20NewsBatch.py

from collections import Counter
import tensorflow as tf
from sklearn.datasets import fetch_20newsgroups
import matplotlib as mplt
mplt.use('agg') # Must be before importing matplotlib.pyplot or pylab!
import matplotlib.pyplot as plt
from string import punctuation
from sklearn.preprocessing import LabelBinarizer
import numpy as np
from nltk.corpus import stopwords
from sklearn.metrics import f1_score, recall_score, precision_score, accuracy_score
from sklearn.model_selection import train_test_split
import nltk
nltk.download('stopwords')



def pre_process():
    newsgroups_data = fetch_20newsgroups(subset='all', remove=('headers', 'footers', 'quotes'))

    words = []
    temp_post_text = []
    print(len(newsgroups_data.data))

    for post in newsgroups_data.data:

        all_text = ''.join([text for text in post if text not in punctuation])
        all_text = all_text.split('\n')
        all_text = ''.join(all_text)
        temp_text = all_text.split(" ")

        for word in temp_text:
            if word.isalpha():
                temp_text[temp_text.index(word)] = word.lower()

        temp_text = [word for word in temp_text if word not in stopwords.words('english')]
        temp_text = list(filter(None, temp_text))
        temp_text = ' '.join([i for i in temp_text if not i.isdigit()])
        words += temp_text.split(" ")
        temp_post_text.append(temp_text)

    # temp_post_text = list(filter(None, temp_post_text))

    dictionary = Counter(words)
    # deleting spacesA
    # del dictionary[""]
    sorted_split_words = sorted(dictionary, key=dictionary.get, reverse=True)
    vocab_to_int = {c: i for i, c in enumerate(sorted_split_words,1)}

    message_ints = []
    for message in temp_post_text:
        temp_message = message.split(" ")
        message_ints.append([vocab_to_int[i] for i in temp_message])

    # maximum message length = 6577
    # message_lens = Counter([len(x) for x in message_ints])

    seq_length = 6577
    num_messages = len(temp_post_text)
    features = np.zeros([num_messages, seq_length], dtype=int)
    for i, row in enumerate(message_ints):
        features[i, :len(row)] = np.array(row)[:seq_length]

    lb = LabelBinarizer()
    lbl = newsgroups_data.target
    labels = np.reshape(lbl, [-1])
    labels = lb.fit_transform(labels)

    sequence_lengths = [len(msg) for msg in message_ints]
    return features, labels, len(sorted_split_words)+1, sequence_lengths


def get_batches(x, y, sql, batch_size=100):
    n_batches = len(x) // batch_size
    x, y = x[:n_batches * batch_size], y[:n_batches * batch_size]
    for ii in range(0, len(x), batch_size):
        yield x[ii:ii + batch_size], y[ii:ii + batch_size], sql[ii:ii+batch_size]



def train_test():
    features, labels, n_words, sequence_length = pre_process()

    train_x, test_x, train_y, test_y = train_test_split(features, labels, test_size=0.2, shuffle=False, random_state=42)

    sequence_length_train = sequence_length[:len(train_y)]
    sequence_length_test= sequence_length[len(train_y):]


    # Defining Hyperparameters

    lstm_layers = 1
    batch_size = 179
    lstm_size = 30
    learning_rate = 0.003
    epoch = 3

    print("learning 32")

    # --------------placeholders-------------------------------------

    # Create the graph object
    graph = tf.Graph()
    # Add nodes to the graph
    with graph.as_default():

        tf.set_random_seed(1)

        inputs_ = tf.placeholder(tf.int32, [None, None], name="inputs")
        # labels_ = tf.placeholder(dtype= tf.int32)
        labels_ = tf.placeholder(tf.float32, [None, None], name="labels")
        sql_in = tf.placeholder(tf.int32, [None], name='sql_in')

        # Size of the embedding vectors (number of units in the embedding layer)
        embed_size = 300

        # generating random values from a uniform distribution (minval included and maxval excluded)
        embedding = tf.Variable(tf.random_uniform((n_words, embed_size), -1, 1))
        embed = tf.nn.embedding_lookup(embedding, inputs_)

        # Your basic LSTM cell
        lstm = tf.contrib.rnn.BasicLSTMCell(lstm_size)

        # Getting an initial state of all zeros
        initial_state = lstm.zero_state(batch_size, tf.float32)

        outputs, final_state = tf.nn.dynamic_rnn(lstm, embed, initial_state=initial_state, sequence_length=sql_in)

        out_batch_size = tf.shape(outputs)[0]
        out_max_length = tf.shape(outputs)[1]
        out_size = int(outputs.get_shape()[2])
        index = tf.range(0, out_batch_size) * out_max_length + (sql_in - 1)
        flat = tf.reshape(outputs, [-1, out_size])
        relevant = tf.gather(flat, index)

        # hidden layer
        hidden = tf.layers.dense(relevant, units=25, activation=tf.nn.relu)

        logit = tf.contrib.layers.fully_connected(hidden, num_outputs=20, activation_fn=None)

        cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=logit, labels=labels_))

        optimizer = tf.train.AdamOptimizer(learning_rate).minimize(cost)

        saver = tf.train.Saver()

    # ----------------------------batch training-----------------------------------------

    with tf.Session(graph=graph) as sess:
        tf.set_random_seed(1)
        sess.run(tf.global_variables_initializer())
        iteration = 1
        for e in range (epoch):
            state = sess.run(initial_state)
            for ii, (x, y, sql) in enumerate(get_batches(np.array(train_x),  np.array(train_y), sequence_length_train, batch_size), 1):

                feed = {inputs_: x,
                        labels_: y,
                        sql_in: sql,
                        initial_state: state}

                loss, states, _ = sess.run([cost, final_state, optimizer], feed_dict=feed)


                if iteration % 5 == 0:
                    print("Epoch: {}/{}".format(e, epoch),
                          "Iteration: {}".format(iteration),
                          "Train loss: {:.3f}".format(loss))
                iteration += 1
        saver.save(sess, "checkpoints/sentiment.ckpt")

     # -----------------testing test set-----------------------------------------
        print("starting testing set")
        argmax_pred_array = []
        argmax_label_array = []
        with tf.Session(graph=graph) as sess:
                tf.set_random_seed(1)
                saver.restore(sess, tf.train.latest_checkpoint('checkpoints'))
                test_state = sess.run(lstm.zero_state(batch_size, tf.float32))

                for ii, (x, y, sql) in enumerate(get_batches(np.array(test_x), np.array(test_y), sequence_length_test, batch_size), 1):
                    feed = {inputs_: x,
                            labels_: y,
                            sql_in: sql,
                            initial_state: test_state}

                    predictions = tf.nn.softmax(logit).eval(feed_dict=feed)

                    for i in range(len(predictions)):
                        argmax_pred_array.append(np.argmax(predictions[i], 0))
                        argmax_label_array.append(np.argmax(y[i], 0))

                    print(len(argmax_pred_array))
                    print(len(argmax_label_array))

                accuracy = accuracy_score(argmax_label_array, argmax_pred_array)

                batch_f1 = f1_score(argmax_label_array, argmax_pred_array, average="macro")

                batch_recall = recall_score(y_true=argmax_label_array, y_pred=argmax_pred_array, average='macro')

                batch_precision = precision_score(argmax_label_array, argmax_pred_array, average='macro')

                print("-----------------testing test set-----------------------------------------")
                print("Test accuracy: {:.3f}".format(accuracy))
                print("F1 Score: {:.3f}".format(batch_f1))
                print("Recall: {:.3f}".format(batch_recall))
                print("Precision: {:.3f}".format(batch_precision))


if __name__ == '__main__':
    train_test()