/
SVD_Quarantine.py
285 lines (255 loc) · 12.3 KB
/
SVD_Quarantine.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
"""
Use SVD with chi-squared analysis to iteratively identify and remove outliers from datasets.
In the time-resolved data, outliers frequently come from variances in x-ray and laser intensity,
as well as from timing errors (I don't think we have any of those in this data though).
"""
from collections import namedtuple
from os import listdir, makedirs, path
import os
import shutil
from sys import argv
import math
import pickle as pkl
import pprint as pprint
import warnings
import matplotlib
matplotlib.use("MacOSX")
from matplotlib import pyplot as plt
import numpy as np
from numpy.linalg import svd
import scipy.stats.mstats
import parse
######################################
# Constants that may relate to geometry or dataset issues.
######################################
QMIN = 0.03
QMAX = 4.28
SKIP_INDEX = 30
OUTLIER_CUTOFF = 2.5
def make_dir(f):
if not os.path.exists(f):
os.makedirs(f)
######################################
# Dataset information for loading in data
######################################
TIMES = [ "562ns", "750ns", "1us", "1.33us", "1.78us", "2.37us", "3.16us", "4.22us", "5.62us",
"7.5us", "10us", "13.3us", "17.8us", "23.7us", "31.6us", "42.2us", "56.2us", "75us", "100us",
"133us", "178us", "237us", "316us", "422us", "562us", "750us", "1ms"]
REPS = range(0,50)
PREFIX_LIST = ["CypA-WT-1",
"CypA-WT-Buffer-1",
"CypA-S99T-2",
"CypA-S99T-Buffer-2",
"CypA-1",
"CypA-Buffer-1",
"CypA-2",
"CypA-Buffer-2",
"CypA-3",
"CypA-Buffer-3",
"CypA-4",
"CypA-Buffer-4",
"CypA-5",
"CypA-Buffer-5",
"CypA-WT-1",
"CypA-WT-Buffer-1",
"CypA-WT-2",
"CypA-WT-Buffer-2",
"CypA-WT-3",
"CypA-WT-Buffer-3",
"CypA-WT-4",
"CypA-WT-Buffer-4",
"CypA-NH-1",
"CypA-NH-Buffer-1",
"CypA-H-1",
"CypA-N-1"]
directories = ["/Volumes/DatumsDepot/2016/Mike/APS_20160701/July_Beamline_Trip/Analysis/common/integration/CypA-WT/CypA-WT-1/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20160701/July_Beamline_Trip/Analysis/common/integration/CypA-WT/CypA-WT-Buffer-1/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20160701/July_Beamline_Trip/Analysis/common/integration/CypA-S99T/CypA-S99T-2/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20160701/July_Beamline_Trip/Analysis/common/integration/CypA-S99T/CypA-S99T-Buffer-2/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-1/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-Buffer-1/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-2/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-Buffer-2/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-3/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-Buffer-3/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-4/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-Buffer-4/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-5/xray_images",
"/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-Buffer-5/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-1/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-Buffer-1/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-2/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-Buffer-2/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-3/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-Buffer-3/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-4/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-Buffer-4/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-NH-1/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-NH-Buffer-1/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-H-1/xray_images",
"/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-N-1/xray_images"]
# PREFIX_LIST = [
# "CypA-WT-1",
# "CypA-WT-Buffer-1",
# "CypA-WT-3"]
# #"/Volumes/DatumsDepot/2016/Mike/APS_20160701/July_Beamline_Trip/Analysis/common/integration/CypA-WT/CypA-WT-1/xray_images",
# directories = [
# "/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-1/xray_images",
# "/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-Buffer-1/xray_images",
# "/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-3/xray_images"]
######################################
# Load reference
######################################
length = 0
reference = parse.parse("/Volumes/DatumsDepot/2017/Mike/APS_20170302/Analysis/WAXS/common/integration/CypA/CypA-WT-Buffer-2/xray_images/CypA-WT-Buffer-2_9_-10us-22.tpkl")
# reference = parse.parse("/Volumes/DatumsDepot/2016/Mike/APS_20161110/Analysis/WAXS/common/integration/CypA/CypA-Buffer-4/xray_images/CypA-Buffer-4_9_-10us-22.tpkl")
# fig,ax = plt.subplots()
# files = []
Dataset = namedtuple('Dataset', ['directory', 'onavg', 'onstd', 'offavg', 'offstd'])
datasets = []
## Iterate the whole process over each directory.
for directory_index, directory in enumerate(directories):
all_vectors = []
subtracted_vectors = {i: [] for i in TIMES}
print "================="
print "Starting directory {}".format(directory)
print "================="
######################################
# Replace files from old quarantines before
######################################
print "Moving quarantined data out of quarantine dir"
quarantine_dir = "{}/_QUARANTINE".format(directory)
print quarantine_dir
if os.path.exists(quarantine_dir):
files = os.listdir(quarantine_dir)
for f in files:
src = "{}/{}".format(quarantine_dir, f)
dst = "{}/{}".format(directory, f)
shutil.move(src,dst)
print "Copying old quarantined data out of quarantine"
old_quarantine_dir = "{}/_QUARANTINED_DATA".format(directory)
print old_quarantine_dir
if os.path.exists(old_quarantine_dir):
files = os.listdir(old_quarantine_dir)
for f in files:
src = "{}/{}".format(old_quarantine_dir, f)
dst = "{}/{}".format(directory, f)
shutil.copy(src,dst)
######################################
# Parse, Load, Scale Data
######################################
print "Loading Data"
for i in REPS:
for index, time in enumerate(TIMES):
try:
onstring = "{0}/{1}_{2}_{3}.tpkl".format(directory, PREFIX_LIST[directory_index], i+1, time)
on = parse.parse(onstring)
# on_scaled = on.scale(reference, qmin=QMIN, qmax=QMAX, approach="algebraic")
on_scaled = on.SA, on.sigSA
if index > 0:
off_count = "-{}".format(index+1)
else:
off_count = ""
offstring = "{0}/{1}_{2}_-10us{3}.tpkl".format(directory, PREFIX_LIST[directory_index], i+1, off_count)
off = parse.parse(offstring)
# off_scaled = off.scale(reference, qmin=QMIN, qmax=QMAX, approach="algebraic")
off_scaled = off.SA, off.sigSA
subtracted = [(on_scaled[0][j] - off_scaled[0][j], np.sqrt(on_scaled[1][j]**2 + off_scaled[1][j]**2)) for j in range(len(on.q))]
all_vectors.append((off_scaled[0][SKIP_INDEX:], onstring, offstring))
all_vectors.append((on_scaled[0][SKIP_INDEX:], onstring, offstring))
subtracted_vectors[time].append(subtracted)
except:
try:
onstring = "{0}/{1}_{2}_{3}_on.tpkl".format(directory, PREFIX_LIST[directory_index], i+1, time)
on = parse.parse(onstring)
# on_scaled = on.scale(reference, qmin=QMIN, qmax=QMAX, approach="algebraic")
on_scaled = on.SA, on.sigSA
if index > 0:
off_count = "-{}".format(index+1)
else:
off_count = ""
offstring = "{0}/{1}_{2}_-10us{3}_on.tpkl".format(directory, PREFIX_LIST[directory_index], i+1, off_count)
off = parse.parse(offstring)
# off_scaled = off.scale(reference, qmin=QMIN, qmax=QMAX, approach="algebraic")
off_scaled = off.SA, off.sigSA
subtracted = [(on_scaled[0][j] - off_scaled[0][j], np.sqrt(on_scaled[1][j]**2 + off_scaled[1][j]**2)) for j in range(len(on.q))]
all_vectors.append((off_scaled[0][SKIP_INDEX:], onstring, offstring))
all_vectors.append((on_scaled[0][SKIP_INDEX:], onstring, offstring))
subtracted_vectors[time].append(subtracted)
except:
print "One or both of the on/off pairs was missing"
print onstring
######################################
# SVD and Outlier Detection
######################################
print "Running SVD"
matrix = np.matrix([i[0] for i in all_vectors]).transpose()
u,s,v = svd(matrix, full_matrices=False)
zero_vector_values_on = v.tolist()[0][1::2]
zero_vector_values_off = v.tolist()[0][0::2]
average_on = np.average(zero_vector_values_on)
std_on = np.std(zero_vector_values_on)
average_off = np.average(zero_vector_values_off)
std_off = np.std(zero_vector_values_off)
print average_on, average_off, std_on, std_off
datasets.append(Dataset(directory, average_on, std_on, average_off, std_off))
quarantine_set = set()
print "Finding Outliers"
for index, value in enumerate(v.tolist()[0]):
# print all_vectors[index][1]
if index % 2 == 0:
if ((value - average_off) / std_off) > OUTLIER_CUTOFF:
# print all_vectors[index][1]
print all_vectors[index][2]
quarantine_set.add(all_vectors[index][1])
quarantine_set.add(all_vectors[index][2])
elif index % 2 == 1:
if math.fabs(((value - average_on) / std_on)) > OUTLIER_CUTOFF:
print all_vectors[index][1]
# print all_vectors[index][2]
quarantine_set.add(all_vectors[index][1])
quarantine_set.add(all_vectors[index][2])
######################################
# Quarantine Outliers
######################################
print "Moving outliers into {}".format(quarantine_dir)
make_dir(quarantine_dir) # I made this directory above
if quarantine_set:
for file in quarantine_set:
basename = path.basename(file)
print basename
dst = "{}/{}".format(quarantine_dir, basename)
try:
shutil.move(file, dst)
except:
warnings.warn("Move from {} to {} failed - please revisit manually".format(file,dst))
quarantine_set.clear()
del quarantine_set
######################################
# Plots
######################################
print "Plotting Output"
fig, ax = plt.subplots()
i = 0
for vector in v.tolist()[0:1]:
# print vector
# ax.plot(range(len(vectors)), [value+i for value in vector], "-")
ax.plot([value+i*0.3 for value in vector], "-")
i+=1
fig.savefig("timepoints_{}.png".format(PREFIX_LIST[directory_index]))
fig2, ax2 = plt.subplots()
j = 0
for vector in u.transpose().tolist()[0:1]:
# print vector
# ax.plot(range(len(vectors)), [value+i for value in vector], "-")
x = [(i+7)*0.0025 for i in range(len(vector))]
ax2.plot(x, [value for value in vector], "-")
j+=0.3
fig2.savefig("Singular_vectors_WT_HD_Unscaled_{}.png".format(PREFIX_LIST[directory_index]))
fig3, ax3= plt.subplots()
ax3.plot([np.log(i) for i in s][0:10], "-")
fig3.savefig("Singular_values_WT_HD_Unscaled_{}.png".format(PREFIX_LIST[directory_index]))
with open("dataset_statistics.csv", 'wb') as f:
f.write("Directory,On Average,On STD,Off Average,Off STD\n")
for d in datasets:
f.write("{},{},{},{},{}\n".format(d.directory, d.onavg, d.onstd, d.offavg, d.offstd))