It's known that graph.R (post) is more accurate than pyserial_pyper.py (capture)

----------------------------
All this without propulsive:

Differences are:

1- End of phase is the same, but start of phase sometimes not.
a) When speed starts with negative values (maybe reduce_curve_by_speed is not working) then it match
b) when speed starts with positive values (reduce_curve... is working) then there are differences in start moment between the two programs

Problem seem to be on pyserial_pyper, that starts on last -1 (of previous curve) instead of on first +1


2.- FIXED!!  when phase start-end is the same in both all the values are the same except mean power

In graph.R 328-330 it's used:
if(eccon == "c")
    meanPower <- mean(kinematics$power)
but in pyserial_pyper.py 204:
myR.run('meanPower <- mean(abs(power))')

then mean power is ABS on concentric at pyserial_pyper, but this is not ok because when accel is lower than -9.81, power should be negative. If you don't want to register this data as negative, the best is to use "propulsive option", and all this negative power will be rejected.


-------------------
Check propulsive stuff (add in pyserial if it is not)

All is fixed now except propulsive stuff

It still was not ok

No in pyserial done the same method of smoothing than in graph.R

1) first remove at start with reduceCurveBySpeed
2) do speed and accel for curve once reducedCurveBySpeed
3) find propulsive (if appropiate)
4) remove at end with propulsive (if appropiate)

it works quite good, but seem sometimes is not very good doing the reduceCurveBySpeed

graph.R is ok because show same values on data written than on paint graph

99% fixed, small differences sometimes