def main():
global NUM_TOP_TOPICS
check_start('./test_doc_guess_algorithm.py <folder> ' +
'<output> <num_topics>', 4)
# Input naming
folder = sys.argv[1]
output = sys.argv[2]
num_topics = int(sys.argv[3])
topic_order_count = 0
topic_feeling_count = 0
topic_offset_count = 0
# File allocation
format_file = folder + '/initial.formatted'
vocab_file = folder + '/initial.vocab'
info_file = folder + '/out/final.other'
with open(format_file) as data_file:
format_lines = data_file.readlines()
with open(info_file) as data_file:
info_lines = data_file.readlines()
NUM_TOP_TOPICS = int(info_lines[0].split(' ')[1])
known_topics = get_known_topics(folder)
print(str(known_topics[0]))
word_dict = get_word_dictionary(folder, vocab_file)
hierarchy_struct = get_hierarchy_struct(folder)
# Cycle through top-level documents
count = 0
for index in range(len(format_lines)):
line = format_lines[index]
line = line.strip()
# known_topics = get_known_topics()
weights = get_document_values(index, hierarchy_struct)
contents = line.split('|~|')[3]
contents = clean_line(contents)
classed_doc = classify(contents, weights, word_dict)
classed_array = get_classed_array(classed_doc)
if index % 1000 == 0 and index != 0:
print('\tTopics correct position ' +
'%d: %s' % (index, str(topic_order_count / index)))
if compare_topics_order(classed_array, known_topics[index], 1):
topic_order_count += 1
count += 1
print('\tTopics correct position %d: %s' % (count, str(topic_order_count / count)))
def main():
check_start('./test_doc_guess_algorithm.py <folder> <output> <num_topics>', 4)
# Input naming
folder = sys.argv[1]
output = sys.argv[2]
num_topics = int(sys.argv[3])
topic_order_count = [0] * num_topics
topic_feeling_count = [0] * num_topics
topic_offset_count = [0] * num_topics
# File allocation
format_file = folder + '/initial.formatted'
vocab_file = folder + '/initial.vocab'
word_assignment_file = folder + '/out/word-assignments.dat'
gamma_file = folder + '/out/final.gamma'
# Read in required data
with open(vocab_file) as data_file:
vocab_lines = data_file.readlines()
with open(word_assignment_file) as data_file:
word_assignment_lines = data_file.readlines()
with open(format_file) as data_file:
format_lines = data_file.readlines()
with open(gamma_file) as data_file:
gamma_lines = data_file.readlines()
vocab_details = get_vocab_details(vocab_lines, word_assignment_lines)
topic_lines = get_document_topics(gamma_lines)
vocab_index = get_vocab_index(vocab_lines)
num_docs = len(format_lines)
for index in range(len(format_lines)):
line = format_lines[index]
line = line.strip()
known_topics = topic_lines[index].strip().split(' ')
document_contents = line.split('|~|')[3]
document_contents = clean_line(document_contents)
classed_doc = classify(document_contents, vocab_details, vocab_index)
if len(classed_doc) == 0:
print(document_contents)
continue
if index % 1000 == 0 and index != 0:
print('Snapshot: ' + str(index))
for i in range(1, num_topics):
print('\tTopics correct position %d: %s' % (i, str(topic_order_count[i] / index)))
print('\tTopics within range %d: %s' % (i, str(topic_feeling_count[i] / index)))
print('\tTopics within offset %d: %s' % (i, str(topic_offset_count[i] / index)))
for i in range(1, len(classed_doc)):
if compare_topics_order(classed_doc, known_topics, i):
topic_order_count[i] += 1
if compare_topic_slice(classed_doc, known_topics, i):
topic_feeling_count[i] += 1
if compare_position_offset(classed_doc, known_topics, i):
topic_offset_count[i] += 1
out_ptr = open(output, 'w')
out_ptr.write('Number of documents: ' + str(num_docs) + '\n')
for i in range(num_topics):
out_ptr.write('Topics correct position %d: %s\n' % (i, str(topic_order_count[i] / num_docs)))
out_ptr.write('Topics within range %d: %s\n' % (i, str(topic_feeling_count[i] / num_docs)))
out_ptr.write('Topics within offset %d: %s\n' % (i, str(topic_offset_count[i] / num_docs)))