import sklearn

from sklearn.feature_extraction.text import TfidfVectorizer

import csv

import re

from sklearn.linear_model import LogisticRegression

#from sklearn.discriminant_analysis import LinearDiscriminantAnalysis

#from sklearn.naive_bayes import MultinomialNB

from scipy.sparse import *

train_x = []

train_y = []

test_x = []

#load two train data files, one test data file, and formatting them

with open("data_set/train_set_x.csv","rb") as csvfile:

reader = csv.reader(csvfile, delimiter=',')

next(reader,None) #skip header of file

for row in reader:

text_deleteurl = re.sub(r"http\S+","", row[1])

text_deletenum=re.sub("\d+","",text_deleteurl)

l=text_deletenum.replace(" ","").lower()

train_x.append(l)

with open("data_set/train_set_y.csv","rb") as csvfile:

reader = csv.reader(csvfile, delimiter=',')

next(reader,None) #skip header of file

for row in reader:

train_y.append(row[1])

with open("data_set/test_set_x.csv","rb") as csvfile:

reader = csv.reader(csvfile, delimiter=',')

next(reader,None) #skip header of file

for row in reader:

text_deletenum=re.sub("\d+","",row[1])

l=text_deletenum.replace(" ","").lower()

test_x.append(l)

#tfidf preprocessing

vec = TfidfVectorizer(decode_error='strict',analyzer='char',min_df=0)

train_x=vec.fit_transform(train_x)

features = vec.get_feature_names()

vec2 = TfidfVectorizer(decode_error='strict',analyzer='char',min_df=0,vocabulary=features)

test_x = vec2.fit_transform(test_x)

#print(test_x)

#print("train_x is a matrix with size : ",train_x.shape[0],train_x.shape[1])

#print("train_y is an array with size: ",len(train_y))

# Fitting Kernel SVM to the Training set

from sklearn.svm import SVC

classifier = SVC(kernel = 'rbf', random_state = 0)

#from sklearn.neighbors import KNeighborsClassifier

#classifier = KNeighborsClassifier(n_neighbors = 5, metric = 'minkowski', p = 2)

classifier.fit(train_x.toarray(), train_y)

# Library function : logisticRegression

#lr_classifier = LogisticRegression(penalty='l2', C=1)

#lr_classifier.fit(train_x, train_y)

# predict on the test file

test_y_pred = classifier.predict(test_x.toarray())

test_y_pred_temp = test_y_pred.tolist()

# write the output to the output file

with open("kNN_output.csv",'wb') as output:

output.write("Id,Category")

output.write("\n")

for i in range(len(test_y_pred_temp)):

output.write(str(i))

output.write(",")

output.write(test_y_pred_temp[i])

output.write("\n")