forked from robevans/minf
/
demos.py
258 lines (227 loc) · 18.1 KB
/
demos.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
__author__ = 'Robert Evans'
import master as m
import numpy as np
import quats
import pca
import plot
import progressbar
import random
import armExercisesDatabase
import parallelSimilarityMatrix
def armExercisesIndividualSimMatrix(db=None, subjectNumber=2):
if db is None:
db = armExercisesDatabase.db(True)
LRclasses = {key:value for (key,value) in zip(db.LDsegs.keys(),[l[subjectNumber] for l in db.LDsegs.values()])}
classes = {}
for key,value in LRclasses.iteritems():
for segment in value:
classes.setdefault(key[:-1], []).append(segment) # Merge left and right hand motions together
averageWeight = [float(sum(t))/len(t) for t in zip(*[[float(sum(t))/len(t) for t in zip(*l)] for l in db.explainedVariances.values()])]
weights = {key:[averageWeight]*len(value) for key,value in classes.iteritems()}
parallelSimilarityMatrix.averageSimilarityMatrix(classes, weights, "Subject C: PCA data distances", savePlot=True)
def armExercisesHDParallelsimMatrix(db=None):
if db is None:
db = armExercisesDatabase.db(True)
LRclasses = {key:value for (key,value) in zip(db.HDsegs.keys(),[sum(l,[]) for l in db.HDsegs.values()])}
classes = {}
for key,value in LRclasses.iteritems():
for segment in value:
classes.setdefault(key[:-1], []).append(segment) # Merge left and right hand motions together
weights = {key:[[1]*np.shape(segments[0])[1]]*len(segments) for (key,segments) in zip(classes.keys(),classes.values())}
parallelSimilarityMatrix.averageSimilarityMatrix(classes, weights, "High Dimensional Arm Cluster Comparisons", savePlot=True)
def armExercisesLDParallelsimMatrix(db=None):
if db is None:
db = armExercisesDatabase.db(True)
LRclasses = {key:value for (key,value) in zip(db.LDsegs.keys(),[sum(l,[]) for l in db.LDsegs.values()])}
classes = {}
for key,value in LRclasses.iteritems():
for segment in value:
classes.setdefault(key[:-1], []).append(segment) # Merge left and right hand motions together
averageWeight = [float(sum(t))/len(t) for t in zip(*[[float(sum(t))/len(t) for t in zip(*l)] for l in db.explainedVariances.values()])]
weights = {key:[averageWeight]*len(value) for key,value in classes.iteritems()}
parallelSimilarityMatrix.averageSimilarityMatrix(classes, weights, "Arm Exercise Clusters - Low Dimensional Distances", savePlot=True)
def armExercisesHDsimMatrix(db=None):
if db is None:
db = armExercisesDatabase.db(True)
LRclasses = {key:random.sample(value,1) for (key,value) in zip(db.HDsegs.keys(),[sum(l,[]) for l in db.HDsegs.values()])}
classes = {}
for key,value in LRclasses.iteritems():
for segment in value:
classes.setdefault(key[:-1], []).append(segment) # Merge left and right hand motions together
weights = {key:[[1]*np.shape(segments[0])[1]]*len(segments) for (key,segments) in zip(classes.keys(),classes.values())}
m.averageSimilarityMatrix(classes, weights, "Distances between clusters of exercises", savePlot=True)
def simMatrixAverageClassDistancesHighDimArmEx(db=None):
if db is None:
db = armExercisesDatabase.db(True)
classes = {key:value[1] for (key,value) in zip(db.data,db.segs.values())}
classes = {key[:-1]:classes[key] for key in classes.keys() if 'l' in key} # Left hand motions only
weights = {key:[[1]*np.shape(segments[0])[1]]*len(segments) for (key,segments) in zip(classes.keys(),classes.values())}
m.averageSimilarityMatrix(classes, weights, "Distances between clusters of exercises with left arm", savePlot=True)
def similarityMatrixForLowDimArmExercisesLeftHand(segmentsFromEach=1, database=None):
if database is None:
data=armExercisesDatabase.db(True)
else:
data = database
segs10l=random.sample(data.segs['10l'][1],segmentsFromEach)
segs20l=random.sample(data.segs['20l'][1],segmentsFromEach)
segs30l=random.sample(data.segs['30l'][1],segmentsFromEach)
segs40l=random.sample(data.segs['40l'][1],segmentsFromEach)
segs50l=random.sample(data.segs['50l'][1],segmentsFromEach)
segs60l=random.sample(data.segs['60l'][1],segmentsFromEach)
segs70l=random.sample(data.segs['70l'][1],segmentsFromEach)
segs80l=random.sample(data.segs['80l'][1],segmentsFromEach)
segs90l=random.sample(data.segs['90l'][1],segmentsFromEach)
names = ["10"]*segmentsFromEach+["20"]*segmentsFromEach+["30"]*segmentsFromEach+["40"]*segmentsFromEach+["50"]*segmentsFromEach+["60"]*segmentsFromEach+["70"]*segmentsFromEach+["80"]*segmentsFromEach+["90"]*segmentsFromEach
weights = [float(sum(t))/float(len(t)) for t in zip(data.pca['10l'][1][1],data.pca['20l'][1][1],data.pca['30l'][1][1],data.pca['50l'][1][1],data.pca['60l'][1][1],data.pca['70l'][1][1],data.pca['80l'][1][1],data.pca['90l'][1][1])]
print "Calculating similarities..."
m.similarityMatrix(segs10l+segs20l+segs30l+segs40l+segs50l+segs60l+segs70l+segs80l+segs90l,names,weights,"Arm exercises, Low Dim",savePlot=True)
def similarityMatrixForExpandingCirclesHighDim():
print "Reading data..."
bar = progressbar.ProgressBar(maxval=12, widgets=[progressbar.Bar('=', '[', ']'), ' ', progressbar.Percentage()])
bar.start(); progress = 0
data0a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius0cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data5a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius5cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data10a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius10cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data15a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius15cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data20a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius20cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data25a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius25cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data0b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius0cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data5b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius5cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data10b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius10cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data15b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius15cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data20b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius20cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data25b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius25cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
bar.finish()
print "Calculating segments..."
bar = progressbar.ProgressBar(maxval=12, widgets=[progressbar.Bar('=', '[', ']'), ' ', progressbar.Percentage()])
bar.start(); progress = 0
segs0a=m.getHighDimensionalSegments(data0a); progress += 1; bar.update(progress)
segs5a=m.getHighDimensionalSegments(data5a); progress += 1; bar.update(progress)
segs10a=m.getHighDimensionalSegments(data10a); progress += 1; bar.update(progress)
segs15a=m.getHighDimensionalSegments(data15a); progress += 1; bar.update(progress)
segs20a=m.getHighDimensionalSegments(data20a)[:-1]; progress += 1; bar.update(progress)
segs25a=m.getHighDimensionalSegments(data25a)[1:]; progress += 1; bar.update(progress)
segs0b=m.getHighDimensionalSegments(data0b)[1:-1]; progress += 1; bar.update(progress)
segs5b=m.getHighDimensionalSegments(data5b)[1:]; progress += 1; bar.update(progress)
segs10b=m.getHighDimensionalSegments(data10b)[:-1]; progress += 1; bar.update(progress)
segs15b=m.getHighDimensionalSegments(data15b)[1:-1]; progress += 1; bar.update(progress)
segs20b=m.getHighDimensionalSegments(data20b)[1:-1]; progress += 1; bar.update(progress)
segs25b=m.getHighDimensionalSegments(data25b)[1:-1]; progress += 1; bar.update(progress)
bar.finish()
names = ["0a"]*len(segs0a)+["5a"]*len(segs5a)+["10a"]*len(segs10a)+["15a"]*len(segs15a)+["20a"]*len(segs20a)+["25a"]*len(segs25a)+["0c"]*len(segs0b)+["5c"]*len(segs5b)+["10c"]*len(segs10b)+["15c"]*len(segs15b)+["20c"]*len(segs20b)+["25c"]*len(segs25b)
#names = ["0a"]*len(segs0a)+["0c"]*len(segs0b)+["5a"]*len(segs5a)+["5c"]*len(segs5b)+["10a"]*len(segs10a)+["10c"]*len(segs10b)+["15a"]*len(segs15a)+["15c"]*len(segs15b)+["20a"]*len(segs20a)+["20c"]*len(segs20b)+["25a"]*len(segs25a)+["25c"]*len(segs25b)
weights = [1]*np.shape((segs0a+segs0b+segs5a+segs5b+segs10a+segs10b+segs15a+segs15b+segs20a+segs20b+segs25a+segs25b)[0])[1]
#names = [0]*len(segs0a)+[10]*len(segs10a)+[20]*len(segs20a)
#weights = [1]*np.shape((segs0a+segs10a+segs20a)[0])[1]
print "Calculating similarities..."
m.similarityMatrix(segs0a+segs5a+segs10a+segs15a+segs20a+segs25a+segs0b+segs5b+segs10b+segs15b+segs20b+segs25b,names,weights,"Similarity Matrix using high dimensional raw sensor data",savePlot=True)
#m.similarityMatrix(segs0a+segs0b+segs5a+segs5b+segs10a+segs10b+segs15a+segs15b+segs20a+segs20b+segs25a+segs25b,names,weights,"Similarity Matrix using high dimensional raw sensor data",savePlot=True)
#m.similarityMatrix(segs0a+segs5a+segs10a+segs15a+segs20a+segs25a,names,weights,"Similarity Matrix using high dimensional raw sensor data",savePlot=True)
#m.similarityMatrix(segs0a+segs10a+segs20a,names,weights,"Similarity Matrix using high dimensional raw sensor data",savePlot=True)
def similarityMatrixForExpandingCirclesLowDim():
print "Reading data..."
bar = progressbar.ProgressBar(maxval=12, widgets=[progressbar.Bar('=', '[', ']'), ' ', progressbar.Percentage()])
bar.start(); progress = 0
data0a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius0cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data5a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius5cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data10a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius10cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data15a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius15cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data20a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius20cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data25a=m.readRaw("captures/mechanicalArm/ArmLength60cm/Anticlockwise/radius25cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data0b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius0cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data5b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius5cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data10b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius10cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data15b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius15cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data20b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius20cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
data25b=m.readRaw("captures/mechanicalArm/ArmLength60cm/Clockwise/radius25cmHorizontal.txt")[26:,4:]; progress += 1; bar.update(progress)
bar.finish()
print "Calculating segments..."
bar = progressbar.ProgressBar(maxval=12, widgets=[progressbar.Bar('=', '[', ']'), ' ', progressbar.Percentage()])
bar.start(); progress = 0
(segs0a,wa0)=m.getLowDimensionalSegments(data0a,n_components=2); progress += 1; bar.update(progress)
(segs5a,wa5)=m.getLowDimensionalSegments(data5a,n_components=2); progress += 1; bar.update(progress)
(segs10a,wa10)=m.getLowDimensionalSegments(data10a,n_components=2); progress += 1; bar.update(progress)
(segs15a,wa15)=m.getLowDimensionalSegments(data15a,n_components=2); progress += 1; bar.update(progress)
(segs20a,wa20)=m.getLowDimensionalSegments(data20a,n_components=2); progress += 1; bar.update(progress)
(segs25a,wa25)=m.getLowDimensionalSegments(data25a,n_components=2); progress += 1; bar.update(progress)
(segs0b,wb0)=m.getLowDimensionalSegments(data0b,n_components=2); progress += 1; bar.update(progress)
(segs5b,wb5)=m.getLowDimensionalSegments(data5b,n_components=2); progress += 1; bar.update(progress)
(segs10b,wb10)=m.getLowDimensionalSegments(data10b,n_components=2); progress += 1; bar.update(progress)
(segs15b,wb15)=m.getLowDimensionalSegments(data15b,n_components=2); progress += 1; bar.update(progress)
(segs20b,wb20)=m.getLowDimensionalSegments(data20b,n_components=2); progress += 1; bar.update(progress)
(segs25b,wb25)=m.getLowDimensionalSegments(data25b,n_components=2); progress += 1; bar.update(progress)
segs20a=segs20a[:-1]
segs25a=segs25a[1:]
segs0b=segs0b[1:-1]
segs5b=segs5b[1:]
segs10b=segs10b[:-1]
segs15b=segs15b[1:-1]
segs20b=segs20b[1:-1]
segs25b=segs25b[1:-1]
names = ["0a"]*len(segs0a)+["5a"]*len(segs5a)+["10a"]*len(segs10a)+["15a"]*len(segs15a)+["20a"]*len(segs20a)+["25a"]*len(segs25a)+["0c"]*len(segs0b)+["5c"]*len(segs5b)+["10c"]*len(segs10b)+["15c"]*len(segs15b)+["20c"]*len(segs20b)+["25c"]*len(segs25b)
weights = [float(sum(t))/float(len(t)) for t in zip(wa0,wa5,wa10,wa15,wa20,wa25,wb0,wb5,wb10,wb15,wb20,wb25)]
bar.finish()
print "Calculating similarities..."
m.similarityMatrix(segs0a+segs5a+segs10a+segs15a+segs20a+segs25a+segs0b+segs5b+segs10b+segs15b+segs20b+segs25b,names,weights,"Similarity Matrix using two principal components",savePlot=True)
def highDimRawSimilarity():
raw_h_large=m.getHighDimensionalSegments(m.readCSVfile("captures/HorizontalArmSpin-Jibran.csv"))
raw_h_small=m.getHighDimensionalSegments(m.readCSVfile("captures/HorizontalArmSpinLittleCircles-Jibran.csv"))
names = ['Big','Big','Big','Big','Big','Big','Big','Small','Small','Small','Small','Small','Small','Small','Small','Small']
weights = [1]*np.shape((raw_h_large+raw_h_small)[0])[1]
m.similarityMatrix(raw_h_large+raw_h_small,names,weights,"Similarity Matrix using high dimensional raw sensor data")
def highDimQuaternionSimilarity(euler=False):
quats_h_large=m.getQuaternionSegmentsByRawData(m.readCSVfile("captures/HorizontalArmSpin-Jibran.csv"),quats.rearrangeQuatsForLatentSpaceAlgorithm(quats.rawDataFileToQuats("captures/raw/HorizontalArmSpin-Jibran"),euler))
quats_h_small=m.getQuaternionSegmentsByRawData(m.readCSVfile("captures/HorizontalArmSpinLittleCircles-Jibran.csv"),quats.rearrangeQuatsForLatentSpaceAlgorithm(quats.rawDataFileToQuats("captures/raw/HorizontalArmSpinLittleCircles-Jibran"),euler))
names = ['Big','Big','Big','Big','Big','Big','Big','Small','Small','Small','Small','Small','Small','Small','Small','Small']
weights = [1]*np.shape((quats_h_large+quats_h_small)[0])[1]
if euler:
title="Similarity Matrix using high dimensional euler angle data"
else:
title="Similarity Matrix using high dimensional quaternion data"
m.similarityMatrix(quats_h_large+quats_h_small,names,weights,"Similarity Matrix using high dimensional euler angle data")
def lowDimQuaternionSimilarity(n_components=3,euler=False):
(quats_h_large,weights_large)=quats.doPCAonQuats("captures/raw/HorizontalArmSpin-Jibran",euler,n_components=n_components)
(quats_h_small,weights_small)=quats.doPCAonQuats("captures/raw/HorizontalArmSpinLittleCircles-Jibran",euler,n_components=n_components)
segments_quats_h_large=m.getQuaternionSegmentsByRawData(m.readCSVfile("captures/HorizontalArmSpin-Jibran.csv"),quats_h_large)
segments_quats_h_small=m.getQuaternionSegmentsByRawData(m.readCSVfile("captures/HorizontalArmSpinLittleCircles-Jibran.csv"),quats_h_small)
names = ['Big','Big','Big','Big','Big','Big','Big','Small','Small','Small','Small','Small','Small','Small','Small','Small']
weights = [1]*np.shape((segments_quats_h_large+segments_quats_h_small)[0])[1] #TODO: AVERAGE THE WEIGHTS!!
if euler:
title="Similarity Matrix using low dimensional euler angle data"
else:
title="Similarity Matrix using low dimensional quaternion data"
m.similarityMatrix(segments_quats_h_large+segments_quats_h_small,names,weights,title)
def lowDimRawSimilarity(n_components=3):
(raw_h_large,weights_large)=m.getLowDimensionalSegments(m.readCSVfile("captures/HorizontalArmSpin-Jibran.csv"),n_components)
(raw_h_small,weights_small)=m.getLowDimensionalSegments(m.readCSVfile("captures/HorizontalArmSpinLittleCircles-Jibran.csv"),n_components)
names = ['Big','Big','Big','Big','Big','Big','Big','Small','Small','Small','Small','Small','Small','Small','Small','Small']
averageWeights = [(a+b)/2.0 for (a,b) in zip(weights_large,weights_small)]
m.similarityMatrix(raw_h_large+raw_h_small,names,averageWeights,"Similarity Matrix using raw sensor data projected to the latent space")
def allMotionslowDimRawSimilarity(n_components=3,savePlot=False,title="Similarity matrix"):
# Load files and get segments
(raw_v1,weights_v1)=m.getLowDimensionalSegments(m.readCSVfile("captures/VerticalArmSpin-Dan.csv"),n_components)
(raw_v2,weights_v2)=m.getLowDimensionalSegments(m.readCSVfile("captures/VerticalArmSpin-Jibran.csv"),n_components)
(raw_h1,weights_h1)=m.getLowDimensionalSegments(m.readCSVfile("captures/HorizontalArmSpin-Dan.csv"),n_components)
(raw_h2,weights_h2)=m.getLowDimensionalSegments(m.readCSVfile("captures/HorizontalArmSpin-Jibran.csv"),n_components)
(raw_h3,weights_h3)=m.getLowDimensionalSegments(m.readCSVfile("captures/HorizontalArmSpinLittleCircles-Jibran.csv"),n_components)
# Discard bad segments
raw_h1 = raw_h1[4:]
raw_v2 = raw_v2[1:5]
# Make name labels
namesV1 = ["V1"]*len(raw_v1)
namesV2 = ["V2"]*len(raw_v2)
namesH1 = ["H1"]*len(raw_h1)
namesH2 = ["H2"]*len(raw_h2)
namesH3 = ["H3"]*len(raw_h3)
names = namesV1+namesV2+namesH1+namesH2+namesH3
# Calculate average weights
averageWeights = [(a+b+c+d+e)/5.0 for (a,b,c,d,e) in zip(weights_v1,weights_v2,weights_h1,weights_h2,weights_h3)]
# Crunch similarity matrix
m.similarityMatrix(raw_v1+raw_v2+raw_h1+raw_h2+raw_h3,names,averageWeights,title,savePlot)
def plotPCA3d(filename):
(lowDims,v)=pca.pcaFromFile(filename,n_components=2)
title = "".join(filename.split('/')[-1].split('.')[:-1])
plot.plotPCA3D(lowDims,title)
def plotSegments():
m.getLowDimensionalSegments(m.readCSVfile("captures/HorizontalArmSpin-Jibran.csv"),n_components=1,plt=True,title="Large horizontal arm spin segments")
m.getLowDimensionalSegments(m.readCSVfile("captures/HorizontalArmSpinLittleCircles-Jibran.csv"),n_components=1,plt=True,title="Small horizontal arm spin segments")