def binary_clustering(feature_vecs, feature_index, **kwargs):
h = {}
J = {}
cluster1 = [] #stores the indices for the first cluster
cluster2 = [] #stores the indices for the second cluster
#will recognize tabu or hybrid
if 'sampler' in kwargs:
sampler_type = kwargs['sampler']
else:
sampler_type = "tabu"
if 'time_limit' in kwargs:
time_limit = kwargs['time_limit']
else:
time_limit = 20
sampler = TabuSampler()
sampler1 = LeapHybridSampler()
for i in feature_index:
for j in feature_index:
if i < j:
J[(i, j)] = np.linalg.norm(feature_vecs[i] - feature_vecs[j])
#Now use a sampler to solve it
if sampler_type == "tabu":
print("Choosing TABU sampler")
sampler = TabuSampler()
sampleset = sampler.sample_ising(h,
J,
num_reads=1,
timeout=time_limit * 1000)
bin_cluster = sampleset.first[0]
else:
print("Choosing the hybrid sampler")
sampler1 = LeapHybridSampler()
sampleset = sampler1.sample_ising(h, J, time_limit=time_limit)
bin_cluster = sampleset.first[0]
# Run the problem on the sampler and print the results
for key in bin_cluster:
#put in cluster 1 if -1, else 2
if bin_cluster[key] == -1:
cluster1.append(key)
elif bin_cluster[key] == 1:
cluster2.append(key)
return cluster1, cluster2
# hybrid solver test with 10 edges, 4 vertices
import dimod
from dwave.system import LeapHybridSampler
import numpy as np
h = np.loadtxt('h.txt')
J = np.loadtxt('J.txt')
sampler = LeapHybridSampler()
results = sampler.sample_ising(h, J, num_reads=100)
out = open('Outputs.txt', 'w')
for campione in results.record:
for Nple in campione:
out.write(str(Nple) + "\t")
out.write("\n")
out.close()
print(results.first)