/
eval_metric_space.py
163 lines (128 loc) · 6.66 KB
/
eval_metric_space.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 eval_metric_space import idx
from init import *
from Analysis import metric_space_indexes as index
from Analysis.RatingCorrelator import calc_distance_matrix
from Analysis.retrieval import Retriever
from Analysis.performance import mean_cross_validated_index
from Network import FileManager
from Analysis.metric_space_indexes import eval_embed_space
from experiments import load_experiments, CrossValidationManager
from Analysis.analysis import smooth
# Setup
experiment_name = 'DirRD-Semi-Supervised'
dset = 'Test'
rating_normalizaion = 'Scale' # 'None', 'Normal', 'Scale'
ratiing_metrics = ['euclidean']
n_groups = 5
runs, run_net_types, run_metrics, run_epochs, run_names, _, _ = load_experiments(experiment_name)
alpha = 0.3
USE_CACHE = False
DUMP_CACHE = False
cv = CrossValidationManager('RET')
configurations = ['{}{}'.format(cv.get_run_id(i)[0], cv.get_run_id(i)[1]) for i in ([0, 1, 3, 4, 7] if dset == 'Valid' else [2, 5, 6, 8, 9])] # [range(10)]
# configurations = range(n_groups)
# configurations = [1]
dset = 'Valid'
#indexes = ['Hubness', 'Symmetry', 'Contrast', 'Concentration', 'Kumari']
indexes = ['Hubness', 'Symmetry'] # 'FeatCorr', 'SampCorr'
M = len(indexes)
# initialize figures
plt.figure("{} Metric Space: {}".format(experiment_name, dset))
p = [None] * len(ratiing_metrics) * M
for i in range(M * len(ratiing_metrics)):
p[i] = plt.subplot(len(ratiing_metrics), M, i + 1)
legend = []
for n in run_names:
legend += [n]
legend += ['']
legend += ['']
# evaluate
Epochs, Idx_hubness, Idx_symmetry, Idx_concentration, Idx_contrast, Idx_kummar, Idx_featCorr, Idx_sampCorr = [], [], [], [], [], [], [], []
for m, metric_ in enumerate(ratiing_metrics):
#print("Begin: {} metric".format(metric))
for run, net_type, r, epochs, metric in zip(runs, run_net_types, range(len(runs)), run_epochs, run_metrics):
plot_data_filename = './Plots/Data/metric-space_{}{}.p'.format(net_type, run)
try:
if USE_CACHE is False:
print('WARNINING - SKIIPING TO CALCULATION')
assert False
combined_epochs, idx_hubness, idx_symmetry, idx_concentration, idx_contrast, idx_kummar, idx_featCorr, idx_sampCorr \
= pickle.load(open(plot_data_filename, 'br'))
#idx_kummar = np.reshape(idx_kummar, (1, np.max(idx_kummar.shape)))
print("Loaded results for {}{}".format(net_type, run))
except:
print("Evaluating classification accuracy for {}{} using {}".format(net_type, run, metric))
Embed = FileManager.Embed(net_type)
embed_source = [Embed(run + 'c{}'.format(c), dset) for c in configurations]
combined_epochs, idx_hubness, idx_symmetry, idx_concentration, idx_contrast, idx_kummar, idx_featCorr, idx_sampCorr = \
eval_embed_space(embed_source, metric, 'euclidean', epochs, dset)
if DUMP_CACHE:
pickle.dump((combined_epochs, idx_hubness, idx_symmetry, idx_concentration, idx_contrast, idx_kummar,
idx_featCorr, idx_sampCorr),
open(plot_data_filename, 'bw'))
else:
print('NO DUMP')
Epochs += [combined_epochs]
Idx_hubness += [idx_hubness]
Idx_symmetry += [idx_symmetry]
Idx_concentration += [idx_concentration]
Idx_contrast += [idx_contrast]
Idx_kummar += [idx_kummar]
Idx_featCorr += [idx_featCorr]
Idx_sampCorr += [idx_sampCorr]
# plot
for epochs, idx_hubness, idx_symmetry, idx_concentration, idx_contrast, idx_kummar, idx_featCorr, idx_sampCorr \
in zip(Epochs, Idx_hubness, Idx_symmetry, Idx_concentration, Idx_contrast, Idx_kummar, Idx_featCorr, Idx_sampCorr):
# hubness
next_plot = 0
if 'Hubness' in indexes:
q = p[M * m + next_plot].plot(epochs, smooth(idx_hubness[0]), marker='x') #
p[M * m + next_plot].plot(epochs, smooth(idx_hubness[0] + idx_hubness[1]), color=q[0].get_color(), ls='--', alpha=alpha)
p[M * m + next_plot].plot(epochs, smooth(idx_hubness[0] - idx_hubness[1]), color=q[0].get_color(), ls='--', alpha=alpha)
next_plot += 1
# symmetry
if 'Symmetry' in indexes:
q = p[M * m + next_plot].plot(epochs, smooth(idx_symmetry[0]))
p[M * m + next_plot].plot(epochs, smooth(idx_symmetry[0] + idx_symmetry[1]), color=q[0].get_color(), ls='--', alpha=alpha)
p[M * m + next_plot].plot(epochs, smooth(idx_symmetry[0] - idx_symmetry[1]), color=q[0].get_color(), ls='--', alpha=alpha)
next_plot += 1
# contrast
if 'Contrast' in indexes:
q = p[M * m + next_plot].plot(epochs, smooth(idx_contrast[0]))
p[M * m + next_plot].plot(epochs, smooth(idx_contrast[0] + idx_contrast[1]), color=q[0].get_color(), ls='--', alpha=alpha)
p[M * m + next_plot].plot(epochs, smooth(idx_contrast[0] - idx_contrast[1]), color=q[0].get_color(), ls='--', alpha=alpha)
next_plot += 1
# concentration
if 'Concentration' in indexes:
q = p[M * m + next_plot].plot(epochs, smooth(idx_concentration[0]))
p[M * m + next_plot].plot(epochs, smooth(idx_concentration[0] + idx_concentration[1]), color=q[0].get_color(), ls='--', alpha=alpha)
p[M * m + next_plot].plot(epochs, smooth(idx_concentration[0] - idx_concentration[1]), color=q[0].get_color(), ls='--', alpha=alpha)
next_plot += 1
# kumar
if 'Kumari' in indexes:
p[M * m + next_plot].plot(epochs, smooth(idx_kummar[0]))
next_plot += 1
if 'FeatCorr' in indexes:
p[M * m + next_plot].plot(epochs, smooth(idx_featCorr[0]))
next_plot += 1
if 'SampCorr' in indexes:
p[M * m + next_plot].plot(epochs, smooth(idx_sampCorr[0]))
next_plot += 1
# labels
#if r == 0: # first column
# p[M * m + 0].axes.yaxis.label.set_text(metric)
if m == 0: # first row
for i, label in enumerate(indexes):
p[M * m + i].axes.title.set_text(label)
#p[M * m + 0].axes.title.set_text('Hubness')
#p[M * m + 1].axes.title.set_text('Symmetry')
#p[M * m + 2].axes.title.set_text('Contrast')
#p[M * m + 3].axes.title.set_text('Concentration')
#p[M * m + 4].axes.title.set_text('Kumari')
if m == len(ratiing_metrics) - 1: # last row
p[M * m + M//2].axes.xaxis.label.set_text('Epochs')
p[1].legend(legend if indexes[-1] is not 'Kumari' else run_names)
for i in range(M * len(ratiing_metrics)):
p[i].grid(which='both', axis='y')
print('Done doMetricSpaceIndexes')
plt.show()