Feature Request: Weighted Tailfit in 2.x

[hschnait]

Dear All

It would be nice to have the possibility of a weighted tail-fit in 2.x. (as mentioned in PR #482 ).
This should work together nicely with the fit_tail() and fit_hermitian_tail() functions.

As the error of Sigma increases with w_n, these two functions mainly fit noise and are rather unusable for my application in their current implementation.

[Wentzell]

Dear Hermann,

Thank you for the feature request! The PR you mention unfortunately got lost during the major
tail rework.

We agree that the capability for weighted tailfits is essential, in particular in view of
the on-going rework of the statistics functionality that is about to be merged into triqs/unstable.
We will let you know as soon as we start to move on this.

Best,

Nils