# -*- coding: utf-8 -*-
"""
Created on Fri Oct 18 22:33:37 2019

@author: Nur Imtiazul Haque
"""
import numpy as np
import pandas as pd
from sklearn.preprocessing import StandardScaler
import sys
import warnings
import math

def modify_value(data):
    data=str(data)
    modified_string=""
    for i in range(len(data)):
        if (data[i]>='0' and data[i]<='9') or data[i]=='.':
            modified_string+=data[i]

    if modified_string=="":
        modified_string="0"
    return modified_string

if not sys.warnoptions:
    warnings.simplefilter("ignore")

def mean_squared_error(theta, X, y):
    estimator = [np.exp(np.dot(theta,x)) for x in X]
    #print(X)
    return .5*sum([math.pow(estimator[i] - y[i],2) for i in range(len(estimator))])

def poisson_regression(X_old,y_old,alpha=0.01,max_iter=4000):
    theta = np.random.rand(4)
    print("Initial Theta: ")
    print(theta)
    iter_mse=[]
    #print(len(X_old))
    #print(len(y_old))
    X = []
    y =[]

    for i in range(len(X_old)):
        #if np.isnan(sum(X_old[i]))==False:
        X.append(X_old[i])
        y.append(y_old[i])
    #print(len(X))
    #print(len(y))
    for k in range(max_iter):
        for j in range(len(theta)):
            update = alpha*sum([(y[i] - np.exp(np.dot(theta,X[i])))*X[i][j] for i in range(len(X))])
            theta[j] = theta[j] + update   #updating theta

        error = mean_squared_error(theta,X,y)
        #print(error)
        iter_mse.append((k, error))
    return error



path = 'nyc-east-river-bicycle-counts.csv'
dataset= pd.read_csv(path)

input_data = dataset.iloc[:, :-1].values
output_data = dataset.iloc[:, 10].values

for row in range(len(input_data)):
    for col in range(3, len(input_data[0])):
        input_data[row][col]=modify_value(input_data[row][col])

prediction_target = [6,7,8,9,10]

for iter in prediction_target:
    X = np.array( [ np.append([1],[i for i in value]) for value in input_data[:,3:6] ]).astype(np.float)
    y = np.array(input_data[:,iter]).astype(np.float)

    ms_error = poisson_regression(X,y)
    print(str(iter) + '--> Mean Square Error: ' + str(ms_error))