def test_add_one_success(self):
gen = BetaPrior()
gen.add_one(mean=0.5, variance=0.2, effective_size=10, label="option1")
assert len(gen.priors) == 1
assert isinstance(gen.priors, dict)
assert isinstance(gen.priors["option1"], dict)
assert gen.priors == {"option1": {"a": 5, "b": 5}}
def test_add_one_value_errors(self):
gen = BetaPrior()
mean_errors = [-1, 0, 1.2]
variance_errors = [-1, 0, 1]
effective_size_errors = [-1, 0]
with pytest.raises(ValueError):
gen.add_one(mean=0.5,
variance=0.5,
effective_size=10,
label="option1")
for item in mean_errors:
gen.add_one(mean=item,
variance=0.1,
effective_size=10,
label="option1")
for item in variance_errors:
gen.add_one(mean=0.5,
variance=item,
effective_size=10,
label="option1")
for item in effective_size_errors:
gen.add_one(mean=0.5,
variance=0.2,
effective_size=item,
label="option1")
class ThompsonSampling:
def __init__(self):
self.beta = BetaPrior()
self.experimento = BernoulliExperiment(priors=self.beta)
def atualizar(self, reward):
self.experimento.add_rewards(reward)
def inicializar(self, estrategias):
for opcao in estrategias:
self.beta.add_one(mean=0.5,
variance=0.2,
effective_size=10,
label=opcao)
def escolher(self):
return self.experimento.choose_arm()
from thompson_sampling.bernoulli import BernoulliExperiment
experiment = BernoulliExperiment(arms=2)
from thompson_sampling.bernoulli import BernoulliExperiment
from thompson_sampling.priors import BetaPrior
pr = BetaPrior()
pr.add_one(mean=0.5, variance=0.2, effective_size=10, label="option1")
pr.add_one(mean=0.6, variance=0.3, effective_size=30, label="option2")
experiment = BernoulliExperiment(priors=pr)
experiment.choose_arm()
#Update rewards
rewards = [{"label":"option1", "reward":1}, {"label":"option2", "reward":0}]
experiment.add_rewards(rewards)