Mind the boundary elements handling in `fill_fftseries` array of `TDIFlyGB`

The waveform generated by TDIFlyGB often needs to be placed into an array representing the analyzed frequency band. This is handled by fill_fftseries()

https://github.com/TriangleDataCenter/Triangle-Simulator/blob/bfaf43fd94274fa4628cb0cc97595f13aa584554/Triangle/TDIFly.py#L435

This involves repeated assignments of the boundary elements. Unfortunately, the numpy standard states:

For advanced assignments, there is in general no guarantee for the iteration order. This means that if an element is set more than once, it is not possible to predict the final result.

On my machine, the repeated assigned element in template_filled ends up taking the last corresponding element of data. This does not cause problem for the first element, but if the last element of data falls to the right side of template_filled, then last element of template_filled is incorrectly assigned that value (Causing a sharp drop in figure below).

This phenomenon may not lead to catastrophic errors in parameter estimation because the waveform at the boundary is weak enough. However, the user had better pay attention to this point, especially when the peak lies on the boundary of the analyzed band.

Start in Triangle/TDIFly.py at fill_fftseries() around line 435 and inspect how boundary indices are assigned with NumPy advanced indexing. Reproduce a case where data extends beyond template_filled, then verify that each boundary element receives the intended value without relying on repeated assignment order. Done means the sharp boundary drop is no longer produced for the reported case.