def traverse_all_bugs(all_bugids, new_bugids, sign, train=True, weaks=None):
'''
:param all_bugids:
:param new_bugids:
:param sign: 用来标识是哪一轮实验，提取好的训练集和测试集活跃度数据会放在以sign明明的文件夹里。
:param train:
:return:
'''
start_time = time.time()
print('start_time:', start_time)
bugs_counts = {}
bug_msg_all, _ = data_helper.get_msg_all()
count = 0
# all_bugids = sorted(bug_msg_all.keys()) # 升序排好
# for id in bug_msg_all.keys():
for id in new_bugids:
print(count)
count += 1
if train:
actives = get_developer_active_sequence(bug_msg_all, all_bugids, id, all_bugids.index(id) - 1, train, weaks)
else:
actives = get_developer_active_sequence(bug_msg_all, all_bugids, id, len(all_bugids) - 1, train, weaks)
bugs_counts[id] = len(actives)
write_active_sequence_to_file(actives, sign, id)
print('计算消耗时间:', time.time() - start_time)
with open('../data/active_counts_{}.txt'.format(time.time()), 'w') as writer:
for key in bugs_counts.keys():
writer.write(str(key) + '\t' + str(bugs_counts[key]) + '\n')
print('最终结束时间:', time.time())
def _writer_little_bug_msg_all(little_bugids):
'''
这个文件是测试用的,是为了查看划分的训练集是否正确
:param little_bugids:
:return:
'''
bug_msg_all, _ = get_msg_all()
with open('../data/eval_little_bug_msg_all.txt', 'w') as writer:
for name, value in little_bugids.items():
for bugid in value:
# writer.write('{}\n'.format('\t'.join(bug_msg_all.get(bugid))))
writer.write('{}\t{}\n'.format(bugid,
bug_msg_all.get(bugid)[0]))
def count_bug_fixed_by_each_developer():
bugid_each_developer = {} # key=developers' name , value=[bug_ids]
bug_msg_all, _ = get_msg_all()
# bug_msg_all = {}
for bugid, value in bug_msg_all.items():
der = value[0] # 获取当前样本的修复者
if der in bugid_each_developer.keys():
bugid_each_developer[der].append(bugid)
else:
bugid_each_developer.setdefault(der, [bugid])
# with open('../data/num_of_bugs_fixed_by_each_developer.txt', 'w') as writer:
# for name in bugid_each_developer.keys():
# writer.write('{}\t{}\n'.format(name, len(bugid_each_developer[name])))
return bugid_each_developer
import numpy as np
import util.PATH as PATH
from util.data_helper import get_msg_all
from util.data_helper import split_dataset_by_time_windows
'''
提取窗口0的词汇表信息
'''
# 字典，key=bugid，value=
bug_msg_all, _ = get_msg_all()
windows = split_dataset_by_time_windows(bug_msg_all)
vocabulary = []
for i in range(len(windows[0])):
print(i)
with open(PATH.path_corpus + str(windows[0][i]), 'r') as reader:
for line in reader.readlines():
if line.strip() not in vocabulary:
vocabulary.append(line.strip())
with open('../data/windows/window_0_vocabulary.txt', 'w') as writer:
for word in vocabulary:
writer.write('{}\n'.format(word))
writer.write('{}\n'.format(word))
def del_invalid_developers(bug_msg_all):
'''
主要是用来检查数据集中是否还存在无效开发者
:param bug_msg_all:
:return:
'''
invalids = [
"nobody", "inbox", "webmaster", "platform", "unassigned", "issues",
"needsconfirm", "swneedsconfirm"
]
invalid_bugs = []
invalid_devs = set()
for key, value in bug_msg_all.items():
if True in list(map(lambda x: x in value[0].lower(),
invalids)): # 该条bug是由无效开发者修复的
invalid_bugs.append(key)
invalid_devs.add(value[0])
print('无效的bug条数={}'.format(len(invalid_bugs)))
print('无效的开发者数目={}'.format(len(invalid_devs)))
return invalid_bugs
if __name__ == '__main__':
bug_msg_all, _ = data_helper.get_msg_all(
) # # key=bug_id value= {assign_to creation_ts delta_ts product component}
# extract_vocabulary(bug_msg_all, PATH.path_vocabulary)
del_invalid_developers(bug_msg_all)
actives = get_developer_active_sequence(bug_msg_all, all_bugids,
id,
len(all_bugids) - 1, train)
bugs_counts[id] = len(actives)
write_active_sequence_to_file(actives, sign, id)
print('计算消耗时间:', time.time() - start_time)
with open('../data/active_counts_{}.txt'.format(time.time()),
'w') as writer:
for key in bugs_counts.keys():
writer.write(str(key) + '\t' + str(bugs_counts[key]) + '\n')
print('最终结束时间:', time.time())
if __name__ == '__main__':
# traverse_all_bugs()
bug_msg_all, _ = data_helper.get_msg_all()
time_windows = data_helper.split_dataset_by_time_windows(bug_msg_all)
for i in [0]:
traverse_all_bugs(time_windows[i], time_windows[i], sign=i,
train=True) # 针对训练集
traverse_all_bugs(time_windows[i],
time_windows[i + 1],
sign=i,
train=False) # 针对测试集
#
# bugids = sorted(bug_msg_all.keys()) # 升序排好
# actives = get_developer_active_sequence(bug_msg_all, bugids, 6533)
# for i in actives:
# print(i)