Fit with custom parametric model fails

[stestoll]

import deerlab as dl
import numpy as np
import matplotlib.pyplot as plt

t = np.linspace(-0.2,3,201)
r = np.linspace(1.5,6,201)
K0 = dl.dipolarkernel(t,r)

def model(par):
    center, width, lam, conc, amp = par
    B = dl.bg_hom3d(t,conc)
    P = dl.dd_gauss(r,[center, width])
    V = amp*((1-lam)+lam*K0@P)*B
    return V

parref = [3.5,0.2,0.4,100,10]
Vref = model(parref)
V = Vref + dl.whitegaussnoise(t,np.max(Vref)*0.02)

par0 = parref
lb = [2, 0.05, 0.01,  10, 1]
ub = [5,   1,    1, 1000, 30]
fit = dl.fitparamodel(V,model,par0,lb,ub)

gives

c:\DeerLab\deerlab\fitparamodel.py:223: UserWarning: The fitted value of parameter #4, is at the lower bound (1.0).
  warnings.warn('The fitted value of parameter #{}, is at the lower bound ({}).'.format(p,lb[p]))

[luisfabib]

This is because the scale is being fitted as well automatically, but you are including an amplitude parameter anyway. If one disables the automatic fit of the scale, the fit works

fit = dl.fitparamodel(V,model,par0,lb,ub,rescale=False)

So there is not really a bug in the function.

[stestoll]

I see, a user error from my side!

What about renaming the rescale keword parameter to fitscale or autoscale? These names would be more consistent with its function.