def __init__(self, project_name):
self.project_name = project_name
self.cs = CosineSimilarity(self.project_name)
self.tf = TermFrequency()
class Process(object):
def __init__(self, project_name):
self.project_name = project_name
self.cs = CosineSimilarity(self.project_name)
self.tf = TermFrequency()
def get_corpus(self, train_time):
request_infos = get_request_title_description(self.project_name,
train_time)
key_list = list()
word_list = list()
for request_info in request_infos:
number = list(request_info.keys())[0]
data = list(request_info.values())[0]
valid_word = delete_stop_words(data)
word = Counter(valid_word)
key_list.append(number)
word_list.append(word)
return key_list, word_list
def get_corpus_result(self, train_time):
key_list, count_list = self.get_corpus(train_time)
start = time.time()
score_dict = {}
for key, count in zip(key_list, count_list):
word_score = {
word: self.tf.tf_idf(word, count, count_list)
for word in count
}
sorted_score = dict(
sorted(word_score.items(), key=lambda x: x[1], reverse=True))
sorted_list = list(sorted_score.keys())
score_dict.update({key: sorted_list})
end = time.time()
print('spend time is {}'.format(end - start))
return score_dict
def cos_score(self, new, past, reviewer, k):
"""
in this function will get pull request similarity.
:param new:
:param past:
:param k: the number of pull request
:return:
"""
review_dict = {}
for number, past_word in past.items():
# print('the number is',number)
lib_tech_past = list(past_word)
past_reviewer = reviewer.get(number)
if not lib_tech_past or not new or not past_reviewer:
continue
score = self.cs.score(new, past_word, number)
for rpv in past_reviewer:
if rpv not in review_dict:
review_dict.setdefault(rpv, score)
else:
past_score = review_dict.get(rpv)
review_dict.update({rpv: past_score + score})
candidate_list_sort = sorted(review_dict.items(),
key=lambda item: item[1],
reverse=True)
# strip zero score
strip_zero_value = list(filter(lambda i: i[1] > 0,
candidate_list_sort))
if len(strip_zero_value) > 0:
split_list = candidate_list_sort
# split_list = candidate_list_sort
result = [i for i in split_list]
else:
result = []
return result
def cos_test_info(self, train_time):
# get word vector
info_dict = self.get_corpus_result(train_time)
# get predict interval
reviewer = get_reviewer(self.project_name)
print(reviewer)
# test_number = get_test_number(self.project_name)
# for i in test_number:
# new_word = info_dict.get(i)
# past_dict = {k: v for k, v in info_dict.items() if (k < i)}
# score_list = self.cos_score(new_word, past_dict, reviewer, 5)
# print(score_list)
return info_dict, reviewer