Fix flow-related functions

[pvillacorta]

This PR is intended to solve some problems with flow experiments. Essentially:

• We need to correct how the interpolation of the reset flags is performed. The get_mask function was incorrectly implemented (there was an unnecessary cumsum, which in fact caused the reset flag to be 1 at less time points than it should).
• We need to ensure that, during the discretization of the sequence, time points corresponding specifically to spin jumps are included. Otherwise, we might miss a jump if no evaluation time instant falls within the interval where the reset bit should equal 1.