-
Notifications
You must be signed in to change notification settings - Fork 0
/
Segment.py
163 lines (150 loc) · 6.18 KB
/
Segment.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
from FlokAlgorithmLocal import FlokAlgorithmLocal, FlokDataFrame
import numpy as np
import pandas as pd
from datetime import datetime
class Segment(FlokAlgorithmLocal):
def calculate_error(st, seq_range):
x = np.arange(seq_range[0], seq_range[1] + 1)
y = np.array(st[seq_range[0]:seq_range[1] + 1])
# 计算error
x_, y_ = x.mean(), y.mean()
num = ((x - x_)*(y - y_)).sum()
den = ((x - x_)**2).sum()
_m = num/den
_b = y_ - _m*x_
error = (abs(_m*x + _b-y)).sum()/len(x)
return error
# 线性拟合
def linear(a):
x = np.arange(0, len(a))
y = np.array(a)
# 返回回归系数
x_, y_ = x.mean(), y.mean()
num = ((x - x_)*(y - y_)).sum()
den = ((x - x_)**2).sum()
_m = num/den
_b = y_ - _m*x_
return list(_m*x + _b)
# 自下而上的拟合，先把数据分成n/2段，然后合并
def Bottom_Up(T, max_error):
seg_piece = [] # 最终返回的分段
seg = [] # 储存分段两头位置
merge_cost = [] # 计算合并带来的误差
for i in range(0, len(T), 2):
seg_piece += [T[i:i + 2]]
seg.append((i, i + 1))
for i in range(0, len(seg_piece) - 2):
merge_cost.insert(i, Segment.calculate_error(
T, (seg[i][0], seg[i + 1][1])))
while min(merge_cost) < max_error:
index = merge_cost.index(min(merge_cost)) # 每次合并误差最小的片段
seg_piece[index] = Segment.linear(
seg_piece[index] + seg_piece[index + 1])
seg[index] = (seg[index][0], seg[index + 1][1])
# 合并完成后删除原值
del seg_piece[index + 1]
del seg[index + 1]
# 更新merge_cost
if index > 1:
merge_cost[index - 1] = Segment.calculate_error(
T, (seg[index - 1][0], seg[index][1]))
if index + 1 < len(merge_cost):
merge_cost[index] = Segment.calculate_error(
T, (seg[index][0], seg[index + 1][1]))
del merge_cost[index + 1]
else:
del merge_cost[index]
seg_piece[-2] = Segment.linear(seg_piece[-2] + seg_piece[-1])
del seg_piece[-1]
return seg_piece
def run(self, inputDataSets, params):
input_data = inputDataSets.get(0)
output_data = input_data
column = input_data.columns[1]
output = params.get("output", 'first')
error = params.get("error", 0.1)
if isinstance(error, str):
error = float(error)
str_ = 'segment({a},\'output\'=\'{b}\',\'error\'=\'{c}\')'.format(
a=column,b=output, c=error)
# 如果输入是等差数列直接输出
if all([((output_data[column][i] - output_data[column][i-1])-(output_data[column][1] - output_data[column][0])) < 1e-10 for i in range(1, len(output_data))]):
if output == 'all':
output_data = output_data
else:
output_data = pd.DataFrame(
{'Time': '1970-01-01 08:00:00.000', str_: output_data[column][0]}, index=[0])
else:
seg_piece = Segment.Bottom_Up(list(output_data[column]), error)
data = []
n = len(seg_piece)
Time = []
if output == 'all':
for i in range(len(seg_piece)):
data += seg_piece[i]
print(data)
print(len(data))
for i in range(len(output_data)):
Time.append(datetime.fromtimestamp((i+1)/1000.0))
Time = pd.Series(
[t.strftime('%Y-%m-%d %H:%M:%S.%f')[:-3] for t in Time])
output_data = pd.DataFrame({'Time': Time, str_: data})
else:
for i in range(len(seg_piece)):
data.append(seg_piece[i][0])
for i in range(len(data)):
Time.append(datetime.fromtimestamp((i+1)/1000.0))
Time = pd.Series(
[t.strftime('%Y-%m-%d %H:%M:%S.%f')[:-3] for t in Time])
output_data = pd.DataFrame({'Time': Time, str_: data})
result = FlokDataFrame()
result.addDF(output_data)
return result
if __name__ == "__main__":
algorithm = Segment()
all_info_1 = {
"input": ["root_test_d1"],
"inputFormat": ["csv"],
"inputLocation": ["local_fs"],
"output": ["./test_out_1.csv"],
"outputFormat": ["csv"],
"outputLocation": ["local_fs"],
"parameters": {'output': 'all', 'error': 0.1}
}
params = all_info_1["parameters"]
inputPaths = all_info_1["input"]
inputTypes = all_info_1["inputFormat"]
inputLocation = all_info_1["inputLocation"]
outputPaths = all_info_1["output"]
outputTypes = all_info_1["outputFormat"]
outputLocation = all_info_1["outputLocation"]
dataSet = algorithm.read(inputPaths, inputTypes,
inputLocation, outputPaths, outputTypes)
from SelectTimeseries import SelectTimeseries
dataSet = SelectTimeseries().run(
dataSet, {"timeseries": "Time,s3"})
result = algorithm.run(dataSet, params)
print(result.get(0))
#algorithm.write(outputPaths, result, outputTypes, outputLocation)
'''
all_info_2 = {
"input": ["./test_in.csv"],
"inputFormat": ["csv"],
"inputLocation": ["local_fs"],
"output": ["./test_out_2.csv"],
"outputFormat": ["csv"],
"outputLocation": ["local_fs"],
"parameters": {"timeseries": "Time,root.test.d2.s2",'output':'first','error':0.1}
}
params = all_info_2["parameters"]
inputPaths = all_info_2["input"]
inputTypes = all_info_2["inputFormat"]
inputLocation = all_info_2["inputLocation"]
outputPaths = all_info_2["output"]
outputTypes = all_info_2["outputFormat"]
outputLocation = all_info_2["outputLocation"]
dataSet = algorithm.read(inputPaths, inputTypes,
inputLocation, outputPaths, outputTypes)
result = algorithm.run(dataSet, params)
algorithm.write(outputPaths, result, outputTypes, outputLocation)
'''