-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathdatabases.py
345 lines (301 loc) · 13.1 KB
/
databases.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
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
# -*- coding: utf-8 -*-
"""
Created on Mon Apr 27 12:18:12 2015
@author: thalita
"""
from base import BaseDatabase
import numpy as np
import scipy.sparse as sp
from utils import oneD
def _get_zero_mean_matrix(matrix, along='users'):
rows, cols = matrix.shape
if along == 'users':
if sp.issparse(matrix):
matrix = sp.csr_matrix(matrix, dtype=np.float)
else:
matrix = np.array(matrix, dtype=np.float)
mean_vals = np.zeros(rows)
for i in range(rows):
if sp.issparse(matrix):
mean = np.mean(matrix[i,:].data)
mean_vals[i] = mean if not np.isnan(mean) else 0
matrix.data[matrix.indptr[i]:matrix.indptr[i+1]] \
-= mean_vals[i]
else:
non_zero_pos = matrix[i, :] > 0
non_zero = matrix[i, non_zero_pos]
mean = np.mean(non_zero)
mean_vals[i] = mean if not np.isnan(mean) else 0
matrix[i, non_zero_pos] -= mean_vals[i]
else: # Along Items
if sp.issparse(matrix):
matrix = sp.csc_matrix(matrix, dtype=np.float)
else:
matrix = np.array(matrix, dtype=np.float)
mean_vals = np.zeros(cols)
for i in range(cols):
if sp.issparse(matrix):
mean = np.mean(matrix[:, i].data)
mean_vals[i] = mean if not np.isnan(mean) else 0
matrix.data[matrix.indptr[i]:matrix.indptr[i+1]] \
-= mean_vals[i]
else:
non_zero_pos = matrix[:, i] > 0
non_zero = matrix[non_zero_pos, i]
mean = np.mean(non_zero)
mean_vals[i] = mean if not np.isnan(mean) else 0
matrix[non_zero_pos, i] -= mean_vals[i]
if sp.issparse(matrix):
matrix = matrix.tocsr()
return matrix, mean_vals
class MatrixDatabase(BaseDatabase):
def __init__(self, matrix):
if sp.issparse(matrix):
matrix = matrix.tocsr()
self.matrix = matrix
self.matrix_csc = None
self.matrix_dok = None
self.thresholded = None
self.thresholded_csc = None
self.zero_mean_matrix = None
self.zero_mean_matrix_csc = None
self.zero_mean_matrix_dok = None
self.means = {}
def _compute_zero_mean(self):
self.zero_mean_matrix = {}
self.zero_mean_matrix['users'], self.means['users'] = \
_get_zero_mean_matrix(self.matrix.copy(), along='users')
self.zero_mean_matrix['items'], self.means['items'] = \
_get_zero_mean_matrix(self.matrix.copy(), along='items')
self.zero_mean_matrix['useritems'], self.means['useritems'] = \
_get_zero_mean_matrix(self.zero_mean_matrix['users'].copy(), along='items')
if sp.issparse(self.matrix):
self.zero_mean_matrix_csc = {}
self.zero_mean_matrix_dok= {}
for name in ['users', 'items', 'useritems']:
self.zero_mean_matrix_csc[name] = self.zero_mean_matrix[name].tocsc()
self.zero_mean_matrix_dok[name] = self.zero_mean_matrix[name].todok()
def get_means(self, along):
if self.zero_mean_matrix is None:
self._compute_zero_mean()
return self.means[along]
def n_users(self):
return self.matrix.shape[0]
def n_items(self):
return self.matrix.shape[1]
def get_matrix(self, zero_mean=False, threshold=None, sparse=False):
if threshold is not None:
if self.thresholded is None:
if sp.issparse(self.matrix):
self.thresholded = self.matrix.copy()
self.thresholded.data = np.array(self.matrix.data > threshold,
dtype=np.float)
if sparse:
return self.thresholded
else:
return self.thresholded.toarray()
else:
self.thresholded = \
np.array(self.matrix > threshold, dtype=np.float)
return self.thresholded
elif zero_mean:
if self.zero_mean_matrix is None:
self._compute_zero_mean()
if sp.issparse(self.matrix) and not sparse:
return self.zero_mean_matrix[zero_mean].toarray(), self.means
else:
return self.zero_mean_matrix[zero_mean], self.means
else:
if sp.issparse(self.matrix):
if sparse:
return self.matrix
else:
return self.matrix.toarray()
else:
return self.matrix
def get_rating(self, user_id, item_id, zero_mean=False):
if zero_mean:
if self.zero_mean_matrix is None:
self._compute_zero_mean()
if sp.issparse(self.matrix):
return self.zero_mean_matrix_dok[zero_mean][user_id, item_id]
else:
return self.zero_mean_matrix[zero_mean][user_id, item_id]
else:
if sp.issparse(self.matrix):
if self.matrix_dok is None:
self.matrix_dok = self.matrix.todok()
return self.matrix_dok[user_id, item_id]
else:
return BaseDatabase.get_rating(self, user_id, item_id)
def set_rating(self, user_id, item_id, rating):
self.matrix[user_id, item_id] = rating
self.matrix_csc = None
self.matrix_dok = None
self.thresholded = None
self.thresholded_csc = None
self.zero_mean_matrix = None
self.zero_mean_matrix_csc = None
self.zero_mean_matrix_dok = None
self.means = {}
def get_user_vector(self, user_id, zero_mean=False, sparse=False):
if zero_mean:
if self.zero_mean_matrix is None:
self._compute_zero_mean()
vector = self.zero_mean_matrix[zero_mean][user_id, :]
else:
vector = BaseDatabase.get_user_vector(self, user_id)
if sp.issparse(self.matrix) and not sparse:
return oneD(vector.toarray())
else:
return vector
def get_rating_list(self, user_id, zero_mean=False):
if sp.issparse(self.matrix):
if zero_mean:
if self.zero_mean_matrix is None:
self._compute_zero_mean()
ratings = self.zero_mean_matrix[user_id, :].data.tolist()
else:
ratings = self.matrix[user_id, :].data.tolist()
items = self.matrix[user_id, :].indices.tolist()
alist = list(zip(ratings, items))
else:
vector = self.get_user_vector(user_id, zero_mean)
alist = [(r, i) for i, r in enumerate(vector) if r != 0]
alist.sort()
return [(i, r) for r, i in alist]
def get_item_vector(self, item_id, zero_mean=False, sparse=False):
if zero_mean:
if self.zero_mean_matrix is None:
self._compute_zero_mean()
if sp.issparse(self.matrix):
if sparse:
self.zero_mean_matrix_csc[zero_mean][:, item_id]
else:
return oneD(self.zero_mean_matrix_csc[zero_mean][:, item_id].toarray())
else:
return self.zero_mean_matrix_csc[zero_mean][:, item_id]
else:
if sp.issparse(self.matrix):
if self.matrix_csc is None:
self.matrix_csc = self.matrix.tocsc()
if sparse:
return self.matrix_csc[:, item_id]
else:
return oneD(self.matrix_csc[:, item_id].toarray())
else:
return BaseDatabase.get_item_vector(self, item_id)
def get_unrated_items(self, user_id):
"return unrated item ids for user"
if sp.issparse(self.matrix):
rated = set(self.matrix[user_id, :].indices)
return [item for item in range(self.n_items())
if item not in rated]
else:
return [idx for idx, rating in enumerate(self.matrix[user_id, :])
if rating == 0]
def get_rated_items(self, user_id):
"return items rated by user_id user"
if sp.issparse(self.matrix):
return self.matrix[user_id,:].indices.tolist()
else:
return [idx for idx, rating in enumerate(self.matrix[user_id, :])
if rating > 0]
def get_rated_users(self, item_id):
"return users who did not rate item_id user"
if sp.issparse(self.matrix):
if self.matrix_csc is None:
self.matrix_csc = self.matrix.tocsc()
return self.matrix_csc[:, item_id].indices.tolist()
else:
return [idx for idx, rating in enumerate(self.matrix[:, item_id])
if rating > 0]
def _test_sparse_matrixdb():
sm = sp.rand(100, 200, density=0.02)
m = sm.toarray()
sdb = MatrixDatabase(sm)
db = MatrixDatabase(m)
assert((db.get_matrix() == sdb.get_matrix()).all())
assert((db.get_matrix(zero_mean='users')[0] == sdb.get_matrix(zero_mean='users')[0]).all())
assert((db.get_matrix(zero_mean='items')[0] == sdb.get_matrix(zero_mean='items')[0]).all())
assert((db.get_matrix(zero_mean='users')[1] == sdb.get_matrix(zero_mean='users')[1]).all())
assert((db.get_matrix(zero_mean='items')[1] == sdb.get_matrix(zero_mean='items')[1]).all())
assert((db.get_matrix(threshold=0)==sdb.get_matrix(threshold=0)).all())
assert((db.get_rating(5,30)==sdb.get_rating(5,30)).all())
assert((db.get_rating(5,30,zero_mean='users')==sdb.get_rating(5,30,zero_mean='users')).all())
assert((db.get_rating(5,30,zero_mean='items')==sdb.get_rating(5,30,zero_mean='items')).all())
assert((db.get_user_vector(5)==sdb.get_user_vector(5)).all())
assert((db.get_user_vector(5,zero_mean='users')==sdb.get_user_vector(5,zero_mean='users')).all())
assert((db.get_user_vector(5,zero_mean='items')==sdb.get_user_vector(5,zero_mean='items')).all())
assert((db.get_item_vector(5)==sdb.get_item_vector(5)).all())
assert(db.get_unrated_items(5)==sdb.get_unrated_items(5))
assert(db.get_rated_items(5)==sdb.get_rated_items(5))
assert(db.get_rated_users(5)==sdb.get_rated_users(5))
class HiddenRatingsDatabase(MatrixDatabase):
def __init__(self, matrix, hidden_coord):
MatrixDatabase.__init__(self, matrix.copy())
if sp.issparse(self.matrix):
self.matrix = self.matrix.tolil()
for u, i in hidden_coord:
self.matrix[u, i] = 0
if sp.issparse(self.matrix):
self.matrix = self.matrix.tocsr()
class SubDatabase(MatrixDatabase):
def __init__(self, matrix, user_indices):
user_indices.sort()
MatrixDatabase.__init__(self, matrix[user_indices, :])
class SubDatabaseOnline(MatrixDatabase):
def __init__(self, matrix, user_indices):
user_indices.sort()
self.user_indices = user_indices
MatrixDatabase.__init__(self, matrix)
def _compute_zero_mean(self):
self.zero_mean_matrix, self.user_means = \
_get_zero_mean_matrix(self.matrix[self.user_indices, :].copy())
def get_matrix(self, zero_mean=False, threshold=False):
if threshold:
if self.thresholded is None:
self.thresholded = \
np.array(self.matrix[self.user_indices, :] > threshold,
dtype=float)
return self.thresholded
if zero_mean:
self._compute_zero_mean()
return self.zero_mean_matrix, self.user_means
else:
return self.matrix[self.user_indices, :]
def n_users(self):
return len(self.user_indices)
def get_ratings(self, user_id, item_id, zero_mean=False):
if zero_mean:
self._compute_zero_mean()
return self.zero_mean_matrix[user_id, item_id]
else:
return self.matrix[self.user_indices[user_id], item_id]
def get_user_vector(self, user_id, zero_mean=False):
"return a 2D array with user ratings"
if zero_mean:
self._compute_zero_mean()
return self.zero_mean_matrix[user_id, :]
else:
return self.matrix[self.user_indices[user_id], :]
def get_item_vector(self, item_id, zero_mean):
"return a 2D array with item ratings"
if zero_mean:
self._compute_zero_mean()
return self.zero_mean_matrix[:, item_id]
else:
return self.matrix[self.user_indices, item_id]
def get_unrated_items(self, user_id):
"return unrated item ids for user"
return [idx for idx, rating in enumerate(self.matrix[user_id, :])
if rating == 0]
def get_rated_items(self, user_id):
"return items rated by user_id user"
return [idx for idx, rating in enumerate(self.matrix[user_id, :])
if rating > 0]
def get_rated_users(self, item_id):
"return users who did rate item_id item"
return [idx for idx, rating
in enumerate(self.matrix[self.user_indices, item_id])
if rating > 0]