Establish conventions for when steps begin and end

[WardLT]

Battery times sere data feature many discontinuities that affect how many statistics are computed. For example, the current can change rapidly between a charge and rest periods. Whether one assumes a step function, linear curve, or something else changes the capacity calculation and any derived quantity (e.g., Coulomb efficiency). We should make a decision about how we should treat these periods between steps.

[WardLT]

My vote is to assume the new behavior begins at the first measurement the next step, and to set the values between the last measurement of the current step and the the next step beginning using linear extrapolation.