-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathjoint_dst.py
314 lines (232 loc) · 8.71 KB
/
joint_dst.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
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
from collections import defaultdict
from random import choices
from util import *
import numpy as np
import bisect
import copy
class SZ_dst:
def __init__(self, i_file, min_val, max_val):
f = open(i_file, "r")
self.all_keys = defaultdict(int)
l = f.readline()
sum_count = 0
total_pr = 0
for l in f:
l = l.strip().split(" ")
if len(l) == 1:
continue
else:
key = int(float(l[0]))
val = float(l[1])
if key >= min_val and key <= max_val:
self.all_keys[key] += val
total_pr += val
p_keys = list(self.all_keys.keys())
vals = []
for k in p_keys:
vals.append(self.all_keys[k])
sum_vals = sum(vals)
vals = [float(x)/sum_vals for x in vals]
self.p_keys = p_keys
self.pr = vals
def sample_keys(self, n):
return choices(self.p_keys, weights=self.pr,k=n)
class POPULARITY_dst:
def __init__(self, i_file, min_val, max_val):
f = open(i_file, "r")
self.popularities = defaultdict(float)
l = f.readline()
key = int(l.strip())
sum_count = 0
for l in f:
l = l.strip().split(" ")
if len(l) == 1:
self.popularities[key] = sum_count
sum_count = 0
key = int(l[0])
if key > max_val:
break
else:
if key >= min_val:
sum_count += float(l[1])
p_keys = list(self.popularities.keys())
p_vals = []
for k in p_keys:
p_vals.append(self.popularities[k])
sum_vals = sum(p_vals)
p_vals = [float(x)/sum_vals for x in p_vals]
self.p_keys = p_keys
self.probabilities = p_vals
def sample_keys(self, n):
return choices(self.p_keys, weights=self.probabilities,k=n)
class POPULARITY_SZ_dst:
def __init__(self, i_file):
pop_sz = defaultdict(lambda : defaultdict(int))
f = open(i_file, "r")
key = "-"
keys_cnt = 0
for l in f:
l = l.strip().split(" ")
if len(l) == 1:
key = int(float(l[0]))
continue
objs = float(l[1])
sz = int(float(l[0]))
pop_sz[key][sz] += objs
f.close()
self.pop_sz_vals = defaultdict(lambda : list)
self.pop_sz_prs = defaultdict(lambda : list)
sum_n_key = 0
for key in pop_sz:
sizes = list(pop_sz[key].keys())
n_key = key
self.pop_sz_vals[n_key] = sizes
sum_n_key += n_key
prs = []
for s in sizes:
prs.append(pop_sz[key][s])
sum_prs = sum(prs)
prs = [float(x)/sum_prs for x in prs]
self.pop_sz_prs[n_key] = prs
self.sample_each_popularity()
def sample_each_popularity(self):
self.samples = defaultdict(list)
self.sampled_sizes = defaultdict(list)
for k in self.pop_sz_prs:
self.sampled_sizes[k] = choices(self.pop_sz_vals[k], weights=self.pop_sz_prs[k], k=10000)
self.samples_index = defaultdict(int)
self.popularities = list(self.pop_sz_prs.keys())
self.popularities.sort()
def findnearest(self, k):
ind = bisect.bisect_left(self.popularities, k)
if ind >= len(self.popularities):
ind = len(self.popularities) - 1
return self.popularities[ind]
def sample(self, k):
if k not in self.samples_index:
k = self.findnearest(k)
curr_index = self.samples_index[k]
if curr_index >= len(self.sampled_sizes[k]):
self.sampled_sizes[k] = choices(self.pop_sz_vals[k], weights=self.pop_sz_prs[k], k=10000)
self.samples_index[k] = 0
curr_index = 0
self.samples_index[k] += 1
return int(self.sampled_sizes[k][curr_index])
class POPULARITY_SZ_dst_backup:
def __init__(self, i_file):
f = open(i_file, "r")
self.pop_sz = defaultdict(lambda: defaultdict(float))
popularities = defaultdict(float)
l = f.readline()
key = int(l.strip())
sum_count = 0
sizes = []
prs = []
for l in f:
l = l.strip().split(" ")
if len(l) == 1:
sum_prs = sum(prs)
for i in range(len(sizes)):
self.pop_sz[key][sizes[i]] = float(prs[i])/sum_prs
key = int(float(l[0]))
sizes = []
prs = []
continue
else:
sz = int(float(l[0]))
pr = float(l[1])
sizes.append(sz)
prs.append(pr)
#self.pop_sz[key][sz] += pr
popularities[key] += pr
f.close()
## Overall popularity distribution
p_keys = list(popularities.keys())
p_vals = []
for k in p_keys:
p_vals.append(popularities[k])
sum_vals = sum(p_vals)
p_vals = [float(x)/sum_vals for x in p_vals]
self.p_keys = p_keys
self.p_vals = p_vals
## Popularity based size distribution
self.pop_sz_keys = defaultdict(lambda : list)
self.pop_sz_prs = defaultdict(lambda : list)
for key in self.pop_sz:
sizes = list(self.pop_sz[key].keys())
self.pop_sz_keys[key] = sizes
prs = []
for s in sizes:
prs.append(self.pop_sz[key][s])
sum_prs = sum(prs)
prs = [float(x)/sum_prs for x in prs]
self.pop_sz_prs[key] = prs
self.sample_each_popularity()
def print_probability(self, p, k):
print(self.pop_sz[p][k])
return
def sample_each_popularity(self):
self.samples = defaultdict(list)
self.sampled_sizes = defaultdict(list)
for k1 in self.pop_sz_prs:
self.sampled_sizes[k1] = choices(self.pop_sz_keys[k1], weights=self.pop_sz_prs[k1], k=10000)
self.samples_index = defaultdict(int)
self.popularities = list(self.pop_sz_prs.keys())
self.popularities.sort()
def findnearest(self, k):
ind = bisect.bisect_left(self.popularities, k)
if ind >= len(self.popularities):
ind = len(self.popularities) - 1
return self.popularities[ind]
def sample(self, k):
if k not in self.samples_index:
k = self.findnearest(k)
curr_index = self.samples_index[k]
if curr_index >= len(self.sampled_sizes[k]):
self.sampled_sizes[k] = choices(self.pop_sz_keys[k], weights=self.pop_sz_prs[k], k=10000)
self.samples_index[k] = 0
curr_index = 0
self.samples_index[k] += 1
return int(self.sampled_sizes[k][curr_index])
def sample_keys(self, n):
return choices(self.p_keys, weights=self.p_vals,k=n)
class SampleFootPrint:
def __init__(self, fd, hr_type, min_val, max_val):
self.sd_keys = []
self.sd_vals = []
self.sd_index = defaultdict(lambda : 0)
self.SD = defaultdict(lambda : 0)
f = open(i_file, "r")
l = f.readline()
l = l.strip().split(" ")
if hr_type == "bhr":
bytes_miss = float(l[-1])
bytes_req = float(l[1])
self.SD[-1] = float(bytes_miss)/bytes_req
else:
reqs_miss = float(l[-2])
reqs = float(l[0])
self.SD[-1] = float(reqs_miss)/reqs
self.sd_index[-1] = 0
total_pr = 0
for l in f:
l = l.strip().split(" ")
sd = int(l[1])
self.SD[sd] += float(l[2])
total_pr += float(l[2])
self.sd_keys = list(self.SD.keys())
self.sd_keys.sort()
i = 1
curr_pr = 0
self.sd_pr = defaultdict()
for sd in self.sd_keys:
self.sd_vals.append(self.SD[sd])
curr_pr += self.SD[sd]
if sd >= 0:
self.sd_pr[sd] = float(curr_pr - self.SD[-1])/(1 - self.SD[-1])
self.sd_index[sd] = i
i += 1
def sample_keys(self, obj_sizes, sampled_sds, n):
return choices(self.sd_keys, weights = self.sd_vals, k = n)
def findPr(self, sd):
return self.sd_pr[sd]