class RosenbrockWithList(Rosenbrock):
"""
A 3 dimensional implementation of the Rosenbrock function
Result at (1,1,1) is 0.
"""
l = parameter.List(np.arange(0, 10),
repeat=True,
doc='list parameter for testing')
x = parameter.Uniform(-10,
10,
1.5,
3.0,
-10,
10,
doc='x value of Rosenbrock function')
y = parameter.Uniform(-10,
10,
1.5,
3.0,
-10,
10,
doc='y value of Rosenbrock function')
z = parameter.Uniform(-10,
10,
1.5,
3.0,
-10,
10,
doc='z value of Rosenbrock function')
class RosenbrockWithConstant(Rosenbrock):
"""
A 3 dimensional implementation of the Rosenbrock function
Result at (1,1,1) is 0.
"""
c = parameter.Constant(0, doc='Constant offset, should stay 0')
x = parameter.Uniform(-10,
10,
1.5,
3.0,
-10,
10,
doc='x value of Rosenbrock function')
y = parameter.Uniform(-10,
10,
1.5,
3.0,
-10,
10,
doc='y value of Rosenbrock function')
z = parameter.Uniform(-10,
10,
1.5,
3.0,
-10,
10,
doc='z value of Rosenbrock function')
class SpotSetupClassAttributes(SpotSetupBase):
"""
A Test case with two parameters as class parameters (a,b)
and 2 given from the parameter function
"""
a = parameter.Uniform(-1, 1)
b = parameter.Uniform(-1, 1)
c = parameter.Uniform(-1, 1)
d = parameter.Uniform(-1, 1)
class SpotSetupMixedParameterList(SpotSetupBase):
"""
A Test case with two parameters as class parameters (a,b)
and 2 given from the parameter function
"""
a = parameter.Uniform(0, 1)
b = parameter.Uniform(1, 2)
def parameters(self):
return parameter.generate(
[parameter.Uniform(name, -1, 1) for name in 'cd'])
def __init__(self, chn_fn=None):
if chn_fn is None:
chn_fn = 'channel_pars.csv'
fn = _join(_par_dir, chn_fn)
fp = open(fn)
csv_rdr = csv.DictReader(fp)
pars = {}
pars_list = []
for row in csv_rdr:
par = row['parameter']
optimize = int(row['optimize'])
optguess = float(row['optguess'])
if optimize == 0: # fixed parameter
pars[par] = optguess
else:
pars[par] = '{%s}' % par
distribution = row['distribution']
if distribution.lower() != 'uniform':
raise NotImplementedError
low = float(row['low'])
high = float(row['high'])
stepsize = row['stepsize']
if len(stepsize) == 0:
stepsize = None
else:
stepsize = float(stepsize)
if stepsize is None:
pars_list.append(
parameter.Uniform(name='chndefs:' + par,
low=low,
high=high,
optguess=optguess))
else:
pars_list.append(
parameter.Uniform(name='chndefs:' + par,
low=low,
high=high,
optguess=optguess,
stepsize=stepsize))
fp.close()
self.pars = pars
self.pars_list = pars_list
def test_find_constant_parameters(self):
flex = parameter.Uniform('flexible', 0, 1)
p = parameter.generate([flex, self.const])
constant_parameters = parameter.find_constant_parameters(p)
self.assertFalse(constant_parameters[0],
'Flexible parameter detected as constant')
self.assertTrue(constant_parameters[1],
'Constant parameter not detected')
def test_uniform_has_correct_statistics(self):
unif = parameter.Uniform("test", 0, 1)
# Generate 10k random numbers
nums = [unif() for _ in range(10000)]
self.assertAlmostEqual(np.mean(nums), (1 - 0) / 2.0, self.tolerance,
"Mean of Unif(0, 1) should be 1/2")
self.assertAlmostEqual(np.var(nums), 1.0 / 12, self.tolerance,
"Variance of Unif(0, 1) should be 1/12")
def paramcontrol(name, calibrate, value, low, high, guess):
d = {
'calibrate': calibrate,
'value': value,
'param': parameter.Uniform(name, low=low, high=high, optguess=guess),
'name': name
}
return d
def test_uniform_has_correct_bounds(self):
unif = parameter.Uniform("test", low=0, high=1)
print(unif.minbound)
self.assertGreaterEqual(unif.minbound, 0)
self.assertLessEqual(unif.maxbound, 1)
def test_uniform_is_callable(self):
unif = parameter.Uniform("test", low=0, high=1)
self.assertTrue(callable(unif),
"Uniform param instance should be callable")
def test_uniform_processes_non_keyword_args(self):
_ = parameter.Uniform("test", 0, 1)
def __init__(self):
self.parameters = [parameter.Uniform(name, -1, 1) for name in 'abcd']
def parameters(self):
return parameter.generate(
[parameter.Uniform(name, -1, 1) for name in 'cd'])
