word2vec_cbow.py

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import math
import random
import re

import numpy as np
from six.moves import urllib
from six.moves import xrange  # pylint: disable=redefined-builtin
import tensorflow as tf

from imdb_sentiment_data import get_dataset
from word2vec_fns import generate_batch, get_mean_context_embeds

vocabulary_size = 50000
def read_and_clean_data(path):
    
    with open(path,"r") as o:
        
        text = o.read()
        
        
       # punc_rem = text.translate(str.maketrans(",", punctuation))
 #       punc_rem = text.translate(None, punctuation)
       punc_rem = re.sub(r'[^\w\s]', '', text)
       lower_words = map(lambda x: x.lower(),punc_rem.split())

    return lower_words
corpus = []
count = []
words = []
folders = [ "pos", "neg"]
for folder in folders:
    for path in glob.glob(folder + "/*.txt"):

        words += read_and_clean_data(path)
        if len(set(words))> vocabulary_size:
            break
    else:
        continue
    break

count = Counter(words)
unique_words = sorted(count.keys())
idxs = range(len(count.keys()))

data, count, dictionary, reverse_dictionary = get_dataset(vocabulary_size)
reverse_dictionary = dict(zip(unique_words, idxs))
dictionary = dict(zip(idxs, unique_words))

#print('Most common words (+UNK)', count[:5])
#print('Sample data', data[:10], [reverse_dictionary[i] for i in data[:10]])

# sanity check the batches
batch_size = 2
check_skip_window = 2      # How many words to consider left and right.
batch, labels = generate_batch(data, batch_size=8, skip_window=check_skip_window)
for i in range(batch_size):
    print(batch[i, :], [reverse_dictionary[batch[i, j]] for j in range(check_skip_window*2)],
          '->', labels[i, 0], reverse_dictionary[labels[i, 0]])



# Step 4: Build and train a skip-gram model.

batch_size = 2
embedding_size = 100  # Dimension of the embedding vector.
skip_window = 1       # How many words to consider left and right.

# We pick a random validation set to sample nearest neighbors. Here we limit the
# validation samples to the words that have a low numeric ID, which by
# construction are also the most frequent.
valid_size = 16     # Random set of words to evaluate similarity on.
valid_window = 100  # Only pick dev samples in the head of the distribution.
valid_examples = np.random.choice(valid_window, valid_size, replace=False)
num_sampled = 64    # Number of negative examples to sample.

graph = tf.Graph()

with graph.as_default():

    # Input data.
    train_inputs = tf.placeholder(tf.int32, shape=[batch_size, 2*skip_window])
    train_labels = tf.placeholder(tf.int32, shape=[batch_size, 1])
    valid_dataset = tf.constant(valid_examples, dtype=tf.int32)

    # Ops and variables pinned to the CPU because of missing GPU implementation
    with tf.device('/cpu:0'):
        # Look up embeddings for inputs.
        embeddings = tf.Variable(
            tf.random_uniform([vocabulary_size, embedding_size], -1.0, 1.0))

        # train_inputs is of shape (batch_size, 2*skip_window)
        mean_context_embeds =\
            get_mean_context_embeds(embeddings, train_inputs)

        # Construct the variables for the NCE loss
        nce_weights = tf.Variable(
            tf.truncated_normal([vocabulary_size, embedding_size],
                                stddev=1.0 / math.sqrt(embedding_size)))
        nce_biases = tf.Variable(tf.zeros([vocabulary_size]))

        # Compute the average NCE loss for the batch.
        # tf.nce_loss automatically draws a new sample of the negative labels each
        # time we evaluate the loss.
        loss = tf.reduce_mean(
            tf.nn.nce_loss(weights=nce_weights,
                           biases=nce_biases,
                           labels=train_labels,
                           inputs=mean_context_embeds,
                           num_sampled=num_sampled,
                           num_classes=vocabulary_size))

        # Construct the SGD optimizer using a learning rate of 1.0.
        optimizer = tf.train.GradientDescentOptimizer(1.0).minimize(loss)

        # Compute the cosine similarity between minibatch examples and all embeddings.
        norm = tf.sqrt(tf.reduce_sum(tf.square(embeddings), 1, keep_dims=True))
        normalized_embeddings = embeddings / norm
        valid_embeddings = tf.nn.embedding_lookup(
            normalized_embeddings, valid_dataset)
        similarity = tf.matmul(
            valid_embeddings, normalized_embeddings, transpose_b=True)

    # Add variable initializer.
    init = tf.global_variables_initializer()

# Step 5: Begin training.
num_steps = 100001

saver = tf.train.Saver({'embedding':embeddings}, max_to_keep=None)


with tf.Session(graph=graph) as session:
    # We must initialize all variables before we use them.
    init.run()
    print('Initialized')

    average_loss = 0
    for step in range(num_steps):
        batch_inputs, batch_labels = generate_batch(data, batch_size, skip_window)
        n_batch_inputs = []
        n_batch_labels = []
        for i in range(batch_size):
            n_batch_inputs.append([reverse_dictionary[batch_inputs[i, j]] for j in range(skip_window * 2)])
            n_batch_labels.append([reverse_dictionary[labels[i, 0]]])
        feed_dict = {train_inputs: n_batch_inputs, train_labels: n_batch_labels}

        # We perform one update step by evaluating the optimizer op (including it
        # in the list of returned values for session.run()
              _, loss_val = session.run([optimizer, loss], feed_dict=feed_dict)
        average_loss += loss_val

        if step % 2000 == 0:
            if step > 0:
                average_loss /= 2000
            # The average loss is an estimate of the loss over the last 2000 batches.
            print('Average loss at step ', step, ': ', average_loss)
            average_loss = 0

        # Note that this is expensive (~20% slowdown if computed every 500 steps)
        if step % 10000 == 0:
            sim = similarity.eval()
            for i in range(valid_size):
                valid_word = dictionary[valid_examples[i]]
                top_k = 8  # number of nearest neighbors
                nearest = (-sim[i, :]).argsort()[1:top_k + 1]
                log_str = 'Nearest to %s:' % valid_word
                for k in range(top_k):
                    close_word = dictionary[nearest[k]]
                    log_str = '%s %s,' % (log_str, close_word)
                print(log_str)
            np.save("CBOW_Embeddings", normalized_embeddings.eval())
            #saver.save(session, 'w2vEmbedding', global_step=step)

    final_embeddings = normalized_embeddings.eval()
    np.save("CBOW_Embeddings", final_embeddings)