def compute_one(
p,
n,
n_alt_plus,
n_alt_minus,
):
B = simple.Simple(n=n, p=p).flatten()
Bplus = simple.Simple(n=n_alt_plus, p=p).flatten()
Bminus = simple.Simple(n=n_alt_minus, p=p).flatten()
print "Computing Plus Alternative"
SPlus = compute_alt(B, Bplus)
print "Computing Minus Alternative"
SMinus = compute_alt(B, Bminus)
return SPlus, SMinus
def SfNplot():
ntrue = 100
nalt0 = 99
nalt1 = 90
S100_1 = simple.Simple(n=ntrue, p=.1).flatten()
S99_1 = simple.Simple(n=nalt0, p=.1).flatten()
S90_1 = simple.Simple(n=nalt1, p=.1).flatten()
S100_5 = simple.Simple(n=ntrue, p=.5).flatten()
S99_5 = simple.Simple(n=nalt0, p=.5).flatten()
S90_5 = simple.Simple(n=nalt1, p=.5).flatten()
S100_9 = simple.Simple(n=ntrue, p=.9).flatten()
S99_9 = simple.Simple(n=nalt0, p=.9).flatten()
S90_9 = simple.Simple(n=nalt1, p=.9).flatten()
fig, ax = my_subplots_for_sfn()
S100_1.compare_3bars(S99_1,
S90_1,
ntrue, (fig, ax[2][0]),
xlab='Number of open channels (k)')
S100_5.compare_3bars(S99_5, S90_5, ntrue, (fig, ax[1][0]))
S100_9.compare_3bars(S99_9, S90_9, ntrue, (fig, ax[0][0]))
se_regions(2, False, (fig, ax[0][1]))
se_regions(1, False, (fig, ax[1][1]))
se_regions(0, False, (fig, ax[2][1]))
se_regions(2, True, (fig, ax[0][1]))
se_regions(1, True, (fig, ax[1][1]))
se_regions(0, True, (fig, ax[2][1]))
venn((30, .1), .1, .05, (fig, ax[0][0]))
ax[0][0].text(18, .085, 'n-10%=90')
ax[0][0].text(41, .105, 'n=100')
ax[0][0].text(4.5, .105, 'n-1=99')
ax[0][0].text(5, .12, 'Number of Channels:')
ax[0][0].text(5, .13, 'Probability of Opening: p=0.9')
ax[1][0].text(5, .13, 'Probability of Opening: p=0.5')
ax[2][0].text(20, .13, 'Probability of Opening: p=0.1')
ax[0][1].text(5, 130, 'Probability of Opening: p=0.9')
ax[0][1].text(5, 120, 'Number of Channels in Alternative:')
ax[0][1].text(10, 110, '(Falsified with 95% Confidence)')
ax[1][1].text(5, 1300, 'Probability of Opening: p=0.5')
ax[2][1].text(5, 13000, 'Probability of Opening: p=0.1')
ax[0][1].set_ylabel('Needed Sample Size')
ax[1][1].set_ylabel('Needed Sample Size')
ax[2][1].set_ylabel('Needed Sample Size')
ax[2][1].set_xlabel('Number of Channels in True Model (n)')
ax[0][1].text(5, 90, 'n-1')
ax[0][1].text(5, 72.5, 'n-10%')
ax[0][1].text(43, 90, 'n+1')
ax[0][1].text(43, 72.5, 'n+10%')
venn2((29, 90), .1, .06, (fig, ax[0][1]))
fig.show()
return fig, ax
def test_hello_world():
import simple
os.chdir('/tmp')
assert simple.Simple().hello_world() == 'Hello world!'
class Simple2:
def __init__(self):
self.info = "SimpleClass2"
class Simple3(simple.Simple):
def __init__(self):
simple.Simple.__init__(self)
text = "text in simple"
assert simple.text == text
_s = simple.Simple()
_s3 = Simple3()
assert _s.info == _s3.info
import recursive_import
_s = recursive_import.myClass()
assert str(_s) == "success!"
import from_import_test.b
assert from_import_test.b.v == 1
import from_import_test.c
assert from_import_test.c.v == 1
# test of keyword "global" in functions of an imported module
# issue 44
funcs = []
for i in [1, 2]:
def f(x=i):
return x
funcs.append(f)
assert funcs[0]() == 1
assert funcs[1]() == 2
# issue 45
import simple
assert simple.Simple().info == "SimpleClass"
# issue 46
class A:
COUNTER = 0
def __init__(self):
self.COUNTER += 1
a = A()
class A:
def test_hello_world(self):
import simple
self.assertEqual(simple.Simple().hello_world('!'), 'Hello world!')
def test_patch(self, patched_hw):
import simple
simple.Simple().hello_world('!')
patched_hw.assert_called_once_with('!')
def tearDown(self):
simple.Simple().finishUp()
def setUp(self):
simple.Simple().getReady()
sy_loss = - tf.reduce_mean(sy_states[-1], name="loss")
sy_trainable = [neg]
sy_opt = tf.train.GradientDescentOptimizer(sy_lr, name="GDOpt")
sy_train_op = sy_opt.minimize(sy_loss, var_list=sy_trainable)
return {"cl_in": x, "cl_label": y, "cl_logit": states[-1],
"cl_prob": cl_prob, "cl_out": cl_out, "cl_op": cl_train_op,
"cl_loss": cl_loss, "cl_acc": accuracy,
"sy_sample": neg, "sy_op": sy_train_op,
"sy_logit": sy_states[-1], "sy_loss": sy_loss}
if __name__ == "__main__":
m = build_arch()
d = simple.Simple()
sess = tf.Session()
sess.run(tf.global_variables_initializer())
neg_data_seq = [sess.run(m["sy_sample"])]
for epoch in range(151):
d.set_neg(neg_data_seq[-1])
plt.plot(neg_data_seq[-1][:,0], neg_data_seq[-1][:,1], ".")
plt.xlim([-10, 10])
plt.ylim([-10, 10])
plt.savefig("figure/"+str(epoch)+".jpg")
plt.show()
for i in range(2000):
