-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtry.py
executable file
·245 lines (238 loc) · 7.4 KB
/
try.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
# -*- coding: utf-8 -*-
"""
Created on Thu Jun 1 21:12:24 2017
@author: WIKKI
"""
import pandas as pd
import gc
import matplotlib.pyplot as plt
import numpy as np
from pandas import Series, DataFrame
from sklearn import preprocessing
from sklearn.cross_validation import train_test_split
from keras.utils import np_utils
gc.collect()
#%%
from keras.utils import np_utils
#%%
Y16Q2 = hdd_all #通用为Q2
#%%
hdd = pd.read_csv('D:\TraceData\harddrive.csv')
hdd = hdd.query('model == "ST4000DM000"')
hdd = hdd.loc[:, ~hdd.isnull().all()]
#%%
frames = [Y16Q1, Y16Q2, Y16Q3, Y16Q4]
result = pd.concat(frames, ignore_index=True)
#%%
if all(Y16Q1.columns == Y16Q4.columns):
print('true')
else:
print('false')
#%%
temp = Y16Q2[Y16Q2['failure']==1]
temp.isnull().any()
#%%
temp.loc[:,'date'] = pd.to_datetime(temp.loc[:,'date'])
temp['day_of_year'] = temp['date'].dt.dayofyear
#%%
temp.plot(kind='scatter', x='day_of_year', y='failure', title='Hard drive failures over time')
plt.show()
#%%
temp = Y16Q2.groupby('serial_number')
temp_last_day = temp.nth(-1)
#%%
uniq_serial_num = pd.Series(temp_last_day.index.unique())
uniq_serial_num.shape
#%%
test_ids = uniq_serial_num.sample(frac=0.25)
train = temp_last_day.query('index not in @test_ids')
test = temp_last_day.query('index in @test_ids')
#%%
train_labels = train['failure']
test_labels = test['failure']
train = train.drop('failure', axis=1)
test = test.drop('failure', axis=1)
#%%
failed = Y16Q2[Y16Q2['failure']==1]
failed_serial_num = failed.loc[:,'serial_number']
#serial_num = pd.Series(failed_serial_num.index.unique())
failed_disk = Y16Q2[Y16Q2['serial_number'].isin(failed_serial_num)]
#%%
failed_disk.loc[:,'date'] = pd.to_datetime(failed_disk.loc[:,'date'])
failed_disk['date'] = failed_disk['date'].dt.dayofyear
#%%
failed_disk_sort = failed_disk.sort_values(by=['serial_number','date'],ascending=[True,False])
failed_disk_sort['status'] = 0
#%%
temp = failed_disk_sort.groupby('serial_number')
temp_last_day = temp.nth(0)
serial_num = temp_last_day.index.tolist()
fail_date = tuple(temp_last_day['date'].tolist())
dict_date = dict(zip(serial_num, fail_date))
#%%
pre = 0
for iter in failed_disk_sort.index:
if failed_disk_sort.loc[iter, 'failure'] == 1:
pre = failed_disk_sort.loc[iter, 'date']
failed_disk_sort.loc[iter, 'status'] = 0
else:
failed_disk_sort.loc[iter, 'status'] = pre - failed_disk_sort.loc[iter, 'date']
#%%
temp = failed_disk_sort.groupby('serial_number')
temp_max_left = temp['status'].max()
#%%
tt = Q2[Q2['serial_number']=='Z305ANSH']
#%%
data = [[2000, 'Ohino', 1.5], [2001, 'Ohino', 1.7], [2002, 'Ohino', 3.6], [2001, 'Nevada', 2.4], [2002, 'Nevada', 2.9]]
data = np.array(data)
df = DataFrame(data, index=['one', 'two', 'three', 'four', 'five'], columns=['year', 'state', 'pop'])
print(df)
#%%
foo = df.values
foo = df.as_matrix()
foo = np.array(df)
print(foo)
foo = df.as_matrix(['year', 'pop'])
print(foo)
#%%
frames = [Q12,Q3,Q4]
Y16 = pd.concat(frames, ignore_index=True)
#%%生成剩余天数与健康状态的对应表
left = range(51)
health = [1]*3
health.extend([2]*3)
health.extend([3]*5)
health.extend([4]*6)
health.extend([5]*7)
health.extend([6]*8)
health.extend([7]*19)
healthmap = dict(zip(left,health))
#%%Y17的index不是从0开始的，要重排以后才能使用map函数
temp = Y17.reset_index(drop=True)
#%%根据剩余天数判定健康状态
temp['status'] = Series(map(lambda x: healthmap[x], temp['left_day']))
#%%
temp = Y16.groupby('serial_number')
temp_max_left = temp['left_day'].min()
#%%
temp = Q4[Q4['serial_number']=='Z300GZ0W']
#%%检查硬盘故障后是否还在产生日志记录
pre = 'begin'
error = []
for iter in Y17.index:
serial = Y17.loc[iter, 'serial_number']
if serial != pre:
pre = serial
if Y17.loc[iter, 'failure'] == 0:
error.append(serial)
#%%找到值始终不变的SMART属性
temp = Y16.dropna()
col = []
for iter in temp.columns:
if temp[iter].value_counts().shape[0] == 1:
col.append(iter)
temp.drop(col,axis=1,inplace=True)
#%%
tt = failed_disk_sort[failed_disk_sort['serial_number']=='Z300GZ0W']
#%%
tem = temp[temp['status'].isnull()]
#%%
#y_data = [c - 1 for c in y_data]
y_test = np_utils.to_categorical(y_data, 6)
#%%
Y16 = pd.read_pickle('/home/jwang/Documents/processed/Y2')
Y17 = pd.read_pickle('/home/jwang/Documents/processed/Y17')
#%%
frames = [Y16, Y17]
Y2 = pd.concat(frames, ignore_index=True)
Y2.to_pickle('/home/jwang/Documents/processed/Y2')
#%%注意Y16的index少掉一位，要重排
dr = ['smart_183_normalized','smart_183_raw','date','failure','left_day']
npdata = Y16.drop(dr,axis=1)
npdata = np.array(npdata)
#%%对各个feature的数值归一化
temp = npdata[:,1:-1]
min_max_1 = preprocessing.MinMaxScaler()
temp = min_max_1.fit_transform(temp)
npdata[:,1:-1] = temp
#%%查找每个硬盘最后一条日志对应的行数
serial = []
idx = []
pre = 'begin'
l = len(npdata)
for i in range(0, l):
if npdata[i,0] != pre:
serial.append(pre)
idx.append(i-1)
pre = npdata[i,0]
idx.append(i)
serial.append(pre)
last_idx = dict(zip(serial,idx))
#%%制作timesteps为20的数据集
now = 'begin'
x_data = []
y_data = []
for i in range(0,l):
if npdata[i,0] != now:
now = npdata[i,0]
last = last_idx[now]
if npdata[i,-1]<7:
end = min(i+19, last) + 1
sliced = npdata[i:end,1:-1]
x_data.append(sliced)
y_data.append(npdata[i,-1])
#%% padding with zeros
data = x_data
for i in range(0,len(data)):
result = data[i]
result = np.pad(result, ((0,20-result.shape[0]),(0,0)), 'constant', constant_values=0)
data[i] = result
x_data = np.array(data)
#%%划分测试集与训练集
X_train, X_test, y_train, y_test = train_test_split(x_data, y_data, test_size=0.2, random_state=42)
#%%
temp = npdata[:,1]
temp.reshape(-1,1)
temp = preprocessing.minmax_scale(temp)
#%%
tt = data[45]
tt = tt[:,1:]
print(tt)
min_max_scaler = preprocessing.MinMaxScaler()
X_train_minmax = min_max_scaler.fit_transform(tt)
#%%
Y16.to_pickle('/home/wikki/Documents/processed/Y16.pkl')
Y17.to_pickle('/home/wikki/Documents/processed/Y17.pkl')
#%%
x_data.save('x_data.npy',a)
#%%
Y16 = pd.read_pickle('/home/wikki/Documents/processed/Y16')
#%%
Y16.to_csv('Y16.csv')
Y17.to_csv('Y17.csv')
#%%
foo = np.array(df)
fo1 = np.array(df)
fo1[0,0] = 2017
ll = []
ll.append(foo)
ll.append(fo1)
#%%
dummy_y = np_utils.to_categorical(y_data, num_classes=6)
dummy_y = dummy_y[:,1:]
#%%
tt = np.array([[ 0.16436073, 0.1616614, 0.19381061, 0.15829226, 0.16597888, 0.15589602],[ 0.16255681, 0.16395631, 0.1929483, 0.16396037, 0.16037005, 0.15620814],[ 0.16461335, 0.16227967, 0.19619974, 0.15777504, 0.16453448, 0.15459777],[ 0.15914862, 0.16232671, 0.20141737, 0.16303335, 0.16228095, 0.15179311],[ 0.15729198, 0.16542646, 0.19662228, 0.16531138, 0.16239175, 0.15295611]])
t1 = tt[:,:5]
t2 = tt[:,5]
#%%
r = [[1,3,3]]
r.append([1,2,5])
r = np.asarray(r)
tt = np.mean(r, axis=0)
#%%
y_true = [0, 1, 2, 0, 1, 2]
y_pred = [0, 2, 1, 0, 0, 1]
r = metrics.f1_score(y_true, y_pred, average=None).tolist()
r = r + metrics.recall_score(y_true, y_pred, average=None).tolist()
r.append(metrics.precision_score(y_true, y_pred, average='macro'))
#classification_report(y_true, y_pred, average='macro')