print "~~~~~~~~~~"
print 'cz = ', cluster_sz
# calculate language vectors
lang_vectors = random_idx.generate_RI_lang(N, RI_letters, cluster_sz, ordered, languages=languages)
total_vec.append(lang_vectors)
# print cosine angles
if ordered == 0:
ord_str = 'unordered!'
else:
ord_str = 'ordered!'
# calculate total vector
final_lang = sum(total_vec)
# calculate variance of cos angle distribution
cosangles = utils.cosangles(final_lang, languages)
vary = utils.var_measure(cosangles)
V[i,j] = vary
print 'N = ' + str(N) + '; k = ' + str(k) + '; letters clusters are ' + str(cluster_sizes) + ', ' + ord_str + '\n'
print "variance of cosine values: " + str(vary)
print '=========='
np.savez('./vars/vars_dump.npz',V=V, Ns=Ns, sparsities=sparsities,ks=ks)
# plot results
#CS = plt.contourf(sparsities,Ns,V, alpha=0.7, cmap=plt.cm.jet)
CS = plt.contourf(np.log10(ks),Ns,V, alpha=0.7, cmap=plt.cm.jet)
CB = plt.colorbar(CS, shrink=0.8, extend='both')
plt.xlabel('log(k)')
plt.ylabel('N')
plt.title('Variance of Cosine Angles Between Vectors')
print "N =", N, ", k =", k, ", ", ord_str for cluster_sz in cluster_sizes: ##print "~~~~~~~~~~" total_vectors = [] # calculate language vectors lang_vectors = random_idx.generate_RI_lang(N, RI_letters, cluster_sz, ordered, languages=languages) #lang_vectors = random_idx.generate_RI_lang_words(N, RI_letters, languages=languages) total_vectors.append(lang_vectors) # print cosine angles #print '==========' # print 'N = ' + str(N) + '; k = ' + str(k) + '; letters clusters are ' + str(cluster_sz) + ', ' + ord_str + '\n' cosangles = utils.cosangles(lang_vectors, languages) variance = utils.var_measure(cosangles) #print "variance of language values: " + str(utils.var_measure(cosangles)) final_lang = sum(total_vectors) ############################### # iterate through test files and calculate correctness print "Now waiting for input" while True: sentence = raw_input().decode(sys.stdin.encoding or locale.getpreferredencoding(True)) unknown_tots = [] sentence = cleaner(sentence) unknown_vector = random_idx.generate_RI_sentence(N, RI_letters, cluster_sz, ordered,sentence) unknown_tots.append(unknown_vector) final_unknown = sum(unknown_tots) likely_lang = utils.find_language(sentence, final_unknown, final_lang, languages,display=0)
lang_vectors = random_idx.generate_RI_lang(N, RI_letters, cluster_sz, ordered, languages=languages) total_vectors.append(lang_vectors) # calculate unknown vector unknown_vector = random_idx.generate_RI_text(N, RI_letters, cluster_sz, ordered,unknown_txt) unknown_tots.append(unknown_vector) # print cosine angles print '==========' if ordered == 0: ord_str = 'unordered!' else: ord_str = 'ordered!' print 'N = ' + str(N) + '; k = ' + str(k) + '; letters clusters are ' + str(cluster_sz) + ', ' + ord_str + '\n' cosangles = utils.cosangles(lang_vectors, languages) variance = utils.var_measure(cosangles) varys.append(variance) print "variance of language values: " + str(utils.var_measure(cosangles)) ######################## ''' # history vectors lang_vectors = random_idx.generate_RI_lang_history(N, RI_letters, languages=languages) total_vectors.append(lang_vectors) unknown_vector = random_idx.generate_RI_text_history(N, RI_letters, unknown_txt) unknown_tots.append(unknown_vector) print "~~~~~~~~~~" print "history vector information" cosangles = utils.cosangles(lang_vectors, languages)
print "N =", N, ", k =", k, ", ", ord_str
for cluster_sz in cluster_sizes:
##print "~~~~~~~~~~"
total_vectors = []
# calculate language vectors
lang_vectors = random_idx.generate_RI_lang(N, RI_letters, cluster_sz, ordered, languages=languages)
#lang_vectors = random_idx.generate_RI_lang_words(N, RI_letters, languages=languages)
total_vectors.append(lang_vectors)
# print cosine angles
#print '=========='
# print 'N = ' + str(N) + '; k = ' + str(k) + '; letters clusters are ' + str(cluster_sz) + ', ' + ord_str + '\n'
cosangles = utils.cosangles(lang_vectors, languages)
variance = utils.var_measure(cosangles)
#print "variance of language values: " + str(utils.var_measure(cosangles))
final_lang = sum(total_vectors)
###############################
# iterate through test files and calculate correctness
test_fn = glob.glob(main_base + test_dir + '/*txt')
total = len(test_fn)
correct = 0
for i in trange(total):
testf = test_fn[i]
actual_lang = re.findall('(\w+)_\d+_p.txt$', testf)[0]
unknown_tots = []
#print len(testf),testf[91:93]