just tests are remaining. sth for flake8 also failed besides #483

[artpelling]

@f-brinkmann This might be ready now.

[f-brinkmann]

Thanks, looks very good already. I only had minor comments on the docstring and testing left. Can you also add a brief test for the shorthand signal.T?

[f-brinkmann]

There seems to be an error in the tests now, but otherwise approved. Thanks for the effort.