## examples/fitting/doc_example3/fit1.lar
run('ReadData.lar')


# create a parameter group for the fit:
params = group(amp1 = param(0.5, min=0, max=1),
               amp2 = param(0.5, min=0, max=1),
               amp3 = param(0.0, min=0, max=1),
               amp4 = param(0.0, min=0, max=1),
               amp5 = param(0.0, min=0, max=1),
               amp6 = param(0.0, min=0, max=1))

# define objective function for fit residual
def sum_standards(pars, data):
    return (pars.amp1*data.s1 + pars.amp2*data.s2 + pars.amp3*data.s3 +
            pars.amp4*data.s4 + pars.amp5*data.s5 + pars.amp6*data.s6 )
#enddef

def resid(pars, data):
    return (data.mu - sum_standards(pars, data))/data.eps
#enddef

# set model here: Amp1 is free,
# Amp2 is constrained to be 1 - Amp1

params.amp1.vary = True
params.amp2.expr = '1 - amp1'

# set uncertainty in data that we'll use to scale the returned residual
data.eps  = 0.001

# perform fit
result = minimize(resid, params, args=(data,))
fit = sum_standards(params, data)
print fit_report(result)


plot(data.energy, data.mu,   label='data', color='blue', marker='+',
     markersize=5, show_legend=True, legend_loc='cr', new=True,
     xlabel='Energy (eV)', title='Sum S1 and S2 to match data',
     xmin=11900, xmax=12000)
plot(data.energy, fit,  label='final', color='red')

plot(data.energy, 10*(data.mu-fit),  label='resid(x10)', color='black')

## end of examples/fitting/doc_example3/fit1.lar