/
test_ar.py
198 lines (165 loc) · 6.34 KB
/
test_ar.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
# -*- coding: utf-8 -*-
from __future__ import division, print_function
import collections
import itertools
import numpy as np
import numpy.matlib
np.matlib = numpy.matlib
import pdb
import scipy.sparse
def get_similarity_matrix_old(um):
um = np.where(um == 0, um, 1) # set all ratings to 1 ???
# um = np.where(um >= 4, 1, 0) # set all high ratings to 1
ucount = um.shape[0]
icount = um.shape[1]
coratings = {i: collections.defaultdict(int) for i in range(icount)}
for u in range(ucount):
print('\r', u+1, '/', ucount, end='')
items = np.nonzero(um[u, :])[0]
for i in itertools.combinations(items, 2):
coratings[i[0]][i[1]] += 1
coratings[i[1]][i[0]] += 1
# save_recommendation_data(coratings, 'coratings')
# coratings = load_recommendation_data('coratings')
not_coratings = {i: collections.defaultdict(int) for i in range(icount)}
for i in coratings.keys():
print('\r', i+1, '/', len(coratings), end='')
not_rated_i = set(np.where(um[:, i] == 0)[0])
for j in coratings[i].keys():
rated_j = set(np.where(um[:, j] == 1)[0])
not_coratings[i][j] = len(not_rated_i & rated_j)
# # save_recommendation_data(not_coratings, 'not_coratings')
# not_coratings = load_recommendation_data('not_coratings')
# # debug helpers
# self.rating_stats(um)
# self.corating_stats(coratings, item_id=0)
# self.ar_simple(um, coratings, 0, 2849)
# self.ar_complex(um, coratings, 0, 2849)
# self.ar_both(um, coratings, 0, 2849)
sims = np.zeros((icount, icount))
for x in range(icount):
is_x = np.sum(um[:, x])
not_x = um.shape[0] - is_x
for y in coratings[x]:
# # (x and y) / x simple version
# denominator = coratings[x][y]
# numerator = is_x
# ((x and y) * !x) / ((!x and y) * x) complex version
denominator = coratings[x][y] * not_x
numerator = not_coratings[x][y] * is_x
if numerator > 0:
sims[x, y] = denominator / numerator
return sims
def get_similarity_matrix(um):
print(1)
um.data = np.ones(um.data.shape[0])
print(2)
coratings = um.T.dot(um).toarray()
np.fill_diagonal(coratings, 0)
um = um.toarray()
print(3)
um_inv = np.copy(um)
print(3.1)
um_inv[um_inv == 0] = 2
print(3.2)
um_inv[um_inv == 1] = 0
print(3.3)
um_inv[um_inv == 2] = 1
print(4)
not_coratings = um_inv.T.dot(um)
print(5)
col_sum = um.sum(axis=0)
not_col_sum = um.shape[0] - col_sum
print(6)
col_sums = np.matlib.repmat(col_sum, coratings.shape[0], 1)
not_col_sums = np.matlib.repmat(not_col_sum, not_coratings.shape[0], 1)
print(7)
numerator = coratings * not_col_sums.T
denominator = not_coratings * col_sums.T
print(8)
sims = numerator / denominator
sims[np.isnan(sims)] = 0
sims[np.isinf(sims)] = 0
return sims
class Test(object):
def __init__(self, path, targets_original):
self.path = path
self.targets_original = targets_original
def compute_stats(self):
STEPS_MAX = 50
self.path_original = self.path[:] # DEBUG
self.stats = np.zeros(STEPS_MAX + 1)
self.path = self.path[2:]
if self.path[-2:] == ['*', '*']:
self.path = self.path[:-2]
diff = len(self.path) - 2 * self.path.count(u'*') - STEPS_MAX - 1
if diff > 0:
self.path = self.path[:-diff]
path = ' '.join(self.path).split('*')
path = [l.strip().split(' ') for l in path]
path = [path[0]] + [p[1:] for p in path[1:]]
del self.targets_original[0]
val = 0
len_sum = -1
for p in path:
self.stats[len_sum:len_sum+len(p)] = val
len_sum += len(p)
val += (1 / len(self.targets_original))
if len_sum < len(self.stats):
fill = self.stats[len_sum - 1]
if path[-1] and path[-1][-1] in self.targets_original[len(path)-1]:
fill = min(fill+1/3, 1.0)
self.stats[len_sum:] = fill
print(self.stats)
pdb.set_trace()
if __name__ == '__main__':
p = [
u'006100345X', u'*',
u'006100345X', u'0', u'0', u'0', u'0399134700', u'*',
u'0399134700', u'0', u'0', u'0006512062', u'*',
u'0006512062', u'0', u'0', u'0', u'0', u'0', u'0060509392'
]
to = [
['006100345X'],
[u'0060198702', u'006093736X', u'0061000027', u'0141001828',
u'0156007754', u'0312084986', u'0345285859', u'0345433491',
u'034544003X', u'0394545370', u'0399134409', u'0399134700',
u'0425144062', u'044020352X', u'0440204429', u'0446343455',
u'0449202496', u'0451205634', u'067091021X', u'0671455990',
u'0767904133', u'0786014245', u'0836218515', u'0836220986',
u'0842342702'],
[u'0006512062', u'0060089555', u'0060155515', u'0060171928',
u'0140185216', u'034542705X', u'0385304943', u'0394531809',
u'0451204530', u'0553050672', u'0553209906', u'0684826127',
u'0743427149', u'0743431030', u'0786866195'],
[u'0060509392', u'0312187106', u'0312261594', u'0330376136',
u'0345469674', u'0373250479', u'0385721234', u'039912764X',
u'0425155404', u'0425183394', u'0446526614', u'0452281679',
u'0553756850', u'0671025708', u'067179356X', u'0743202562',
u'0767907809']
]
t = Test(p, to)
t.compute_stats()
print(t.stats)
pdb.set_trace()
# um_dense = np.array([ # simple test case
# [5, 1, 0, 2, 2, 4, 3, 2],
# [1, 5, 2, 5, 5, 1, 1, 4],
# [2, 0, 3, 5, 4, 1, 2, 4],
# [4, 3, 5, 3, 0, 5, 3, 0],
# [2, 0, 1, 3, 0, 2, 5, 3],
# [4, 1, 0, 1, 0, 4, 3, 2],
# [4, 2, 1, 1, 0, 5, 4, 1],
# [5, 2, 2, 0, 2, 5, 4, 1],
# [4, 3, 3, 0, 0, 4, 3, 0]
# ])
#
# um_sparse = scipy.sparse.csr_matrix(um_dense)
# dataset = 'bookcrossing'
# # dataset = 'movielens'
# # dataset = 'imdb'
# um_dense = np.load('data/' + dataset + '/recommendation_data/RatingBasedRecommender_um.obj.npy')
# um_sparse = np.load('data/' + dataset + '/recommendation_data/RatingBasedRecommender_um_sparse.obj.npy').item()
# sims_new = get_similarity_matrix(um_sparse)
# # sims_old = get_similarity_matrix_old(um_dense)
# pdb.set_trace()