paranormal-or-skeptic/run.py

import lzma
import gensim.downloader
import numpy as np
import torch

FEAUTERES = 100

def predict(data):
    nn_model.eval()
    predictions = []
    for i in range(len(data)):
        X = data[i]
        X = torch.tensor(X.astype(np.float32))
        Y_predictions = nn_model(X)
        if Y_predictions[0] > 0.5:
            predictions.append("1")
        else:
            predictions.append("0")
    return predictions

def vectorize(data):
    vectorized_data = np.array([np.array([word2vec_model.wv[i] for i in ls if i in words]) for ls in data])
    avarage_vector = []
    for vector in vectorized_data:
        if vector.size:
            avarage_vector.append(vector.mean(axis=0))
        else:
            avarage_vector.append(np.zeros(100, dtype=float))
    return avarage_vector


def generate_out(folder_path):
    print('Generating out')
    X_dev = []
    with lzma.open(f"{folder_path}/in.tsv.xz", 'r') as file:
        for line in file:
            line = line.strip()
            line = line.decode("utf-8")
            tabs = line.rsplit('\t')
            content = tabs[0]
            pre_processed = gensim.utils.simple_preprocess(content)
            X_dev.append(pre_processed)

    print("step 5")
    X_dev = vectorize(X_dev)
    prediction = predict(X_dev)

    print("step 6")
    f = open(f"{folder_path}/out.tsv", "a")
    for p in prediction:
        f.write(str(p) + '\n')
    f.close()


def get_loss_acc(model, X_dataset, Y_dataset):
    loss_score = 0
    acc_score = 0
    items_total = 0
    model.eval()
    for i in range(0, Y_dataset.shape[0], BATCH_SIZE):
        X = X_dataset[i:i+BATCH_SIZE]
        X = torch.tensor(X.astype(np.float32))
        Y = Y_dataset[i:i+BATCH_SIZE]
        Y = torch.tensor(Y.astype(np.float32)).reshape(-1,1)
        Y_predictions = model(X)
        acc_score += torch.sum((Y_predictions > 0.5) == Y).item()
        items_total += Y.shape[0] 

        loss = criterion(Y_predictions, Y)

        loss_score += loss.item() * Y.shape[0] 
    return (loss_score / items_total), (acc_score / items_total)


class NeuralNetworkModel(torch.nn.Module):

    def __init__(self):
        super(NeuralNetworkModel, self).__init__()
        self.fc1 = torch.nn.Linear(FEAUTERES,500)
        self.fc2 = torch.nn.Linear(500,1)

    def forward(self, x):
        x = self.fc1(x)
        x = torch.relu(x)
        x = self.fc2(x)
        x = torch.sigmoid(x)
        return x



if __name__ == "__main__":
    X = []
    Y = []
    with lzma.open('train/in.tsv.xz', 'r') as file:
        for line in file:
            line = line.strip()
            line = line.decode("utf-8")
            tabs = line.rsplit('\t')
            content = tabs[0]
            pre_processed = gensim.utils.simple_preprocess(content)
            X.append(pre_processed)
    print("step 1")
    with open('train/expected.tsv', 'r') as file:
        for line in file:
            line = line.strip()
            Y.append(int(line))

    X_train = X
    Y_train = Y
    print("step 2")


    print('Word to vec start')

    word2vec_model = gensim.models.Word2Vec(X_train, vector_size=100, window=5, min_count=2)
    print('Created model')
    words = set(word2vec_model.wv.index_to_key)
    print('Created set of worlds')

    X_train = vectorize(X_train)
    X_train = np.array(X_train)
    Y_train = np.array(Y_train)
    print('Vectorized data')


    print('Word to vec ended')
    print("step 3")
    # model = LogisticRegression()
    # model.fit(X_vectorized, Y)
    nn_model = NeuralNetworkModel()
    BATCH_SIZE = 5
    criterion = torch.nn.BCELoss()
    optimizer = torch.optim.SGD(nn_model.parameters(), lr = 0.1)

    for epoch in range(7):
        loss_score = 0
        acc_score = 0
        items_total = 0
        nn_model.train()
        for i in range(0, Y_train.shape[0], BATCH_SIZE):
            X = X_train[i:i+BATCH_SIZE]
            X = torch.tensor(X.astype(np.float32))
            Y = Y_train[i:i+BATCH_SIZE]
            Y = torch.tensor(Y.astype(np.float32)).reshape(-1,1)
            Y_predictions = nn_model(X)
            acc_score += torch.sum((Y_predictions > 0.5) == Y).item()
            items_total += Y.shape[0] 

            optimizer.zero_grad()
            loss = criterion(Y_predictions, Y)
            loss.backward()
            optimizer.step()


            loss_score += loss.item() * Y.shape[0] 
        print('Printing')
        print(epoch)
        print(get_loss_acc(nn_model, X_train, Y_train))

    # display(get_loss_acc(nn_model, X_dev, Y_dev))

    print("step 4")
    generate_out('dev-0')
    # generate_out('dev-1')
    generate_out('test-A')