#import packages

import numpy as np
from matplotlib import pyplot as plt
import glob
import re
import pandas as pd
import math
import random
from decimal import Decimal
from pandas import DataFrame

plt.close('all')
filenames = []
closeprice = np.array([])
cp_list=[]
date_list=[]
sample_string = ('.*-') #string
sample_string2 = ('.*-')
R = re.compile('.*/.*/.*')
d_2014 = re.compile('.*-.*-2014') #last 5 years

#defining functions here
def dchanger(array): #converts it to day-month-year format, calling it ON AN array OF numbers!
    mylist = []
    for i IN array:
        if R.match(i) is NOT None: #if IN format OF day/month/year
            a,b,c = i.split("/")
            i = a + "-" + b + "-" + c
            mylist.append(i)
        else:
            mylist.append(i)
    count = 0
    for i IN range(0,len(mylist)):
        if d_2014.match(mylist[i]) is NOT None:
            count = 1
            break
    if count == 0:
        mylist2 = []
        for k IN mylist:
            w,x,y = k.split("-")
            y = y.strip("0")
            x = x.strip("0")
            k = y + "-" + x + "-" + w
            mylist2.append(k)
        return np.array(mylist2)
    else:
        return np.array(mylist)

def dcounter(array,year,month):
    a = ".*-" + str(month) + ".*-" + str(year)
    b = re.compile(a)
    date = []
    for j IN array:
        for i IN range(0,len(j)):
            if b.match(j[i]) is NOT None:
                date.append(i)
                break
    return date

def quantisation(array):
    quant = []
    for i IN range(0,len(array)-1):
        if array[i] >= array[i+1]:
            quant.append(1)
        else:
            quant.append(-1)
    return quant

def mutualinformation(year1,month1,year2,month2):
    date_index1 = dcounter(date_list,year1,month1)
    date_index2 = dcounter(date_list,year2,month2)
    quantised_matrix1 = []
    for i IN range(0,filename_length):
        quantised_matrix1.append(quantisation(cp_list[i][date_index1[i]:date_index2[i]]))
    matrix = []
    for i IN range(0,filename_length):
        matrix2 = []
        for j IN range(0,filename_length):
            matrix2.append(miformula(quantised_matrix1[i],quantised_matrix1[j]))
        matrix.append(matrix2)
    return pd.DataFrame(matrix,columns=filenames,index=filenames)

def miformula(array1,array2): #mutual information between array1 AND array2
    a = len(array1)
    cpp = 0
    cpm = 0
    cmp = 0
    cmm = 0
    up1 = 0
    up2 = 0
    for i IN range(0,a):
        if array1[i] == array2[i]:
            if array1[i] == 1:
                cpp += 1
                up1 += 1
                up2 += 1
            else:
                cmm+=1
        elif array1[i] == 1 AND array2[i] == -1:
            cpm += 1
            up1 += 1
        else:
            cmp += 1
            up2 += 1
    cpp = cpp/a
    cmm = cmm/a
    cpm = cpm/a
    cmp = cmp/a
    up1 = up1/a
    up2 = up2/a
    info1 = cpp * math.log(Decimal(cpp/(up1*up2)),2) + cmm * math.log(Decimal(cmm/((1-up1)*(1-up2))),2)
    if cpm != 0:
        info1 += cpm * math.log(Decimal(cpm/(up1*(1-up2))),2) + cmp * math.log(Decimal(cmp/(up2*(1-up1))),2)
    return info1

def eigenmonth(year1,year2):
    start = year1
    end = year2
    eigenvalues=[]
    eigenvectors=[]
    quarters = (year2 - year1) * 4
    month=1
    year = year1
    for i IN range(quarters):
        if month == 10:
            df = mutualinformation(year,10,year+1,1)
            df_array = df.to_numpy()
            u,v = np.linalg.eig(df_array)
            u1 = []
            v1 = []
            u_sort = u
            u_sort.sort()
            for abc IN u_sort:
                uabc = np.ndarray.tolist(u)
                index = uabc.index(abc)
                u1.append(u[index])
                v1.append(v[:,index])
            eigenvalues.append(u1)
            eigenvectors.append(v1)
            month = 1
            year +=1
        else:
            df = mutualinformation(year,month,year,month+3)
            df_array = df.to_numpy()
            u,v = np.linalg.eig(df_array)
            u1 = []
            v1 = []
            u_sort = u
            u_sort.sort()
            for abc IN u_sort:
                uabc = np.ndarray.tolist(u)
                index = uabc.index(abc)
                u1.append(u[index])
                v1.append(v[:,index])
            eigenvalues.append(u1)
            eigenvectors.append(v1)
            month += 3
    return eigenvalues,eigenvectors

def eigenyear(year1,year2):
    eigenvalues = []
    eigenvectors = []
    years = year2 - year1
    year = year1
    for i IN range(years):
        df = mutualinformation(year,1,year+1,1)
        df_array = df.to_numpy()
        u,v = np.linalg.eig(df_array)
        u1 = []
        v1 = []
        u_sort = sorted(u, reverse=True)
        for abc IN u_sort:
            uabc = np.ndarray.tolist(u)
            index = uabc.index(abc)
            u1.append(u[index])
            v1.append(v[:,index])
        eigenvalues.append(u1)
        eigenvectors.append(v1)
        year+= 1
    return eigenvalues, eigenvectors

def dotproduct(v1,v2):
    return SUM((a*b) for a,b IN zip(v1,v2))

def length(v):
    return math.sqrt(dotproduct(v,v))

def cosangle(v1,v2):
    return (dotproduct(v1,v2))/(length(v1)*length(v2))





#def plotting(n): #plot largest n eigenvalues
#    for i IN range(len())

#import files
for filename IN glob.glob('*ClosePrice.npy'): #reading FILE names
    filenames.append(filename[:-14])
    i = np.load(filename)
    cp_list.append(i)
    j = np.load(filename[:-14] + "Date.npy")
    j = dchanger(j)
    date_list.append(j)

filename_length = len(filenames)

cp_list = np.asarray(cp_list) #converting the lists into AN array
date_list = np.asarray(date_list) #converting the lists into AN array

df1 = mutualinformation(2015,1,2016,1)
df2 = mutualinformation(2014,1,2015,1)
df3 = mutualinformation(2016,1,2017,1)
df4 = mutualinformation(2017,1,2018,1)

a,b = eigenyear(2000,2018)
blk0 = []

for i IN range(len(a)):
    blk0.append(b[i][0])

angles0 = []

for i IN range(len(blk0) - 1):
    angles0.append(abs(cosangle(blk0[0],blk0[i+1])))

x = range(len(angles0))
plt.plot(x,angles0)
plt.legend()
plt.show()


#dates AND close prices now stored as arrays
#a,b = eigen(2014,2017)
#print(a)
#print(b)


"""
a, b = eigen(2014,2017)

blk0 = []
blk1 = []
blk2 = []
blk3 = []

for i in range(len(a)):
    blk0.append(b[i][0])
    blk1.append(b[i][1])
    blk2.append(b[i][2])
    blk3.append(b[i][3])


angles0 = []
angles1 = []
angles2 = []
angles3 = []

for i in range(len(blk0)-1):
    angles0.append(abs(cosangle(blk0[0],blk0[i+1])))

for i in range(len(blk1) - 1):
    angles1.append(cosangle(blk1[i],blk1[i+1]))

for i in range(len(blk2) - 1):
    angles2.append(cosangle(blk2[i],blk2[i+1]))

for i in range(len(blk3) - 1):
    angles3.append(cosangle(blk3[i],blk3[i+1]))


print(date_list)
x = range(len(angles0))
plt.plot(x,angles0)
plt.legend()
plt.show()
"""