#!/usr/bin/env python
# coding: utf-8

# In[171]:


import pandas as pd
from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB


# In[172]:


f = open("train\in.tsv", "r", encoding='utf-8')
lines = f.readlines()
x_train = pd.DataFrame(lines)
x_train.rename(columns = {0 : 'text'}, inplace = True)


# In[173]:


y_train = pd.read_csv('train\expected.tsv', sep='\t', names=['paranormal'], encoding='utf-8')


# In[174]:


tfidf_vectorizer=TfidfVectorizer(max_df=0.95, max_features=500)
tfidf_vectorizer.fit_transform(x_train['text'].values)
x_train_prepared = tfidf_vectorizer.transform(x_train['text'].values).toarray()


# In[175]:


mnb = MultinomialNB()
model_mnb = mnb.fit(x_train_prepared, y_train.values.ravel())


# In[176]:


f = open("dev-0\in.tsv", "r", encoding='utf-8')
lines = f.readlines()
x_dev = pd.DataFrame(lines)
x_dev.rename(columns = {0 : 'text'}, inplace = True)

x_dev_prepared = tfidf_vectorizer.transform(x_dev['text'].values).toarray()

y_dev = pd.read_csv('dev-0\expected.tsv', sep='\t', names=['paranormal'], encoding='utf-8')

y_dev_pred = model_mnb.predict(x_dev_prepared)


# In[177]:


file = open('dev-0\out.tsv', 'w')
for y in y_dev_pred:
    file.write(f'{y}\n')
file.close()


# In[179]:


f = open("test-A\in.tsv", "r", encoding='utf-8')
lines = f.readlines()
x_test = pd.DataFrame(lines)
x_test.rename(columns = {0 : 'text'}, inplace = True)

x_test_prepared = tfidf_vectorizer.transform(x_test['text'].values).toarray()

y_test_pred = model_mnb.predict(x_test_prepared)


# In[180]:


file = open('test-A\out.tsv', 'w')
for y in y_test_pred:
    file.write(f'{y}\n')
file.close()