TODO : - how to make UV attributes like std, skew, accessible ? using q.value.skew ? - multiplication precedence : x = mcerp.Normal(1, 2)*s doesn't work because mcerp can't cast quantity object to UV

mcerp

See : - github : https://github.com/tisimst/mcerp - online doc : https://pythonhosted.org/mcerp/

Introduction on mcerp

UV : uncertain Variable

liste of distributions : https://pythonhosted.org/mcerp/distribution_constructors.html

import mcerp
mcerp.npts = 12_000

x = mcerp.Normal(1, 3, tag="toto")
y = mcerp.Normal(2, 5)
print(x)
print(y)
print(x.mean)
print(x.var)
print(x.kurt)
print(x.skew)
print(mcerp.npts)
print(x.tag)

x.describe()
y.describe()
x.plot()
y.plot()
print(x.tag)

z = x + y

z.describe()
z.plot()

print(2*z)
print(z*2)
print(2**z)
print(z**2)
print(2/z)
print(z/2)
#print(2//z)
#print(z//2)

Support with physipy

import physipy 
from physipy import m, cd, s, Quantity, Dimension

Construction

x = Quantity(mcerp.Normal(1, 2), Dimension("L"))
x

# using multiplication
# x = mcerp.Normal(1, 2)*s : NotUpcast: <class 'physipy.quantity.quantity.Quantity'> cannot be converted to a number with uncertainty
x = m *mcerp.Normal(1, 2)
x

Basic Operation

# with scalar
scalar = 2*m
print(x+scalar)
print(x-scalar)
print(x*scalar)
print(x/scalar)
# print(x**s) : NotUpcast: <class 'physipy.quantity.quantity.Quantity'> cannot be converted to a number with uncertainty

print(scalar+x)
print(scalar-x)
print(scalar*x)
print(scalar/x)
# print(x**s) : NotUpcast: <class 'physipy.quantity.quantity.Quantity'> cannot be converted to a number with uncertainty

# with other uv
x = Quantity(mcerp.Normal(1, 2), Dimension("L"))
y = Quantity(mcerp.Normal(1, 2), Dimension("L"))

print(x+y)
print(x-y)
print(x*y)
print(x/y)
# print(x**y) : NotUpcast: <class 'physipy.quantity.quantity.Quantity'> cannot be converted to a number with uncertainty

print(y+x)
print(y-x)
print(y*x)
print(y/x)
# print(y**x) : NotUpcast: <class 'physipy.quantity.quantity.Quantity'> cannot be converted to a number with uncertainty

Functionnalities

Use q.value. to acces mcerp attributes

x.value.describe()
x.value.skew
print(x.value.stats)

Examples

from mcerp import N, U, Gamma, Beta, Exp, H

x1 = m * N(24, 1)
x2 = m * N(37, 4)
x3 = s * Exp(2)

# x1.mean
print("x1.mean")
print(x1.mean) # falls back on UV value, hence drops the unit
# x1.mean() : fails because relies on numpy.mean

# x1.var
print("x1.var")
print(x1.var) # falls back on UV value, hence drops the nit

# x1.skew
print("x1.skew")
print(x1.skew) # falls back on UV value, hence drops the nit

# x1.kurt
print("x1.kurt")
print(x1.kurt)

# x1.stats
print("x1.stats")
print(x1.stats)

# Z = (x1*x2**2)/(15*(1.5 + x3))
Z = (x1*x2**2)/(15*(1.5*s + x3))
print(Z)

# Z.describe()
Z.describe()

Distributions

mu = 1
sigma = 0.1
x = s * mcerp.Normal(mu, sigma)

# y = mcerp.Normal(mu*s, sigma*s) # fails because "assert sigma > 0, "

import scipy.stats as ss
x = s * mcerp.uv(ss.norm(loc=10, scale=1))
x

distrib = ss.norm(loc=10*s, scale=1*s)
distrib.mean() # unit is stripped, not mcerp's fault

y = mcerp.uv(distrib)
y

Plotting

# x1.plot() : fails because uses Quantity.plot, which doesn't handle mcerp

x1.value.plot()

# fails because of matploltib version on normed 
x1.value.plot(hist=True)

rvs1 = m*N(5, 10)
rvs2 = m*N(5, 10) + m*N(0, 0.2)
rvs3 = m*N(8, 10) + m*N(0, 0.2)

from scipy.stats import ttest_rel
tstat, pval = ttest_rel(rvs1._mcpts, rvs2._mcpts)
pval

rvs1<rvs2

tstat, pval = ttest_rel(rvs1._mcpts, rvs3._mcpts)
pval

float(tstat)

rvs1<rvs3

x = N(0, 1)
y = N(0, 10)
x<y

x>y

x==y

x1==x1

n1 = s*N(0, 1)
n2 = s*N(0, 1)
n1==n2

print(Z*Z)
print(Z**2)
Z*Z == Z**2

h = s*H(50, 5, 10)
h==4

h<=3*s

n = s*N(0, 1)
n==0*s

n==0.5*s

n==1.2345*s