def mode(labels):
return Counter(labels).most_common(1)[0][0]
def __init__(self):
self.word2idx = {}
self.idx2word = []
self.counter = Counter()
self.total = 0
def count_words(sentence):
res = sentence.replace("\n", "").lower()
return Counter(re.findall(r"[a-z0-9]+(?:'[a-z]+)?", res))
for i in range(len(output)): #按时间窗聚类
word_list = [output[i][j][0]
for j in range(len(output[i]))] #每个时间窗的突发词list
vecs = []
for k in range(num, num + count_times[i]):
vec = np.array([
word_list[j] in splitword[k] for j in range(len(word_list))
]) #突发词是否在每个文本的分词
vecs.append(vec.astype(int))
clf = Clusters(vecs)
tag = clf.classify(len(set(tags)))
tags.extend(tag)
print('第%s月聚类结果):' % (dates[i]))
result = Counter(tag)
print(result.most_common(min(5, len(result))))
if len(word_list) == 1:
classify = word_list[:]
else:
vecs_kw = []
for j in range(len(word_list)):
vec_kw = np.array([
word_list[j] in splitword[m]
for m in range(num, num + count_times[i])
])
vecs_kw.append(vec_kw.astype(int))
#计算AP
# ap = AffinityPropagation(preference=-26).fit(vecs_kw)
# cluster_centers_indices = ap.cluster_centers_indices_ # 预测出的中心点的索引，如[123,23,34]
def bulid_initial_char_prob(message):
initial_char_prob = Counter()
for word in message:
initial_char_prob[word[0]] += 1
return initial_char_prob
async def submit(self, req: Request) -> Response:
"""Handle submission.xml containing submissions to server.
First submission info is parsed and then for every action in submission
(add/modify/validate) corresponding operation is performed.
Finally submission info itself is added.
:param req: Multipart POST request with submission.xml and files
:raises: HTTPBadRequest if request is missing some parameters or cannot be processed
:returns: XML-based receipt from submission
"""
files = await _extract_xml_upload(req)
schema_types = Counter(file[1] for file in files)
if "submission" not in schema_types:
reason = "There must be a submission.xml file in submission."
LOG.error(reason)
raise web.HTTPBadRequest(reason=reason)
if schema_types["submission"] > 1:
reason = "You should submit only one submission.xml file."
LOG.error(reason)
raise web.HTTPBadRequest(reason=reason)
submission_xml = files[0][0]
submission_json = XMLToJSONParser().parse("submission", submission_xml)
# Check what actions should be performed, collect them to dictionary
actions: Dict[str, List] = {}
for action_set in submission_json["actions"]["action"]:
for action, attr in action_set.items():
if not attr:
reason = f"""You also need to provide necessary
information for submission action.
Now {action} was provided without any
extra information."""
LOG.error(reason)
raise web.HTTPBadRequest(reason=reason)
LOG.debug(f"submission has action {action}")
if attr["schema"] in actions:
set = []
set.append(actions[attr["schema"]])
set.append(action)
actions[attr["schema"]] = set
else:
actions[attr["schema"]] = action
# Go through parsed files and do the actual action
results: List[Dict] = []
db_client = req.app["db_client"]
for file in files:
content_xml = file[0]
schema_type = file[1]
if schema_type == "submission":
LOG.debug("file has schema of submission type, continuing ...")
continue # No need to use submission xml
action = actions[schema_type]
if isinstance(action, List):
for item in action:
result = await self._execute_action(schema_type, content_xml, db_client, item)
results.append(result)
else:
result = await self._execute_action(schema_type, content_xml, db_client, action)
results.append(result)
body = json.dumps(results)
LOG.info(f"Processed a submission of {len(results)} actions.")
return web.Response(body=body, status=200, content_type="application/json")
