def exe_extract_topic_words(argv):
from nltk.tokenize import word_tokenize;
topic_path, word_list_path = argv;
trec_format = StandardFormat();
word_list_file = open(word_list_path, 'w');
topics = trec_format.read(topic_path);
word_set = set();
for topic_id, topic_text in topics.items():
words = map(lambda word: word.lower(), word_tokenize(topic_text));
word_set.update(words);
word_list_file.write('\n'.join(word_set));
word_list_file.close();
