def ideal(self, *gens, **kwds):
"""
Create an ideal in this polynomial ring.
"""
do_coerce = False
if len(gens) == 1:
from sage.rings.ideal import is_Ideal
if is_Ideal(gens[0]):
if gens[0].ring() is self:
return gens[0]
gens = gens[0].gens()
elif isinstance(gens[0], (list, tuple)):
gens = gens[0]
if not self._has_singular:
# pass through
MPolynomialRing_generic.ideal(self,gens,**kwds)
if is_SingularElement(gens):
gens = list(gens)
do_coerce = True
if is_Macaulay2Element(gens):
gens = list(gens)
do_coerce = True
elif not isinstance(gens, (list, tuple)):
gens = [gens]
if ('coerce' in kwds and kwds['coerce']) or do_coerce:
gens = [self(x) for x in gens] # this will even coerce from singular ideals correctly!
return multi_polynomial_ideal.MPolynomialIdeal(self, gens, **kwds)
def ideal(self, *gens, **kwds):
"""
Create an ideal in this polynomial ring.
"""
do_coerce = False
if len(gens) == 1:
from sage.rings.ideal import is_Ideal
if is_Ideal(gens[0]):
if gens[0].ring() is self:
return gens[0]
gens = gens[0].gens()
elif isinstance(gens[0], (list, tuple)):
gens = gens[0]
if not self._has_singular:
# pass through
MPolynomialRing_generic.ideal(self, gens, **kwds)
if is_SingularElement(gens):
gens = list(gens)
do_coerce = True
if is_Macaulay2Element(gens):
gens = list(gens)
do_coerce = True
elif not isinstance(gens, (list, tuple)):
gens = [gens]
if ('coerce' in kwds and kwds['coerce']) or do_coerce:
gens = [self(x) for x in gens
] # this will even coerce from singular ideals correctly!
return multi_polynomial_ideal.MPolynomialIdeal(self, gens, **kwds)
def __call__(self, x, check=True):
"""
Convert ``x`` to an element of this multivariate polynomial ring,
possibly non-canonically.
EXAMPLES:
We create a Macaulay2 multivariate polynomial via ideal
arithmetic, then convert it into R.
::
sage: R.<x,y> = PolynomialRing(QQ, 2) # optional
sage: I = R.ideal([x^3 + y, y]) # optional
sage: S = I._macaulay2_() # optional
sage: T = S*S*S # optional
sage: U = T.gens().entries().flatten() # optional
sage: f = U[2]; f # optional
x^6*y+2*x^3*y^2+y^3
sage: R(repr(f)) # optional
x^6*y + 2*x^3*y^2 + y^3
Some other subtle conversions. We create polynomial rings in 2
variables over the rationals, integers, and a finite field.
::
sage: R.<x,y> = QQ[]
sage: S.<x,y> = ZZ[]
sage: T.<x,y> = GF(7)[]
We convert from integer polynomials to rational polynomials,
and back::
sage: f = R(S.0^2 - 4*S.1^3); f
-4*y^3 + x^2
sage: parent(f)
Multivariate Polynomial Ring in x, y over Rational Field
sage: parent(S(f))
Multivariate Polynomial Ring in x, y over Integer Ring
We convert from polynomials over the finite field.
::
sage: f = R(T.0^2 - 4*T.1^3); f
3*y^3 + x^2
sage: parent(f)
Multivariate Polynomial Ring in x, y over Rational Field
We dump and load the polynomial ring S::
sage: S2 = loads(dumps(S))
sage: S2 == S
True
Coerce works and gets the right parent.
::
sage: parent(S2._coerce_(S.0)) is S2
True
Conversion to reduce modulo a prime between rings with different
variable names::
sage: R.<x,y> = PolynomialRing(QQ,2)
sage: S.<a,b> = PolynomialRing(GF(7),2)
sage: f = x^2 + 2/3*y^3
sage: S(f)
3*b^3 + a^2
Conversion from symbolic variables::
sage: x,y,z = var('x,y,z')
sage: R = QQ[x,y,z]
sage: type(x)
<type 'sage.symbolic.expression.Expression'>
sage: type(R(x))
<type 'sage.rings.polynomial.multi_polynomial_libsingular.MPolynomial_libsingular'>
sage: f = R(x^3 + y^3 - z^3); f
x^3 + y^3 - z^3
sage: type(f)
<type 'sage.rings.polynomial.multi_polynomial_libsingular.MPolynomial_libsingular'>
sage: parent(f)
Multivariate Polynomial Ring in x, y, z over Rational Field
A more complicated symbolic and computational mix. Behind the
scenes Singular and Maxima are doing the real work.
::
sage: R = QQ[x,y,z]
sage: f = (x^3 + y^3 - z^3)^10; f
(x^3 + y^3 - z^3)^10
sage: g = R(f); parent(g)
Multivariate Polynomial Ring in x, y, z over Rational Field
sage: (f - g).expand()
0
It intelligently handles conversions from polynomial rings in a subset
of the variables too.
::
sage: R = GF(5)['x,y,z']
sage: S = ZZ['y']
sage: R(7*S.0)
2*y
sage: T = ZZ['x,z']
sage: R(2*T.0 + 6*T.1 + T.0*T.1^2)
x*z^2 + 2*x + z
::
sage: R = QQ['t,x,y,z']
sage: S.<x> = ZZ['x']
sage: T.<z> = S['z']
sage: T
Univariate Polynomial Ring in z over Univariate Polynomial Ring in x over Integer Ring
sage: f = (x+3*z+5)^2; f
9*z^2 + (6*x + 30)*z + x^2 + 10*x + 25
sage: R(f)
x^2 + 6*x*z + 9*z^2 + 10*x + 30*z + 25
Arithmetic with a constant from a base ring::
sage: R.<u,v> = QQ[]
sage: S.<x,y> = R[]
sage: u^3*x^2 + v*y
u^3*x^2 + v*y
Stacked polynomial rings convert into constants if possible. First,
the univariate case::
sage: R.<x> = QQ[]
sage: S.<u,v> = R[]
sage: S(u + 2)
u + 2
sage: S(u + 2).degree()
1
sage: S(x + 3)
x + 3
sage: S(x + 3).degree()
0
Second, the multivariate case::
sage: R.<x,y> = QQ[]
sage: S.<u,v> = R[]
sage: S(x + 2*y)
x + 2*y
sage: S(u + 2*v)
u + 2*v
Conversion from strings::
sage: R.<x,y> = QQ[]
sage: R('x+(1/2)*y^2')
1/2*y^2 + x
sage: S.<u,v> = ZZ[]
sage: S('u^2 + u*v + v^2')
u^2 + u*v + v^2
Foreign polynomial rings convert into the highest ring; the point
here is that an element of T could convert to an element of R or an
element of S; it is anticipated that an element of T is more likely
to be "the right thing" and is historically consistent.
::
sage: R.<x,y> = QQ[]
sage: S.<u,v> = R[]
sage: T.<a,b> = QQ[]
sage: S(a + b)
u + v
TESTS:
Check if we still allow nonsense :trac:`7951`::
sage: P = PolynomialRing(QQ, 0, '')
sage: P('pi')
Traceback (most recent call last):
...
TypeError: Unable to coerce pi (<class 'sage.symbolic.constants.Pi'>) to Rational
Check that it is possible to convert strings to iterated polynomial
rings :trac:`13327`::
sage: Rm = QQ["a"]["b, c"]
sage: Rm("a*b")
a*b
sage: parent(_) is Rm
True
"""
from sage.rings.polynomial.multi_polynomial_element import MPolynomial_polydict
import sage.rings.polynomial.polynomial_element as polynomial_element
# handle constants that coerce into self.base_ring() first, if possible
if isinstance(x, Element) and x.parent() is self.base_ring():
# A Constant multi-polynomial
return self({self._zero_tuple:x})
try:
y = self.base_ring()._coerce_(x)
return MPolynomial_polydict(self, {self._zero_tuple:y})
except TypeError:
pass
from multi_polynomial_libsingular import MPolynomial_libsingular
if isinstance(x, MPolynomial_polydict):
P = x.parent()
if P is self:
return x
elif P == self:
return MPolynomial_polydict(self, x.element().dict())
elif self.base_ring().has_coerce_map_from(P):
# it might be in the base ring (i.e. a poly ring over a poly ring)
c = self.base_ring()(x)
return MPolynomial_polydict(self, {self._zero_tuple:c})
elif len(P.variable_names()) == len(self.variable_names()):
# Map the variables in some crazy way (but in order,
# of course). This is here since R(blah) is supposed
# to be "make an element of R if at all possible with
# no guarantees that this is mathematically solid."
K = self.base_ring()
D = x.element().dict()
for i, a in D.iteritems():
D[i] = K(a)
return MPolynomial_polydict(self, D)
elif set(P.variable_names()).issubset(set(self.variable_names())) and self.base_ring().has_coerce_map_from(P.base_ring()):
# If the named variables are a superset of the input, map the variables by name
return MPolynomial_polydict(self, self._extract_polydict(x))
else:
return MPolynomial_polydict(self, x._mpoly_dict_recursive(self.variable_names(), self.base_ring()))
elif isinstance(x, MPolynomial_libsingular):
P = x.parent()
if P == self:
return MPolynomial_polydict(self, x.dict())
elif self.base_ring().has_coerce_map_from(P):
# it might be in the base ring (i.e. a poly ring over a poly ring)
c = self.base_ring()(x)
return MPolynomial_polydict(self, {self._zero_tuple:c})
elif len(P.variable_names()) == len(self.variable_names()):
# Map the variables in some crazy way (but in order,
# of course). This is here since R(blah) is supposed
# to be "make an element of R if at all possible with
# no guarantees that this is mathematically solid."
K = self.base_ring()
D = x.dict()
for i, a in D.iteritems():
D[i] = K(a)
return MPolynomial_polydict(self, D)
elif set(P.variable_names()).issubset(set(self.variable_names())) and self.base_ring().has_coerce_map_from(P.base_ring()):
# If the named variables are a superset of the input, map the variables by name
return MPolynomial_polydict(self, self._extract_polydict(x))
else:
return MPolynomial_polydict(self, x._mpoly_dict_recursive(self.variable_names(), self.base_ring()))
elif isinstance(x, polynomial_element.Polynomial):
return MPolynomial_polydict(self, x._mpoly_dict_recursive(self.variable_names(), self.base_ring()))
elif isinstance(x, PolyDict):
return MPolynomial_polydict(self, x)
elif isinstance(x, fraction_field_element.FractionFieldElement) and x.parent().ring() == self:
if x.denominator() == 1:
return x.numerator()
else:
raise TypeError("unable to coerce since the denominator is not 1")
elif is_SingularElement(x) and self._has_singular:
self._singular_().set_ring()
try:
return x.sage_poly(self)
except TypeError:
raise TypeError("unable to coerce singular object")
elif hasattr(x, '_polynomial_'):
return x._polynomial_(self)
elif isinstance(x, str):
try:
from sage.misc.sage_eval import sage_eval
return self(sage_eval(x, self.gens_dict_recursive()))
except NameError as e:
raise TypeError("unable to convert string")
elif is_Macaulay2Element(x):
try:
s = x.sage_polystring()
if len(s) == 0:
raise TypeError
# NOTE: It's CRUCIAL to use the eval command as follows,
# i.e., with the gen dict as the third arg and the second
# empty. Otherwise pickling won't work after calls to this eval!!!
# This took a while to figure out!
return self(eval(s, {}, self.gens_dict()))
except (AttributeError, TypeError, NameError, SyntaxError):
raise TypeError("Unable to coerce macaulay2 object")
return MPolynomial_polydict(self, x)
if isinstance(x, dict):
return MPolynomial_polydict(self, x)
else:
c = self.base_ring()(x)
return MPolynomial_polydict(self, {self._zero_tuple:c})
def __call__(self, x, check=True):
"""
Convert ``x`` to an element of this multivariate polynomial ring,
possibly non-canonically.
EXAMPLES:
We create a Macaulay2 multivariate polynomial via ideal
arithmetic, then convert it into R.
::
sage: R.<x,y> = PolynomialRing(QQ, 2)
sage: I = R.ideal([x^3 + y, y])
sage: S = I._macaulay2_() # optional - macaulay2
sage: T = S*S*S # optional - macaulay2
sage: U = T.gens().entries().flatten() # optional - macaulay2
sage: f = U[2]; f # optional - macaulay2
x^6*y+2*x^3*y^2+y^3
sage: R(repr(f)) # optional - macaulay2
x^6*y + 2*x^3*y^2 + y^3
Some other subtle conversions. We create polynomial rings in 2
variables over the rationals, integers, and a finite field.
::
sage: R.<x,y> = QQ[]
sage: S.<x,y> = ZZ[]
sage: T.<x,y> = GF(7)[]
We convert from integer polynomials to rational polynomials,
and back::
sage: f = R(S.0^2 - 4*S.1^3); f
-4*y^3 + x^2
sage: parent(f)
Multivariate Polynomial Ring in x, y over Rational Field
sage: parent(S(f))
Multivariate Polynomial Ring in x, y over Integer Ring
We convert from polynomials over the finite field.
::
sage: f = R(T.0^2 - 4*T.1^3); f
3*y^3 + x^2
sage: parent(f)
Multivariate Polynomial Ring in x, y over Rational Field
We dump and load the polynomial ring S::
sage: S2 = loads(dumps(S))
sage: S2 == S
True
Coerce works and gets the right parent.
::
sage: parent(S2._coerce_(S.0)) is S2
True
Conversion to reduce modulo a prime between rings with different
variable names::
sage: R.<x,y> = PolynomialRing(QQ,2)
sage: S.<a,b> = PolynomialRing(GF(7),2)
sage: f = x^2 + 2/3*y^3
sage: S(f)
3*b^3 + a^2
Conversion from symbolic variables::
sage: x,y,z = var('x,y,z')
sage: R = QQ['x,y,z']
sage: type(x)
<type 'sage.symbolic.expression.Expression'>
sage: type(R(x))
<type 'sage.rings.polynomial.multi_polynomial_libsingular.MPolynomial_libsingular'>
sage: f = R(x^3 + y^3 - z^3); f
x^3 + y^3 - z^3
sage: type(f)
<type 'sage.rings.polynomial.multi_polynomial_libsingular.MPolynomial_libsingular'>
sage: parent(f)
Multivariate Polynomial Ring in x, y, z over Rational Field
A more complicated symbolic and computational mix. Behind the
scenes Singular and Maxima are doing the real work.
::
sage: R = QQ['x,y,z']
sage: f = (x^3 + y^3 - z^3)^10; f
(x^3 + y^3 - z^3)^10
sage: g = R(f); parent(g)
Multivariate Polynomial Ring in x, y, z over Rational Field
sage: (f - g).expand()
0
It intelligently handles conversions from polynomial rings in a subset
of the variables too.
::
sage: R = GF(5)['x,y,z']
sage: S = ZZ['y']
sage: R(7*S.0)
2*y
sage: T = ZZ['x,z']
sage: R(2*T.0 + 6*T.1 + T.0*T.1^2)
x*z^2 + 2*x + z
::
sage: R = QQ['t,x,y,z']
sage: S.<x> = ZZ['x']
sage: T.<z> = S['z']
sage: T
Univariate Polynomial Ring in z over Univariate Polynomial Ring in x over Integer Ring
sage: f = (x+3*z+5)^2; f
9*z^2 + (6*x + 30)*z + x^2 + 10*x + 25
sage: R(f)
x^2 + 6*x*z + 9*z^2 + 10*x + 30*z + 25
Arithmetic with a constant from a base ring::
sage: R.<u,v> = QQ[]
sage: S.<x,y> = R[]
sage: u^3*x^2 + v*y
u^3*x^2 + v*y
Stacked polynomial rings convert into constants if possible. First,
the univariate case::
sage: R.<x> = QQ[]
sage: S.<u,v> = R[]
sage: S(u + 2)
u + 2
sage: S(u + 2).degree()
1
sage: S(x + 3)
x + 3
sage: S(x + 3).degree()
0
Second, the multivariate case::
sage: R.<x,y> = QQ[]
sage: S.<u,v> = R[]
sage: S(x + 2*y)
x + 2*y
sage: S(u + 2*v)
u + 2*v
Conversion from strings::
sage: R.<x,y> = QQ[]
sage: R('x+(1/2)*y^2')
1/2*y^2 + x
sage: S.<u,v> = ZZ[]
sage: S('u^2 + u*v + v^2')
u^2 + u*v + v^2
Foreign polynomial rings convert into the highest ring; the point
here is that an element of T could convert to an element of R or an
element of S; it is anticipated that an element of T is more likely
to be "the right thing" and is historically consistent.
::
sage: R.<x,y> = QQ[]
sage: S.<u,v> = R[]
sage: T.<a,b> = QQ[]
sage: S(a + b)
u + v
TESTS:
Check if we still allow nonsense (see :trac:`7951`)::
sage: P = PolynomialRing(QQ, 0, '')
sage: P('pi')
Traceback (most recent call last):
...
TypeError: Unable to coerce pi (<class 'sage.symbolic.constants.Pi'>) to Rational
Check that it is possible to convert strings to iterated
polynomial rings (see :trac:`13327`)::
sage: Rm = QQ["a"]["b, c"]
sage: Rm("a*b")
a*b
sage: parent(_) is Rm
True
Check that conversion from PARI works correctly (see
:trac:`17974`)::
sage: A.<a> = PolynomialRing(QQ)
sage: B.<d,e> = PolynomialRing(A)
sage: f = pari(a*d)
sage: B(f)
a*d
sage: A.<a,b> = PolynomialRing(QQ)
sage: B.<d,e> = PolynomialRing(A)
sage: f = pari(a*d)
sage: B(f)
a*d
It is possible to convert `f` into `B` by using ``f.sage()``,
but this requires specifying a ``locals`` argument::
sage: f
d*a
sage: f.sage()
Traceback (most recent call last):
...
NameError: name 'd' is not defined
sage: f.sage(locals={'a': a, 'd': d})
a*d
"""
from sage.rings.polynomial.multi_polynomial_element import MPolynomial_polydict
import sage.rings.polynomial.polynomial_element as polynomial_element
# handle constants that coerce into self.base_ring() first, if possible
if isinstance(x, Element) and x.parent() is self.base_ring():
# A Constant multi-polynomial
return self({self._zero_tuple: x})
try:
y = self.base_ring()._coerce_(x)
return MPolynomial_polydict(self, {self._zero_tuple: y})
except TypeError:
pass
from multi_polynomial_libsingular import MPolynomial_libsingular
if isinstance(x, MPolynomial_polydict):
P = x.parent()
if P is self:
return x
elif P == self:
return MPolynomial_polydict(self, x.element().dict())
elif self.base_ring().has_coerce_map_from(P):
# it might be in the base ring (i.e. a poly ring over a poly ring)
c = self.base_ring()(x)
return MPolynomial_polydict(self, {self._zero_tuple: c})
elif len(P.variable_names()) == len(self.variable_names()):
# Map the variables in some crazy way (but in order,
# of course). This is here since R(blah) is supposed
# to be "make an element of R if at all possible with
# no guarantees that this is mathematically solid."
K = self.base_ring()
D = x.element().dict()
for i, a in D.iteritems():
D[i] = K(a)
return MPolynomial_polydict(self, D)
elif set(P.variable_names()).issubset(set(self.variable_names(
))) and self.base_ring().has_coerce_map_from(P.base_ring()):
# If the named variables are a superset of the input, map the variables by name
return MPolynomial_polydict(self, self._extract_polydict(x))
else:
return MPolynomial_polydict(
self,
x._mpoly_dict_recursive(self.variable_names(),
self.base_ring()))
elif isinstance(x, MPolynomial_libsingular):
P = x.parent()
if P == self:
return MPolynomial_polydict(self, x.dict())
elif self.base_ring().has_coerce_map_from(P):
# it might be in the base ring (i.e. a poly ring over a poly ring)
c = self.base_ring()(x)
return MPolynomial_polydict(self, {self._zero_tuple: c})
elif len(P.variable_names()) == len(self.variable_names()):
# Map the variables in some crazy way (but in order,
# of course). This is here since R(blah) is supposed
# to be "make an element of R if at all possible with
# no guarantees that this is mathematically solid."
K = self.base_ring()
D = x.dict()
for i, a in D.iteritems():
D[i] = K(a)
return MPolynomial_polydict(self, D)
elif set(P.variable_names()).issubset(set(self.variable_names(
))) and self.base_ring().has_coerce_map_from(P.base_ring()):
# If the named variables are a superset of the input, map the variables by name
return MPolynomial_polydict(self, self._extract_polydict(x))
else:
return MPolynomial_polydict(
self,
x._mpoly_dict_recursive(self.variable_names(),
self.base_ring()))
elif isinstance(x, polynomial_element.Polynomial):
return MPolynomial_polydict(
self,
x._mpoly_dict_recursive(self.variable_names(),
self.base_ring()))
elif isinstance(x, PolyDict):
return MPolynomial_polydict(self, x)
elif isinstance(x, fraction_field_element.FractionFieldElement
) and x.parent().ring() == self:
if x.denominator() == 1:
return x.numerator()
else:
raise TypeError(
"unable to coerce since the denominator is not 1")
elif is_SingularElement(x) and self._has_singular:
self._singular_().set_ring()
try:
return x.sage_poly(self)
except TypeError:
raise TypeError("unable to coerce singular object")
elif hasattr(x, '_polynomial_'):
return x._polynomial_(self)
elif isinstance(x, str):
from sage.misc.sage_eval import sage_eval
try:
x = sage_eval(x, self.gens_dict_recursive())
except NameError:
raise TypeError("unable to evaluate {!r} in {}".format(
x, self))
return self(x)
elif is_Macaulay2Element(x):
try:
s = x.sage_polystring()
if len(s) == 0:
raise TypeError
# NOTE: It's CRUCIAL to use the eval command as follows,
# i.e., with the gen dict as the third arg and the second
# empty. Otherwise pickling won't work after calls to this eval!!!
# This took a while to figure out!
return self(eval(s, {}, self.gens_dict()))
except (AttributeError, TypeError, NameError, SyntaxError):
raise TypeError("Unable to coerce macaulay2 object")
return MPolynomial_polydict(self, x)
elif isinstance(x, pari_gen) and x.type() == 't_POL':
# This recursive approach is needed because PARI
# represents multivariate polynomials as iterated
# univariate polynomials. Below, v is the variable
# with highest priority, and the x[i] are expressions
# in the remaining variables.
v = self.gens_dict_recursive()[str(x.variable())]
return sum(self(x[i]) * v**i for i in xrange(x.poldegree() + 1))
if isinstance(x, dict):
return MPolynomial_polydict(self, x)
else:
c = self.base_ring()(x)
return MPolynomial_polydict(self, {self._zero_tuple: c})
class MPolynomialRing_polydict( MPolynomialRing_macaulay2_repr, PolynomialRing_singular_repr, MPolynomialRing_generic):
"""
Multivariable polynomial ring.
EXAMPLES::
sage: R = PolynomialRing(Integers(12), 'x', 5); R
Multivariate Polynomial Ring in x0, x1, x2, x3, x4 over Ring of integers modulo 12
sage: loads(R.dumps()) == R
True
"""
def __init__(self, base_ring, n, names, order):
from sage.rings.polynomial.polynomial_singular_interface import can_convert_to_singular
order = TermOrder(order,n)
MPolynomialRing_generic.__init__(self, base_ring, n, names, order)
# Construct the generators
v = [0 for _ in xrange(n)]
one = base_ring(1);
self._gens = []
C = self._poly_class()
for i in xrange(n):
v[i] = 1 # int's!
self._gens.append(C(self, {tuple(v):one}))
v[i] = 0
self._gens = tuple(self._gens)
self._zero_tuple = tuple(v)
self._has_singular = can_convert_to_singular(self)
# This polynomial ring should belong to Algebras(base_ring).
# Algebras(...).parent_class, which was called from MPolynomialRing_generic.__init__,
# tries to provide a conversion from the base ring, if it does not exist.
# This is for algebras that only do the generic stuff in their initialisation.
# But here, we want to use PolynomialBaseringInjection. Hence, we need to
# wipe the memory and construct the conversion from scratch.
if n:
from sage.rings.polynomial.polynomial_element import PolynomialBaseringInjection
base_inject = PolynomialBaseringInjection(base_ring, self)
self.register_coercion(base_inject)
def _monomial_order_function(self):
return self.__monomial_order_function
def _poly_class(self):
from sage.rings.polynomial.multi_polynomial_element import MPolynomial_polydict
return MPolynomial_polydict
def __cmp__(left, right):
if not is_MPolynomialRing(right):
return cmp(type(left),type(right))
else:
return cmp((left.base_ring(), left.ngens(), left.variable_names(), left.term_order()),
(right.base_ring(), right.ngens(), right.variable_names(), right.term_order()))
def __call__(self, x, check=True):
"""
Convert ``x`` to an element of this multivariate polynomial ring,
possibly non-canonically.
EXAMPLES:
We create a Macaulay2 multivariate polynomial via ideal
arithmetic, then convert it into R.
::
sage: R.<x,y> = PolynomialRing(QQ, 2) # optional
sage: I = R.ideal([x^3 + y, y]) # optional
sage: S = I._macaulay2_() # optional
sage: T = S*S*S # optional
sage: U = T.gens().entries().flatten() # optional
sage: f = U[2]; f # optional
x^6*y+2*x^3*y^2+y^3
sage: R(repr(f)) # optional
x^6*y + 2*x^3*y^2 + y^3
Some other subtle conversions. We create polynomial rings in 2
variables over the rationals, integers, and a finite field.
::
sage: R.<x,y> = QQ[]
sage: S.<x,y> = ZZ[]
sage: T.<x,y> = GF(7)[]
We convert from integer polynomials to rational polynomials,
and back::
sage: f = R(S.0^2 - 4*S.1^3); f
-4*y^3 + x^2
sage: parent(f)
Multivariate Polynomial Ring in x, y over Rational Field
sage: parent(S(f))
Multivariate Polynomial Ring in x, y over Integer Ring
We convert from polynomials over the finite field.
::
sage: f = R(T.0^2 - 4*T.1^3); f
3*y^3 + x^2
sage: parent(f)
Multivariate Polynomial Ring in x, y over Rational Field
We dump and load the polynomial ring S::
sage: S2 = loads(dumps(S))
sage: S2 == S
True
Coerce works and gets the right parent.
::
sage: parent(S2._coerce_(S.0)) is S2
True
Conversion to reduce modulo a prime between rings with different
variable names::
sage: R.<x,y> = PolynomialRing(QQ,2)
sage: S.<a,b> = PolynomialRing(GF(7),2)
sage: f = x^2 + 2/3*y^3
sage: S(f)
3*b^3 + a^2
Conversion from symbolic variables::
sage: x,y,z = var('x,y,z')
sage: R = QQ[x,y,z]
sage: type(x)
<type 'sage.symbolic.expression.Expression'>
sage: type(R(x))
<type 'sage.rings.polynomial.multi_polynomial_libsingular.MPolynomial_libsingular'>
sage: f = R(x^3 + y^3 - z^3); f
x^3 + y^3 - z^3
sage: type(f)
<type 'sage.rings.polynomial.multi_polynomial_libsingular.MPolynomial_libsingular'>
sage: parent(f)
Multivariate Polynomial Ring in x, y, z over Rational Field
A more complicated symbolic and computational mix. Behind the
scenes Singular and Maxima are doing the real work.
::
sage: R = QQ[x,y,z]
sage: f = (x^3 + y^3 - z^3)^10; f
(x^3 + y^3 - z^3)^10
sage: g = R(f); parent(g)
Multivariate Polynomial Ring in x, y, z over Rational Field
sage: (f - g).expand()
0
It intelligently handles conversions from polynomial rings in a subset
of the variables too.
::
sage: R = GF(5)['x,y,z']
sage: S = ZZ['y']
sage: R(7*S.0)
2*y
sage: T = ZZ['x,z']
sage: R(2*T.0 + 6*T.1 + T.0*T.1^2)
x*z^2 + 2*x + z
::
sage: R = QQ['t,x,y,z']
sage: S.<x> = ZZ['x']
sage: T.<z> = S['z']
sage: T
Univariate Polynomial Ring in z over Univariate Polynomial Ring in x over Integer Ring
sage: f = (x+3*z+5)^2; f
9*z^2 + (6*x + 30)*z + x^2 + 10*x + 25
sage: R(f)
x^2 + 6*x*z + 9*z^2 + 10*x + 30*z + 25
Arithmetic with a constant from a base ring::
sage: R.<u,v> = QQ[]
sage: S.<x,y> = R[]
sage: u^3*x^2 + v*y
u^3*x^2 + v*y
Stacked polynomial rings convert into constants if possible. First,
the univariate case::
sage: R.<x> = QQ[]
sage: S.<u,v> = R[]
sage: S(u + 2)
u + 2
sage: S(u + 2).degree()
1
sage: S(x + 3)
x + 3
sage: S(x + 3).degree()
0
Second, the multivariate case::
sage: R.<x,y> = QQ[]
sage: S.<u,v> = R[]
sage: S(x + 2*y)
x + 2*y
sage: S(u + 2*v)
u + 2*v
Conversion from strings::
sage: R.<x,y> = QQ[]
sage: R('x+(1/2)*y^2')
1/2*y^2 + x
sage: S.<u,v> = ZZ[]
sage: S('u^2 + u*v + v^2')
u^2 + u*v + v^2
Foreign polynomial rings convert into the highest ring; the point
here is that an element of T could convert to an element of R or an
element of S; it is anticipated that an element of T is more likely
to be "the right thing" and is historically consistent.
::
sage: R.<x,y> = QQ[]
sage: S.<u,v> = R[]
sage: T.<a,b> = QQ[]
sage: S(a + b)
u + v
TESTS:
Check if we still allow nonsense :trac:`7951`::
sage: P = PolynomialRing(QQ, 0, '')
sage: P('pi')
Traceback (most recent call last):
...
TypeError: Unable to coerce pi (<class 'sage.symbolic.constants.Pi'>) to Rational
Check that it is possible to convert strings to iterated polynomial
rings :trac:`13327`::
sage: Rm = QQ["a"]["b, c"]
sage: Rm("a*b")
a*b
sage: parent(_) is Rm
True
"""
from sage.rings.polynomial.multi_polynomial_element import MPolynomial_polydict
import sage.rings.polynomial.polynomial_element as polynomial_element
# handle constants that coerce into self.base_ring() first, if possible
if isinstance(x, Element) and x.parent() is self.base_ring():
# A Constant multi-polynomial
return self({self._zero_tuple:x})
try:
y = self.base_ring()._coerce_(x)
return MPolynomial_polydict(self, {self._zero_tuple:y})
except TypeError:
pass
from multi_polynomial_libsingular import MPolynomial_libsingular
if isinstance(x, MPolynomial_polydict):
P = x.parent()
if P is self:
return x
elif P == self:
return MPolynomial_polydict(self, x.element().dict())
elif self.base_ring().has_coerce_map_from(P):
# it might be in the base ring (i.e. a poly ring over a poly ring)
c = self.base_ring()(x)
return MPolynomial_polydict(self, {self._zero_tuple:c})
elif len(P.variable_names()) == len(self.variable_names()):
# Map the variables in some crazy way (but in order,
# of course). This is here since R(blah) is supposed
# to be "make an element of R if at all possible with
# no guarantees that this is mathematically solid."
K = self.base_ring()
D = x.element().dict()
for i, a in D.iteritems():
D[i] = K(a)
return MPolynomial_polydict(self, D)
elif set(P.variable_names()).issubset(set(self.variable_names())) and self.base_ring().has_coerce_map_from(P.base_ring()):
# If the named variables are a superset of the input, map the variables by name
return MPolynomial_polydict(self, self._extract_polydict(x))
else:
return MPolynomial_polydict(self, x._mpoly_dict_recursive(self.variable_names(), self.base_ring()))
elif isinstance(x, MPolynomial_libsingular):
P = x.parent()
if P == self:
return MPolynomial_polydict(self, x.dict())
elif self.base_ring().has_coerce_map_from(P):
# it might be in the base ring (i.e. a poly ring over a poly ring)
c = self.base_ring()(x)
return MPolynomial_polydict(self, {self._zero_tuple:c})
elif len(P.variable_names()) == len(self.variable_names()):
# Map the variables in some crazy way (but in order,
# of course). This is here since R(blah) is supposed
# to be "make an element of R if at all possible with
# no guarantees that this is mathematically solid."
K = self.base_ring()
D = x.dict()
for i, a in D.iteritems():
D[i] = K(a)
return MPolynomial_polydict(self, D)
elif set(P.variable_names()).issubset(set(self.variable_names())) and self.base_ring().has_coerce_map_from(P.base_ring()):
# If the named variables are a superset of the input, map the variables by name
return MPolynomial_polydict(self, self._extract_polydict(x))
else:
return MPolynomial_polydict(self, x._mpoly_dict_recursive(self.variable_names(), self.base_ring()))
elif isinstance(x, polynomial_element.Polynomial):
return MPolynomial_polydict(self, x._mpoly_dict_recursive(self.variable_names(), self.base_ring()))
elif isinstance(x, PolyDict):
return MPolynomial_polydict(self, x)
elif isinstance(x, fraction_field_element.FractionFieldElement) and x.parent().ring() == self:
if x.denominator() == 1:
return x.numerator()
else:
raise TypeError, "unable to coerce since the denominator is not 1"
elif is_SingularElement(x) and self._has_singular:
self._singular_().set_ring()
try:
return x.sage_poly(self)
except TypeError:
raise TypeError, "unable to coerce singular object"
elif hasattr(x, '_polynomial_'):
return x._polynomial_(self)
elif isinstance(x, str):
try:
from sage.misc.sage_eval import sage_eval
return self(sage_eval(x, self.gens_dict_recursive()))
except NameError, e:
raise TypeError, "unable to convert string"
elif is_Macaulay2Element(x):
try:
s = x.sage_polystring()
if len(s) == 0:
raise TypeError
# NOTE: It's CRUCIAL to use the eval command as follows,
# i.e., with the gen dict as the third arg and the second
# empty. Otherwise pickling won't work after calls to this eval!!!
# This took a while to figure out!
return self(eval(s, {}, self.gens_dict()))
except (AttributeError, TypeError, NameError, SyntaxError):
raise TypeError, "Unable to coerce macaulay2 object"
return MPolynomial_polydict(self, x)