def runproblogsampling(s, n=5, output='html'):
model = PrologString(s)
samples = plsample.sample(model, n=n, tuples=True)
result = ''
for sample in samples:
result += ','.join(
str(Term(query[0], *query[1:-1])) for query in sample) + '<br/>'
if output == 'html':
return '<pre>{}</pre>'.format(result)
return result
def runproblogsampling(s, n=5, output="html"):
model = PrologString(s)
samples = plsample.sample(model, n=n, tuples=True)
result = ""
for sample in samples:
result += (
",".join(str(Term(query[0], *query[1:-1])) for query in sample) + "<br/>"
)
if output == "html":
return "<pre>{}</pre>".format(result)
return result
def create_interpretations(p, n):
result = sample.sample(p, n, format='dict')
resultNew = []
for i in range(n):
interpr = next(result)
toRemove = []
for j in interpr:
if random.uniform(0, 1) <= 0.3:
toRemove.append(j)
for j in toRemove:
dict.pop(interpr, j)
resultNew.append(interpr)
return resultNew
def generate_dataset(model, size, missing_rate):
datapoints = []
result = sample.sample(model, n=size, format='dict')
for s in result:
datapoints.append(s)
output = pd.DataFrame(datapoints)
output = output.applymap(lambda x: 1 if x == True else 0)
# applying missing rate
if missing_rate != 0:
# transform values greater than 1 to percentage
if missing_rate >= 1:
missing_rate = float(missing_rate) / 100.
# missing values are represented as NaN
output = output.applymap(lambda x: x
if random() > missing_rate else float('NaN'))
return output
def missing_sample(modeltext, missing_rate):
"""Returns a relational observation sampled from modeltext (which should be a structure in
ProbLog syntax).
"""
model = PrologString(modeltext)
result = sample.sample(model, n=1, format='dict')
output = ""
for s in result:
for q in s:
value = s[q]
if random() > missing_rate / 100:
if value == True:
output += "evidence({},true).\n".format(str(q))
else:
output += "evidence({},false).\n".format(str(q))
else:
output += "evidence({},none).\n".format(str(q))
return output
def generate_interpretations(problog_filename, m):
pl = PrologFile(problog_filename)
interpretations = []
for interpretation in sample(pl, n=m, as_evidence=True):
temp = []
# drop each observation with a probability of 30%
for i in interpretation.splitlines():
if random.random() <= 0.7:
temp.append(i)
interpretations.append("\n".join(temp))
interpretations_file = os.path.abspath(
os.path.join(os.path.dirname(__file__), "files",
"interpretations.txt"))
with open(interpretations_file, "w") as out:
sep = os.linesep + separator_1 + os.linesep
out.write(sep.join(interpretations))
return interpretations
node(n4, 5, [n1]). node(n5, 5, [n2]). node(n6, 5, [n2]). link(n1,n3,40,5). % link(source node, target node, latency, bw) link(n1,n2,10,5). link(n1,n4,10,5). link(n5,n2,10,5). link(n6,n2,10,5). route(s1, path(n5, n1, [n5, n2]), 45, 100). route(s1, path(n3, n1, [n3]), 45, 20). route(s1, path(n6, n1, [n6, n2]), 45, 20). route(s2, path(n4, n4, []), 45, 20). query(nop(s1)). """ # result = get_evaluatable().create_from(PrologString(model)).evaluate() model = PrologString(modeltext) result2 = get_evaluatable().create_from(model).evaluate() print(result2) result = sample.sample(model, n=3, format='dict') print(result) # https: // github.com / yuce / pyswip / issues / 48
my_uniform(0,10)::a.
0.5::b.
c :- value(a, A), A >= 3; b.
query(a).
query(b).
query(c).
"""
modeltext_new = """
0.5::b.
c :- value(a, A), A >= 3; b.
query(a).
query(b).
query(c).
"""
s = SimpleProgram()
unif = Term('my_uniform')(Constant(0),Constant(10))
a = Term('a', p=unif)
s += a
for clause in PrologString(modeltext_new):
s += clause
model = PrologString(modeltext)
# Pass the mapping between name and function using the distributions parameter.
x = list(PrologString(modeltext).__iter__())
x2 = list(s.__iter__())
result = list(sample.sample(model, n=3, format='dict', distributions={'my_uniform': integer_uniform}))
result2 = list(sample.sample(s, n=3, format='dict', distributions={'my_uniform': integer_uniform}))
def sample_model(model_text):
model = PrologString(model_text)
result = sample.sample(model, n=10, format='dict')
for i in result:
print(i)
def get_samples(modeltext, sample_size):
model = PrologString(modeltext)
result = sample.sample(model, n=sample_size, format='dict')
return result