Bug in stacked weights computation

[combet]

Running the demo_mock_ensemble notebook, the output of cell 11 has the values of the n_src column in the W_l column. I think (but need to double check) the latter should correspond to the sum of the individual weights w_ls computed in cell 6.

I don't know if the problem has always been there or has been introduced in a recent change. But this is a bug and needs to be looked at (I noticed it while implementing things in TXPipe...)

[combet]

OK - it was not actually a bug but simply a missing use_weights options in the notebook, that I also forgot when adding code to TXPipe. All good I think