Resampling algorithm could be more efficient for large datasets

[philipstarkey]

Right now the resampling algorithm is passed the entire trace for resampling even if it is taking just a subset during the resampling process. We should be able to slice the traces first before passing them in to the algorithm which I think would speed up the resampling algorithm when dealing with traces with a large umber of points (e.g. order 2e6)

[philipstarkey]

Note: I've logged this here, rather than labscript-suite/labscript-c-extensions as I think the slicing can be done in runviewer without needing to change the C extension.