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

le=preprocessing.LabelEncoder()

def open_file_xz(file: str):
    with lzma.open(file, mode='rt') as f:
        return f.readlines()

def open_file(file: str):
    with open(file) as f:
        return f.readlines()

def train_model(train_in, train_exp):
    train_exp = le.fit_transform(train_exp)
    pipeline = make_pipeline(TfidfVectorizer(),MultinomialNB())
    model=pipeline.fit(train_in, train_exp)
    return model

def make_prediction(catalog: str):
    train_in = open_file_xz("train/in.tsv.xz")
    train_exp = open_file("train/expected.tsv")
    to_be_predict = open_file_xz(catalog + "/in.tsv.xz")

    model = train_model(train_in,train_exp)
    
    prediction=model.predict(to_be_predict)
    np.savetxt(catalog + "/out.tsv", prediction, fmt='%d', delimiter='\n')

make_prediction("dev-0")
make_prediction("test-A")