def test_labels_with_time_and_arguments(self):
#Label with time and args
l = L('Gx', (0, 1), time=1.2, args=('1.4', '1.7'))
self.assertEqual(l.time, 1.2)
self.assertEqual(l.args, ('1.4', '1.7'))
self.assertEqual(tuple(l), ('Gx', 4, '1.4', '1.7', 0, 1))
l2 = L(('Gx', ';1.4', ';1.7', 0, 1, '!1.25'))
self.assertEqual(tuple(l2), ('Gx', 4, '1.4', '1.7', 0, 1))
l3 = L(('Gx', ';', '1.4', ';', '1.7', 0, 1, '!', 1.3))
self.assertEqual(tuple(l3), ('Gx', 4, '1.4', '1.7', 0, 1))
self.assertTrue(l == l2 == l3)
#Time without args
l = L('Gx', (0, 1), time=1.2)
self.assertEqual(l.time, 1.2)
self.assertEqual(l.args, ())
self.assertEqual(tuple(l), ('Gx', 0, 1))
#Args without time
l = L('Gx', (0, 1), args=('1.4', ))
self.assertEqual(l.time, 0)
self.assertEqual(l.args, ('1.4', ))
self.assertEqual(tuple(l), ('Gx', 3, '1.4', 0, 1))
def test_circuit_init(self):
#Check that parallel operation labels get converted to circuits properly
opstr = Circuit(((('Gx', 0), ('Gy', 1)), ('Gcnot', 0, 1)))
c = Circuit(layer_labels=opstr, num_lines=2)
print(c._labels)
self.assertEqual(c._labels, (L(
(('Gx', 0), ('Gy', 1))), L('Gcnot', (0, 1))))
def test_only_nonzero_time_is_printed(self):
l = L('GrotX', (0, 1), args=('1.4', ))
self.assertEqual(
str(l), "GrotX;1.4:0:1"
) # make sure we don't print time when it's not given (i.e. zero)
self.assertEqual(l.time, 0.0) # BUT l.time is still 0, not None
l = L('GrotX', (0, 1), args=('1.4', ), time=0.2)
self.assertEqual(str(l), "GrotX;1.4:0:1!0.2"
) # make sure we do print time when it's nonzero
self.assertEqual(l.time, 0.2)
def test_label_time_is_not_hashed(self):
#Ensure that time is not considered in the equality (or hashing) of labels - it's a
# tag-along "comment" that does not change the real value of a Label.
l1 = L('Gx', time=1.2)
l2 = L('Gx')
self.assertEqual(l1, l2)
self.assertTrue(l1.time != l2.time)
l1 = L('Gx', (0, ), time=1.2)
l2 = L('Gx', (0, ))
self.assertEqual(l1, l2)
self.assertTrue(l1.time != l2.time)
def test_layerlizzard(self):
#Test this here b/c auto-gators are associated with parallel operation labels
availability = {'Gcnot': [(0, 1)]}
mdl = pc.build_cloudnoise_model_from_hops_and_weights(
2, ['Gx', 'Gy', 'Gcnot'], {},
None,
availability,
None,
"line",
maxIdleWeight=1,
maxhops=1,
extraWeight1Hops=0,
extraGateWeight=1,
sparse=True,
sim_type="map",
parameterization="H+S")
# mdl[('Gx',0)].factorops # Composed([fullTargetOp,fullIdleErr,fullLocalErr])
self.assertEqual(
set(mdl.get_primitive_op_labels()),
set([
L('Gx', 0),
L('Gy', 0),
L('Gx', 1),
L('Gy', 1),
L('Gcnot', (0, 1))
]))
#But we can *compute* with circuits containing parallel labels...
parallelLbl = L([('Gx', 0), ('Gy', 1)])
with self.assertRaises(KeyError):
mdl.operation_blks[parallelLbl]
opstr = pygsti.obj.Circuit((parallelLbl, ))
probs = mdl.probs(opstr)
print(probs)
expected = {
('00', ): 0.25,
('01', ): 0.25,
('10', ): 0.25,
('11', ): 0.25
}
for k, v in probs.items():
self.assertAlmostEqual(v, expected[k])
def testLabels(self):
labels = []
labels.append(L('Gx', 0)) # a LabelTup
labels.append(L('Gx', (0, 1))) # a LabelTup
labels.append(L(('Gx', 0, 1))) # a LabelTup
labels.append(L('Gx')) # a LabelStr
labels.append(L('Gx', None)) #still a LabelStr
labels.append(L([('Gx', 0),
('Gy', 0)])) # a LabelTupTup of LabelTup objs
labels.append(L(
(('Gx', None), ('Gy', None)))) # a LabelTupTup of LabelStr objs
labels.append(L([('Gx', 0)])) # just a LabelTup b/c only one component
labels.append(L([L('Gx'), L('Gy')])) # a LabelTupTup of LabelStrs
labels.append(L(L('Gx'))) # Init from another label
for l in labels:
native = l.tonative()
print(l, " (", type(l), "): native =", native)
if isinstance(l, label.LabelTupTup):
print(
" comps: ", ", ".join(
["%s (%s)" % (c, str(type(c))) for c in l.components]))
j = jsoncodec.encode_obj(l, False)
#print("Json: ",j)
l3 = jsoncodec.decode_obj(j, False)
#print("Unjsoned ", l3, " a ",type(l3))
self.assertEqual(type(l), type(l3))
def test_label_methods():
def test_to_native(label):
native = label.tonative()
# TODO assert correctness
from_native = L(native)
assert label == from_native
def test_pickle(label):
s = pickle.dumps(label)
l2 = pickle.loads(s)
assert type(label) == type(l2)
def test_json_encode(label):
j = jsoncodec.encode_obj(label, False)
l2 = jsoncodec.decode_obj(j, False)
assert type(label) == type(l2)
labels = [
L('Gx', 0), # a LabelTup
L('Gx', (0, 1)), # a LabelTup
L(('Gx', 0, 1)), # a LabelTup
L('Gx'), # a LabelStr
L('Gx', None), # still a LabelStr
L([('Gx', 0), ('Gy', 0)]), # a LabelTupTup of LabelTup objs
L((('Gx', None), ('Gy', None))), # a LabelTupTup of LabelStr objs
L([('Gx', 0)]), # just a LabelTup b/c only one component
L([L('Gx'), L('Gy')]), # a LabelTupTup of LabelStrs
L(L('Gx')) # Init from another label
]
for lbl in labels:
yield test_to_native, lbl
yield test_pickle, lbl
yield test_json_encode, lbl
def test_to_native(label):
native = label.tonative()
# TODO assert correctness
from_native = L(native)
assert label == from_native