import pandas as pd
from sklearn.linear_model import LinearRegression
from sklearn.feature_extraction.text import TfidfVectorizer


def linear_regression(train_path, predict_path, out_path):
    print("Wczytywanie zbioru treningowego")
    train = pd.read_csv(train_path, sep='\t', header=None)
    X_train = train[4]
    Y_train = train[0]

    print("Wczytywanie pliku do predykcji")
    pred_x = []
    with open(predict_path, encoding='utf-8') as f:
        for line in f:
            pred_x.append(line)

    print("Wektoryzacja")
    vectorizer = TfidfVectorizer()
    X_train = vectorizer.fit_transform(X_train)
    pred_x = vectorizer.transform(pred_x)

    print("Uczenie modelu")
    model = LinearRegression()
    model.fit(X_train, Y_train)

    print("Predykcja wyników")
    pred_y = model.predict(pred_x)


    print("Zapis do pliku")
    pd.DataFrame(pred_y).to_csv(out_path, header=False, index=None)

linear_regression("train/train.tsv", "dev-0/in.tsv", "dev-0/out.tsv")
linear_regression("train/train.tsv", "dev-1/in.tsv", "dev-1/out.tsv")
linear_regression("train/train.tsv", "test-A/in.tsv", "test-A/out.tsv")