You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

1793 lines
53 KiB

"""
Main Random Variables Module
Defines abstract random variable type.
Contains interfaces for probability space object (PSpace) as well as standard
operators, P, E, sample, density, where, quantile
See Also
========
sympy.stats.crv
sympy.stats.frv
sympy.stats.rv_interface
"""
from __future__ import annotations
from functools import singledispatch
from math import prod
from sympy.core.add import Add
from sympy.core.basic import Basic
from sympy.core.containers import Tuple
from sympy.core.expr import Expr
from sympy.core.function import (Function, Lambda)
from sympy.core.logic import fuzzy_and
from sympy.core.mul import Mul
from sympy.core.relational import (Eq, Ne)
from sympy.core.singleton import S
from sympy.core.symbol import (Dummy, Symbol)
from sympy.core.sympify import sympify
from sympy.functions.special.delta_functions import DiracDelta
from sympy.functions.special.tensor_functions import KroneckerDelta
from sympy.logic.boolalg import (And, Or)
from sympy.matrices.expressions.matexpr import MatrixSymbol
from sympy.tensor.indexed import Indexed
from sympy.utilities.lambdify import lambdify
from sympy.core.relational import Relational
from sympy.core.sympify import _sympify
from sympy.sets.sets import FiniteSet, ProductSet, Intersection
from sympy.solvers.solveset import solveset
from sympy.external import import_module
from sympy.utilities.decorator import doctest_depends_on
from sympy.utilities.exceptions import sympy_deprecation_warning
from sympy.utilities.iterables import iterable
x = Symbol('x')
@singledispatch
def is_random(x):
return False
@is_random.register(Basic)
def _(x):
atoms = x.free_symbols
return any(is_random(i) for i in atoms)
class RandomDomain(Basic):
"""
Represents a set of variables and the values which they can take.
See Also
========
sympy.stats.crv.ContinuousDomain
sympy.stats.frv.FiniteDomain
"""
is_ProductDomain = False
is_Finite = False
is_Continuous = False
is_Discrete = False
def __new__(cls, symbols, *args):
symbols = FiniteSet(*symbols)
return Basic.__new__(cls, symbols, *args)
@property
def symbols(self):
return self.args[0]
@property
def set(self):
return self.args[1]
def __contains__(self, other):
raise NotImplementedError()
def compute_expectation(self, expr):
raise NotImplementedError()
class SingleDomain(RandomDomain):
"""
A single variable and its domain.
See Also
========
sympy.stats.crv.SingleContinuousDomain
sympy.stats.frv.SingleFiniteDomain
"""
def __new__(cls, symbol, set):
assert symbol.is_Symbol
return Basic.__new__(cls, symbol, set)
@property
def symbol(self):
return self.args[0]
@property
def symbols(self):
return FiniteSet(self.symbol)
def __contains__(self, other):
if len(other) != 1:
return False
sym, val = tuple(other)[0]
return self.symbol == sym and val in self.set
class MatrixDomain(RandomDomain):
"""
A Random Matrix variable and its domain.
"""
def __new__(cls, symbol, set):
symbol, set = _symbol_converter(symbol), _sympify(set)
return Basic.__new__(cls, symbol, set)
@property
def symbol(self):
return self.args[0]
@property
def symbols(self):
return FiniteSet(self.symbol)
class ConditionalDomain(RandomDomain):
"""
A RandomDomain with an attached condition.
See Also
========
sympy.stats.crv.ConditionalContinuousDomain
sympy.stats.frv.ConditionalFiniteDomain
"""
def __new__(cls, fulldomain, condition):
condition = condition.xreplace({rs: rs.symbol
for rs in random_symbols(condition)})
return Basic.__new__(cls, fulldomain, condition)
@property
def symbols(self):
return self.fulldomain.symbols
@property
def fulldomain(self):
return self.args[0]
@property
def condition(self):
return self.args[1]
@property
def set(self):
raise NotImplementedError("Set of Conditional Domain not Implemented")
def as_boolean(self):
return And(self.fulldomain.as_boolean(), self.condition)
class PSpace(Basic):
"""
A Probability Space.
Explanation
===========
Probability Spaces encode processes that equal different values
probabilistically. These underly Random Symbols which occur in SymPy
expressions and contain the mechanics to evaluate statistical statements.
See Also
========
sympy.stats.crv.ContinuousPSpace
sympy.stats.frv.FinitePSpace
"""
is_Finite = None # type: bool
is_Continuous = None # type: bool
is_Discrete = None # type: bool
is_real = None # type: bool
@property
def domain(self):
return self.args[0]
@property
def density(self):
return self.args[1]
@property
def values(self):
return frozenset(RandomSymbol(sym, self) for sym in self.symbols)
@property
def symbols(self):
return self.domain.symbols
def where(self, condition):
raise NotImplementedError()
def compute_density(self, expr):
raise NotImplementedError()
def sample(self, size=(), library='scipy', seed=None):
raise NotImplementedError()
def probability(self, condition):
raise NotImplementedError()
def compute_expectation(self, expr):
raise NotImplementedError()
class SinglePSpace(PSpace):
"""
Represents the probabilities of a set of random events that can be
attributed to a single variable/symbol.
"""
def __new__(cls, s, distribution):
s = _symbol_converter(s)
return Basic.__new__(cls, s, distribution)
@property
def value(self):
return RandomSymbol(self.symbol, self)
@property
def symbol(self):
return self.args[0]
@property
def distribution(self):
return self.args[1]
@property
def pdf(self):
return self.distribution.pdf(self.symbol)
class RandomSymbol(Expr):
"""
Random Symbols represent ProbabilitySpaces in SymPy Expressions.
In principle they can take on any value that their symbol can take on
within the associated PSpace with probability determined by the PSpace
Density.
Explanation
===========
Random Symbols contain pspace and symbol properties.
The pspace property points to the represented Probability Space
The symbol is a standard SymPy Symbol that is used in that probability space
for example in defining a density.
You can form normal SymPy expressions using RandomSymbols and operate on
those expressions with the Functions
E - Expectation of a random expression
P - Probability of a condition
density - Probability Density of an expression
given - A new random expression (with new random symbols) given a condition
An object of the RandomSymbol type should almost never be created by the
user. They tend to be created instead by the PSpace class's value method.
Traditionally a user does not even do this but instead calls one of the
convenience functions Normal, Exponential, Coin, Die, FiniteRV, etc....
"""
def __new__(cls, symbol, pspace=None):
from sympy.stats.joint_rv import JointRandomSymbol
if pspace is None:
# Allow single arg, representing pspace == PSpace()
pspace = PSpace()
symbol = _symbol_converter(symbol)
if not isinstance(pspace, PSpace):
raise TypeError("pspace variable should be of type PSpace")
if cls == JointRandomSymbol and isinstance(pspace, SinglePSpace):
cls = RandomSymbol
return Basic.__new__(cls, symbol, pspace)
is_finite = True
is_symbol = True
is_Atom = True
_diff_wrt = True
pspace = property(lambda self: self.args[1])
symbol = property(lambda self: self.args[0])
name = property(lambda self: self.symbol.name)
def _eval_is_positive(self):
return self.symbol.is_positive
def _eval_is_integer(self):
return self.symbol.is_integer
def _eval_is_real(self):
return self.symbol.is_real or self.pspace.is_real
@property
def is_commutative(self):
return self.symbol.is_commutative
@property
def free_symbols(self):
return {self}
class RandomIndexedSymbol(RandomSymbol):
def __new__(cls, idx_obj, pspace=None):
if pspace is None:
# Allow single arg, representing pspace == PSpace()
pspace = PSpace()
if not isinstance(idx_obj, (Indexed, Function)):
raise TypeError("An Function or Indexed object is expected not %s"%(idx_obj))
return Basic.__new__(cls, idx_obj, pspace)
symbol = property(lambda self: self.args[0])
name = property(lambda self: str(self.args[0]))
@property
def key(self):
if isinstance(self.symbol, Indexed):
return self.symbol.args[1]
elif isinstance(self.symbol, Function):
return self.symbol.args[0]
@property
def free_symbols(self):
if self.key.free_symbols:
free_syms = self.key.free_symbols
free_syms.add(self)
return free_syms
return {self}
@property
def pspace(self):
return self.args[1]
class RandomMatrixSymbol(RandomSymbol, MatrixSymbol): # type: ignore
def __new__(cls, symbol, n, m, pspace=None):
n, m = _sympify(n), _sympify(m)
symbol = _symbol_converter(symbol)
if pspace is None:
# Allow single arg, representing pspace == PSpace()
pspace = PSpace()
return Basic.__new__(cls, symbol, n, m, pspace)
symbol = property(lambda self: self.args[0])
pspace = property(lambda self: self.args[3])
class ProductPSpace(PSpace):
"""
Abstract class for representing probability spaces with multiple random
variables.
See Also
========
sympy.stats.rv.IndependentProductPSpace
sympy.stats.joint_rv.JointPSpace
"""
pass
class IndependentProductPSpace(ProductPSpace):
"""
A probability space resulting from the merger of two independent probability
spaces.
Often created using the function, pspace.
"""
def __new__(cls, *spaces):
rs_space_dict = {}
for space in spaces:
for value in space.values:
rs_space_dict[value] = space
symbols = FiniteSet(*[val.symbol for val in rs_space_dict.keys()])
# Overlapping symbols
from sympy.stats.joint_rv import MarginalDistribution
from sympy.stats.compound_rv import CompoundDistribution
if len(symbols) < sum(len(space.symbols) for space in spaces if not
isinstance(space.distribution, (
CompoundDistribution, MarginalDistribution))):
raise ValueError("Overlapping Random Variables")
if all(space.is_Finite for space in spaces):
from sympy.stats.frv import ProductFinitePSpace
cls = ProductFinitePSpace
obj = Basic.__new__(cls, *FiniteSet(*spaces))
return obj
@property
def pdf(self):
p = Mul(*[space.pdf for space in self.spaces])
return p.subs({rv: rv.symbol for rv in self.values})
@property
def rs_space_dict(self):
d = {}
for space in self.spaces:
for value in space.values:
d[value] = space
return d
@property
def symbols(self):
return FiniteSet(*[val.symbol for val in self.rs_space_dict.keys()])
@property
def spaces(self):
return FiniteSet(*self.args)
@property
def values(self):
return sumsets(space.values for space in self.spaces)
def compute_expectation(self, expr, rvs=None, evaluate=False, **kwargs):
rvs = rvs or self.values
rvs = frozenset(rvs)
for space in self.spaces:
expr = space.compute_expectation(expr, rvs & space.values, evaluate=False, **kwargs)
if evaluate and hasattr(expr, 'doit'):
return expr.doit(**kwargs)
return expr
@property
def domain(self):
return ProductDomain(*[space.domain for space in self.spaces])
@property
def density(self):
raise NotImplementedError("Density not available for ProductSpaces")
def sample(self, size=(), library='scipy', seed=None):
return {k: v for space in self.spaces
for k, v in space.sample(size=size, library=library, seed=seed).items()}
def probability(self, condition, **kwargs):
cond_inv = False
if isinstance(condition, Ne):
condition = Eq(condition.args[0], condition.args[1])
cond_inv = True
elif isinstance(condition, And): # they are independent
return Mul(*[self.probability(arg) for arg in condition.args])
elif isinstance(condition, Or): # they are independent
return Add(*[self.probability(arg) for arg in condition.args])
expr = condition.lhs - condition.rhs
rvs = random_symbols(expr)
dens = self.compute_density(expr)
if any(pspace(rv).is_Continuous for rv in rvs):
from sympy.stats.crv import SingleContinuousPSpace
from sympy.stats.crv_types import ContinuousDistributionHandmade
if expr in self.values:
# Marginalize all other random symbols out of the density
randomsymbols = tuple(set(self.values) - frozenset([expr]))
symbols = tuple(rs.symbol for rs in randomsymbols)
pdf = self.domain.integrate(self.pdf, symbols, **kwargs)
return Lambda(expr.symbol, pdf)
dens = ContinuousDistributionHandmade(dens)
z = Dummy('z', real=True)
space = SingleContinuousPSpace(z, dens)
result = space.probability(condition.__class__(space.value, 0))
else:
from sympy.stats.drv import SingleDiscretePSpace
from sympy.stats.drv_types import DiscreteDistributionHandmade
dens = DiscreteDistributionHandmade(dens)
z = Dummy('z', integer=True)
space = SingleDiscretePSpace(z, dens)
result = space.probability(condition.__class__(space.value, 0))
return result if not cond_inv else S.One - result
def compute_density(self, expr, **kwargs):
rvs = random_symbols(expr)
if any(pspace(rv).is_Continuous for rv in rvs):
z = Dummy('z', real=True)
expr = self.compute_expectation(DiracDelta(expr - z),
**kwargs)
else:
z = Dummy('z', integer=True)
expr = self.compute_expectation(KroneckerDelta(expr, z),
**kwargs)
return Lambda(z, expr)
def compute_cdf(self, expr, **kwargs):
raise ValueError("CDF not well defined on multivariate expressions")
def conditional_space(self, condition, normalize=True, **kwargs):
rvs = random_symbols(condition)
condition = condition.xreplace({rv: rv.symbol for rv in self.values})
pspaces = [pspace(rv) for rv in rvs]
if any(ps.is_Continuous for ps in pspaces):
from sympy.stats.crv import (ConditionalContinuousDomain,
ContinuousPSpace)
space = ContinuousPSpace
domain = ConditionalContinuousDomain(self.domain, condition)
elif any(ps.is_Discrete for ps in pspaces):
from sympy.stats.drv import (ConditionalDiscreteDomain,
DiscretePSpace)
space = DiscretePSpace
domain = ConditionalDiscreteDomain(self.domain, condition)
elif all(ps.is_Finite for ps in pspaces):
from sympy.stats.frv import FinitePSpace
return FinitePSpace.conditional_space(self, condition)
if normalize:
replacement = {rv: Dummy(str(rv)) for rv in self.symbols}
norm = domain.compute_expectation(self.pdf, **kwargs)
pdf = self.pdf / norm.xreplace(replacement)
# XXX: Converting symbols from set to tuple. The order matters to
# Lambda though so we shouldn't be starting with a set here...
density = Lambda(tuple(domain.symbols), pdf)
return space(domain, density)
class ProductDomain(RandomDomain):
"""
A domain resulting from the merger of two independent domains.
See Also
========
sympy.stats.crv.ProductContinuousDomain
sympy.stats.frv.ProductFiniteDomain
"""
is_ProductDomain = True
def __new__(cls, *domains):
# Flatten any product of products
domains2 = []
for domain in domains:
if not domain.is_ProductDomain:
domains2.append(domain)
else:
domains2.extend(domain.domains)
domains2 = FiniteSet(*domains2)
if all(domain.is_Finite for domain in domains2):
from sympy.stats.frv import ProductFiniteDomain
cls = ProductFiniteDomain
if all(domain.is_Continuous for domain in domains2):
from sympy.stats.crv import ProductContinuousDomain
cls = ProductContinuousDomain
if all(domain.is_Discrete for domain in domains2):
from sympy.stats.drv import ProductDiscreteDomain
cls = ProductDiscreteDomain
return Basic.__new__(cls, *domains2)
@property
def sym_domain_dict(self):
return {symbol: domain for domain in self.domains
for symbol in domain.symbols}
@property
def symbols(self):
return FiniteSet(*[sym for domain in self.domains
for sym in domain.symbols])
@property
def domains(self):
return self.args
@property
def set(self):
return ProductSet(*(domain.set for domain in self.domains))
def __contains__(self, other):
# Split event into each subdomain
for domain in self.domains:
# Collect the parts of this event which associate to this domain
elem = frozenset([item for item in other
if sympify(domain.symbols.contains(item[0]))
is S.true])
# Test this sub-event
if elem not in domain:
return False
# All subevents passed
return True
def as_boolean(self):
return And(*[domain.as_boolean() for domain in self.domains])
def random_symbols(expr):
"""
Returns all RandomSymbols within a SymPy Expression.
"""
atoms = getattr(expr, 'atoms', None)
if atoms is not None:
comp = lambda rv: rv.symbol.name
l = list(atoms(RandomSymbol))
return sorted(l, key=comp)
else:
return []
def pspace(expr):
"""
Returns the underlying Probability Space of a random expression.
For internal use.
Examples
========
>>> from sympy.stats import pspace, Normal
>>> X = Normal('X', 0, 1)
>>> pspace(2*X + 1) == X.pspace
True
"""
expr = sympify(expr)
if isinstance(expr, RandomSymbol) and expr.pspace is not None:
return expr.pspace
if expr.has(RandomMatrixSymbol):
rm = list(expr.atoms(RandomMatrixSymbol))[0]
return rm.pspace
rvs = random_symbols(expr)
if not rvs:
raise ValueError("Expression containing Random Variable expected, not %s" % (expr))
# If only one space present
if all(rv.pspace == rvs[0].pspace for rv in rvs):
return rvs[0].pspace
from sympy.stats.compound_rv import CompoundPSpace
from sympy.stats.stochastic_process import StochasticPSpace
for rv in rvs:
if isinstance(rv.pspace, (CompoundPSpace, StochasticPSpace)):
return rv.pspace
# Otherwise make a product space
return IndependentProductPSpace(*[rv.pspace for rv in rvs])
def sumsets(sets):
"""
Union of sets
"""
return frozenset().union(*sets)
def rs_swap(a, b):
"""
Build a dictionary to swap RandomSymbols based on their underlying symbol.
i.e.
if ``X = ('x', pspace1)``
and ``Y = ('x', pspace2)``
then ``X`` and ``Y`` match and the key, value pair
``{X:Y}`` will appear in the result
Inputs: collections a and b of random variables which share common symbols
Output: dict mapping RVs in a to RVs in b
"""
d = {}
for rsa in a:
d[rsa] = [rsb for rsb in b if rsa.symbol == rsb.symbol][0]
return d
def given(expr, condition=None, **kwargs):
r""" Conditional Random Expression.
Explanation
===========
From a random expression and a condition on that expression creates a new
probability space from the condition and returns the same expression on that
conditional probability space.
Examples
========
>>> from sympy.stats import given, density, Die
>>> X = Die('X', 6)
>>> Y = given(X, X > 3)
>>> density(Y).dict
{4: 1/3, 5: 1/3, 6: 1/3}
Following convention, if the condition is a random symbol then that symbol
is considered fixed.
>>> from sympy.stats import Normal
>>> from sympy import pprint
>>> from sympy.abc import z
>>> X = Normal('X', 0, 1)
>>> Y = Normal('Y', 0, 1)
>>> pprint(density(X + Y, Y)(z), use_unicode=False)
2
-(-Y + z)
-----------
___ 2
\/ 2 *e
------------------
____
2*\/ pi
"""
if not is_random(condition) or pspace_independent(expr, condition):
return expr
if isinstance(condition, RandomSymbol):
condition = Eq(condition, condition.symbol)
condsymbols = random_symbols(condition)
if (isinstance(condition, Eq) and len(condsymbols) == 1 and
not isinstance(pspace(expr).domain, ConditionalDomain)):
rv = tuple(condsymbols)[0]
results = solveset(condition, rv)
if isinstance(results, Intersection) and S.Reals in results.args:
results = list(results.args[1])
sums = 0
for res in results:
temp = expr.subs(rv, res)
if temp == True:
return True
if temp != False:
# XXX: This seems nonsensical but preserves existing behaviour
# after the change that Relational is no longer a subclass of
# Expr. Here expr is sometimes Relational and sometimes Expr
# but we are trying to add them with +=. This needs to be
# fixed somehow.
if sums == 0 and isinstance(expr, Relational):
sums = expr.subs(rv, res)
else:
sums += expr.subs(rv, res)
if sums == 0:
return False
return sums
# Get full probability space of both the expression and the condition
fullspace = pspace(Tuple(expr, condition))
# Build new space given the condition
space = fullspace.conditional_space(condition, **kwargs)
# Dictionary to swap out RandomSymbols in expr with new RandomSymbols
# That point to the new conditional space
swapdict = rs_swap(fullspace.values, space.values)
# Swap random variables in the expression
expr = expr.xreplace(swapdict)
return expr
def expectation(expr, condition=None, numsamples=None, evaluate=True, **kwargs):
"""
Returns the expected value of a random expression.
Parameters
==========
expr : Expr containing RandomSymbols
The expression of which you want to compute the expectation value
given : Expr containing RandomSymbols
A conditional expression. E(X, X>0) is expectation of X given X > 0
numsamples : int
Enables sampling and approximates the expectation with this many samples
evalf : Bool (defaults to True)
If sampling return a number rather than a complex expression
evaluate : Bool (defaults to True)
In case of continuous systems return unevaluated integral
Examples
========
>>> from sympy.stats import E, Die
>>> X = Die('X', 6)
>>> E(X)
7/2
>>> E(2*X + 1)
8
>>> E(X, X > 3) # Expectation of X given that it is above 3
5
"""
if not is_random(expr): # expr isn't random?
return expr
kwargs['numsamples'] = numsamples
from sympy.stats.symbolic_probability import Expectation
if evaluate:
return Expectation(expr, condition).doit(**kwargs)
return Expectation(expr, condition)
def probability(condition, given_condition=None, numsamples=None,
evaluate=True, **kwargs):
"""
Probability that a condition is true, optionally given a second condition.
Parameters
==========
condition : Combination of Relationals containing RandomSymbols
The condition of which you want to compute the probability
given_condition : Combination of Relationals containing RandomSymbols
A conditional expression. P(X > 1, X > 0) is expectation of X > 1
given X > 0
numsamples : int
Enables sampling and approximates the probability with this many samples
evaluate : Bool (defaults to True)
In case of continuous systems return unevaluated integral
Examples
========
>>> from sympy.stats import P, Die
>>> from sympy import Eq
>>> X, Y = Die('X', 6), Die('Y', 6)
>>> P(X > 3)
1/2
>>> P(Eq(X, 5), X > 2) # Probability that X == 5 given that X > 2
1/4
>>> P(X > Y)
5/12
"""
kwargs['numsamples'] = numsamples
from sympy.stats.symbolic_probability import Probability
if evaluate:
return Probability(condition, given_condition).doit(**kwargs)
return Probability(condition, given_condition)
class Density(Basic):
expr = property(lambda self: self.args[0])
def __new__(cls, expr, condition = None):
expr = _sympify(expr)
if condition is None:
obj = Basic.__new__(cls, expr)
else:
condition = _sympify(condition)
obj = Basic.__new__(cls, expr, condition)
return obj
@property
def condition(self):
if len(self.args) > 1:
return self.args[1]
else:
return None
def doit(self, evaluate=True, **kwargs):
from sympy.stats.random_matrix import RandomMatrixPSpace
from sympy.stats.joint_rv import JointPSpace
from sympy.stats.matrix_distributions import MatrixPSpace
from sympy.stats.compound_rv import CompoundPSpace
from sympy.stats.frv import SingleFiniteDistribution
expr, condition = self.expr, self.condition
if isinstance(expr, SingleFiniteDistribution):
return expr.dict
if condition is not None:
# Recompute on new conditional expr
expr = given(expr, condition, **kwargs)
if not random_symbols(expr):
return Lambda(x, DiracDelta(x - expr))
if isinstance(expr, RandomSymbol):
if isinstance(expr.pspace, (SinglePSpace, JointPSpace, MatrixPSpace)) and \
hasattr(expr.pspace, 'distribution'):
return expr.pspace.distribution
elif isinstance(expr.pspace, RandomMatrixPSpace):
return expr.pspace.model
if isinstance(pspace(expr), CompoundPSpace):
kwargs['compound_evaluate'] = evaluate
result = pspace(expr).compute_density(expr, **kwargs)
if evaluate and hasattr(result, 'doit'):
return result.doit()
else:
return result
def density(expr, condition=None, evaluate=True, numsamples=None, **kwargs):
"""
Probability density of a random expression, optionally given a second
condition.
Explanation
===========
This density will take on different forms for different types of
probability spaces. Discrete variables produce Dicts. Continuous
variables produce Lambdas.
Parameters
==========
expr : Expr containing RandomSymbols
The expression of which you want to compute the density value
condition : Relational containing RandomSymbols
A conditional expression. density(X > 1, X > 0) is density of X > 1
given X > 0
numsamples : int
Enables sampling and approximates the density with this many samples
Examples
========
>>> from sympy.stats import density, Die, Normal
>>> from sympy import Symbol
>>> x = Symbol('x')
>>> D = Die('D', 6)
>>> X = Normal(x, 0, 1)
>>> density(D).dict
{1: 1/6, 2: 1/6, 3: 1/6, 4: 1/6, 5: 1/6, 6: 1/6}
>>> density(2*D).dict
{2: 1/6, 4: 1/6, 6: 1/6, 8: 1/6, 10: 1/6, 12: 1/6}
>>> density(X)(x)
sqrt(2)*exp(-x**2/2)/(2*sqrt(pi))
"""
if numsamples:
return sampling_density(expr, condition, numsamples=numsamples,
**kwargs)
return Density(expr, condition).doit(evaluate=evaluate, **kwargs)
def cdf(expr, condition=None, evaluate=True, **kwargs):
"""
Cumulative Distribution Function of a random expression.
optionally given a second condition.
Explanation
===========
This density will take on different forms for different types of
probability spaces.
Discrete variables produce Dicts.
Continuous variables produce Lambdas.
Examples
========
>>> from sympy.stats import density, Die, Normal, cdf
>>> D = Die('D', 6)
>>> X = Normal('X', 0, 1)
>>> density(D).dict
{1: 1/6, 2: 1/6, 3: 1/6, 4: 1/6, 5: 1/6, 6: 1/6}
>>> cdf(D)
{1: 1/6, 2: 1/3, 3: 1/2, 4: 2/3, 5: 5/6, 6: 1}
>>> cdf(3*D, D > 2)
{9: 1/4, 12: 1/2, 15: 3/4, 18: 1}
>>> cdf(X)
Lambda(_z, erf(sqrt(2)*_z/2)/2 + 1/2)
"""
if condition is not None: # If there is a condition
# Recompute on new conditional expr
return cdf(given(expr, condition, **kwargs), **kwargs)
# Otherwise pass work off to the ProbabilitySpace
result = pspace(expr).compute_cdf(expr, **kwargs)
if evaluate and hasattr(result, 'doit'):
return result.doit()
else:
return result
def characteristic_function(expr, condition=None, evaluate=True, **kwargs):
"""
Characteristic function of a random expression, optionally given a second condition.
Returns a Lambda.
Examples
========
>>> from sympy.stats import Normal, DiscreteUniform, Poisson, characteristic_function
>>> X = Normal('X', 0, 1)
>>> characteristic_function(X)
Lambda(_t, exp(-_t**2/2))
>>> Y = DiscreteUniform('Y', [1, 2, 7])
>>> characteristic_function(Y)
Lambda(_t, exp(7*_t*I)/3 + exp(2*_t*I)/3 + exp(_t*I)/3)
>>> Z = Poisson('Z', 2)
>>> characteristic_function(Z)
Lambda(_t, exp(2*exp(_t*I) - 2))
"""
if condition is not None:
return characteristic_function(given(expr, condition, **kwargs), **kwargs)
result = pspace(expr).compute_characteristic_function(expr, **kwargs)
if evaluate and hasattr(result, 'doit'):
return result.doit()
else:
return result
def moment_generating_function(expr, condition=None, evaluate=True, **kwargs):
if condition is not None:
return moment_generating_function(given(expr, condition, **kwargs), **kwargs)
result = pspace(expr).compute_moment_generating_function(expr, **kwargs)
if evaluate and hasattr(result, 'doit'):
return result.doit()
else:
return result
def where(condition, given_condition=None, **kwargs):
"""
Returns the domain where a condition is True.
Examples
========
>>> from sympy.stats import where, Die, Normal
>>> from sympy import And
>>> D1, D2 = Die('a', 6), Die('b', 6)
>>> a, b = D1.symbol, D2.symbol
>>> X = Normal('x', 0, 1)
>>> where(X**2<1)
Domain: (-1 < x) & (x < 1)
>>> where(X**2<1).set
Interval.open(-1, 1)
>>> where(And(D1<=D2, D2<3))
Domain: (Eq(a, 1) & Eq(b, 1)) | (Eq(a, 1) & Eq(b, 2)) | (Eq(a, 2) & Eq(b, 2))
"""
if given_condition is not None: # If there is a condition
# Recompute on new conditional expr
return where(given(condition, given_condition, **kwargs), **kwargs)
# Otherwise pass work off to the ProbabilitySpace
return pspace(condition).where(condition, **kwargs)
@doctest_depends_on(modules=('scipy',))
def sample(expr, condition=None, size=(), library='scipy',
numsamples=1, seed=None, **kwargs):
"""
A realization of the random expression.
Parameters
==========
expr : Expression of random variables
Expression from which sample is extracted
condition : Expr containing RandomSymbols
A conditional expression
size : int, tuple
Represents size of each sample in numsamples
library : str
- 'scipy' : Sample using scipy
- 'numpy' : Sample using numpy
- 'pymc' : Sample using PyMC
Choose any of the available options to sample from as string,
by default is 'scipy'
numsamples : int
Number of samples, each with size as ``size``.
.. deprecated:: 1.9
The ``numsamples`` parameter is deprecated and is only provided for
compatibility with v1.8. Use a list comprehension or an additional
dimension in ``size`` instead. See
:ref:`deprecated-sympy-stats-numsamples` for details.
seed :
An object to be used as seed by the given external library for sampling `expr`.
Following is the list of possible types of object for the supported libraries,
- 'scipy': int, numpy.random.RandomState, numpy.random.Generator
- 'numpy': int, numpy.random.RandomState, numpy.random.Generator
- 'pymc': int
Optional, by default None, in which case seed settings
related to the given library will be used.
No modifications to environment's global seed settings
are done by this argument.
Returns
=======
sample: float/list/numpy.ndarray
one sample or a collection of samples of the random expression.
- sample(X) returns float/numpy.float64/numpy.int64 object.
- sample(X, size=int/tuple) returns numpy.ndarray object.
Examples
========
>>> from sympy.stats import Die, sample, Normal, Geometric
>>> X, Y, Z = Die('X', 6), Die('Y', 6), Die('Z', 6) # Finite Random Variable
>>> die_roll = sample(X + Y + Z)
>>> die_roll # doctest: +SKIP
3
>>> N = Normal('N', 3, 4) # Continuous Random Variable
>>> samp = sample(N)
>>> samp in N.pspace.domain.set
True
>>> samp = sample(N, N>0)
>>> samp > 0
True
>>> samp_list = sample(N, size=4)
>>> [sam in N.pspace.domain.set for sam in samp_list]
[True, True, True, True]
>>> sample(N, size = (2,3)) # doctest: +SKIP
array([[5.42519758, 6.40207856, 4.94991743],
[1.85819627, 6.83403519, 1.9412172 ]])
>>> G = Geometric('G', 0.5) # Discrete Random Variable
>>> samp_list = sample(G, size=3)
>>> samp_list # doctest: +SKIP
[1, 3, 2]
>>> [sam in G.pspace.domain.set for sam in samp_list]
[True, True, True]
>>> MN = Normal("MN", [3, 4], [[2, 1], [1, 2]]) # Joint Random Variable
>>> samp_list = sample(MN, size=4)
>>> samp_list # doctest: +SKIP
[array([2.85768055, 3.38954165]),
array([4.11163337, 4.3176591 ]),
array([0.79115232, 1.63232916]),
array([4.01747268, 3.96716083])]
>>> [tuple(sam) in MN.pspace.domain.set for sam in samp_list]
[True, True, True, True]
.. versionchanged:: 1.7.0
sample used to return an iterator containing the samples instead of value.
.. versionchanged:: 1.9.0
sample returns values or array of values instead of an iterator and numsamples is deprecated.
"""
iterator = sample_iter(expr, condition, size=size, library=library,
numsamples=numsamples, seed=seed)
if numsamples != 1:
sympy_deprecation_warning(
f"""
The numsamples parameter to sympy.stats.sample() is deprecated.
Either use a list comprehension, like
[sample(...) for i in range({numsamples})]
or add a dimension to size, like
sample(..., size={(numsamples,) + size})
""",
deprecated_since_version="1.9",
active_deprecations_target="deprecated-sympy-stats-numsamples",
)
return [next(iterator) for i in range(numsamples)]
return next(iterator)
def quantile(expr, evaluate=True, **kwargs):
r"""
Return the :math:`p^{th}` order quantile of a probability distribution.
Explanation
===========
Quantile is defined as the value at which the probability of the random
variable is less than or equal to the given probability.
.. math::
Q(p) = \inf\{x \in (-\infty, \infty) : p \le F(x)\}
Examples
========
>>> from sympy.stats import quantile, Die, Exponential
>>> from sympy import Symbol, pprint
>>> p = Symbol("p")
>>> l = Symbol("lambda", positive=True)
>>> X = Exponential("x", l)
>>> quantile(X)(p)
-log(1 - p)/lambda
>>> D = Die("d", 6)
>>> pprint(quantile(D)(p), use_unicode=False)
/nan for Or(p > 1, p < 0)
|
| 1 for p <= 1/6
|
| 2 for p <= 1/3
|
< 3 for p <= 1/2
|
| 4 for p <= 2/3
|
| 5 for p <= 5/6
|
\ 6 for p <= 1
"""
result = pspace(expr).compute_quantile(expr, **kwargs)
if evaluate and hasattr(result, 'doit'):
return result.doit()
else:
return result
def sample_iter(expr, condition=None, size=(), library='scipy',
numsamples=S.Infinity, seed=None, **kwargs):
"""
Returns an iterator of realizations from the expression given a condition.
Parameters
==========
expr: Expr
Random expression to be realized
condition: Expr, optional
A conditional expression
size : int, tuple
Represents size of each sample in numsamples
numsamples: integer, optional
Length of the iterator (defaults to infinity)
seed :
An object to be used as seed by the given external library for sampling `expr`.
Following is the list of possible types of object for the supported libraries,
- 'scipy': int, numpy.random.RandomState, numpy.random.Generator
- 'numpy': int, numpy.random.RandomState, numpy.random.Generator
- 'pymc': int
Optional, by default None, in which case seed settings
related to the given library will be used.
No modifications to environment's global seed settings
are done by this argument.
Examples
========
>>> from sympy.stats import Normal, sample_iter
>>> X = Normal('X', 0, 1)
>>> expr = X*X + 3
>>> iterator = sample_iter(expr, numsamples=3) # doctest: +SKIP
>>> list(iterator) # doctest: +SKIP
[12, 4, 7]
Returns
=======
sample_iter: iterator object
iterator object containing the sample/samples of given expr
See Also
========
sample
sampling_P
sampling_E
"""
from sympy.stats.joint_rv import JointRandomSymbol
if not import_module(library):
raise ValueError("Failed to import %s" % library)
if condition is not None:
ps = pspace(Tuple(expr, condition))
else:
ps = pspace(expr)
rvs = list(ps.values)
if isinstance(expr, JointRandomSymbol):
expr = expr.subs({expr: RandomSymbol(expr.symbol, expr.pspace)})
else:
sub = {}
for arg in expr.args:
if isinstance(arg, JointRandomSymbol):
sub[arg] = RandomSymbol(arg.symbol, arg.pspace)
expr = expr.subs(sub)
def fn_subs(*args):
return expr.subs({rv: arg for rv, arg in zip(rvs, args)})
def given_fn_subs(*args):
if condition is not None:
return condition.subs({rv: arg for rv, arg in zip(rvs, args)})
return False
if library in ('pymc', 'pymc3'):
# Currently unable to lambdify in pymc
# TODO : Remove when lambdify accepts 'pymc' as module
fn = lambdify(rvs, expr, **kwargs)
else:
fn = lambdify(rvs, expr, modules=library, **kwargs)
if condition is not None:
given_fn = lambdify(rvs, condition, **kwargs)
def return_generator_infinite():
count = 0
_size = (1,)+((size,) if isinstance(size, int) else size)
while count < numsamples:
d = ps.sample(size=_size, library=library, seed=seed) # a dictionary that maps RVs to values
args = [d[rv][0] for rv in rvs]
if condition is not None: # Check that these values satisfy the condition
# TODO: Replace the try-except block with only given_fn(*args)
# once lambdify works with unevaluated SymPy objects.
try:
gd = given_fn(*args)
except (NameError, TypeError):
gd = given_fn_subs(*args)
if gd != True and gd != False:
raise ValueError(
"Conditions must not contain free symbols")
if not gd: # If the values don't satisfy then try again
continue
yield fn(*args)
count += 1
def return_generator_finite():
faulty = True
while faulty:
d = ps.sample(size=(numsamples,) + ((size,) if isinstance(size, int) else size),
library=library, seed=seed) # a dictionary that maps RVs to values
faulty = False
count = 0
while count < numsamples and not faulty:
args = [d[rv][count] for rv in rvs]
if condition is not None: # Check that these values satisfy the condition
# TODO: Replace the try-except block with only given_fn(*args)
# once lambdify works with unevaluated SymPy objects.
try:
gd = given_fn(*args)
except (NameError, TypeError):
gd = given_fn_subs(*args)
if gd != True and gd != False:
raise ValueError(
"Conditions must not contain free symbols")
if not gd: # If the values don't satisfy then try again
faulty = True
count += 1
count = 0
while count < numsamples:
args = [d[rv][count] for rv in rvs]
# TODO: Replace the try-except block with only fn(*args)
# once lambdify works with unevaluated SymPy objects.
try:
yield fn(*args)
except (NameError, TypeError):
yield fn_subs(*args)
count += 1
if numsamples is S.Infinity:
return return_generator_infinite()
return return_generator_finite()
def sample_iter_lambdify(expr, condition=None, size=(),
numsamples=S.Infinity, seed=None, **kwargs):
return sample_iter(expr, condition=condition, size=size,
numsamples=numsamples, seed=seed, **kwargs)
def sample_iter_subs(expr, condition=None, size=(),
numsamples=S.Infinity, seed=None, **kwargs):
return sample_iter(expr, condition=condition, size=size,
numsamples=numsamples, seed=seed, **kwargs)
def sampling_P(condition, given_condition=None, library='scipy', numsamples=1,
evalf=True, seed=None, **kwargs):
"""
Sampling version of P.
See Also
========
P
sampling_E
sampling_density
"""
count_true = 0
count_false = 0
samples = sample_iter(condition, given_condition, library=library,
numsamples=numsamples, seed=seed, **kwargs)
for sample in samples:
if sample:
count_true += 1
else:
count_false += 1
result = S(count_true) / numsamples
if evalf:
return result.evalf()
else:
return result
def sampling_E(expr, given_condition=None, library='scipy', numsamples=1,
evalf=True, seed=None, **kwargs):
"""
Sampling version of E.
See Also
========
P
sampling_P
sampling_density
"""
samples = list(sample_iter(expr, given_condition, library=library,
numsamples=numsamples, seed=seed, **kwargs))
result = Add(*samples) / numsamples
if evalf:
return result.evalf()
else:
return result
def sampling_density(expr, given_condition=None, library='scipy',
numsamples=1, seed=None, **kwargs):
"""
Sampling version of density.
See Also
========
density
sampling_P
sampling_E
"""
results = {}
for result in sample_iter(expr, given_condition, library=library,
numsamples=numsamples, seed=seed, **kwargs):
results[result] = results.get(result, 0) + 1
return results
def dependent(a, b):
"""
Dependence of two random expressions.
Two expressions are independent if knowledge of one does not change
computations on the other.
Examples
========
>>> from sympy.stats import Normal, dependent, given
>>> from sympy import Tuple, Eq
>>> X, Y = Normal('X', 0, 1), Normal('Y', 0, 1)
>>> dependent(X, Y)
False
>>> dependent(2*X + Y, -Y)
True
>>> X, Y = given(Tuple(X, Y), Eq(X + Y, 3))
>>> dependent(X, Y)
True
See Also
========
independent
"""
if pspace_independent(a, b):
return False
z = Symbol('z', real=True)
# Dependent if density is unchanged when one is given information about
# the other
return (density(a, Eq(b, z)) != density(a) or
density(b, Eq(a, z)) != density(b))
def independent(a, b):
"""
Independence of two random expressions.
Two expressions are independent if knowledge of one does not change
computations on the other.
Examples
========
>>> from sympy.stats import Normal, independent, given
>>> from sympy import Tuple, Eq
>>> X, Y = Normal('X', 0, 1), Normal('Y', 0, 1)
>>> independent(X, Y)
True
>>> independent(2*X + Y, -Y)
False
>>> X, Y = given(Tuple(X, Y), Eq(X + Y, 3))
>>> independent(X, Y)
False
See Also
========
dependent
"""
return not dependent(a, b)
def pspace_independent(a, b):
"""
Tests for independence between a and b by checking if their PSpaces have
overlapping symbols. This is a sufficient but not necessary condition for
independence and is intended to be used internally.
Notes
=====
pspace_independent(a, b) implies independent(a, b)
independent(a, b) does not imply pspace_independent(a, b)
"""
a_symbols = set(pspace(b).symbols)
b_symbols = set(pspace(a).symbols)
if len(set(random_symbols(a)).intersection(random_symbols(b))) != 0:
return False
if len(a_symbols.intersection(b_symbols)) == 0:
return True
return None
def rv_subs(expr, symbols=None):
"""
Given a random expression replace all random variables with their symbols.
If symbols keyword is given restrict the swap to only the symbols listed.
"""
if symbols is None:
symbols = random_symbols(expr)
if not symbols:
return expr
swapdict = {rv: rv.symbol for rv in symbols}
return expr.subs(swapdict)
class NamedArgsMixin:
_argnames: tuple[str, ...] = ()
def __getattr__(self, attr):
try:
return self.args[self._argnames.index(attr)]
except ValueError:
raise AttributeError("'%s' object has no attribute '%s'" % (
type(self).__name__, attr))
class Distribution(Basic):
def sample(self, size=(), library='scipy', seed=None):
""" A random realization from the distribution """
module = import_module(library)
if library in {'scipy', 'numpy', 'pymc3', 'pymc'} and module is None:
raise ValueError("Failed to import %s" % library)
if library == 'scipy':
# scipy does not require map as it can handle using custom distributions.
# However, we will still use a map where we can.
# TODO: do this for drv.py and frv.py if necessary.
# TODO: add more distributions here if there are more
# See links below referring to sections beginning with "A common parametrization..."
# I will remove all these comments if everything is ok.
from sympy.stats.sampling.sample_scipy import do_sample_scipy
import numpy
if seed is None or isinstance(seed, int):
rand_state = numpy.random.default_rng(seed=seed)
else:
rand_state = seed
samps = do_sample_scipy(self, size, rand_state)
elif library == 'numpy':
from sympy.stats.sampling.sample_numpy import do_sample_numpy
import numpy
if seed is None or isinstance(seed, int):
rand_state = numpy.random.default_rng(seed=seed)
else:
rand_state = seed
_size = None if size == () else size
samps = do_sample_numpy(self, _size, rand_state)
elif library in ('pymc', 'pymc3'):
from sympy.stats.sampling.sample_pymc import do_sample_pymc
import logging
logging.getLogger("pymc").setLevel(logging.ERROR)
try:
import pymc
except ImportError:
import pymc3 as pymc
with pymc.Model():
if do_sample_pymc(self):
samps = pymc.sample(draws=prod(size), chains=1, compute_convergence_checks=False,
progressbar=False, random_seed=seed, return_inferencedata=False)[:]['X']
samps = samps.reshape(size)
else:
samps = None
else:
raise NotImplementedError("Sampling from %s is not supported yet."
% str(library))
if samps is not None:
return samps
raise NotImplementedError(
"Sampling for %s is not currently implemented from %s"
% (self, library))
def _value_check(condition, message):
"""
Raise a ValueError with message if condition is False, else
return True if all conditions were True, else False.
Examples
========
>>> from sympy.stats.rv import _value_check
>>> from sympy.abc import a, b, c
>>> from sympy import And, Dummy
>>> _value_check(2 < 3, '')
True
Here, the condition is not False, but it does not evaluate to True
so False is returned (but no error is raised). So checking if the
return value is True or False will tell you if all conditions were
evaluated.
>>> _value_check(a < b, '')
False
In this case the condition is False so an error is raised:
>>> r = Dummy(real=True)
>>> _value_check(r < r - 1, 'condition is not true')
Traceback (most recent call last):
...
ValueError: condition is not true
If no condition of many conditions must be False, they can be
checked by passing them as an iterable:
>>> _value_check((a < 0, b < 0, c < 0), '')
False
The iterable can be a generator, too:
>>> _value_check((i < 0 for i in (a, b, c)), '')
False
The following are equivalent to the above but do not pass
an iterable:
>>> all(_value_check(i < 0, '') for i in (a, b, c))
False
>>> _value_check(And(a < 0, b < 0, c < 0), '')
False
"""
if not iterable(condition):
condition = [condition]
truth = fuzzy_and(condition)
if truth == False:
raise ValueError(message)
return truth == True
def _symbol_converter(sym):
"""
Casts the parameter to Symbol if it is 'str'
otherwise no operation is performed on it.
Parameters
==========
sym
The parameter to be converted.
Returns
=======
Symbol
the parameter converted to Symbol.
Raises
======
TypeError
If the parameter is not an instance of both str and
Symbol.
Examples
========
>>> from sympy import Symbol
>>> from sympy.stats.rv import _symbol_converter
>>> s = _symbol_converter('s')
>>> isinstance(s, Symbol)
True
>>> _symbol_converter(1)
Traceback (most recent call last):
...
TypeError: 1 is neither a Symbol nor a string
>>> r = Symbol('r')
>>> isinstance(r, Symbol)
True
"""
if isinstance(sym, str):
sym = Symbol(sym)
if not isinstance(sym, Symbol):
raise TypeError("%s is neither a Symbol nor a string"%(sym))
return sym
def sample_stochastic_process(process):
"""
This function is used to sample from stochastic process.
Parameters
==========
process: StochasticProcess
Process used to extract the samples. It must be an instance of
StochasticProcess
Examples
========
>>> from sympy.stats import sample_stochastic_process, DiscreteMarkovChain
>>> from sympy import Matrix
>>> T = Matrix([[0.5, 0.2, 0.3],[0.2, 0.5, 0.3],[0.2, 0.3, 0.5]])
>>> Y = DiscreteMarkovChain("Y", [0, 1, 2], T)
>>> next(sample_stochastic_process(Y)) in Y.state_space
True
>>> next(sample_stochastic_process(Y)) # doctest: +SKIP
0
>>> next(sample_stochastic_process(Y)) # doctest: +SKIP
2
Returns
=======
sample: iterator object
iterator object containing the sample of given process
"""
from sympy.stats.stochastic_process_types import StochasticProcess
if not isinstance(process, StochasticProcess):
raise ValueError("Process must be an instance of Stochastic Process")
return process.sample()