def main(): #argv
print("Usage: python ./main.py <number_of_topics> <maxiteration>")
# load stop words list from file
stopwordsfile = open("stopwords.txt", "r", encoding='utf-8')
for word in stopwordsfile: # a stop word in each line
word = word.replace("\n", '')
word = word.replace("\r\n", '')
STOP_WORDS_SET.add(word)
corpus = plsa.Corpus() # instantiate corpus
# iterate over the files in the directory.
document_paths = ['./texts/grimm_fairy_tales', './texts/tech_blog_posts', './texts/nyt']
# document_paths = ['./test/']
for document_path in document_paths:
for document_file in glob.glob(os.path.join(document_path, '*.txt')):
print(document_file)
document = plsa.Document(document_file) # instantiate document
document.split(STOP_WORDS_SET) # tokenize
corpus.add_document(document) # push onto corpus documents list
corpus.build_vocabulary()
print("Vocabulary size:" + str(len(corpus.vocabulary)))
print("Number of documents:" + str(len(corpus.documents)))
#print("here:", argv)
number_of_topics = 5 #int(argv[1])
max_iterations = 20 #int(argv[2])
corpus.plsa(number_of_topics, max_iterations)
# print corpus.document_topic_prob
# print corpus.topic_word_prob
# cPickle.dump(corpus, open('./models/corpus.pickle', 'w'))
print_topic_word_distribution(corpus, number_of_topics, 20, "./topic-word.txt")
print_document_topic_distribution(corpus, number_of_topics, 10, "./document-topic.txt")
def main(argv):
print "Usage: python ./main.py <number_of_topics> <maxiteration>"
# load stop words list from file
stopwordsfile = open("stopwords.txt", "r")
for word in stopwordsfile: # a stop word in each line
word = word.replace("\n", '')
word = word.replace("\r\n", '')
STOP_WORDS_SET.add(word)
corpus = plsa.Corpus() # instantiate corpus
# iterate over the files in the directory.
document_paths = ['./texts/grimm_fairy_tales', './texts/tech_blog_posts', './texts/nyt']
#document_paths = ['./test/']
for document_path in document_paths:
for document_file in glob.glob(os.path.join(document_path, '*.txt')):
document = plsa.Document(document_file) # instantiate document
document.split(STOP_WORDS_SET) # tokenize
corpus.add_document(document) # push onto corpus documents list
corpus.build_vocabulary() # Construct a list of unique words in the corpus.
print "Vocabulary size:" + str(len(corpus.vocabulary))
print "Number of documents:" + str(len(corpus.documents))
number_of_topics = int(argv[1])
max_iterations = int(argv[2])
corpus.plsa(number_of_topics, max_iterations)
#print corpus.document_topic_prob
#print corpus.topic_word_prob
#print top-k(20) word which occurs frequent in topic-(number_of_topics)
print_topic_word_distribution(corpus, number_of_topics, 20, "./topic-word.txt")
#
print_document_topic_distribution(corpus, number_of_topics, 10, "./document-topic.txt")
def plsa(data_paths=[PATH_TO_RAW_DATA],
topics_num=TOPICS_NUM,
write_results_to=FILE_DEFAULT_PLSA_RESULTS):
""" Performs topic modeling with PLSA method. """
import plsa
import glob
import os
corpus = plsa.Corpus() # instantiate corpus
document_paths = data_paths
for document_path in document_paths:
for document_file in glob.glob(os.path.join(document_path, '*.txt')):
document = plsa.Document(document_file) # instantiate document
document.split(list_stopwords(lang='en')) # tokenize
corpus.add_document(document) # push onto corpus documents list
corpus.build_vocabulary()
corpus.plsa(TOPICS_NUM, 1)
V = len(corpus.vocabulary)
assert (TOPICS_NUM < V)
f = open(write_results_to, "w")
for k in range(TOPICS_NUM):
word_prob = corpus.topic_word_prob[k, :]
word_index_prob = []
for i in range(V):
word_index_prob.append([i, word_prob[i]])
word_index_prob = sorted(word_index_prob,
key=itemgetter(1),
reverse=True) # sort by word count
f.write("Topic #" + str(k) + ":\n")
for i in range(TOPICS_NUM):
index = word_index_prob[i][0]
f.write(corpus.vocabulary[index] + " ")
f.write("\n")
f.close()
def main(argv):
try:
document_topk = int(argv[1])
topic_topk = int(argv[2])
number_of_topics = int(argv[3])
max_iterations = int(argv[4])
if document_topk > number_of_topics:
raise Exception
except:
print "Usage: python ./main.py <document_topk> <topic_topk> <number_of_topics> <maxiteration> "
print "Necessary condition: document_topk < number_of_topics"
sys.exit(0)
# load stop words list from file
stopwordsfile = open("stopwords.txt", "r")
for word in stopwordsfile: # a stop word in each line
word = word.replace("\n", '')
word = word.replace("\r\n", '')
STOP_WORDS_SET.add(word)
stopwordsfile.close()
corpus = plsa.Corpus() # instantiate corpus
# iterate over the files in the directory.
document_paths =['./texts/txt']
#document_paths = ['./test/']
for document_path in document_paths:
for document_file in glob.glob(os.path.join(document_path, '*.txt')):
document = plsa.Document(document_file) # instantiate document
document.split(STOP_WORDS_SET) # tokenize
corpus.add_document(document) # push onto corpus documents list
corpus.build_vocabulary()
print "Vocabulary size:" + str(len(corpus.vocabulary))
print "Number of documents:" + str(len(corpus.documents))
corpus.plsa(number_of_topics, max_iterations)
# My Code from here! [Saurabh]
corpus.calculate_lts()
corpus.calculate_lte()
corpus.calculate_stat_lte()
datapath = ['./texts/txt']
doc_id = 0
line_number = 1
for document_path in datapath:
for document_file in glob.glob(os.path.join(document_path, '*.txt')):
sentenceList = []
f = open(document_file, "r")
print "Reading file ... ",os.path.basename(document_file)
print "Calculating sentence score..."
line_number = 1
for line in f:
bScore = BIGRAM_WEIGHT*bigramScore(line)
tScore = TRIGRAM_WEIGHT*trigramScore(line)
sScore = STAT_WEIGHT*statScore(line,doc_id,corpus)
sentenceList.append([line, bScore + tScore + sScore, line_number])
line_number = line_number+ 1
l = int(len(sentenceList)*TOP)
print ("Extracting the top %0.2f percent sentences for summarization." %(TOP*100))
sentenceList = sorted(sentenceList, key = lambda x: x[1], reverse = True)[:l]
sentenceList = sorted(sentenceList, key = lambda x: x[2], reverse = False)[:l]
new_filename = "./dataset/summaries/set2"+os.path.basename(document_file).split('.')[0] + "Summaries.txt"
print "Writing to file...", os.path.basename(new_filename)
s = ""
with open(new_filename, "w+") as f2:
for ele in sentenceList:
s += ele[0].encode('utf-8') + "\n"
f2.write(s)
f2.close()
print ("-------------------------------------------------------------")
f.close()
doc_id += 1
