Numpy functions on Syncopy Data objects

[tensionhead]

Using functions like np.sum oder np.mean on spy.Analogdata does nothing, yet there is still a in place selection message emitted.
Reproducing code example:

mock_up_data = [np.arange(6, dtype=float).reshape(3,2) for _ in range(10)]
my_data = spy.AnalogData(mock_up_data)
my_data = np.sum(my_data)
my_data.show()

will output the unchanged original data:

array([[0., 1.],
       [2., 3.],
       [4., 5.],
       [0., 1.],
       ...

[pantaray]

With recent fixes in SpyArithmetic the output of np.sum(my_data) is simply:

In [4]: np.sum(my_data)
Out[4]: 
Syncopy AnalogData object with fields

            cfg : dictionary with keys ''
        channel : [2] element <class 'numpy.ndarray'>
      container : None
           data : 10 trials of length 3.0 defined on [30 x 2] float64 Dataset of size 0.00 MB
         dimord : time by channel
       filename : /home/fuertingers/.spy/spy_a319_6b9c3f40.analog
           mode : r+
     sampleinfo : [10 x 2] element <class 'numpy.ndarray'>
     samplerate : 1.0
      selection : None
            tag : None
           time : 10 element list
      trialinfo : [10 x 0] element <class 'numpy.ndarray'>
         trials : 10 element iterable

Use `.log` to see object history

which is consistent with, e.g.,

In [6]: np.sum([1])
Out[6]: 1

However, adding my_data to itself produces

In [5]: new_data = np.sum([my_data, my_data])
Syncopy <+> WARNING: Found existing in-place selection in operand. Shapes and trial counts of base and operand objects have to match up!
Syncopy <+> INFO: Allocating 0.00 MB AnalogData object on disk for result of + operation
100% |██████████████████████████████████| 10/10 [00:00<00:00]
In [8]: new_data.data[()]
Out[8]: 
array([[ 0.,  2.],
       [ 4.,  6.],
       [ 8., 10.],
       [ 0.,  2.],
       [ 4.,  6.],
       [ 8., 10.],
       [ 0.,  2.],
       [ 4.,  6.],
       [ 8., 10.],
       [ 0.,  2.],
       [ 4.,  6.],
       [ 8., 10.],
       [ 0.,  2.],
       [ 4.,  6.],
       [ 8., 10.],
       [ 0.,  2.],
       [ 4.,  6.],
       [ 8., 10.],
       [ 0.,  2.],
       [ 4.,  6.],
       [ 8., 10.],
       [ 0.,  2.],
       [ 4.,  6.],
       [ 8., 10.],
       [ 0.,  2.],
       [ 4.,  6.],
       [ 8., 10.],
       [ 0.,  2.],
       [ 4.,  6.],
       [ 8., 10.]])

Same behavior for np.prod et al. So I guess this is consistent now. I think we close this one @tensionhead ?

[tensionhead]

Phew, ok this has gone under completely.. yet is related to #376 where this topic will be generally addressed, so I am closing this.