#!/usr/bin/env python3
import numpy as np
import csv
import matplotlib.pyplot as plt
import math

xcolumn=0 #column with readings
ycolumn=4 #column with calibration data
ncoefs=4 #number of coefficients

#mit log
#low (0), coeffs=4,, <2.7,  ab <=2.3 benutzbar
#high (1), coeffs=5, >2  ab >=1 knapp benutzbar, ab 2 besser

xvalues=[]
yvalues=[]

with open('20180413_calibration.csv', 'r') as csvfile:
    csvreader = csv.reader(csvfile, delimiter=';')
    firstrow=True
    for row in csvreader:
        xvalue=row[xcolumn]
        yvalue=row[ycolumn]
        if len(xvalue)>0 and len(yvalue)>0 and not firstrow:
            xvalue=math.log(float(xvalue))
            yvalue=float(yvalue)
            if yvalue<2.7:
                #print(""+str(xvalue)+" - "+str(yvalue))
                xvalues.append(xvalue)
                yvalues.append(yvalue)
        firstrow=False

    coefs=np.polyfit(xvalues,yvalues,ncoefs) #fit polynomial curve
    print(coefs) #coef 0 is the one with highest polynomial

    xtest=np.arange(max(xvalues),step=0.001) #x values for test visualization
    ytest=np.polyval(coefs, xtest) #calculate y values with polynomial function
    #ytest=[coefs[3]+coefs[2]*pow(x,1)+coefs[1]*pow(x,2)+coefs[0]*pow(x,3) for x in xtest]

    #for i in range(200,4000,10):
    #    yv=np.polyval(coefs, i)
    #    print(str(i)+";"+str(yv))
    #exit()

    plt.scatter(xvalues,yvalues,s=0.7,c='g') #plot sample data
    plt.plot(xtest,ytest,c='r') #plot approximated curve
    plt.xlabel('LDR Value')
    plt.ylabel('Ev')

    #plt.xlim(0,4096) #ldr
    plt.ylim(-3,20) #ev

    #limits with lux
    #plt.xlim(0,4096) #ldr
    #plt.xlim(0,500) #ldr
    #plt.ylim(0,1280000) #ev
    #plt.ylim(0,220) #ev
    plt.show()