def test_invar_discrete():
from scipy.stats import randint
invar = InVar('randint',
ndraws=1000,
dist=randint,
distkwargs={
'low': 1,
'high': 5
},
seed=invarseeds[0],
samplemethod=SampleMethod.RANDOM)
assert invar.stats().mean == pytest.approx(2.538)
def plot_testing():
from scipy.stats import norm
invar = InVar('norm',
ndraws=1000,
dist=norm,
distkwargs={
'loc': 10,
'scale': 4
},
seed=invarseeds[1],
samplemethod=SampleMethod.RANDOM,
firstcaseismedian=False)
fig, ax = invar.plot()
plt.show()
def invar_custom_dist():
return InVar('custom',
ndraws=1000,
dist=custom_dist,
distkwargs=dict(),
ninvar=1,
samplemethod=SampleMethod.RANDOM,
seed=invarseeds[2],
firstcaseismedian=False)
def invar_lognorm_random():
from scipy.stats import lognorm
return InVar('lognorm',
ndraws=1000,
dist=lognorm,
distkwargs={
's': lognorm_sigma,
'scale': np.exp(lognorm_mu)
},
seed=invarseeds[1],
samplemethod=SampleMethod.RANDOM,
firstcaseismedian=False)
def invar_norm_random():
from scipy.stats import norm
return InVar('norm',
ndraws=1000,
dist=norm,
distkwargs={
'loc': 10,
'scale': 4
},
seed=invarseeds[1],
samplemethod=SampleMethod.RANDOM,
firstcaseismedian=False)
def test_invar_norm_sobol_warning():
from scipy.stats import norm
with pytest.warns(UserWarning, match='Infinite value'):
InVar('norm',
ndraws=1000,
dist=norm,
distkwargs={
'loc': 10,
'scale': 4
},
ninvar=1,
samplemethod=SampleMethod.SOBOL,
seed=invarseeds[1],
firstcaseismedian=False)
def test_invar_nummap():
invar = InVar('map',
ndraws=10,
dist=custom_dist,
distkwargs=dict(),
ninvar=1,
nummap={
1: 'a',
5: 'e',
6: 'f'
},
samplemethod=SampleMethod.RANDOM,
seed=invarseeds[3],
firstcaseismedian=False)
assert invar.vals == ['f', 'e', 'f', 'f', 'a', 'e', 'e', 'a', 'e', 'e']
def case():
seed = 74494861
from scipy.stats import norm
invar = {
'Test':
InVar('Test',
ndraws=10,
ninvar=1,
dist=norm,
distkwargs={
'loc': 10,
'scale': 4
},
seed=seed,
firstcaseismedian=True)
}
case = Case(ncase=0,
ismedian=False,
invars=invar,
constvals=dict(),
seed=seed)
return case
def plot_testing():
import numpy as np
import matplotlib.pyplot as plt
from scipy.stats import randint, norm
from monaco.mc_var import InVar, OutVar
from monaco.mc_plot import (plot, plot_hist, plot_cdf,
plot_cov_corr, plot_integration_convergence,
plot_integration_error, plot_sensitivities)
from monaco.mc_enums import SampleMethod
generator = np.random.RandomState(74494861)
invarseeds = generator.randint(0, 2**31-1, size=10)
plt.close('all')
invars = dict()
nummap = {1: 'a', 2: 'b', 3: 'c', 4: 'd', 5: 'e'}
invars['randint'] = InVar(name='randint', ndraws=1000,
dist=randint, distkwargs={'low': 1, 'high': 6},
nummap=nummap,
samplemethod=SampleMethod.RANDOM, seed=invarseeds[0])
invars['norm'] = InVar(name='norm', ndraws=1000,
dist=norm, distkwargs={'loc': 10, 'scale': 4},
samplemethod=SampleMethod.RANDOM, seed=invarseeds[1])
invars['norm'].addVarStat(stat='orderstatTI',
statkwargs={'p': 0.75, 'c': 0.50, 'bound': '2-sided'})
invars['norm'].addVarStat(stat='orderstatP',
statkwargs={'p': 0.5, 'c': 0.9999, 'bound': 'all'})
invars['norm2'] = InVar(name='norm2', ndraws=1000,
dist=norm, distkwargs={'loc': 10, 'scale': 4},
samplemethod=SampleMethod.RANDOM, seed=invarseeds[2])
outvars = dict()
outvars['test'] = OutVar(name='test', vals=[1, 0, 2, 2], firstcaseismedian=True)
outvars['test'].sensitivity_ratios = {'a': 0.15, 'b': 0.35, 'c': 0.4}
f, (ax1, ax2) = plt.subplots(2, 1)
plot_hist(invars['randint'], ax=ax1, orientation='horizontal') # plot_hist
plot_cdf(invars['randint'], ax=ax2) # plot_cdf
plot(invars['norm'], title='norm', invar_space=InVarSpace.PCTS) # plot_hist
plot_hist(outvars['test'], orientation='horizontal', rug_plot=False) # plot_hist
plot_cdf(invars['norm'], orientation='horizontal') # plot_cdf
plot_cdf(outvars['test']) # plot_cdf
plot_sensitivities(outvars['test'], sensitivities='ratios', sort=True) # plot_sensitivities
plot(invars['randint'], invars['norm'],
cases=None, highlight_cases=range(10, 30),
rug_plot=True, cov_plot=True, cov_p=[0.90, 0.95, 0.99]) # plot_2d_scatter
plot(invars['randint'], invars['norm'], invars['norm2'],
cases=[], highlight_cases=range(10, 30)) # plot_3d_scatter
plot_2d_scatter(invars['randint'], invars['norm'], invars['norm2']) # plot_2d_scatter
v = np.array([-2, -1, 2, 3, 4, 5])
var0 = OutVar(name='testscalar', vals=[1, 2, 3, 4, 5], firstcaseismedian=True)
var1 = OutVar(name='testx', vals=[v, v, v, v, v], firstcaseismedian=True)
var2 = OutVar(name='testy', vals=[1*v, 2*v, 0*v, -1*v, -2*v], firstcaseismedian=True)
var3 = OutVar(name='testz', vals=[1*v, 2*v, 0*v, -1*v, -2*v], firstcaseismedian=True)
var2.addVarStat(stat='sigma', statkwargs={'sig': 3})
var2.addVarStat(stat='sigma', statkwargs={'sig': -3})
var2.addVarStat(stat='orderstatTI', statkwargs={'p': 0.6, 'c': 0.50, 'bound': '2-sided'})
var2.addVarStat(stat='mean')
plot(var2, highlight_cases=None) # plot_2d_line
plot(var1, var2, highlight_cases=[0, 1]) # plot_2d_line
plot(var0, var1) # plot_2p5d_line
plot(var1, var2, var3, highlight_cases=[0, ]) # plot_3d_line
plot(var0, var1, var2) # plot_2p5d_line
m = np.eye(3)
var4 = OutVar(name='testm', vals=[1*m, 2*m, 0*m, -1*m, -2*m])
plot(var4, highlight_cases=[]) # plot_3d_line
n = np.eye(2)
var5 = OutVar(name='testn', vals=[1*n, 2*n, 0*n, -1*n, -2*n])
plot(var0, var5) # plot_3d_line
plot_cov_corr(np.array([[2, 0.1111, np.nan], [-0.19, -1, np.nan], [np.nan, np.nan, np.nan]]),
['Test1', 'Test2', 'Test3'])
invars['randint2'] = InVar(name='randint2', ndraws=1000,
dist=randint, distkwargs={'low': 0, 'high': 2},
samplemethod=SampleMethod.RANDOM, seed=invarseeds[3])
plot_integration_convergence(invars['randint2'], volume=1, dimension=1,
refval=0.5, conf=0.95)
plot_integration_error(invars['randint2'], volume=1, dimension=1,
refval=0.5, conf=0.95)
plt.show()
def plot_testing():
import numpy as np
from scipy.stats import norm
import matplotlib.pyplot as plt
from monaco.mc_var import InVar
from monaco.mc_multi_plot import multi_plot, multi_plot_grid_rect
from monaco.mc_enums import SampleMethod
plt.close('all')
generator = np.random.RandomState(74494861)
invarseeds = generator.randint(0, 2**31 - 1, size=10)
invars = dict()
invars['norm1'] = InVar('norm1',
ndraws=1000,
dist=norm,
distkwargs={
'loc': 1,
'scale': 5
},
seed=invarseeds[0],
samplemethod=SampleMethod.RANDOM)
invars['norm2'] = InVar('norm2',
ndraws=1000,
dist=norm,
distkwargs={
'loc': 10,
'scale': 4
},
seed=invarseeds[1],
samplemethod=SampleMethod.RANDOM)
invars['norm3'] = InVar('norm3',
ndraws=1000,
dist=norm,
distkwargs={
'loc': 5,
'scale': 2
},
seed=invarseeds[3],
samplemethod=SampleMethod.RANDOM)
fig, axs = multi_plot([invars['norm1'], invars['norm2']],
highlight_cases=range(10, 30),
rug_plot=True,
cov_plot=True,
cov_p=0.95,
title='test') # multi_plot_2d_scatter_hist
fig, axs = multi_plot([invars['norm1'], invars['norm2'], invars['norm3']],
highlight_cases=range(10, 30),
rug_plot=True,
cov_plot=True,
cov_p=0.95,
title='test') # multi_plot_grid_tri
fig, axs = multi_plot_grid_rect([invars['norm1'], invars['norm2']],
invars['norm3'],
highlight_cases=range(10, 30),
rug_plot=True,
cov_plot=True,
cov_p=0.95,
title='test') # multi_plot_grid_rect
plt.show(block=True)