def test_predicate_node_sanity():
with kb_context("test_predicate_node_sanity"):
BOOP, BOOPType = define_fact("BOOP",{"A": "string", "B" : "number"})
kb = KnowledgeBase()
pn = get_alpha_predicate_node(BOOPType,"B", "<",9)
pnc = cast(pn, BasePredicateNodeType)
png = cast(pn, GenericAlphaPredicateNodeType)
attrs_pn = np.array([struct_get_attr_offset(pnc,x[0]) for x in base_subscriber_fields+basepredicate_node_fields])
attrs_pnc = np.array([struct_get_attr_offset(pn, x[0]) for x in base_subscriber_fields+alpha_predicate_node_fields])
attrs_png = np.array([struct_get_attr_offset(png,x[0]) for x in base_subscriber_fields+alpha_predicate_node_fields])
assert np.array_equal(attrs_pn,attrs_pnc[:len(attrs_pn)])
assert np.array_equal(attrs_pn,attrs_png[:len(attrs_pn)])
assert np.array_equal(attrs_pnc,attrs_png)
ld = generate_link_data(pnc, kb)
assert len(filter_alpha(png, ld, np.arange(5)))==0
pn = get_beta_predicate_node(BOOPType,"B", "<", BOOPType,"B")
pnc = cast(pn, BasePredicateNodeType)
png = cast(pn, GenericBetaPredicateNodeType)
ld = generate_link_data(pnc, kb)
attrs_pn = np.array([struct_get_attr_offset(pnc,x[0]) for x in base_subscriber_fields+basepredicate_node_fields])
attrs_pnc = np.array([struct_get_attr_offset(pn, x[0]) for x in base_subscriber_fields+beta_predicate_node_fields])
attrs_png = np.array([struct_get_attr_offset(png,x[0]) for x in base_subscriber_fields+beta_predicate_node_fields])
assert np.array_equal(attrs_pn,attrs_pnc[:len(attrs_pn)])
assert np.array_equal(attrs_pn,attrs_png[:len(attrs_pn)])
assert np.array_equal(attrs_pnc,attrs_png)
assert len(filter_beta(png, ld, np.arange(5), np.arange(5))) == 0
def _benchmark_setup():
with kb_context("test_predicate_node"):
kb = KnowledgeBase()
pn = get_alpha_predicate_node(BOOPType,"B", "<",50)
ld = generate_link_data(pn,kb)
# kb.add_subscriber(pn)
return (kb, pn, ld), {}
def _beta_setup():
with kb_context("test_predicate_node"):
kb = KnowledgeBase()
pn = get_beta_predicate_node(BOOPType,"B", "<", BOOPType,"B")
# kb.add_subscriber(pn)
ld = generate_link_data(pn, kb)
idrecs = np.empty((100,),dtype=np.int64)
for i in range(100):
idrecs[i] = kb.declare(BOOP("?",i))
return (kb, pn, ld, np.arange(100), idrecs), {}
def link_pterm_instance(pterm, kb):
link_data = generate_link_data(pterm.pred_node, kb)
# if(copy): pterm = pterm_copy(pterm)
pterm.link_data = link_data
pterm.is_linked = True
return pterm
def test_alpha_predicate_node():
with kb_context("test_alpha_predicate_node"):
BOOP, BOOPType = define_fact("BOOP",{"A": "string", "B" : "number"})
kb = KnowledgeBase()
pn = get_alpha_predicate_node(BOOPType,"B", "<",9)
ld = generate_link_data(pn, kb)
# kb.add_subscriber(pn)
x = BOOP("x",7)
y = BOOP("y",11)
z = BOOP("z",8)
assert len(ld.truth_values) == 0
kb.declare(x)
kb.declare(y)
kb.declare(z)
# njit_update(pn)
inds = filter_alpha(pn, ld, np.arange(3))
assert np.array_equal(inds, [0,2])
assert all(ld.truth_values[:3,0] == [1,0,1])
kb.modify(x,"B", 100)
kb.modify(y,"B", 3)
kb.modify(z,"B", 88)
#Checks doesn't change before update
assert all(ld.truth_values[:3,0] == [1,0,1])
inds = filter_alpha(pn, ld, np.arange(3))
assert np.array_equal(inds, [1])
assert all(ld.truth_values[:3,0] == [0,1,0])
kb.retract(x)
kb.retract(y)
kb.retract(z)
#Checks doesn't change before update
assert all(ld.truth_values[:3,0] == [0,1,0])
inds = filter_alpha(pn, ld, np.arange(3))
print(inds)
assert np.array_equal(inds, [])
# print(ld.truth_values)
#Checks that retracted facts show up as u1.nan = 0XFF
assert all(ld.truth_values[:3,0] == [0xFF,0xFF,0xFF])
kb.declare(x)
kb.declare(y)
kb.declare(z)
kb.modify(z,"A","Z")
kb.modify(x,"B",0)
kb.modify(y,"B",0)
kb.modify(z,"B",0)
inds = filter_alpha(pn, ld, np.arange(3))
assert np.array_equal(inds, [0,1,2])
assert all(ld.truth_values[:3,0] == [1,1,1])
def test_beta_predicate_node_2_typed():
with kb_context("test_beta_predicate_node_2_typed"):
BOOP1, BOOP1Type = define_fact("BOOP1",{"A": "number", "B" : "string"})
BOOP2, BOOP2Type = define_fact("BOOP2",{"A": "string", "B" : "number"})
kb = KnowledgeBase()
pn = get_beta_predicate_node(BOOP1Type,"A", "<", BOOP2Type,"B")
ld = generate_link_data(pn, kb)
# kb.add_subscriber(pn)
x1,x2 = BOOP1(7,"x"), BOOP2("x",7.5) #<- slightly different
y1,y2 = BOOP1(11,"y"), BOOP2("y",11)
z1,z2 = BOOP1(8,"z"), BOOP2("z",8)
assert len(ld.truth_values) == 0
kb.declare(x1); kb.declare(x2)
kb.declare(y1); kb.declare(y2)
kb.declare(z1); kb.declare(z2)
inds = filter_beta(pn, ld, np.arange(3), np.arange(3))
# print(inds)
# print(ld.truth_values)
assert np.array_equal(inds, [[0,0],[0, 1], [0, 2], [2, 1]])
assert all(ld.truth_values[0,:3] == [1,1,1])
assert all(ld.truth_values[1,:3] == [0,0,0])
assert all(ld.truth_values[2,:3] == [0,1,0])
q1, q2 = BOOP1(-7,"q"), BOOP2("q",-7)
r1, r2 = BOOP1(-11,"r"), BOOP2("r",-11)
t1, t2 = BOOP1(-8,"t"), BOOP2("t",-8)
kb.declare(q1); kb.declare(q2)
kb.declare(r1); kb.declare(r2)
kb.declare(t1); kb.declare(t2)
inds = filter_beta(pn, ld, np.arange(6), np.arange(6))
# print(inds)
# print(ld.truth_values[:6,:6])
assert np.array_equal(inds,[[0,0],[0,1],[0,2],[2,1],[3,0],[3,1],
[3,2],[4,0],[4,1],[4,2],[4,3],[4,5],[5,0],[5,1],[5,2],[5,3]])
assert all(ld.truth_values[0,:6] == [1,1,1,0,0,0])
assert all(ld.truth_values[1,:6] == [0,0,0,0,0,0])
assert all(ld.truth_values[2,:6] == [0,1,0,0,0,0])
assert all(ld.truth_values[3,:6] == [1,1,1,0,0,0])
assert all(ld.truth_values[4,:6] == [1,1,1,1,0,1])
assert all(ld.truth_values[5,:6] == [1,1,1,1,0,0])
kb.modify(r1,"A", 0); kb.modify(r2,"B", 0)
kb.modify(y1,"A", 0); kb.modify(y2,"B", 0)
inds = filter_beta(pn, ld, np.arange(6), np.arange(6))
# print(ld.truth_values[:6,:6])
assert all(ld.truth_values[0,:6] == [1,0,1,0,0,0])
assert all(ld.truth_values[1,:6] == [1,0,1,0,0,0])
assert all(ld.truth_values[2,:6] == [0,0,0,0,0,0])
assert all(ld.truth_values[3,:6] == [1,1,1,0,1,0])
assert all(ld.truth_values[4,:6] == [1,0,1,0,0,0])
assert all(ld.truth_values[5,:6] == [1,1,1,1,1,0])
kb.retract(r1); kb.retract(r2)
kb.retract(y1); kb.retract(y2)
inds = filter_beta(pn, ld, np.arange(6), np.arange(6))
# print(ld.truth_values[:6,:6])
assert all(ld.truth_values[0,:6] == [1,0,1,0,0,0])
assert all(ld.truth_values[1,:6] == [0,0,0,0,0,0])
assert all(ld.truth_values[2,:6] == [0,0,0,0,0,0])
assert all(ld.truth_values[3,:6] == [1,0,1,0,0,0])
assert all(ld.truth_values[4,:6] == [0,0,0,0,0,0])
assert all(ld.truth_values[5,:6] == [1,0,1,1,0,0])
def test_beta_predicate_node_1_typed():
with kb_context("test_beta_predicate_node_1_typed"):
BOOP, BOOPType = define_fact("BOOP",{"A": "string", "B" : "number"})
kb = KnowledgeBase()
pn = get_beta_predicate_node(BOOPType,"B", "<", BOOPType,"B")
# png = cast(pn, GenericBetaPredicateNodeType)
ld = generate_link_data(pn, kb)
# kb.add_subscriber(pn)
x = BOOP("x",7)
y = BOOP("y",11)
z = BOOP("z",8)
assert len(ld.truth_values) == 0
kb.declare(x)
kb.declare(y)
kb.declare(z)
# njit_update(pn)
inds = filter_beta(pn, ld, np.arange(3), np.arange(3))
# print("B")
# print(inds)
assert np.array_equal(inds, [[0, 1], [0, 2], [2, 1]])
# print(ld.truth_values)
assert all(ld.truth_values[0,:3] == [0,1,1])
assert all(ld.truth_values[1,:3] == [0,0,0])
assert all(ld.truth_values[2,:3] == [0,1,0])
q = BOOP("q",-7)
r = BOOP("r",-11)
t = BOOP("t",-8)
kb.declare(q)
kb.declare(r)
kb.declare(t)
inds = filter_beta(pn, ld, np.arange(6), np.arange(6))
# print(inds)
assert np.array_equal(inds,[[0,1],[0,2],[2,1],[3,0],[3,1],
[3,2],[4,0],[4,1],[4,2],[4,3],[4,5],[5,0],[5,1],[5,2],[5,3]])
# njit_update(pn)
# print(ld.truth_values[:6,:6])
assert all(ld.truth_values[0,:6] == [0,1,1,0,0,0])
assert all(ld.truth_values[1,:6] == [0,0,0,0,0,0])
assert all(ld.truth_values[2,:6] == [0,1,0,0,0,0])
assert all(ld.truth_values[3,:6] == [1,1,1,0,0,0])
assert all(ld.truth_values[4,:6] == [1,1,1,1,0,1])
assert all(ld.truth_values[5,:6] == [1,1,1,1,0,0])
kb.modify(r,"B", 0)
kb.modify(y,"B", 0)
inds = filter_beta(pn, ld, np.arange(6), np.arange(6))
# njit_update(pn)
# print(ld.truth_values[:6,:6])
assert all(ld.truth_values[0,:6] == [0,0,1,0,0,0])
assert all(ld.truth_values[1,:6] == [1,0,1,0,0,0])
assert all(ld.truth_values[2,:6] == [0,0,0,0,0,0])
assert all(ld.truth_values[3,:6] == [1,1,1,0,1,0])
assert all(ld.truth_values[4,:6] == [1,0,1,0,0,0])
assert all(ld.truth_values[5,:6] == [1,1,1,1,1,0])
kb.retract(r)
kb.retract(y)
inds = filter_beta(pn, ld, np.arange(6), np.arange(6))
# njit_update(pn)
# print(ld.truth_values[:6,:6])
assert all(ld.truth_values[0,:6] == [0,0,1,0,0,0])
assert all(ld.truth_values[1,:6] == [0,0,0,0,0,0])
assert all(ld.truth_values[2,:6] == [0,0,0,0,0,0])
assert all(ld.truth_values[3,:6] == [1,0,1,0,0,0])
assert all(ld.truth_values[4,:6] == [0,0,0,0,0,0])
assert all(ld.truth_values[5,:6] == [1,0,1,1,0,0])