In [ ]:

    

%pylab notebook
%load_ext autoreload
%autoreload 2
from fitGui import FitGui


    

In [ ]:

    

def sfgn1(x, nr, phase, amplitude, pos, width):
    '''NR background and one lorenzian peak

    Parameters
    ---------- 
    x : array
        wavenumbers
    nr : Non Resonant background (amplitude)
    phase : Phase of the non resonant background
    amplitude : number
        Amplitude
    pos : number
    width : width of the lorenzian (FWHM)

    Returns
    -------
    array with the same shape as x
    '''

    # Non resonant part
    ChiNR = nr * (cos(phase) + 1j * sin(phase))

    # Resonent part
    ## Prepare parameters
    A = amplitude
    delta = pos - x
    gamma = width / 2
    
    ## The actural lorenzian distribution
    ChiR_i = A * gamma / (delta**2 + gamma**2)
    ChiR_r = A * delta / (delta**2 + gamma**2)
    ChiR = ChiR_r + 1j * ChiR_i

    # The physical Chi
    Chi = ChiR + ChiNR

    # Doing it this way seems to be the fastest
    return square(Chi.real) + square(Chi.imag)


    

In [ ]:

    

# Load some example data
data = load('./example_peak.npz')

# Data must be sorted. Otherwise roi will do shit.
x = data['wavenumber'][::-1]

# y_data for the fit
y = data['normalized'].squeeze()[::-1] * 1000 # 1000 so values are ~1. This helps the fitting algorithm

# Uncertaincy of the y_data for the fit
y_err = data['dnormalized'].squeeze()[::-1] * 1000

# fit function
fit_func = sfgn1

f = FitGui(x, y, fit_func, y_err)

# Set debug level for further output 0 to 3 are allowed levels
#f._debug = 3

fig, ax = subplots()
plt.ylim(0, 10)
plt.xlim(1400, 1800)
f()


    

In [ ]:

    

fig, ax = subplots()
f.chi2.show();


    

In [ ]:

    

import sfg2d
import pandas
import iminuit
import probfit


    

In [ ]:

    

urea8m = sfg2d.io.load_npz_to_Scan('/home/malte/MeasurmentData/Dolly1/2016/10/19/urea_002_v1_avg.npz')
urea7m = sfg2d.io.load_npz_to_Scan('/home/malte/MeasurmentData/Dolly1/2016/10/20/urea_002_03_avg.npz')
urea6m = sfg2d.io.load_npz_to_Scan('/home/malte/MeasurmentData/Dolly1/2016/10/20/urea_002_02_avg.npz')
urea4m = sfg2d.io.load_npz_to_Scan('/home/malte/MeasurmentData/Dolly1/2016/10/19/urea_002_01_v1_avg.npz')

data = urea8m 

# Region of Interest in coordinates of x
roi = slice(1540, 1660)

# x_data for the fit
x = data.df['normalized'][roi].index.get_values()

# y_data for the fit
y = data.df['normalized'][roi].get_values() * 1000

# Uncertaincy of the y_data for the fit
y_err = data.df['dnormalized'][roi].get_values() * 1000
g
# fit function
fit_func = sfg2d.utils.static.sfgn1

f = FitGui(x, y, fit_func, y_err)

# Set debug level for further output 0 to 3 are allowed levels
#f._debug = 3

fig, ax = subplots()
f()


    

In [ ]:

    

fig, ax = subplots()
f.chi2.show()


    

In [ ]:

    

urea8m = sfg2d.io.load_npz_to_Scan('/home/malte/MeasurmentData/Dolly1/2016/10/19/urea_002_v1_avg.npz')
urea7m = sfg2d.io.load_npz_to_Scan('/home/malte/MeasurmentData/Dolly1/2016/10/20/urea_002_03_avg.npz')
urea6m = sfg2d.io.load_npz_to_Scan('/home/malte/MeasurmentData/Dolly1/2016/10/20/urea_002_02_avg.npz')
urea4m = sfg2d.io.load_npz_to_Scan('/home/malte/MeasurmentData/Dolly1/2016/10/19/urea_002_01_v1_avg.npz')

urea8m.fitarg = sfg2d.io.load_fitarg('/home/malte/MeasurmentData/Dolly1/2016/10/19/urea_002_v1_fit.text')
urea7m.fitarg = sfg2d.io.load_fitarg('/home/malte/MeasurmentData/Dolly1/2016/10/20/urea_002_03_fit.text')
urea6m.fitarg = sfg2d.io.load_fitarg('/home/malte/MeasurmentData/Dolly1/2016/10/20/urea_002_02_fit.text')
urea4m.fitarg = sfg2d.io.load_fitarg('/home/malte/MeasurmentData/Dolly1/2016/10/19/urea_002_01_v1_fit.text')

urea8m.fit = fit = iminuit.Minuit(sfg2d.utils.static.sfgn1, **urea8m.fitarg, pedantic=False)


def sfg1d(x, fit_func, fitarg):
    kwargs = iminuit.describe(sfg2d.utils.static.sfgn1)[1:]
    
    
    
    sfg2d.utils.static.sfgn1(x, )


    

In [ ]:

    

urea8m.fit.values


    

In [ ]:

    

fit = iminuit.Minuit(sfg2d.utils.static.sfgn1, **fitarg)


    

In [ ]:

    

fig, ax = subplots()

plot(urea8m.df.normalized * 1000)
plot(urea7m.df.normalized * 1000)
plot(urea6m.df.normalized * 1000)
plot(urea4m.df.normalized * 1000)

xlim(1400, 1800)
ylim(0, 10)


    

In [ ]: