def main():
slot2Id = getSlot2Id()
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = BertForTokenClassification.from_pretrained(
config.pretrained_model_name_or_path, num_labels=len(slot2Id))
model.to(device)
x, y = processData()
train_dataloader, val_dataloader = getDataLoader(x, y)
train(model, device, train_dataloader, val_dataloader, config.epochs,
config.max_grad_norm)
def promethoes():
test_data = b'{"receiver":"test","status":"firing","alerts":[{"status":"firing","labels":{"alertname":"\xe4\xb8\xbb\xe6\x9c\xba\xe8\xbf\x9c\xe7\xa8\x8b\xe8\xbf\x9e\xe6\x8e\xa5\xe5\xa4\xb1\xe8\xb4\xa5","host":"db-01","instance":"127.0.0.1:9066","job":"service_guard","service":"dbas"},"annotations":{"description":"db-01\xe4\xb8\xbb\xe6\x9c\xba\xe8\xbf\x9c\xe7\xa8\x8b\xe8\xbf\x9e\xe6\x8e\xa5\xe5\xa4\xb1\xe8\xb4\xa5.","summary":"db-01\xe4\xb8\xbb\xe6\x9c\xba\xe8\xbf\x9c\xe7\xa8\x8b\xe8\xbf\x9e\xe6\x8e\xa5\xe5\xa4\xb1\xe8\xb4\xa5."},"startsAt":"2019-04-03T16:22:20.137532326+08:00","endsAt":"0001-01-01T00:00:00Z","generatorURL":"http://mix-app-131-11:9090/graph?g0.expr=guard_remote_conn_up+%3D%3D+0\\u0026g0.tab=1"}],"groupLabels":{},"commonLabels":{"service":"dbas"},"commonAnnotations":{},"externalURL":"http://mix-app-131-11:9093","version":"4","groupKey":"{}:{}"}\n'
print("========收到告警==========")
# request.data = test_data
processed_alerts = process_data.processData(request.data)
for alert in processed_alerts:
status = alert["status"]
if status == "resolved":
alarm = "告警解除---"
else:
alarm = "告警触发---"
content_msg = alarm + alert["content"]
key = alert["receiver"]
alarmType = alarm + "自监控告警"
print("告警消息为：" + content_msg)
utils(content_msg, key, key, "", alarmType)
# if res:
# log_out.log_out("告警:"+" 时间: "+alert["time"]+" 内容: "+ alert["content"])
time.sleep(2)
print("========告警已发送=========")
return "告警成功"
import process_data as pd
import csv
from sklearn.naive_bayes import GaussianNB
from sklearn.model_selection import GridSearchCV
trainfile08 = 'train_2008.csv'
testfile08 = 'test_2008.csv'
perc = 75
X, Y, Xtest = pd.processData(trainfile08, testfile08, perc)
parameters = {}
def classifyAll():
model = GaussianNB()
clf = GridSearchCV(model, parameters, verbose=20)
clf = clf.fit(X, Y)
with open('NaiveBayes.csv', 'w') as csv_file:
writer = csv.writer(csv_file, delimiter=',', lineterminator='\n')
for key, value in clf.cv_results_.items():
writer.writerow([key, value])
print(clf.cv_results_)
return clf
def classifyOne():
clf = GaussianNB(priors=None)
clf = clf.fit(X, Y)
return clf
import csv
import numpy as np
from sklearn.ensemble import AdaBoostClassifier
from sklearn.model_selection import GridSearchCV
from sklearn.tree import DecisionTreeClassifier
from sklearn.feature_selection import RFECV, RFE
import process_data as pd
trainfile08 = "train_2008.csv"
testfile08 = "test_2008.csv"
testfile12 = "test_2012.csv"
perc = 100
X, Y, Xtest = pd.processData(trainfile08, testfile08, perc)
_, _, Xtest2 = pd.processData(trainfile08, testfile12, perc)
print("Data processed")
def classifyRFE():
clf = AdaBoostClassifier(
base_estimator=DecisionTreeClassifier(
max_depth=1,
max_features='log2'),
n_estimators=750
)
print("Performing RFECV")
selector = RFECV(clf, verbose=100, step=10)
selector = selector.fit(X, Y)
# Print results to file
with open("logs\\adaBDT_rfe_results_3.txt", "w") as fle:
temp_train = cvModel.predict([X_train, X_train_angle])
y_train_pred_log += temp_train.reshape(temp_train.shape[0])
y_test_pred_log = y_test_pred_log / K
y_train_pred_log = y_train_pred_log / K
print('Train Log Loss Validation= ', log_loss(y_train, y_train_pred_log))
print('Valid Log Loss Validation= ', log_loss(y_train, y_valid_pred_log))
return y_test_pred_log
### Input Model Name ###
### See README for different models ###
model = 'vgg16'
X_train, X_train_angle, X_test, X_test_angle, y_train = processData(
"train.json", "test.json", model)
predictions = transferCV(X_train,
X_train_angle,
X_test,
X_test_angle,
y_train,
model=model,
finetune=False,
finetune_layer=15)
test = pd.read_json("test.json")
predictions_df = test[['id']].copy()
predictions_df['is_iceberg'] = predictions
