def test_one_batch_history(self):
composition = [[['aj'], ['bj']], [['bj'], ['cj']], [['cj'], ['aj']]]
results = [[1, 0], [1, 0], [1, 0]]
times = [1, 1, 1]
priors = dict()
for k in ["aj", "bj", "cj"]:
priors[k] = ttt.Player(ttt.Gaussian(25., 25.0 / 3), 25.0 / 6,
0.15 * 25.0 / 3)
h1 = ttt.History(composition, results, times, priors)
self.assertAlmostEqual(h1.batches[0].posterior("aj").mu, 22.904, 3)
self.assertAlmostEqual(h1.batches[0].posterior("aj").sigma, 6.010, 3)
self.assertAlmostEqual(h1.batches[0].posterior("cj").mu, 25.110, 3)
self.assertAlmostEqual(h1.batches[0].posterior("cj").sigma, 5.866, 3)
step, i = h1.convergence()
self.assertAlmostEqual(h1.batches[0].posterior("aj").mu, 25.000, 3)
self.assertAlmostEqual(h1.batches[0].posterior("aj").sigma, 5.419, 3)
self.assertAlmostEqual(h1.batches[0].posterior("cj").mu, 25.000, 3)
self.assertAlmostEqual(h1.batches[0].posterior("cj").sigma, 5.419, 3)
priors = dict()
for k in ["aj", "bj", "cj"]:
priors[k] = ttt.Player(ttt.Gaussian(25., 25.0 / 3), 25.0 / 6,
25.0 / 300)
h2 = ttt.History(composition, results, [1, 2, 3], priors)
self.assertAlmostEqual(h2.batches[2].posterior("aj").mu, 22.904, 3)
self.assertAlmostEqual(h2.batches[2].posterior("aj").sigma, 6.011, 3)
self.assertAlmostEqual(h2.batches[2].posterior("cj").mu, 25.111, 3)
self.assertAlmostEqual(h2.batches[2].posterior("cj").sigma, 5.867, 3)
step2, i2 = h2.convergence()
self.assertAlmostEqual(h2.batches[2].posterior("aj").mu, 24.999, 3)
self.assertAlmostEqual(h2.batches[2].posterior("aj").sigma, 5.420, 3)
self.assertAlmostEqual(h2.batches[2].posterior("cj").mu, 25.001, 3)
self.assertAlmostEqual(h2.batches[2].posterior("cj").sigma, 5.420, 3)
def test_one_batch_history(self):
composition = [[['aj'], ['bj']], [['bj'], ['cj']], [['cj'], ['aj']]]
results = [[0, 1], [0, 1], [0, 1]]
bache = [1, 1, 1]
h1 = ttt.History(composition, results, bache)
self.assertAlmostEqual(h1.batches[0].posterior("aj").mu, 22.904, 2)
self.assertAlmostEqual(h1.batches[0].posterior("aj").sigma, 6.010, 2)
self.assertAlmostEqual(h1.batches[0].posterior("cj").mu, 25.110, 2)
self.assertAlmostEqual(h1.batches[0].posterior("cj").sigma, 5.866, 2)
step, i = h1.convergence()
self.assertAlmostEqual(h1.batches[0].posterior("aj").mu, 25.000, 2)
self.assertAlmostEqual(h1.batches[0].posterior("aj").sigma, 5.419, 2)
self.assertAlmostEqual(h1.batches[0].posterior("cj").mu, 25.000, 2)
self.assertAlmostEqual(h1.batches[0].posterior("cj").sigma, 5.419, 2)
h2 = ttt.History(composition, results, [1, 2, 3])
self.assertAlmostEqual(h2.batches[2].posterior("aj").mu, 22.904, 2)
self.assertAlmostEqual(h2.batches[2].posterior("aj").sigma, 6.012, 2)
self.assertAlmostEqual(h2.batches[2].posterior("cj").mu, 25.110, 2)
self.assertAlmostEqual(h2.batches[2].posterior("cj").sigma, 5.867, 2)
step2, i2 = h2.convergence()
self.assertAlmostEqual(h2.batches[2].posterior("aj").mu, 24.997, 2)
self.assertAlmostEqual(h2.batches[2].posterior("aj").sigma, 5.421, 2)
self.assertAlmostEqual(h2.batches[2].posterior("cj").mu, 25.000, 2)
self.assertAlmostEqual(h2.batches[2].posterior("cj").sigma, 5.420, 2)
def test_env_0_TTT(self):
composition = [[["a"], ["b"]], [["a"], ["c"]], [["b"], ["c"]]]
results = [[1, 0], [0, 1], [1, 0]]
h = ttt.History(composition=composition,
results=results,
mu=0.0,
sigma=6.0,
beta=1.0,
gamma=0.05)
step, i = h.convergence(iterations=100)
self.assertAlmostEqual(h.batches[0].posterior("a").mu, 0.001, 3)
self.assertAlmostEqual(h.batches[0].posterior("a").sigma, 2.395, 3)
self.assertAlmostEqual(h.batches[0].posterior("b").mu, -0.001, 3)
self.assertAlmostEqual(h.batches[0].posterior("b").sigma, 2.396, 3)
self.assertAlmostEqual(h.batches[2].posterior("b").mu, 0.001, 3)
self.assertAlmostEqual(h.batches[2].posterior("b").sigma, 2.396, 3)
composition = [[["a"], ["b"]], [["c"], ["a"]], [["b"], ["c"]]]
h = ttt.History(composition=composition,
mu=0.0,
sigma=6.0,
beta=1.0,
gamma=0.05)
step, i = h.convergence(iterations=100)
self.assertAlmostEqual(h.batches[0].posterior("a").mu, 0.001, 3)
self.assertAlmostEqual(h.batches[0].posterior("a").sigma, 2.395, 3)
self.assertAlmostEqual(h.batches[0].posterior("b").mu, -0.001, 3)
self.assertAlmostEqual(h.batches[0].posterior("b").sigma, 2.396, 3)
self.assertAlmostEqual(h.batches[2].posterior("b").mu, 0.001, 3)
self.assertAlmostEqual(h.batches[2].posterior("b").sigma, 2.396, 3)
def test_teams(self):
composition = [[["a", "b"], ["c", "d"]], [["e", "f"], ["b", "c"]],
[["a", "d"], ["e", "f"]]]
results = [[1, 0], [0, 1], [1, 0]]
h = ttt.History(composition=composition,
results=results,
mu=0.0,
sigma=6.0,
beta=1.0,
gamma=0.0)
step, i = h.convergence()
self.assertAlmostEqual(h.batches[0].posterior("a").mu,
h.batches[0].posterior("b").mu, 3)
self.assertAlmostEqual(h.batches[0].posterior("a").sigma,
h.batches[0].posterior("b").sigma, 3)
self.assertAlmostEqual(h.batches[0].posterior("c").mu,
h.batches[0].posterior("d").mu, 3)
self.assertAlmostEqual(h.batches[0].posterior("c").sigma,
h.batches[0].posterior("d").sigma, 3)
self.assertAlmostEqual(h.batches[1].posterior("e").mu,
h.batches[1].posterior("f").mu, 3)
self.assertAlmostEqual(h.batches[1].posterior("e").sigma,
h.batches[1].posterior("f").sigma, 3)
self.assertAlmostEqual(h.batches[0].posterior("a").mu, 4.085, 3)
self.assertAlmostEqual(h.batches[0].posterior("a").sigma, 5.107, 3)
self.assertAlmostEqual(h.batches[0].posterior("c").mu, -0.533, 3)
self.assertAlmostEqual(h.batches[0].posterior("c").sigma, 5.107, 3)
self.assertAlmostEqual(h.batches[2].posterior("e").mu, -3.552, 3)
self.assertAlmostEqual(h.batches[2].posterior("e").sigma, 5.155, 3)
def gamma(self):
composition = [[["a"], ["b"]], [["a"], ["b"]]]
results = [[1, 0], [1, 0]]
h = ttt.History(composition=composition,
results=results,
mu=0.0,
sigma=6.0,
beta=1.0,
gamma=0.0)
mu0, sigma0 = h.batches[1].skills['a'].forward
self.assertAlmostEqual(mu0, 3.33907896)
self.assertAlmostEqual(sigma0, 4.98503276)
h = ttt.History(composition=composition,
results=results,
mu=0.0,
sigma=6.0,
beta=1.0,
gamma=10.0)
mu10, sigma10 = h.batches[1].skills['a'].forward
self.assertAlmostEqual(mu10, 3.33907896)
self.assertAlmostEqual(sigma10, 11.1736543)
#Observaci'on:
# El paquete trueskill python agrega gamma antes de la partida
# devuelve, trueskill.Player(mu=6.555, sigma=9.645)
h = ttt.History(composition=composition,
results=results,
mu=0.0,
sigma=math.sqrt(6.0**2 + 10**2),
beta=1.0,
gamma=10.0)
mu100, sigma100 = h.batches[0].posterior("a")
self.assertAlmostEqual(mu100, 6.555467)
self.assertAlmostEqual(sigma100, 9.6449906)
def test_trueSkill_Through_Time(self):
events = [[["a"], ["b"]], [["a"], ["c"]], [["b"], ["c"]]]
results = [[0, 1], [1, 0], [0, 1]]
h = ttt.History(events, results, [1, 2, 3])
h.batches[0].posteriors()
step, i = h.convergence()
self.assertAlmostEqual(h.batches[0].posterior("a").mu, 25.0002673, 5)
self.assertAlmostEqual(h.batches[0].posterior("a").sigma, 5.41950697,
5)
self.assertAlmostEqual(h.batches[0].posterior("b").mu, 24.9986633, 5)
self.assertAlmostEqual(h.batches[0].posterior("b").sigma, 5.41968377,
5)
self.assertAlmostEqual(h.batches[2].posterior("b").mu, 25.0029304, 5)
self.assertAlmostEqual(h.batches[2].posterior("b").sigma, 5.42076739,
5)
def test_learning_curve(self):
composition = [[['aj'], ['bj']], [['bj'], ['cj']], [['cj'], ['aj']]]
results = [[0, 1], [0, 1], [0, 1]]
h = ttt.History(composition, results, [5, 6, 7])
h.convergence()
lc = h.learning_curves()
for a in lc:
self.assertEqual(
all(lc[a][i][0] < lc[a][i + 1][0]
for i in range(len(lc[a]) - 1)), True)
self.assertEqual(lc["aj"][0][0], 5)
self.assertEqual(lc["aj"][-1][0], 7)
self.assertAlmostEqual(lc["aj"][-1][1].mu, 24.997, 2)
self.assertAlmostEqual(lc["aj"][-1][1].sigma, 5.421, 2)
self.assertAlmostEqual(lc["cj"][-1][1].mu, 25.000, 2)
self.assertAlmostEqual(lc["cj"][-1][1].sigma, 5.420, 2)
def test_learning_curve(self):
composition = [[["aj"], ["bj"]], [["bj"], ["cj"]], [["cj"], ["aj"]]]
results = [[1, 0], [1, 0], [1, 0]]
priors = dict()
for k in ["aj", "bj", "cj"]:
priors[k] = ttt.Player(ttt.Gaussian(25., 25.0 / 3), 25.0 / 6,
25.0 / 300)
h = ttt.History(composition, results, [5, 6, 7], priors)
h.convergence()
lc = h.learning_curves()
self.assertEqual(lc["aj"][0][0], 5)
self.assertEqual(lc["aj"][-1][0], 7)
self.assertAlmostEqual(lc["aj"][-1][1].mu, 24.999, 3)
self.assertAlmostEqual(lc["aj"][-1][1].sigma, 5.420, 3)
self.assertAlmostEqual(lc["cj"][-1][1].mu, 25.001, 3)
self.assertAlmostEqual(lc["cj"][-1][1].sigma, 5.420, 3)
def test_memory_Size(self):
def summarysize(obj):
import sys
from types import ModuleType, FunctionType
from gc import get_referents
BLACKLIST = type, ModuleType, FunctionType
if isinstance(obj, BLACKLIST):
raise TypeError('getsize() does not take argument of type: ' +
str(type(obj)))
seen_ids = set()
size = 0
objects = [obj]
while objects:
need_referents = []
for obj in objects:
if not isinstance(obj,
BLACKLIST) and id(obj) not in seen_ids:
seen_ids.add(id(obj))
size += sys.getsizeof(obj)
need_referents.append(obj)
objects = get_referents(*need_referents)
return size
composition = [[["a"], ["b"]], [["a"], ["c"]], [["b"], ["c"]]]
results = [[1, 0], [0, 1], [1, 0]]
h = ttt.History(composition=composition,
results=results,
times=[0, 10, 20],
mu=0.0,
sigma=6.0,
beta=1.0,
gamma=0.05)
self.assertEqual(summarysize(h) < 13000, True)
self.assertEqual(
summarysize(h.batches) - summarysize(h.agents) < 10000, True)
self.assertEqual(summarysize(h.agents) < 3000, True)
self.assertEqual(
summarysize(h.batches[1]) - summarysize(h.agents) < 3500, True)
self.assertEqual(summarysize(h) < summarysize(composition) * 15, True)
self.assertEqual(summarysize(h.batches[0].skills) < 1650, True)
self.assertEqual(summarysize(h.batches[0].events) < 1850, True)
def test_history_init(self):
events = [[["a"], ["b"]], [["a"], ["c"]], [["b"], ["c"]]]
results = [[0, 1], [1, 0], [0, 1]]
h = ttt.History(events, results, [1, 2, 3])
self.assertEqual(
not ttt.gr_tuple(h.batches[1].max_step, 1e-6)
and not ttt.gr_tuple(h.batches[1].max_step, 1e-6), True)
p0 = h.batches[0].posteriors()
self.assertAlmostEqual(p0["a"].mu, 29.205, 3)
self.assertAlmostEqual(p0["a"].sigma, 7.19448, 3)
observed = h.batches[1].prior_forward["a"].N.sigma
expected = math.sqrt((ttt.GAMMA * 1)**2 +
h.batches[0].posterior("a").sigma**2)
self.assertAlmostEqual(observed, expected)
observed = h.batches[1].posterior("a")
g = ttt.Game([[h.batches[1].prior_forward["a"]],
[h.batches[1].prior_forward["c"]]], [1, 0])
[expected], [c] = g.posteriors
self.assertAlmostEqual(observed.mu, expected.mu, 3)
self.assertAlmostEqual(observed.sigma, expected.sigma, 3)
def test_trueSkill_Through_Time(self):
composition = [[["a"], ["b"]], [["a"], ["c"]], [["b"], ["c"]]]
results = [[1, 0], [0, 1], [1, 0]]
priors = dict()
for k in ["a", "b", "c"]:
priors[k] = ttt.Player(ttt.Gaussian(25., 25.0 / 3), 25.0 / 6,
25.0 / 300)
h = ttt.History(composition, results, [], priors)
step, i = h.convergence()
self.assertEqual(h.batches[2].skills["b"].elapsed, 1)
self.assertEqual(h.batches[2].skills["c"].elapsed, 1)
self.assertAlmostEqual(h.batches[0].posterior("a").mu, 25.0002673, 5)
self.assertAlmostEqual(h.batches[0].posterior("a").sigma, 5.41938162,
5)
self.assertAlmostEqual(h.batches[0].posterior("b").mu, 24.999465, 5)
self.assertAlmostEqual(h.batches[0].posterior("b").sigma, 5.419425831,
5)
self.assertAlmostEqual(h.batches[2].posterior("b").mu, 25.00053219, 5)
self.assertAlmostEqual(h.batches[2].posterior("b").sigma, 5.419696790,
5)
def test_sigma_beta_0(self):
composition = [[["a", "a_b", "b"], ["c", "c_d", "d"]],
[["e", "e_f", "f"], ["b", "b_c", "c"]],
[["a", "a_d", "d"], ["e", "e_f", "f"]]]
results = [[1, 0], [0, 1], [1, 0]]
priors = dict()
for k in ["a_b", "c_d", "e_f", "b_c", "a_d", "e_f"]:
priors[k] = ttt.Player(ttt.Gaussian(mu=0.0, sigma=1e-7),
beta=0.0,
gamma=0.2)
h = ttt.History(composition=composition,
results=results,
priors=priors,
mu=0.0,
sigma=6.0,
beta=1.0,
gamma=0.0)
step, i = h.convergence()
self.assertAlmostEqual(h.batches[0].posterior("a_b").mu, 0.0, 4)
self.assertAlmostEqual(h.batches[0].posterior("a_b").sigma, 0.0, 4)
self.assertAlmostEqual(h.batches[2].posterior("e_f").mu, -0.002, 4)
self.assertAlmostEqual(h.batches[2].posterior("e_f").sigma, 0.2, 4)
def test_env_TTT(self):
composition = [[["a"], ["b"]], [["a"], ["c"]], [["b"], ["c"]]]
results = [[1, 0], [0, 1], [1, 0]]
h = ttt.History(composition=composition,
results=results,
mu=25.,
sigma=25.0 / 3,
beta=25.0 / 6,
gamma=25.0 / 300)
step, i = h.convergence()
self.assertEqual(h.batches[2].skills["b"].elapsed, 1)
self.assertEqual(h.batches[2].skills["c"].elapsed, 1)
self.assertAlmostEqual(h.batches[0].posterior("a").mu, 25.0002673, 5)
self.assertAlmostEqual(h.batches[0].posterior("a").sigma, 5.41938162,
5)
self.assertAlmostEqual(h.batches[0].posterior("b").mu, 24.999465, 5)
self.assertAlmostEqual(h.batches[0].posterior("b").sigma, 5.419425831,
5)
self.assertAlmostEqual(h.batches[2].posterior("b").mu, 25.00053219, 5)
self.assertAlmostEqual(h.batches[2].posterior("b").sigma, 5.419696790,
5)
def test_history_init(self):
composition = [[["aa"], ["b"]], [["aa"], ["c"]], [["b"], ["c"]]]
results = [[1, 0], [0, 1], [1, 0]]
priors = dict()
for k in ["aa", "b", "c"]:
priors[k] = ttt.Player(ttt.Gaussian(25., 25.0 / 3), 25.0 / 6,
0.15 * 25.0 / 3)
h = ttt.History(composition, results, [1, 2, 3], priors)
p0 = h.batches[0].posteriors()
self.assertAlmostEqual(p0["aa"].mu, 29.205, 3)
self.assertAlmostEqual(p0["aa"].sigma, 7.19448, 3)
observed = h.batches[1].skills["aa"].forward.sigma
gamma = 0.15 * 25.0 / 3
expected = math.sqrt((gamma * 1)**2 +
h.batches[0].posterior("aa").sigma**2)
self.assertAlmostEqual(observed, expected)
observed = h.batches[1].posterior("aa")
[expected], [c] = ttt.Game(h.batches[1].within_priors(0),
[0, 1]).posteriors()
self.assertAlmostEqual(observed.mu, expected.mu, 3)
self.assertAlmostEqual(observed.sigma, expected.sigma, 3)
for w, b, h, s in zip(df.white, df.black, df.handicap, df.width)
]
batch = list(df.date)
count = 0
for i in range(len(batch) - 1):
if batch[i] == batch[i + 1]:
batch[i] = count
else:
batch[i] = count
count += 1
batch[-1] = count
print("Composicion: ", composition[:5])
print("Resultado: ", results[:5])
print(" ")
print('Starting TrueSkill')
historyM = thM.History(composition, results, batch)
#print('TrueSkill', historyM.batches[0].posteriors())
pickle_file = open('./trueSkill.pickle', 'wb')
print('Saving TrueSkill data')
pickle.dump(historyM.learning_curves(), pickle_file)
pickle_file.close()
print('Saving evidence')
evidence = {}
for i in range(len(historyM.batches)):
evidence[i] = historyM.batches[i].evidences
evidence_file = open('./trueSkillEvidence.pickle', 'wb')
pickle.dump(evidence, evidence_file)
evidence_file.close()
print('Done!, starting TTT converge')
step, i = historyM.convergence()
from importlib import reload
reload(ttt)
from collections import defaultdict
df = pd.read_csv('./simulacion.csv')
df = df[:2500]
#prior_dict = defaultdict(lambda:env.Rating(0,25/3,0,1/100))
#for h_key in set([(h,s) for h, s in zip(df.handicap, df.width) ]):
# prior_dict[h_key]
#dict(prior_dict)
results = list(df.black_win.map(lambda x: [0, 1] if x else [1, 0]))
composition = [[[str(w)], [str(b)]] for w, b in zip(df.white, df.black)]
batch = range(0, len(results))
h = ttt.History(composition, results, batch)
#history.trueSkill()
h.convergence()
w_mean = [print(w) for t, w, b in zip(h.times, df.white, df.black)]
w_mean = [
h.batches[t].posterior(w).mu
for t, w, b in zip(h.times, df.white, df.black)
]
b_mean = [
h.batches[t].posterior(b).mu
for t, w, b in zip(h.times, df.white, df.black)
]
w_std = [
h.batches[t].posterior(w).sigma
for t, w, b in zip(h.times, df.white, df.black)
