show.py

from numpy import loadtxt, zeros, ones, array, linspace, logspace
from pylab import scatter, show, title, xlabel, ylabel, plot, contour
import pickle
import csv
import numpy as np
from scipy.optimize import curve_fit
x1=[]

def func(x, p1,p2):
  return p1*np.cos(p2*x) + p2*np.sin(p1*x)
loc_submission = "train_mod.csv"
#Load the dataset
out =[]
with open('test.csv') as f:
     l = 0 
     for line in f:
	if l==0:
           head= line
        l = l+1
        if l>1:
           line = line.split(',')
           count = 0
           count1 = 0
           fill = 0.000001
           length = -1
           line1 = line[28:147]
           for i in line1:
              length = length + 1
              if i == '':
                 line1[length] = fill
                #print fill
                #count = count + 1
              else:
                 fill = line1[length]
           sub =0
           lin =[]
           for i in line1:
              sub = sub + float(i);
              lin.append(sub)
          # print len(lin)        
           #line[26:209] = lin
          
           out.append(lin)
           print len(lin)
           x1=[]
           #x1 = list(range(119))
           for i in xrange(119):
               x1.append(i+2)
           x1 = np.array(x1)
           print x1   
           print len(x1)
	   y = lin
           y = np.array(y) 
           #popt, pcov = curve_fit(func, x1, y,p0=(1.0,0.2),maxfev=1000)
           try:
              popt, pcov = curve_fit(func, x1, y,p0=(1.0,0.2),maxfev=5000)
           except RuntimeError:
              print("Error - curve_fit failed")
           p1 = popt[0]
	   p2 = popt[1]
           v = 120
           for i in xrange(60):
               x =v+i    
               lin.append(p1*np.cos(p2*x) + p2*np.sin(p1*x)) 
#Plot the data

           x1 = list(range(179))
           y = lin   
           scatter(x1, y, marker='o', c='b')
           title('Stock returns')
           xlabel('Daywise')
           ylabel('Return value')
     	   show()
#with open(loc_submission, "w") as outfile:
#head = head.split(',')
#head = head[26:147]
   #  outfile.writelines(head)
    # outfile.write("\n")
    # pickle.dump(out,outfile)   
     ##  print>>outfile,i   

resultFile = open("output_train.csv",'wb')
resultFile.writelines(head)
#resultFile.write("\n")
wr = csv.writer(resultFile, dialect='excel')
wr.writerows(out)