Repair functionality "findvalue"
The findvalue functionality could use some overhaul. I tried to use it to plot the voltage at which the sensor cells reach a certain current. I ran into several problems:
-
The finding algorithm operates on 1/value², and not on the z values themselves. So it propably was intended to work on CV data. Not a big problem, one can use the flag --notrafo to work on the z values themselves. However, this behaviour should be documented.
-
When I use a threshold which is larger than all z values, the algorithm returns NaN. Not sure if this makes sense. For IV curves, I'd rather put the maximum voltage there, because it would mean that the cell is stable at least up to that voltage.
-
The algorithm uses a spline for interpolation. This spline is very sensitive to outliers. In my test data I have several IV curves, where the very first point is ridiculously large. Is there a possibility to have the spline ignore points which would lead to a high local curvature?
-
The algorithm doesn't adhere to limits of the fit range, unlike the other fit functions. Such a functionality would be a viable workaround for point 3).