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

# In[46]:


import sklearn
from sklearn.pipeline import make_pipeline
from sklearn.feature_extraction.text import TfidfVectorizer
import numpy as np
from sklearn.naive_bayes import MultinomialNB
from sklearn.preprocessing import LabelEncoder 


# In[47]:


def getInput(path):
    with open(path,encoding='utf-8') as f:
          return f.readlines()


# In[48]:


get_ipython().system('pwd')


# In[49]:


train_in=getInput('./train/in.tsv')
train_expected=getInput('./train/expected.tsv')
test_in=getInput('./test-A/in.tsv')
dev_in=getInput('./dev-0/in.tsv')
dev_expected=getInput('./dev-0/expected.tsv')


# In[50]:


pipeline = make_pipeline(TfidfVectorizer(),MultinomialNB())
encTransform = LabelEncoder().fit_transform(train_expected)
model = pipeline.fit(train_in, encTransform)
dev_predicted = model.predict(dev_in)
test_predicted = model.predict(test_in)


# In[ ]:





# In[54]:


with open('./dev-0/out.tsv', "w") as result:
            for out in dev_predicted:
                result.write(str(out) + '\n')
with open('./test-A/out.tsv', "w") as result:
            for out in test_predicted:
                result.write(str(out) + '\n')                


# In[55]:


get_ipython().system('jupyter nbconvert --to script Naiwny_bayes.ipynb')


# In[ ]: