[PyMcaMath] fix numpy.dot usage in PCA Tools

[vasole]

closes #1161

[vasole]

I am not satisfied with this fix. Everywhere in PyMca it is assumed the dot product returns a scalar.

Probably the issue is on the shape of the input arrays.

[vasole]

Now it is better. However, issue #1162 is found almost everywhere in PyMca.

[woutdenolf]

For reference:

tmpData.shape = 1, nChannels
...
numpy.dot(tmpData, eigenvectors[j])

is like

>>> numpy.zeros((1,1024)).dot(numpy.ones((1024,)))
array([0.])

which is to be expected

(1×n)(n)→(1)

So the solution is to squeeze the first dimension before doing the matrix multiplication

>>> numpy.zeros((1024,)).dot(numpy.ones((1024,)))
np.float64(0.0)

[woutdenolf]

Note that this could have been detected earlier if we run the tests with -W error

>>> import numpy
>>> arr = numpy.zeros((5,5))
>>> arr[0,0] = numpy.asarray([0])
<stdin>:1: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract a single element from your array before performing this operation. (Deprecated NumPy 1.25.)

This is why we always run with -W error. I'll make an issue for that.

[vasole]

I prefer reshape or ravel rather than a math operation. I guess it will be faster or at least use less memory.