def _check(self, results, linear, quad, offset=0, num_reads=1):
# Did we get the right number of samples?
self.assertEqual(num_reads, sum(results.num_occurrences))
# verify num_occurrences sum corresponds to num_reads
self.assertEqual(100, sum(results.num_occurrences))
# Make sure energies are correct in raw results
for energy, state in zip(results.energies, results.samples):
self.assertEqual(
energy, evaluate_ising(linear, quad, state, offset=offset))
# skip sampleset test if dimod is not installed
if not dimod:
return
# Make sure the sampleset matches raw results
for record, energy, num_occurrences, state in \
zip(results.sampleset.record,
results['energies'],
results['num_occurrences'],
results['solutions']):
recalc_energy = evaluate_ising(linear, quad, state, offset=offset)
self.assertEqual(energy, recalc_energy)
self.assertEqual(energy, float(record.energy))
self.assertEqual(num_occurrences, int(record.num_occurrences))
self.assertEqual(state, list(record.sample))
def _check(self, results, linear, quad, offset=0, num_reads=1):
# Did we get the right number of samples?
self.assertEqual(num_reads, sum(results.num_occurrences))
# verify .occurrences property still works, although is deprecated
with warnings.catch_warnings():
warnings.simplefilter("ignore")
self.assertEqual(100, sum(results.occurrences))
# Make sure energies are correct in raw results
for energy, state in zip(results.energies, results.samples):
self.assertEqual(
energy, evaluate_ising(linear, quad, state, offset=offset))
# skip sampleset test if dimod is not installed
if not dimod:
return
# Make sure the sampleset matches raw results
for record, energy, num_occurrences, state in \
zip(results.sampleset.record,
results['energies'],
results['num_occurrences'],
results['solutions']):
recalc_energy = evaluate_ising(linear, quad, state, offset=offset)
self.assertEqual(energy, recalc_energy)
self.assertEqual(energy, float(record.energy))
self.assertEqual(num_occurrences, int(record.num_occurrences))
self.assertEqual(state, list(record.sample))
def test_wait_many(self):
"""Submit a batch of problems then use `wait_multiple` to wait on all of them."""
with Client(**config) as client:
solver = client.get_solver()
linear, quad = generate_random_ising_problem(solver)
result_list = []
for _ in range(100):
results = solver.sample_ising(linear, quad, num_reads=40)
result_list.append([results, linear, quad])
dwave.cloud.computation.Future.wait_multiple(
[f[0] for f in result_list])
for results, _, _ in result_list:
self.assertTrue(results.done())
for results, linear, quad in result_list:
# Did we get the right number of samples?
self.assertEqual(40, sum(results.occurrences))
# Make sure the number of occurrences and energies are all correct
for energy, state in zip(results.energies, results.samples):
self.assertAlmostEqual(energy,
evaluate_ising(linear, quad, state))
def test_cancel_batch(self):
"""Submit batch of problems, then cancel them."""
with Client(**config) as client:
solver = client.get_solver()
linear, quad = generate_random_ising_problem(solver)
max_num_reads = max(
solver.properties.get('num_reads_range', [1, 100]))
result_list = []
for _ in range(1000):
results = solver.sample_ising(linear,
quad,
num_reads=max_num_reads)
result_list.append([results, linear, quad])
[r[0].cancel() for r in result_list]
for results, linear, quad in result_list:
# Responses must be canceled or correct
try:
# Did we get the right number of samples?
self.assertEqual(max_num_reads, sum(results.occurrences))
# Make sure the number of occurrences and energies are all correct
for energy, state in zip(results.energies,
results.samples):
self.assertAlmostEqual(
energy, evaluate_ising(linear, quad, state))
except CanceledFutureError:
pass
def test_as_completed(self):
"""Submit a batch of problems then use `as_completed` to iterate over
all of them."""
# Connect
with Client(**config) as client:
solver = client.get_solver()
# Build a problem
linear = [0] * (max(solver.nodes) + 1)
for index in solver.nodes:
linear[index] = random.choice([-1, 1])
quad = {
key: random.choice([-1, 1])
for key in solver.undirected_edges
}
# Sample the solution 100x40 times
computations = [
solver.sample_ising(linear, quad, num_reads=40)
for _ in range(100)
]
# Go over computations, one by one, as they're done and check they're OK
for computation in dwave.cloud.computation.Future.as_completed(
computations):
self.assertTrue(computation.done())
self.assertTrue(40 == sum(computation.occurrences))
for energy, state in zip(computation.energies,
computation.samples):
self.assertTrue(
energy == evaluate_ising(linear, quad, state))
def test_submit_batch(self):
"""Submit batch of problems."""
# Connect
with Client(**config) as client:
solver = client.get_solver()
result_list = []
for _ in range(100):
# Build a linear problem
linear = [0] * (max(solver.nodes) + 1)
for index in solver.nodes:
linear[index] = random.choice([-1, 1])
# Build a
quad = {
key: random.choice([-1, 1])
for key in solver.undirected_edges
}
results = solver.sample_ising(linear, quad, num_reads=10)
result_list.append([results, linear, quad])
for results, linear, quad in result_list:
# Did we get the right number of samples?
self.assertTrue(10 == sum(results.occurrences))
# Make sure the number of occurrences and energies are all correct
for energy, state in zip(results.energies, results.samples):
self.assertTrue(
energy == evaluate_ising(linear, quad, state))
def test_wait_many(self):
"""Submit a batch of problems then use `wait_multiple` to wait on all of them."""
# Connect
with Client(**config) as client:
solver = client.get_solver()
# Build a linear problem
linear = [0] * (max(solver.nodes) + 1)
for index in solver.nodes:
linear[index] = random.choice([-1, 1])
# Build a
quad = {key: random.choice([-1, 1]) for key in solver.undirected_edges}
result_list = []
for _ in range(100):
results = solver.sample_ising(linear, quad, num_reads=40)
result_list.append([results, linear, quad])
dwave.cloud.computation.Future.wait_multiple([f[0] for f in result_list])
for results, _, _ in result_list:
self.assertTrue(results.done())
for results, linear, quad in result_list:
# Did we get the right number of samples?
self.assertTrue(40 == sum(results.occurrences))
# Make sure the number of occurrences and energies are all correct
for energy, state in zip(results.energies, results.samples):
self.assertTrue(energy == evaluate_ising(linear, quad, state))
def _check(self, results, linear, quad, num):
# Did we get the right number of samples?
self.assertTrue(100 == sum(results.occurrences))
# Make sure the number of occurrences and energies are all correct
for energy, state in zip(results.energies, results.samples):
self.assertTrue(energy == evaluate_ising(linear, quad, state))
def _submit_and_check(self, solver, linear, quad, **kwargs):
results = solver.sample_ising(linear, quad, num_reads=100, **kwargs)
# Did we get the right number of samples?
self.assertEqual(100, sum(results.num_occurrences))
# verify .occurrences property still works, although is deprecated
with warnings.catch_warnings():
warnings.simplefilter("ignore")
self.assertEqual(100, sum(results.occurrences))
# offset is optional
offset = kwargs.get('offset', 0)
# Make sure the number of occurrences and energies are all correct
for energy, state in zip(results.energies, results.samples):
self.assertAlmostEqual(
energy, evaluate_ising(linear, quad, state, offset=offset))
# label is optional
label = kwargs.get('label')
if label is not None:
self.assertEqual(results.label, label)
return results
def test_submit_bqm_qubo_ok_reply(self):
"""Handle a normal query and response."""
qubo_msg_diff = dict(type="qubo")
qubo_answer_diff = {
'energies': 'AAAAAAAAAAA=',
'solutions': 'AA==',
'active_variables': 'AAAAAAQAAAA='
}
# each thread can have its instance of a session because
# the mocked responses are stateless
def create_mock_session(client):
session = mock.Mock()
session.post = lambda a, _: choose_reply(a, {
'problems/': '[%s]' % complete_no_answer_reply(
'123', 'abc123')})
session.get = lambda a: choose_reply(a, {
'problems/123/': complete_reply(
'123', 'abc123', answer=qubo_answer_diff, msg=qubo_msg_diff)})
return session
with mock.patch.object(Client, 'create_session', create_mock_session):
with Client('endpoint', 'token') as client:
solver = Solver(client, solver_data('abc123'))
qubo = {(0, 0): 4.0, (0, 4): -4, (4, 4): 4.0}
offset = -2.0
params = dict(num_reads=100)
results = solver.sample_qubo(qubo, offset, **params)
# make sure energies are correct in raw results
for energy, sample in zip(results.energies, results.samples):
self.assertEqual(energy, evaluate_ising({}, qubo, sample, offset=offset))
def _submit_and_check(self, solver, linear, quad, **param):
results = solver.sample_ising(linear, quad, num_reads=100, **param)
# Did we get the right number of samples?
self.assertEqual(100, sum(results.occurrences))
# Make sure the number of occurrences and energies are all correct
for energy, state in zip(results.energies, results.samples):
self.assertAlmostEqual(energy, evaluate_ising(linear, quad, state))
return results
def test_submit_bqm_problem(self):
"""Submit a problem with all supported coefficients set."""
with Client(**config) as client:
solver = client.get_solver()
linear, quad = generate_random_ising_problem(solver)
bqm = dimod.BinaryQuadraticModel.from_ising(linear, quad)
results = solver.sample_bqm(bqm, num_reads=100)
# Did we get the right number of samples?
self.assertEqual(100, sum(results.occurrences))
# Make sure the number of occurrences and energies are all correct
for energy, state in zip(results.energies, results.samples):
self.assertAlmostEqual(energy, evaluate_ising(linear, quad, state))
def test_as_completed(self):
"""Submit a batch of problems then use `as_completed` to iterate over
all of them."""
with Client(**config) as client:
solver = client.get_solver()
linear, quad = generate_random_ising_problem(solver)
# Sample the solution 100x40 times
computations = [solver.sample_ising(linear, quad, num_reads=40) for _ in range(100)]
# Go over computations, one by one, as they're done and check they're OK
for computation in dwave.cloud.computation.Future.as_completed(computations):
self.assertTrue(computation.done())
self.assertEqual(40, sum(computation.occurrences))
for energy, state in zip(computation.energies, computation.samples):
self.assertAlmostEqual(energy, evaluate_ising(linear, quad, state))
def test_submit_batch(self):
"""Submit batch of problems."""
with Client(**config) as client:
solver = client.get_solver()
result_list = []
for _ in range(100):
linear, quad = generate_random_ising_problem(solver)
results = solver.sample_ising(linear, quad, num_reads=10)
result_list.append([results, linear, quad])
for results, linear, quad in result_list:
# Did we get the right number of samples?
self.assertEqual(10, sum(results.occurrences))
# Make sure the number of occurrences and energies are all correct
for energy, state in zip(results.energies, results.samples):
self.assertAlmostEqual(energy, evaluate_ising(linear, quad, state))
def test_cancel_batch(self):
"""Submit batch of problems, then cancel them."""
# Connect
with Client(**config) as client:
solver = client.get_solver()
# Build a linear problem
linear = [0] * (max(solver.nodes) + 1)
for index in solver.nodes:
linear[index] = random.choice([-1, 1])
quad = {
key: random.choice([-1, 1])
for key in solver.undirected_edges
}
max_num_reads = max(
solver.properties.get('num_reads_range', [1, 100]))
result_list = []
for _ in range(1000):
results = solver.sample_ising(linear,
quad,
num_reads=max_num_reads)
result_list.append([results, linear, quad])
[r[0].cancel() for r in result_list]
for results, linear, quad in result_list:
# Responses must be canceled or correct
try:
# Did we get the right number of samples?
self.assertTrue(max_num_reads == sum(results.occurrences))
# Make sure the number of occurrences and energies are all correct
for energy, state in zip(results.energies,
results.samples):
self.assertTrue(
energy == evaluate_ising(linear, quad, state))
except CanceledFutureError:
pass