-
Notifications
You must be signed in to change notification settings - Fork 7
/
Copy pathrandom_idx.py
executable file
·277 lines (222 loc) · 8.6 KB
/
random_idx.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
# random_idx.py
# creates random index vectors for a number of languages
# libraries
import sys
import numpy as np
import string
import utils
import pandas as pd
import os
alphabet = string.lowercase + " "
lang_dir = './preprocessed_texts/'
cluster_cache = {}
def generate_letter_id_vectors(N, k, alph=alphabet):
# build row-wise k-sparse random index matrix
# each row is random index vector for letter
num_letters = len(alphabet)
RI_letters = np.zeros((num_letters,N))
for i in xrange(num_letters):
rand_idx = np.random.permutation(N)
RI_letters[i,rand_idx[0:k]] = 1
RI_letters[i,rand_idx[k:2*k]] = -1
return RI_letters
# def generate_id(RI_letters,alph=alphabet,cluster_sz=1, ordered=0):
# # generate id vectors of clusters from "alphabet" with size "cluster_sz"
#
# # generate clusters
# if ordered == 0:
# clusters = utils.generate_unordered_clusters(alph,cluster_sz=cluster_sz)
# else:
# clusters = utils.generate_ordered_clusters(alph,cluster_sz=cluster_sz)
#
# M = len(clusters) # number of letter clusters
# num_letters,N = RI_letters.shape
# #RI_letters = generate_letter_id_vectors(N, k, alphabet)
#
# RI = np.zeros((M,N))
# for i in xrange(M):
# # calculate repeats
# cluster = clusters[i]
# RI[i,:] = id_vector(N, cluster, alphabet, RI_letters, ordered=ordered)
# dictionary = {}
# for i in range(len(clusters)):
# dictionary[clusters[i]] = RI[i]
# return dictionary
def id_vector(N, cluster, alphabet, RI_letters,ordered=0):
if cluster == '':
return 0
if ordered == 0:
# unordered clusters
cluster = ''.join(sorted(cluster))
#if it's already calculated, just return it
if cluster in cluster_cache:
return cluster_cache[cluster]
vector = np.zeros(N)
first = cluster[0]
repeats = 0
for char in cluster:
if first == char:
repeats += 1
if repeats == len(cluster):
# check if cluster all same letter
letter_idx = alphabet.find(first)
#print first, RI_letters[letter_idx,:]
vector = RI_letters[letter_idx,:]
else:
# letters = list(cluster)
# prod = np.ones((1,N))
# for letter in letters:
# letter_idx = alphabet.find(letter)
# prod = np.multiply(prod, RI_letters[letter_idx,:])
# vector = prod
# ordered clusters
letters = list(cluster)
prod = np.ones((1,N))
roller = len(letters)-1
for letter in letters:
# working code
#letter_idx = alphabet.find(letter)
#prod = np.multiply(prod, RI_letters[letter_idx,:])
#prod = np.roll(prod,1)
# testing permutations
letter_idx = alphabet.find(letter)
prod = np.multiply(prod, np.roll(RI_letters[letter_idx, :],roller))
roller -= 1
vector = prod
if len(cluster_cache) > 100000:
cluster_cache.clear()
print "clearing cache"
cluster_cache[cluster] = vector
return vector
def generate_RI_str(N, RI_letters, cluster_sz, ordered, string, alph=alphabet):
# generate RI vector for "text_name"
# assumes text_name has .txt
text_vector = np.zeros((1,N))
for char_num in xrange(len(string)):
if char_num < cluster_sz:
continue
else:
# build cluster
cluster = ''
for j in xrange(cluster_sz):
cluster = string[char_num - j] + cluster
text_vector += id_vector(N, cluster, alph,RI_letters, ordered)
return text_vector
def generate_RI_text(N, RI_letters, cluster_sz, ordered, text_name, alph=alphabet):
# generate RI vector for "text_name"
# assumes text_name has .txt
text_vector = np.zeros((1, N))
text = utils.load_text_spaces(text_name)
for char_num in xrange(len(text)):
if char_num < cluster_sz:
continue
else:
# build cluster
cluster = ''
for j in xrange(cluster_sz):
cluster = text[char_num - j] + cluster
text_vector += id_vector(N, cluster, alph,RI_letters, ordered)
return text_vector
def generate_RI_sentence(N, RI_letters, cluster_sz, ordered, text, alph=alphabet):
# generate RI vector for "text_name"
# assumes text_name has .txt
text_vector = np.zeros((1, N))
##text = utils.load_text_spaces(text_name)
for char_num in xrange(len(text)):
if char_num < cluster_sz:
continue
else:
# build cluster
cluster = ''
for j in xrange(cluster_sz):
cluster = text[char_num - j] + cluster
text_vector += id_vector(N, cluster, alph,RI_letters, ordered)
return text_vector
#Not really fast. Theoretically faster, but not for real (not using cache)
def generate_RI_text_fast(N, RI_letters, cluster_sz, ordered, text_name, alph=alphabet):
text_vector = np.zeros((1, N))
text = utils.load_text(text_name)
cluster2 = ''
vector = np.ones((1,N))
for char_num in xrange(len(text)):
cluster = cluster + text[char_num]
if len(cluster) < cluster_sz:
continue
elif len(cluster) > cluster_sz:
prev_letter = cluster[0]
prev_letter_idx = alphabet.find(letter)
inverse = np.roll(RI_letters[prev_letter_idx,:], cluster_sz-1)
vector = np.multiply(vector, inverse)
vector = np.roll(vector, 1)
letter = text[char_num]
letter_idx = alphabet.find(letter)
vector = np.multiply(vector, RI_letters[letter_idx,:])
cluster = cluster[1:]
else: # (len(cluster) == cluster_size), happens once
letters = list(cluster)
for letter in letters:
vector = np.roll(vector,1)
letter_idx = alphabet.find(letter)
vector = np.multiply(vector, RI_letters[letter_idx,:])
text_vector += vector
return text_vector
def generate_RI_text_words(N, RI_letters, text_name, alph=alphabet):
# generate RI vector for "text_name"
# assumes text_name has .txt
text_vector = np.zeros((1, N))
text = utils.load_text_spaces(text_name)
cluster = ''
for char_num in xrange(len(text)):
char = text[char_num]
if char == ' ':
text_vector += id_vector(N, cluster, alph, RI_letters)
# reset cluster
cluster = ''
else:
cluster += text[char_num]
return text_vector
def generate_RI_text_history(N, RI_letters, text_name, alph=alphabet):
# generate RI vector for "text_name"
# assumes text_name has .txt
text_vector = np.zeros((1, N))
history_vector = np.zeros((1,N))
text = utils.load_text(text_name)
for char_num in xrange(len(text)):
char = text[char_num]
letter_idx = alphabet.find(char)
history_vector = 0.75*history_vector + RI_letters[letter_idx,:]
text_vector += history_vector
return text_vector
def generate_RI_lang(N,RI_letters, cluster_sz, ordered, languages=None):
cluster_cache.clear()
if languages == None:
languages = ['english','german','norwegian','finnish']
num_lang = len(languages)
lang_vectors = np.zeros((num_lang,N))
for i in xrange(num_lang):
# load text one at a time (to save mem), English, German, Norwegian
print 'loading ' + languages[i]
lang_vectors[i,:] = generate_RI_text(N, RI_letters, cluster_sz, ordered, lang_dir + languages[i] + '.txt')
return lang_vectors
def generate_RI_lang_history(N,RI_letters, languages=None):
cluster_cache.clear()
if languages == None:
languages = ['english','german','norwegian','finnish']
num_lang = len(languages)
lang_vectors = np.zeros((num_lang,N))
for i in xrange(num_lang):
print 'loading ' + languages[i]
# load text one at a time (to save mem), English, German, Norwegian
lang_vectors[i,:] = generate_RI_text_history(N, RI_letters, lang_dir + languages[i] + '.txt')
return lang_vectors
def generate_RI_lang_words(N, RI_letters, languages=None):
cluster_cache.clear()
if languages == None:
languages = ['english','german','norwegian','finnish']
num_lang = len(languages)
lang_vectors = np.zeros((num_lang,N))
for i in xrange(num_lang):
print 'loading ' + languages[i]
# load text one at a time (to save mem), English, German, Norwegian
lang_vectors[i,:] = generate_RI_text_words(N, RI_letters, lang_dir + languages[i] + '.txt')
return lang_vectors