# Field of Algebraic Numbers#

AUTHOR:

• Carl Witty (2007-01-27): initial version

• Carl Witty (2007-10-29): massive rewrite to support complex as well as real numbers

This is an implementation of the algebraic numbers (the complex numbers which are the zero of a polynomial in $$\ZZ[x]$$; in other words, the algebraic closure of $$\QQ$$, with an embedding into $$\CC$$). All computations are exact. We also include an implementation of the algebraic reals (the intersection of the algebraic numbers with $$\RR$$). The field of algebraic numbers $$\QQbar$$ is available with abbreviation QQbar; the field of algebraic reals has abbreviation AA.

As with many other implementations of the algebraic numbers, we try hard to avoid computing a number field and working in the number field; instead, we use floating-point interval arithmetic whenever possible (basically whenever we need to prove non-equalities), and resort to symbolic computation only as needed (basically to prove equalities).

Algebraic numbers exist in one of the following forms:

• a rational number

• the sum, difference, product, or quotient of algebraic numbers

• the negation, inverse, absolute value, norm, real part, imaginary part, or complex conjugate of an algebraic number

• a particular root of a polynomial, given as a polynomial with algebraic coefficients together with an isolating interval (given as a RealIntervalFieldElement) which encloses exactly one root, and the multiplicity of the root

• a polynomial in one generator, where the generator is an algebraic number given as the root of an irreducible polynomial with integral coefficients and the polynomial is given as a NumberFieldElement.

An algebraic number can be coerced into ComplexIntervalField (or RealIntervalField, for algebraic reals); every algebraic number has a cached interval of the highest precision yet calculated.

In most cases, computations that need to compare two algebraic numbers compute them with 128-bit precision intervals; if this does not suffice to prove that the numbers are different, then we fall back on exact computation.

Note that division involves an implicit comparison of the divisor against zero, and may thus trigger exact computation.

Also, using an algebraic number in the leading coefficient of a polynomial also involves an implicit comparison against zero, which again may trigger exact computation.

Note that we work fairly hard to avoid computing new number fields; to help, we keep a lattice of already-computed number fields and their inclusions.

EXAMPLES:

sage: sqrt(AA(2)) > 0
True
sage: (sqrt(5 + 2*sqrt(QQbar(6))) - sqrt(QQbar(3)))^2 == 2
True
sage: AA((sqrt(5 + 2*sqrt(6)) - sqrt(3))^2) == 2
True


For a monic cubic polynomial $$x^3 + bx^2 + cx + d$$ with roots $$s1$$, $$s2$$, $$s3$$, the discriminant is defined as $$(s1-s2)^2(s1-s3)^2(s2-s3)^2$$ and can be computed as $$b^2c^2 - 4b^3d - 4c^3 + 18bcd - 27d^2$$. We can test that these definitions do give the same result:

sage: def disc1(b, c, d):
....:     return b^2*c^2 - 4*b^3*d - 4*c^3 + 18*b*c*d - 27*d^2
sage: def disc2(s1, s2, s3):
....:     return ((s1-s2)*(s1-s3)*(s2-s3))^2
sage: x = polygen(AA)
sage: p = x*(x-2)*(x-4)
sage: cp = AA.common_polynomial(p)
sage: d, c, b, _ = p.list()
sage: s1 = AA.polynomial_root(cp, RIF(-1, 1))
sage: s2 = AA.polynomial_root(cp, RIF(1, 3))
sage: s3 = AA.polynomial_root(cp, RIF(3, 5))
sage: disc1(b, c, d) == disc2(s1, s2, s3)
True
sage: p = p + 1
sage: cp = AA.common_polynomial(p)
sage: d, c, b, _ = p.list()
sage: s1 = AA.polynomial_root(cp, RIF(-1, 1))
sage: s2 = AA.polynomial_root(cp, RIF(1, 3))
sage: s3 = AA.polynomial_root(cp, RIF(3, 5))
sage: disc1(b, c, d) == disc2(s1, s2, s3)
True
sage: p = (x-sqrt(AA(2)))*(x-AA(2).nth_root(3))*(x-sqrt(AA(3)))
sage: cp = AA.common_polynomial(p)
sage: d, c, b, _ = p.list()
sage: s1 = AA.polynomial_root(cp, RIF(1.4, 1.5))
sage: s2 = AA.polynomial_root(cp, RIF(1.7, 1.8))
sage: s3 = AA.polynomial_root(cp, RIF(1.2, 1.3))
sage: disc1(b, c, d) == disc2(s1, s2, s3)
True


We can convert from symbolic expressions:

sage: QQbar(sqrt(-5))
2.236067977499790?*I
sage: AA(sqrt(2) + sqrt(3))
3.146264369941973?
sage: QQbar(I)
I
sage: QQbar(I * golden_ratio)
1.618033988749895?*I
sage: AA(golden_ratio)^2 - AA(golden_ratio)
1
sage: QQbar((-8)^(1/3))
1.000000000000000? + 1.732050807568878?*I
sage: AA((-8)^(1/3))
-2
sage: QQbar((-4)^(1/4))
1 + 1*I
sage: AA((-4)^(1/4))
Traceback (most recent call last):
...
ValueError: Cannot coerce algebraic number with non-zero imaginary part to algebraic real


The coercion, however, goes in the other direction, since not all symbolic expressions are algebraic numbers:

sage: QQbar(sqrt(2)) + sqrt(3)
sqrt(3) + 1.414213562373095?
sage: QQbar(sqrt(2) + QQbar(sqrt(3)))
3.146264369941973?


Note the different behavior in taking roots: for AA we prefer real roots if they exist, but for QQbar we take the principal root:

sage: AA(-1)^(1/3)
-1
sage: QQbar(-1)^(1/3)
0.500000000000000? + 0.866025403784439?*I


However, implicit coercion from $$\QQ[I]$$ is only allowed when it is equipped with a complex embedding:

sage: i.parent()
Number Field in I with defining polynomial x^2 + 1 with I = 1*I
sage: QQbar(1) + i
I + 1

sage: QQbar(1) + im
Traceback (most recent call last):
...
TypeError: unsupported operand parent(s) for +: 'Algebraic Field' and
'Number Field in im with defining polynomial x^2 + 1'


However, we can explicitly coerce from the abstract number field $$\QQ[I]$$. (Technically, this is not quite kosher, since we do not know whether the field generator is supposed to map to $$+I$$ or $$-I$$. We assume that for any quadratic field with polynomial $$x^2+1$$, the generator maps to $$+I$$.):

sage: pythag = QQbar(3/5 + 4*im/5); pythag
4/5*I + 3/5
sage: pythag.abs() == 1
True


We can implicitly coerce from algebraic reals to algebraic numbers:

sage: a = QQbar(1); a, a.parent()
(1, Algebraic Field)
sage: b = AA(1); b, b.parent()
(1, Algebraic Real Field)
sage: c = a + b; c, c.parent()
(2, Algebraic Field)


sage: phi = (1 + sqrt(AA(5))) / 2
sage: phi^2 == phi + 1
True
sage: tau = (1 - sqrt(AA(5))) / 2
sage: tau^2 == tau + 1
True
sage: phi + tau == 1
True
sage: tau < 0
True

sage: rt23 = sqrt(AA(2/3))
sage: rt35 = sqrt(AA(3/5))
sage: rt25 = sqrt(AA(2/5))
sage: rt23 * rt35 == rt25
True


The Sage rings AA and QQbar can decide equalities between radical expressions (over the reals and complex numbers respectively):

sage: a = AA((2/(3*sqrt(3)) + 10/27)^(1/3) - 2/(9*(2/(3*sqrt(3)) + 10/27)^(1/3)) + 1/3)
sage: a
1.000000000000000?
sage: a == 1
True


Algebraic numbers which are known to be rational print as rationals; otherwise they print as intervals (with 53-bit precision):

sage: AA(2)/3
2/3
sage: QQbar(5/7)
5/7
sage: QQbar(1/3 - 1/4*I)
-1/4*I + 1/3
sage: two = QQbar(4).nth_root(4)^2; two
2.000000000000000?
sage: two == 2; two
True
2
sage: phi
1.618033988749895?


We can find the real and imaginary parts of an algebraic number (exactly):

sage: r = QQbar.polynomial_root(x^5 - x - 1, CIF(RIF(0.1, 0.2), RIF(1.0, 1.1))); r
0.1812324444698754? + 1.083954101317711?*I
sage: r.real()
0.1812324444698754?
sage: r.imag()
1.083954101317711?
sage: r.minpoly()
x^5 - x - 1
sage: r.real().minpoly()
x^10 + 3/16*x^6 + 11/32*x^5 - 1/64*x^2 + 1/128*x - 1/1024
sage: r.imag().minpoly()  # long time (10s on sage.math, 2013)
x^20 - 5/8*x^16 - 95/256*x^12 - 625/1024*x^10 - 5/512*x^8 - 1875/8192*x^6 + 25/4096*x^4 - 625/32768*x^2 + 2869/1048576


We can find the absolute value and norm of an algebraic number exactly. (Note that we define the norm as the product of a number and its complex conjugate; this is the algebraic definition of norm, if we view QQbar as AA[I].):

sage: R.<x> = QQ[]
sage: r = (x^3 + 8).roots(QQbar, multiplicities=False); r
1.000000000000000? + 1.732050807568878?*I
sage: r.abs() == 2
True
sage: r.norm() == 4
True
sage: (r+QQbar(I)).norm().minpoly()
x^2 - 10*x + 13
sage: r = AA.polynomial_root(x^2 - x - 1, RIF(-1, 0)); r
-0.618033988749895?
sage: r.abs().minpoly()
x^2 + x - 1


We can compute the multiplicative order of an algebraic number:

sage: QQbar(-1/2 + I*sqrt(3)/2).multiplicative_order()
3
sage: QQbar(-sqrt(3)/2 + I/2).multiplicative_order()
12
sage: (QQbar.zeta(23)**5).multiplicative_order()
23


The paper “ARPREC: An Arbitrary Precision Computation Package” by Bailey, Yozo, Li and Thompson discusses this result. Evidently it is difficult to find, but we can easily verify it.

sage: alpha = QQbar.polynomial_root(x^10 + x^9 - x^7 - x^6 - x^5 - x^4 - x^3 + x + 1, RIF(1, 1.2))
sage: lhs = alpha^630 - 1
sage: rhs_num = (alpha^315 - 1) * (alpha^210 - 1) * (alpha^126 - 1)^2 * (alpha^90 - 1) * (alpha^3 - 1)^3 * (alpha^2 - 1)^5 * (alpha - 1)^3
sage: rhs_den = (alpha^35 - 1) * (alpha^15 - 1)^2 * (alpha^14 - 1)^2 * (alpha^5 - 1)^6 * alpha^68
sage: rhs = rhs_num / rhs_den
sage: lhs
2.642040335819351?e44
sage: rhs
2.642040335819351?e44
sage: lhs - rhs
0.?e29
sage: lhs == rhs
True
sage: lhs - rhs
0
sage: lhs._exact_value()
-10648699402510886229334132989629606002223831*a^9 + 23174560249100286133718183712802529035435800*a^8 - 27259790692625442252605558473646959458901265*a^7 + 21416469499004652376912957054411004410158065*a^6 - 14543082864016871805545108986578337637140321*a^5 + 6458050008796664339372667222902512216589785*a^4 + 3052219053800078449122081871454923124998263*a^3 - 14238966128623353681821644902045640915516176*a^2 + 16749022728952328254673732618939204392161001*a - 9052854758155114957837247156588012516273410 where a^10 - a^9 + a^7 - a^6 + a^5 - a^4 + a^3 - a + 1 = 0 and a in -1.176280818259918?


Given an algebraic number, we can produce a string that will reproduce that algebraic number if you type the string into Sage. We can see that until exact computation is triggered, an algebraic number keeps track of the computation steps used to produce that number:

sage: rt2 = AA(sqrt(2))
sage: rt3 = AA(sqrt(3))
sage: n = (rt2 + rt3)^5; n
308.3018001722975?
sage: sage_input(n)
R.<x> = AA[]
v1 = AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951))) + AA.polynomial_root(AA.common_polynomial(x^2 - 3), RIF(RR(1.7320508075688772), RR(1.7320508075688774)))
v2 = v1*v1
v2*v2*v1


But once exact computation is triggered, the computation tree is discarded, and we get a way to produce the number directly:

sage: n == 109*rt2 + 89*rt3
True
sage: sage_input(n)
R.<y> = QQ[]
v = AA.polynomial_root(AA.common_polynomial(y^4 - 4*y^2 + 1), RIF(-RR(1.9318516525781366), -RR(1.9318516525781364)))
-109*v^3 + 89*v^2 + 327*v - 178


We can also see that some computations (basically, those which are easy to perform exactly) are performed directly, instead of storing the computation tree:

sage: z3_3 = QQbar.zeta(3) * 3
sage: z4_4 = QQbar.zeta(4) * 4
sage: z5_5 = QQbar.zeta(5) * 5
sage: sage_input(z3_3 * z4_4 * z5_5)
R.<y> = QQ[]
3*QQbar.polynomial_root(AA.common_polynomial(y^2 + y + 1), CIF(RIF(-RR(0.50000000000000011), -RR(0.49999999999999994)), RIF(RR(0.8660254037844386), RR(0.86602540378443871))))*QQbar(4*I)*(5*QQbar.polynomial_root(AA.common_polynomial(y^4 + y^3 + y^2 + y + 1), CIF(RIF(RR(0.3090169943749474), RR(0.30901699437494745)), RIF(RR(0.95105651629515353), RR(0.95105651629515364)))))


Note that the verify=True argument to sage_input will always trigger exact computation, so running sage_input twice in a row on the same number will actually give different answers. In the following, running sage_input on n will also trigger exact computation on rt2, as you can see by the fact that the third output is different than the first:

sage: rt2 = AA(sqrt(2))
sage: n = rt2^2
sage: sage_input(n, verify=True)
# Verified
R.<x> = AA[]
v = AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951)))
v*v
sage: sage_input(n, verify=True)
# Verified
AA(2)
sage: n = rt2^2
sage: sage_input(n, verify=True)
# Verified
AA(2)


Just for fun, let’s try sage_input on a very complicated expression. The output of this example changed with the rewriting of polynomial multiplication algorithms in github issue #10255:

sage: rt2 = sqrt(AA(2))
sage: rt3 = sqrt(QQbar(3))
sage: x = polygen(QQbar)
sage: nrt3 = AA.polynomial_root((x-rt2)*(x+rt3), RIF(-2, -1))
sage: one = AA.polynomial_root((x-rt2)*(x-rt3)*(x-nrt3)*(x-1-rt3-nrt3), RIF(0.9, 1.1))
sage: one
1.000000000000000?
sage: sage_input(one, verify=True)
# Verified
R1.<x> = QQbar[]
R2.<y> = QQ[]
v = AA.polynomial_root(AA.common_polynomial(y^4 - 4*y^2 + 1), RIF(-RR(1.9318516525781366), -RR(1.9318516525781364)))
AA.polynomial_root(AA.common_polynomial(x^4 + QQbar(v^3 - 3*v - 1)*x^3 + QQbar(-v^3 + 3*v - 3)*x^2 + QQbar(-3*v^3 + 9*v + 3)*x + QQbar(3*v^3 - 9*v)), RIF(RR(0.99999999999999989), RR(1.0000000000000002)))
sage: one
1


We can pickle and unpickle algebraic fields (and they are globally unique):

sage: loads(dumps(AlgebraicField())) is AlgebraicField()
True
True


We can pickle and unpickle algebraic numbers:

sage: loads(dumps(QQbar(10))) == QQbar(10)
True
True
True

sage: t = QQbar(sqrt(2)); type(t._descr)
<class 'sage.rings.qqbar.ANRoot'>
True

sage: t.exactify(); type(t._descr)
<class 'sage.rings.qqbar.ANExtensionElement'>
True

sage: t = ~QQbar(sqrt(2)); type(t._descr)
<class 'sage.rings.qqbar.ANUnaryExpr'>
True

sage: t = QQbar(sqrt(2)) + QQbar(sqrt(3)); type(t._descr)
<class 'sage.rings.qqbar.ANBinaryExpr'>
sage: loads(dumps(t)) == QQbar(sqrt(2)) + QQbar(sqrt(3))
True


We can convert elements of QQbar and AA into the following types: float, complex, RDF, CDF, RR, CC, RIF, CIF, ZZ, and QQ, with a few exceptions. (For the arbitrary-precision types, RR, CC, RIF, and CIF, it can convert into a field of arbitrary precision.)

Converting from QQbar to a real type (float, RDF, RR, RIF, ZZ, or QQ) succeeds only if the QQbar is actually real (has an imaginary component of exactly zero). Converting from either AA or QQbar to ZZ or QQ succeeds only if the number actually is an integer or rational. If conversion fails, a ValueError will be raised.

Here are examples of all of these conversions:

sage: all_vals = [AA(42), AA(22/7), AA(golden_ratio), QQbar(-13), QQbar(89/55), QQbar(-sqrt(7)), QQbar.zeta(5)]
sage: def convert_test_all(ty):
....:     def convert_test(v):
....:         try:
....:             return ty(v)
....:         except (TypeError, ValueError):
....:             return None
....:     return [convert_test(_) for _ in all_vals]
sage: convert_test_all(float)
[42.0, 3.1428571428571432, 1.618033988749895, -13.0, 1.6181818181818182, -2.6457513110645907, None]
sage: convert_test_all(complex)
[(42+0j), (3.1428571428571432+0j), (1.618033988749895+0j), (-13+0j), (1.6181818181818182+0j), (-2.6457513110645907+0j), (0.30901699437494745+0.9510565162951536j)]
sage: convert_test_all(RDF)
[42.0, 3.1428571428571432, 1.618033988749895, -13.0, 1.6181818181818182, -2.6457513110645907, None]
sage: convert_test_all(CDF)
[42.0, 3.1428571428571432, 1.618033988749895, -13.0, 1.6181818181818182, -2.6457513110645907, 0.30901699437494745 + 0.9510565162951536*I]
sage: convert_test_all(RR)
[42.0000000000000, 3.14285714285714, 1.61803398874989, -13.0000000000000, 1.61818181818182, -2.64575131106459, None]
sage: convert_test_all(CC)
[42.0000000000000, 3.14285714285714, 1.61803398874989, -13.0000000000000, 1.61818181818182, -2.64575131106459, 0.309016994374947 + 0.951056516295154*I]
sage: convert_test_all(RIF)
[42, 3.142857142857143?, 1.618033988749895?, -13, 1.618181818181819?, -2.645751311064591?, None]
sage: convert_test_all(CIF)
[42, 3.142857142857143?, 1.618033988749895?, -13, 1.618181818181819?, -2.645751311064591?, 0.3090169943749474? + 0.9510565162951536?*I]
sage: convert_test_all(ZZ)
[42, None, None, -13, None, None, None]
sage: convert_test_all(QQ)
[42, 22/7, None, -13, 89/55, None, None]


Compute the exact coordinates of a 34-gon (the formulas used are from Weisstein, Eric W. “Trigonometry Angles–Pi/17.” and can be found at http://mathworld.wolfram.com/TrigonometryAnglesPi17.html):

sage: rt17 = AA(17).sqrt()
sage: rt2 = AA(2).sqrt()
sage: eps = (17 + rt17).sqrt()
sage: epss = (17 - rt17).sqrt()
sage: delta = rt17 - 1
sage: alpha = (34 + 6*rt17 + rt2*delta*epss - 8*rt2*eps).sqrt()
sage: beta = 2*(17 + 3*rt17 - 2*rt2*eps - rt2*epss).sqrt()
sage: x = rt2*(15 + rt17 + rt2*(alpha + epss)).sqrt()/8
sage: y = rt2*(epss**2 - rt2*(alpha + epss)).sqrt()/8

sage: cx, cy = 1, 0
sage: for i in range(34):
....:    cx, cy = x*cx-y*cy, x*cy+y*cx
sage: cx
1.000000000000000?
sage: cy
0.?e-15

sage: ax = polygen(AA)
sage: x2 = AA.polynomial_root(256*ax**8 - 128*ax**7 - 448*ax**6 + 192*ax**5 + 240*ax**4 - 80*ax**3 - 40*ax**2 + 8*ax + 1, RIF(0.9829, 0.983))
sage: y2 = (1-x2**2).sqrt()
sage: x - x2
0.?e-18
sage: y - y2
0.?e-17


Ideally, in the above example we should be able to test x == x2 and y == y2 but this is currently infinitely long.

class sage.rings.qqbar.ANBinaryExpr(left, right, op)#

Bases: ANDescr

Initialize this ANBinaryExpr.

EXAMPLES:

sage: t = QQbar(sqrt(2)) + QQbar(sqrt(3)); type(t._descr) # indirect doctest    # optional - sage.symbolic
<class 'sage.rings.qqbar.ANBinaryExpr'>

exactify()#
handle_sage_input(sib, coerce, is_qqbar)#

Produce an expression which will reproduce this value when evaluated, and an indication of whether this value is worth sharing (always True for ANBinaryExpr).

EXAMPLES:

sage: sage_input(2 + sqrt(AA(2)), verify=True)
# Verified
R.<x> = AA[]
2 + AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951)))
sage: sage_input(sqrt(AA(2)) + 2, verify=True)
# Verified
R.<x> = AA[]
AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951))) + 2
sage: sage_input(2 - sqrt(AA(2)), verify=True)
# Verified
R.<x> = AA[]
2 - AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951)))
sage: sage_input(2 / sqrt(AA(2)), verify=True)
# Verified
R.<x> = AA[]
2/AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951)))
sage: sage_input(2 + (-1*sqrt(AA(2))), verify=True)
# Verified
R.<x> = AA[]
2 - AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951)))
sage: sage_input(2*sqrt(AA(2)), verify=True)
# Verified
R.<x> = AA[]
2*AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951)))
sage: rt2 = sqrt(AA(2))
sage: one = rt2/rt2
sage: n = one+3
sage: sage_input(n)
R.<x> = AA[]
v = AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951)))
v/v + 3
sage: one == 1
True
sage: sage_input(n)
1 + AA(3)
sage: rt3 = QQbar(sqrt(3))                                                  # optional - sage.symbolic
sage: one = rt3/rt3                                                         # optional - sage.symbolic
sage: n = sqrt(AA(2)) + one
sage: one == 1                                                              # optional - sage.symbolic
True
sage: sage_input(n)
R.<x> = AA[]
QQbar.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951))) + 1
sage: from sage.rings.qqbar import *
sage: from sage.misc.sage_input import SageInputBuilder
sage: sib = SageInputBuilder()
sage: binexp = ANBinaryExpr(AA(3), AA(5), operator.mul)
sage: binexp.handle_sage_input(sib, False, False)
({binop:* {atomic:3} {call: {atomic:AA}({atomic:5})}}, True)
sage: binexp.handle_sage_input(sib, False, True)
({call: {atomic:QQbar}({binop:* {atomic:3} {call: {atomic:AA}({atomic:5})}})}, True)

is_complex()#

Whether this element is complex. Does not trigger exact computation, so may return True even if the element is real.

EXAMPLES:

sage: x = (QQbar(sqrt(-2)) / QQbar(sqrt(-5)))._descr                        # optional - sage.symbolic
sage: x.is_complex()                                                        # optional - sage.symbolic
True

class sage.rings.qqbar.ANDescr#

Bases: SageObject

An AlgebraicNumber or AlgebraicReal is a wrapper around an ANDescr object. ANDescr is an abstract base class, which should never be directly instantiated; its concrete subclasses are ANRational, ANBinaryExpr, ANUnaryExpr, ANRoot, and ANExtensionElement. ANDescr and all of its subclasses are for internal use, and should not be used directly.

abs(n)#

Absolute value of self.

EXAMPLES:

sage: a = QQbar(sqrt(2))
sage: b = a._descr
sage: b.abs(a)
<sage.rings.qqbar.ANUnaryExpr object at ...>

conjugate(n)#

Complex conjugate of self.

EXAMPLES:

sage: a = QQbar(sqrt(-7))
sage: b = a._descr
sage: b.conjugate(a)
<sage.rings.qqbar.ANUnaryExpr object at ...>

imag(n)#

Imaginary part of self.

EXAMPLES:

sage: a = QQbar(sqrt(-7))
sage: b = a._descr
sage: b.imag(a)
<sage.rings.qqbar.ANUnaryExpr object at ...>

invert(n)#

1/self.

EXAMPLES:

sage: a = QQbar(sqrt(2))
sage: b = a._descr
sage: b.invert(a)
<sage.rings.qqbar.ANUnaryExpr object at ...>

is_simple()#

Check whether this descriptor represents a value with the same algebraic degree as the number field associated with the descriptor.

This returns True if self is an ANRational, or a minimal ANExtensionElement.

EXAMPLES:

sage: from sage.rings.qqbar import ANRational
sage: ANRational(1/2).is_simple()
True
sage: rt2 = AA(sqrt(2))
sage: rt3 = AA(sqrt(3))
sage: rt2b = rt3 + rt2 - rt3
sage: rt2.exactify()
sage: rt2._descr.is_simple()
True
sage: rt2b.exactify()
sage: rt2b._descr.is_simple()
False
sage: rt2b.simplify()
sage: rt2b._descr.is_simple()
True

neg(n)#

Negation of self.

EXAMPLES:

sage: a = QQbar(sqrt(2))
sage: b = a._descr
sage: b.neg(a)
<sage.rings.qqbar.ANUnaryExpr object at ...>

norm(n)#

Field norm of self from $$\overline{\QQ}$$ to its real subfield $$\mathbf{A}$$, i.e.~the square of the usual complex absolute value.

EXAMPLES:

sage: a = QQbar(sqrt(-7))
sage: b = a._descr
sage: b.norm(a)
<sage.rings.qqbar.ANUnaryExpr object at ...>

real(n)#

Real part of self.

EXAMPLES:

sage: a = QQbar(sqrt(-7))
sage: b = a._descr
sage: b.real(a)
<sage.rings.qqbar.ANUnaryExpr object at ...>

class sage.rings.qqbar.ANExtensionElement(generator, value)#

Bases: ANDescr

The subclass of ANDescr that represents a number field element in terms of a specific generator. Consists of a polynomial with rational coefficients in terms of the generator, and the generator itself, an AlgebraicGenerator.

abs(n)#

Return the absolute value of self (square root of the norm).

EXAMPLES:

sage: a = QQbar(sqrt(-2)) + QQbar(sqrt(-3))                                 # optional - sage.symbolic
sage: a.exactify()                                                          # optional - sage.symbolic
sage: b = a._descr                                                          # optional - sage.symbolic
sage: type(b)                                                               # optional - sage.symbolic
<class 'sage.rings.qqbar.ANExtensionElement'>
sage: b.abs(a)                                                              # optional - sage.symbolic
Root 3.146264369941972342? of x^2 - 9.89897948556636?

conjugate(n)#

Complex conjugate of self.

EXAMPLES:

sage: a = QQbar(sqrt(-2)) + QQbar(sqrt(-3))                                 # optional - sage.symbolic
sage: a.exactify()                                                          # optional - sage.symbolic
sage: b = a._descr                                                          # optional - sage.symbolic
sage: type(b)                                                               # optional - sage.symbolic
<class 'sage.rings.qqbar.ANExtensionElement'>
sage: c = b.conjugate(None); c  # random (not uniquely represented)         # optional - sage.symbolic
1/3*a^3 - 1/3*a^2 + a + 1 where a^4 - 2*a^3 + a^2 + 6*a + 3 = 0
and a in 1.724744871391589? - 1.573132184970987?*I


Internally, complex conjugation is implemented by taking the same abstract field element but conjugating the complex embedding of the field:

sage: c.generator() == b.generator().conjugate()                            # optional - sage.symbolic
True
sage: c.field_element_value() == b.field_element_value()                    # optional - sage.symbolic
True


The parameter is ignored:

sage: (b.conjugate("random").generator() == c.generator()                   # optional - sage.symbolic
....:  and b.conjugate("random").field_element_value() == c.field_element_value())
True

exactify()#

Return an exact representation of self.

Since self is already exact, just return self.

EXAMPLES:

sage: x = polygen(ZZ, 'x')
sage: v = (x^2 - x - 1).roots(ring=AA, multiplicities=False)._descr.exactify()
sage: type(v)
<class 'sage.rings.qqbar.ANExtensionElement'>
sage: v.exactify() is v
True

field_element_value()#

Return the underlying number field element.

EXAMPLES:

sage: x = polygen(ZZ, 'x')
sage: v = (x^2 - x - 1).roots(ring=AA, multiplicities=False)._descr.exactify()
sage: v.field_element_value()
a

generator()#

Return the AlgebraicGenerator object corresponding to self.

EXAMPLES:

sage: x = polygen(ZZ, 'x')
sage: v = (x^2 - x - 1).roots(ring=AA, multiplicities=False)._descr.exactify()
sage: v.generator()
Number Field in a with defining polynomial y^2 - y - 1 with a in 1.618033988749895?

handle_sage_input(sib, coerce, is_qqbar)#

Produce an expression which will reproduce this value when evaluated, and an indication of whether this value is worth sharing (always True for ANExtensionElement).

EXAMPLES:

sage: I = QQbar(I)
sage: sage_input(3+4*I, verify=True)
# Verified
QQbar(3 + 4*I)
sage: v = QQbar.zeta(3) + QQbar.zeta(5)
sage: v - v == 0
True
sage: sage_input(vector(QQbar, (4-3*I, QQbar.zeta(7))), verify=True)
# Verified
R.<y> = QQ[]
vector(QQbar, [4 - 3*I, QQbar.polynomial_root(AA.common_polynomial(y^6 + y^5 + y^4 + y^3 + y^2 + y + 1), CIF(RIF(RR(0.62348980185873348), RR(0.62348980185873359)), RIF(RR(0.7818314824680298), RR(0.78183148246802991))))])
sage: sage_input(v, verify=True)
# Verified
R.<y> = QQ[]
v = QQbar.polynomial_root(AA.common_polynomial(y^8 - y^7 + y^5 - y^4 + y^3 - y + 1), CIF(RIF(RR(0.91354545764260087), RR(0.91354545764260098)), RIF(RR(0.40673664307580015), RR(0.40673664307580021))))
v^5 + v^3
sage: v = QQbar(sqrt(AA(2)))
sage: v.exactify()
sage: sage_input(v, verify=True)
# Verified
R.<y> = QQ[]
QQbar(AA.polynomial_root(AA.common_polynomial(y^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951))))
sage: from sage.rings.qqbar import *
sage: from sage.misc.sage_input import SageInputBuilder
sage: sib = SageInputBuilder()
sage: extel = ANExtensionElement(QQbar_I_generator, QQbar_I_generator.field().gen() + 1)
sage: extel.handle_sage_input(sib, False, True)
({call: {atomic:QQbar}({binop:+ {atomic:1} {atomic:I}})}, True)

invert(n)#

Reciprocal of self.

EXAMPLES:

sage: a = QQbar(sqrt(-2)) + QQbar(sqrt(-3))                                 # optional - sage.symbolic
sage: a.exactify()                                                          # optional - sage.symbolic
sage: b = a._descr                                                          # optional - sage.symbolic
sage: type(b)                                                               # optional - sage.symbolic
<class 'sage.rings.qqbar.ANExtensionElement'>
sage: c = b.invert(None); c  # random (not uniquely represented)            # optional - sage.symbolic
-7/3*a^3 + 19/3*a^2 - 7*a - 9 where a^4 - 2*a^3 + a^2 + 6*a + 3 = 0
and a in 1.724744871391589? + 1.573132184970987?*I
sage: (c.generator() == b.generator()                                       # optional - sage.symbolic
....:  and c.field_element_value() * b.field_element_value() == 1)
True


The parameter is ignored:

sage: (b.invert("random").generator() == c.generator()                      # optional - sage.symbolic
....:  and b.invert("random").field_element_value() == c.field_element_value())
True

is_complex()#

Return True if the number field that defines this element is not real.

This does not imply that the element itself is definitely non-real, as in the example below.

EXAMPLES:

sage: rt2 = QQbar(sqrt(2))                                                  # optional - sage.symbolic
sage: rtm3 = QQbar(sqrt(-3))                                                # optional - sage.symbolic
sage: x = rtm3 + rt2 - rtm3                                                 # optional - sage.symbolic
sage: x.exactify()                                                          # optional - sage.symbolic
sage: y = x._descr                                                          # optional - sage.symbolic
sage: type(y)                                                               # optional - sage.symbolic
<class 'sage.rings.qqbar.ANExtensionElement'>
sage: y.is_complex()                                                        # optional - sage.symbolic
True
sage: x.imag() == 0                                                         # optional - sage.symbolic
True

is_simple()#

Check whether this descriptor represents a value with the same algebraic degree as the number field associated with the descriptor.

For ANExtensionElement elements, we check this by comparing the degree of the minimal polynomial to the degree of the field.

EXAMPLES:

sage: rt2 = AA(sqrt(2))                                                     # optional - sage.symbolic
sage: rt3 = AA(sqrt(3))                                                     # optional - sage.symbolic
sage: rt2b = rt3 + rt2 - rt3                                                # optional - sage.symbolic
sage: rt2.exactify()                                                        # optional - sage.symbolic
sage: rt2._descr                                                            # optional - sage.symbolic
a where a^2 - 2 = 0 and a in 1.414213562373095?
sage: rt2._descr.is_simple()                                                # optional - sage.symbolic
True

sage: rt2b.exactify()                                                       # optional - sage.symbolic
sage: rt2b._descr                                                           # optional - sage.symbolic
a^3 - 3*a where a^4 - 4*a^2 + 1 = 0 and a in -0.5176380902050415?
sage: rt2b._descr.is_simple()                                               # optional - sage.symbolic
False

minpoly()#

Compute the minimal polynomial of this algebraic number.

EXAMPLES:

sage: x = polygen(ZZ, 'x')
sage: v = (x^2 - x - 1).roots(ring=AA, multiplicities=False)._descr.exactify()
sage: type(v)
<class 'sage.rings.qqbar.ANExtensionElement'>
sage: v.minpoly()
x^2 - x - 1

neg(n)#

Negation of self.

EXAMPLES:

sage: a = QQbar(sqrt(-2)) + QQbar(sqrt(-3))                                 # optional - sage.symbolic
sage: a.exactify()                                                          # optional - sage.symbolic
sage: b = a._descr                                                          # optional - sage.symbolic
sage: type(b)                                                               # optional - sage.symbolic
<class 'sage.rings.qqbar.ANExtensionElement'>
sage: c = b.neg(None); c  # random (not uniquely represented)               # optional - sage.symbolic
-1/3*a^3 + 1/3*a^2 - a - 1 where a^4 - 2*a^3 + a^2 + 6*a + 3 = 0
and a in 1.724744871391589? + 1.573132184970987?*I
sage: (c.generator() == b.generator()                                       # optional - sage.symbolic
....:  and c.field_element_value() + b.field_element_value() == 0)
True


The parameter is ignored:

sage: (b.neg("random").generator() == c.generator()                         # optional - sage.symbolic
....:  and b.neg("random").field_element_value() == c.field_element_value())
True

norm(n)#

Norm of self (square of complex absolute value)

EXAMPLES:

sage: a = QQbar(sqrt(-2)) + QQbar(sqrt(-3))                                 # optional - sage.symbolic
sage: a.exactify()                                                          # optional - sage.symbolic
sage: b = a._descr                                                          # optional - sage.symbolic
sage: type(b)                                                               # optional - sage.symbolic
<class 'sage.rings.qqbar.ANExtensionElement'>
sage: b.norm(a)                                                             # optional - sage.symbolic
<sage.rings.qqbar.ANUnaryExpr object at ...>

rational_argument(n)#

If the argument of self is $$2\pi$$ times some rational number in $$[1/2, -1/2)$$, return that rational; otherwise, return None.

EXAMPLES:

sage: a = QQbar(sqrt(-2)) + QQbar(sqrt(3))                                  # optional - sage.symbolic
sage: a.exactify()                                                          # optional - sage.symbolic
sage: b = a._descr                                                          # optional - sage.symbolic
sage: type(b)                                                               # optional - sage.symbolic
<class 'sage.rings.qqbar.ANExtensionElement'>
sage: b.rational_argument(a) is None                                        # optional - sage.symbolic
True
sage: x = polygen(QQ)
sage: a = (x^4 + 1).roots(QQbar, multiplicities=False)
sage: a.exactify()
sage: b = a._descr
sage: b.rational_argument(a)
-3/8

simplify(n)#

Compute an exact representation for this descriptor, in the smallest possible number field.

INPUT:

• n – The element of AA or QQbar corresponding to this descriptor.

EXAMPLES:

sage: rt2 = AA(sqrt(2))                                                     # optional - sage.symbolic
sage: rt3 = AA(sqrt(3))                                                     # optional - sage.symbolic
sage: rt2b = rt3 + rt2 - rt3                                                # optional - sage.symbolic
sage: rt2b.exactify()                                                       # optional - sage.symbolic
sage: rt2b._descr                                                           # optional - sage.symbolic
a^3 - 3*a where a^4 - 4*a^2 + 1 = 0 and a in -0.5176380902050415?
sage: rt2b._descr.simplify(rt2b)                                            # optional - sage.symbolic
a where a^2 - 2 = 0 and a in 1.414213562373095?

class sage.rings.qqbar.ANRational(x)#

Bases: ANDescr

The subclass of ANDescr that represents an arbitrary rational. This class is private, and should not be used directly.

abs(n)#

Absolute value of self.

EXAMPLES:

sage: a = QQbar(3)
sage: b = a._descr
sage: b.abs(a)
3

angle()#

Return a rational number $$q \in (-1/2, 1/2]$$ such that self is a rational multiple of $$e^{2\pi i q}$$. Always returns 0, since this element is rational.

EXAMPLES:

sage: QQbar(3)._descr.angle()
0
sage: QQbar(-3)._descr.angle()
0
sage: QQbar(0)._descr.angle()
0

exactify()#

Calculate self exactly. Since self is a rational number, return self.

EXAMPLES:

sage: a = QQbar(1/3)._descr
sage: a.exactify() is a
True

generator()#

Return an AlgebraicGenerator object associated to this element. Returns the trivial generator, since self is rational.

EXAMPLES:

sage: QQbar(0)._descr.generator()
Trivial generator

handle_sage_input(sib, coerce, is_qqbar)#

Produce an expression which will reproduce this value when evaluated, and an indication of whether this value is worth sharing (always False, for rationals).

EXAMPLES:

sage: sage_input(QQbar(22/7), verify=True)
# Verified
QQbar(22/7)
sage: sage_input(-AA(3)/5, verify=True)
# Verified
AA(-3/5)
sage: sage_input(vector(AA, (0, 1/2, 1/3)), verify=True)
# Verified
vector(AA, [0, 1/2, 1/3])
sage: from sage.rings.qqbar import *
sage: from sage.misc.sage_input import SageInputBuilder
sage: sib = SageInputBuilder()
sage: rat = ANRational(9/10)
sage: rat.handle_sage_input(sib, False, True)
({call: {atomic:QQbar}({binop:/ {atomic:9} {atomic:10}})}, False)

invert(n)#

1/self.

EXAMPLES:

sage: a = QQbar(3)
sage: b = a._descr
sage: b.invert(a)
1/3

is_complex()#

Return False, since rational numbers are real

EXAMPLES:

sage: QQbar(1/7)._descr.is_complex()
False

is_simple()#

Checks whether this descriptor represents a value with the same algebraic degree as the number field associated with the descriptor.

This is always true for rational numbers.

EXAMPLES:

sage: AA(1/2)._descr.is_simple()
True

minpoly()#

Return the min poly of self over $$\QQ$$.

EXAMPLES:

sage: QQbar(7)._descr.minpoly()
x - 7

neg(n)#

Negation of self.

EXAMPLES:

sage: a = QQbar(3)
sage: b = a._descr
sage: type(b)
<class 'sage.rings.qqbar.ANRational'>
sage: b.neg(a)
-3

rational_argument(n)#

Return the argument of self divided by $$2 \pi$$, or None if this element is 0.

EXAMPLES:

sage: QQbar(3)._descr.rational_argument(None)
0
sage: QQbar(-3)._descr.rational_argument(None)
1/2
sage: QQbar(0)._descr.rational_argument(None) is None
True

scale()#

Return a rational number $$r$$ such that self is equal to $$r e^{2 \pi i q}$$ for some $$q \in (-1/2, 1/2]$$. In other words, just return self as a rational number.

EXAMPLES:

sage: QQbar(-3)._descr.scale()
-3

class sage.rings.qqbar.ANRoot(poly, interval, multiplicity=1)#

Bases: ANDescr

The subclass of ANDescr that represents a particular root of a polynomial with algebraic coefficients. This class is private, and should not be used directly.

conjugate(n)#

Complex conjugate of this ANRoot object.

EXAMPLES:

sage: a = (x^2 + 23).roots(ring=QQbar, multiplicities=False)             # optional - sage.symbolic
sage: b = a._descr                                                          # optional - sage.symbolic
sage: type(b)                                                               # optional - sage.symbolic
<class 'sage.rings.qqbar.ANRoot'>
sage: c = b.conjugate(a); c                                                 # optional - sage.symbolic
<sage.rings.qqbar.ANUnaryExpr object at ...>
sage: c.exactify()                                                          # optional - sage.symbolic
-2*a + 1 where a^2 - a + 6 = 0 and a in 0.50000000000000000? - 2.397915761656360?*I

exactify()#

Return either an ANRational or an ANExtensionElement with the same value as this number.

EXAMPLES:

sage: from sage.rings.qqbar import ANRoot
sage: x = polygen(QQbar)
sage: two = ANRoot((x-2)*(x-sqrt(QQbar(2))), RIF(1.9, 2.1))
sage: two.exactify()
2
sage: strange = ANRoot(x^2 + sqrt(QQbar(3))*x - sqrt(QQbar(2)), RIF(-0, 1))
sage: strange.exactify()
a where a^8 - 6*a^6 + 5*a^4 - 12*a^2 + 4 = 0 and a in 0.6051012265139511?

handle_sage_input(sib, coerce, is_qqbar)#

Produce an expression which will reproduce this value when evaluated, and an indication of whether this value is worth sharing (always True for ANRoot).

EXAMPLES:

sage: sage_input((AA(3)^(1/2))^(1/3), verify=True)
# Verified
R.<x> = AA[]
AA.polynomial_root(AA.common_polynomial(x^3 - AA.polynomial_root(AA.common_polynomial(x^2 - 3), RIF(RR(1.7320508075688772), RR(1.7320508075688774)))), RIF(RR(1.2009369551760025), RR(1.2009369551760027)))


These two examples are too big to verify quickly. (Verification would create a field of degree 28.):

sage: sage_input((sqrt(AA(3))^(5/7))^(9/4))
R.<x> = AA[]
v1 = AA.polynomial_root(AA.common_polynomial(x^2 - 3), RIF(RR(1.7320508075688772), RR(1.7320508075688774)))
v2 = v1*v1
v3 = AA.polynomial_root(AA.common_polynomial(x^7 - v2*v2*v1), RIF(RR(1.4804728524798112), RR(1.4804728524798114)))
v4 = v3*v3
v5 = v4*v4
AA.polynomial_root(AA.common_polynomial(x^4 - v5*v5*v3), RIF(RR(2.4176921938267877), RR(2.4176921938267881)))
sage: sage_input((sqrt(QQbar(-7))^(5/7))^(9/4))
R.<x> = QQbar[]
v1 = QQbar.polynomial_root(AA.common_polynomial(x^2 + 7), CIF(RIF(RR(0)), RIF(RR(2.6457513110645903), RR(2.6457513110645907))))
v2 = v1*v1
v3 = QQbar.polynomial_root(AA.common_polynomial(x^7 - v2*v2*v1), CIF(RIF(RR(0.8693488875796217), RR(0.86934888757962181)), RIF(RR(1.8052215661454434), RR(1.8052215661454436))))
v4 = v3*v3
v5 = v4*v4
QQbar.polynomial_root(AA.common_polynomial(x^4 - v5*v5*v3), CIF(RIF(-RR(3.8954086044650791), -RR(3.8954086044650786)), RIF(RR(2.7639398015408925), RR(2.7639398015408929))))
sage: x = polygen(QQ)
sage: sage_input(AA.polynomial_root(x^2-x-1, RIF(1, 2)), verify=True)
# Verified
R.<y> = QQ[]
AA.polynomial_root(AA.common_polynomial(y^2 - y - 1), RIF(RR(1.6180339887498947), RR(1.6180339887498949)))
sage: sage_input(QQbar.polynomial_root(x^3-5, CIF(RIF(-3, 0), RIF(0, 3))), verify=True)
# Verified
R.<y> = QQ[]
QQbar.polynomial_root(AA.common_polynomial(y^3 - 5), CIF(RIF(-RR(0.85498797333834853), -RR(0.85498797333834842)), RIF(RR(1.4808826096823642), RR(1.4808826096823644))))
sage: from sage.rings.qqbar import *
sage: from sage.misc.sage_input import SageInputBuilder
sage: sib = SageInputBuilder()
sage: rt = ANRoot(x^3 - 2, RIF(0, 4))
sage: rt.handle_sage_input(sib, False, True)
({call: {getattr: {atomic:QQbar}.polynomial_root}({call: {getattr: {atomic:AA}.common_polynomial}({binop:- {binop:** {gen:y {constr_parent: {subscr: {atomic:QQ}[{atomic:'y'}]} with gens: ('y',)}} {atomic:3}} {atomic:2}})}, {call: {atomic:RIF}({call: {atomic:RR}({atomic:1.259921049894873})}, {call: {atomic:RR}({atomic:1.2599210498948732})})})},
True)

is_complex()#

Whether this is a root in $$\overline{\QQ}$$ (rather than $$\mathbf{A}$$). Note that this may return True even if the root is actually real, as the second example shows; it does not trigger exact computation to see if the root is real.

EXAMPLES:

sage: x = polygen(QQ)
sage: (x^2 - x - 1).roots(ring=AA, multiplicities=False)._descr.is_complex()
False
sage: (x^2 - x - 1).roots(ring=QQbar, multiplicities=False)._descr.is_complex()
True

refine_interval(interval, prec)#

Takes an interval which is assumed to enclose exactly one root of the polynomial (or, with multiplicity=k, exactly one root of the $$k-1$$-st derivative); and a precision, in bits.

Tries to find a narrow interval enclosing the root using interval arithmetic of the given precision. (No particular number of resulting bits of precision is guaranteed.)

Uses a combination of Newton’s method (adapted for interval arithmetic) and bisection. The algorithm will converge very quickly if started with a sufficiently narrow interval.

EXAMPLES:

sage: from sage.rings.qqbar import ANRoot
sage: x = polygen(AA)
sage: rt2 = ANRoot(x^2 - 2, RIF(0, 2))
sage: rt2.refine_interval(RIF(0, 2), 75)
1.4142135623730950488017?

class sage.rings.qqbar.ANUnaryExpr(arg, op)#

Bases: ANDescr

Initialize this ANUnaryExpr.

EXAMPLES:

sage: t = ~QQbar(sqrt(2)); type(t._descr)  # indirect doctest               # optional - sage.symbolic
<class 'sage.rings.qqbar.ANUnaryExpr'>

exactify()#

Trigger exact computation of self.

EXAMPLES:

sage: v = (-QQbar(sqrt(2)))._descr                                          # optional - sage.symbolic
sage: type(v)                                                               # optional - sage.symbolic
<class 'sage.rings.qqbar.ANUnaryExpr'>
sage: v.exactify()                                                          # optional - sage.symbolic
-a where a^2 - 2 = 0 and a in 1.414213562373095?

handle_sage_input(sib, coerce, is_qqbar)#

Produce an expression which will reproduce this value when evaluated, and an indication of whether this value is worth sharing (always True for ANUnaryExpr).

EXAMPLES:

sage: sage_input(-sqrt(AA(2)), verify=True)
# Verified
R.<x> = AA[]
-AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951)))
sage: sage_input(~sqrt(AA(2)), verify=True)
# Verified
R.<x> = AA[]
~AA.polynomial_root(AA.common_polynomial(x^2 - 2), RIF(RR(1.4142135623730949), RR(1.4142135623730951)))
sage: sage_input(sqrt(QQbar(-3)).conjugate(), verify=True)
# Verified
R.<x> = QQbar[]
QQbar.polynomial_root(AA.common_polynomial(x^2 + 3), CIF(RIF(RR(0)), RIF(RR(1.7320508075688772), RR(1.7320508075688774)))).conjugate()
sage: sage_input(QQbar.zeta(3).real(), verify=True)
# Verified
R.<y> = QQ[]
QQbar.polynomial_root(AA.common_polynomial(y^2 + y + 1), CIF(RIF(-RR(0.50000000000000011), -RR(0.49999999999999994)), RIF(RR(0.8660254037844386), RR(0.86602540378443871)))).real()
sage: sage_input(QQbar.zeta(3).imag(), verify=True)
# Verified
R.<y> = QQ[]
QQbar.polynomial_root(AA.common_polynomial(y^2 + y + 1), CIF(RIF(-RR(0.50000000000000011), -RR(0.49999999999999994)), RIF(RR(0.8660254037844386), RR(0.86602540378443871)))).imag()
sage: sage_input(abs(sqrt(QQbar(-3))), verify=True)
# Verified
R.<x> = QQbar[]
abs(QQbar.polynomial_root(AA.common_polynomial(x^2 + 3), CIF(RIF(RR(0)), RIF(RR(1.7320508075688772), RR(1.7320508075688774)))))
sage: sage_input(sqrt(QQbar(-3)).norm(), verify=True)
# Verified
R.<x> = QQbar[]
QQbar.polynomial_root(AA.common_polynomial(x^2 + 3), CIF(RIF(RR(0)), RIF(RR(1.7320508075688772), RR(1.7320508075688774)))).norm()
sage: sage_input(QQbar(QQbar.zeta(3).real()), verify=True)
# Verified
R.<y> = QQ[]
QQbar(QQbar.polynomial_root(AA.common_polynomial(y^2 + y + 1), CIF(RIF(-RR(0.50000000000000011), -RR(0.49999999999999994)), RIF(RR(0.8660254037844386), RR(0.86602540378443871)))).real())
sage: from sage.rings.qqbar import *
sage: from sage.misc.sage_input import SageInputBuilder
sage: sib = SageInputBuilder()
sage: unexp = ANUnaryExpr(sqrt(AA(2)), '~')
sage: unexp.handle_sage_input(sib, False, False)
({unop:~ {call: {getattr: {atomic:AA}.polynomial_root}({call: {getattr: {atomic:AA}.common_polynomial}({binop:- {binop:** {gen:x {constr_parent: {subscr: {atomic:AA}[{atomic:'x'}]} with gens: ('x',)}} {atomic:2}} {atomic:2}})}, {call: {atomic:RIF}({call: {atomic:RR}({atomic:1.4142135623730949})}, {call: {atomic:RR}({atomic:1.4142135623730951})})})}},
True)
sage: unexp.handle_sage_input(sib, False, True)
({call: {atomic:QQbar}({unop:~ {call: {getattr: {atomic:AA}.polynomial_root}({call: {getattr: {atomic:AA}.common_polynomial}({binop:- {binop:** {gen:x {constr_parent: {subscr: {atomic:AA}[{atomic:'x'}]} with gens: ('x',)}} {atomic:2}} {atomic:2}})}, {call: {atomic:RIF}({call: {atomic:RR}({atomic:1.4142135623730949})}, {call: {atomic:RR}({atomic:1.4142135623730951})})})}})},
True)

is_complex()#

Return whether or not this element is complex. Note that this is a data type check, and triggers no computations – if it returns False, the element might still be real, it just doesn’t know it yet.

EXAMPLES:

sage: t = AA(sqrt(2))                                                       # optional - sage.symbolic
sage: s = (-t)._descr                                                       # optional - sage.symbolic
sage: s                                                                     # optional - sage.symbolic
<sage.rings.qqbar.ANUnaryExpr object at ...>
sage: s.is_complex()                                                        # optional - sage.symbolic
False
sage: QQbar(-sqrt(2))._descr.is_complex()                                   # optional - sage.symbolic
True

class sage.rings.qqbar.AlgebraicField#

The field of all algebraic complex numbers.

algebraic_closure()#

Return the algebraic closure of this field.

As this field is already algebraically closed, just returns self.

EXAMPLES:

sage: QQbar.algebraic_closure()
Algebraic Field

completion(p, prec, extras={})#

Return the completion of self at the place $$p$$.

Only implemented for $$p = \infty$$ at present.

INPUT:

• p – either a prime (not implemented at present) or Infinity

• prec – precision of approximate field to return

• extras – (optional) a dict of extra keyword arguments for the RealField constructor

EXAMPLES:

sage: QQbar.completion(infinity, 500)
Complex Field with 500 bits of precision
sage: QQbar.completion(infinity, prec=53, extras={'type':'RDF'})
Complex Double Field
sage: QQbar.completion(infinity, 53) is CC
True
sage: QQbar.completion(3, 20)
Traceback (most recent call last):
...
NotImplementedError

construction()#

Return a functor that constructs self (used by the coercion machinery).

EXAMPLES:

sage: QQbar.construction()
(AlgebraicClosureFunctor, Rational Field)

gen(n=0)#

Return the $$n$$-th element of the tuple returned by gens().

EXAMPLES:

sage: QQbar.gen(0)
I
sage: QQbar.gen(1)
Traceback (most recent call last):
...
IndexError: n must be 0

gens()#

Return a set of generators for this field.

As this field is not finitely generated over its prime field, we opt for just returning I.

EXAMPLES:

sage: QQbar.gens()
(I,)

ngens()#

Return the size of the tuple returned by gens().

EXAMPLES:

sage: QQbar.ngens()
1

polynomial_root(poly, interval, multiplicity=1)#

Given a polynomial with algebraic coefficients and an interval enclosing exactly one root of the polynomial, constructs an algebraic real representation of that root.

The polynomial need not be irreducible, or even squarefree; but if the given root is a multiple root, its multiplicity must be specified. (IMPORTANT NOTE: Currently, multiplicity-$$k$$ roots are handled by taking the $$(k-1)$$-st derivative of the polynomial. This means that the interval must enclose exactly one root of this derivative.)

The conditions on the arguments (that the interval encloses exactly one root, and that multiple roots match the given multiplicity) are not checked; if they are not satisfied, an error may be thrown (possibly later, when the algebraic number is used), or wrong answers may result.

Note that if you are constructing multiple roots of a single polynomial, it is better to use QQbar.common_polynomial to get a shared polynomial.

EXAMPLES:

sage: x = polygen(QQbar)
sage: phi = QQbar.polynomial_root(x^2 - x - 1, RIF(0, 2)); phi
1.618033988749895?
sage: p = (x-1)^7 * (x-2)
sage: r = QQbar.polynomial_root(p, RIF(9/10, 11/10), multiplicity=7)
sage: r; r == 1
1
True
sage: p = (x-phi)*(x-sqrt(QQbar(2)))
sage: r = QQbar.polynomial_root(p, RIF(1, 3/2))
sage: r; r == sqrt(QQbar(2))
1.414213562373095?
True

random_element(poly_degree=2, *args, **kwds)#

Return a random algebraic number.

INPUT:

• poly_degree - default: 2 - degree of the random polynomial over the integers of which the returned algebraic number is a root. This is not necessarily the degree of the minimal polynomial of the number. Increase this parameter to achieve a greater diversity of algebraic numbers, at a cost of greater computation time. You can also vary the distribution of the coefficients but that will not vary the degree of the extension containing the element.

• args, kwds - arguments and keywords passed to the random number generator for elements of ZZ, the integers. See random_element() for details, or see example below.

OUTPUT:

An element of QQbar, the field of algebraic numbers (see sage.rings.qqbar).

ALGORITHM:

A polynomial with degree between 1 and poly_degree, with random integer coefficients is created. A root of this polynomial is chosen at random. The default degree is 2 and the integer coefficients come from a distribution heavily weighted towards $$0, \pm 1, \pm 2$$.

EXAMPLES:

sage: a = QQbar.random_element()
sage: a                         # random
0.2626138748742799? + 0.8769062830975992?*I
sage: a in QQbar
True

sage: b = QQbar.random_element(poly_degree=20)
sage: b                         # random
-0.8642649077479498? - 0.5995098147478391?*I
sage: b in QQbar
True


Parameters for the distribution of the integer coefficients of the polynomials can be passed on to the random element method for integers. For example, current default behavior of this method returns zero about 15% of the time; if we do not include zero as a possible coefficient, there will never be a zero constant term, and thus never a zero root.

sage: z = [QQbar.random_element(x=1, y=10) for _ in range(20)]
sage: QQbar(0) in z
False


If you just want real algebraic numbers you can filter them out. Using an odd degree for the polynomials will ensure some degree of success.

sage: r = []
sage: while len(r) < 3:
....:   x = QQbar.random_element(poly_degree=3)
....:   if x in AA:
....:     r.append(x)
sage: (len(r) == 3) and all(z in AA for z in r)
True

zeta(n=4)#

Return a primitive $$n$$’th root of unity, specifically $$\exp(2*\pi*i/n)$$.

INPUT:

• n (integer) – default 4

EXAMPLES:

sage: QQbar.zeta(1)
1
sage: QQbar.zeta(2)
-1
sage: QQbar.zeta(3)
-0.500000000000000? + 0.866025403784439?*I
sage: QQbar.zeta(4)
I
sage: QQbar.zeta()
I
sage: QQbar.zeta(5)
0.3090169943749474? + 0.9510565162951536?*I
sage: QQbar.zeta(3000)
0.999997806755380? + 0.002094393571219374?*I

class sage.rings.qqbar.AlgebraicField_common#

Common base class for the classes AlgebraicRealField and AlgebraicField.

characteristic()#

Return the characteristic of this field.

Since this class is only used for fields of characteristic 0, this always returns 0.

EXAMPLES:

sage: AA.characteristic()
0

common_polynomial(poly)#

Given a polynomial with algebraic coefficients, return a wrapper that caches high-precision calculations and factorizations. This wrapper can be passed to polynomial_root() in place of the polynomial.

Using common_polynomial() makes no semantic difference, but will improve efficiency if you are dealing with multiple roots of a single polynomial.

EXAMPLES:

sage: x = polygen(ZZ)
sage: p = AA.common_polynomial(x^2 - x - 1)
sage: phi = AA.polynomial_root(p, RIF(1, 2))
sage: tau = AA.polynomial_root(p, RIF(-1, 0))
sage: phi + tau == 1
True
sage: phi * tau == -1
True

sage: x = polygen(SR)                                                       # optional - sage.symbolic
sage: p = (x - sqrt(-5)) * (x - sqrt(3)); p                                 # optional - sage.symbolic
x^2 + (-sqrt(3) - sqrt(-5))*x + sqrt(3)*sqrt(-5)
sage: p = QQbar.common_polynomial(p)                                        # optional - sage.symbolic
sage: a = QQbar.polynomial_root(p, CIF(RIF(-0.1, 0.1), RIF(2, 3))); a       # optional - sage.symbolic
0.?e-18 + 2.236067977499790?*I
sage: b = QQbar.polynomial_root(p, RIF(1, 2)); b                            # optional - sage.symbolic
1.732050807568878?


These “common polynomials” can be shared between real and complex roots:

sage: p = AA.common_polynomial(x^3 - x - 1)
sage: r1 = AA.polynomial_root(p, RIF(1.3, 1.4)); r1
1.324717957244746?
sage: r2 = QQbar.polynomial_root(p, CIF(RIF(-0.7, -0.6), RIF(0.5, 0.6))); r2
-0.6623589786223730? + 0.5622795120623013?*I

default_interval_prec()#

Return the default interval precision used for root isolation.

EXAMPLES:

sage: AA.default_interval_prec()
64

options = Current options for AlgebraicField   - display_format: decimal#
order()#

Return the cardinality of self.

Since this class is only used for fields of characteristic 0, always returns Infinity.

EXAMPLES:

sage: QQbar.order()
+Infinity

class sage.rings.qqbar.AlgebraicGenerator(field, root)#

Bases: SageObject

An AlgebraicGenerator represents both an algebraic number $$\alpha$$ and the number field $$\QQ[\alpha]$$. There is a single AlgebraicGenerator representing $$\QQ$$ (with $$\alpha=0$$).

The AlgebraicGenerator class is private, and should not be used directly.

conjugate()#

If this generator is for the algebraic number $$\alpha$$, return a generator for the complex conjugate of $$\alpha$$.

EXAMPLES:

sage: from sage.rings.qqbar import AlgebraicGenerator
sage: x = polygen(QQ); f = x^4 + x + 17
sage: nf = NumberField(f,name='a')
sage: b = f.roots(QQbar)
sage: root = b._descr
sage: gen = AlgebraicGenerator(nf, root)
sage: gen.conjugate()
Number Field in a with defining polynomial x^4 + x + 17 with a in -1.436449997483091? + 1.374535713065812?*I

field()#

Return the number field attached to self.

EXAMPLES:

sage: from sage.rings.qqbar import qq_generator
sage: qq_generator.field()
Rational Field

is_complex()#

Return True if this is a generator for a non-real number field.

EXAMPLES:

sage: z7 = QQbar.zeta(7)
sage: g = z7._descr._generator
sage: g.is_complex()
True

sage: from sage.rings.qqbar import ANRoot, AlgebraicGenerator
sage: y = polygen(QQ, 'y')
sage: x = polygen(QQbar)
sage: nf = NumberField(y^2 - y - 1, name='a', check=False)
sage: root = ANRoot(x^2 - x - 1, RIF(1, 2))
sage: gen = AlgebraicGenerator(nf, root)
sage: gen.is_complex()
False

is_trivial()#

Return True iff this is the trivial generator (alpha == 1), which does not actually extend the rationals.

EXAMPLES:

sage: from sage.rings.qqbar import qq_generator
sage: qq_generator.is_trivial()
True

pari_field()#

Return the PARI field attached to this generator.

EXAMPLES:

sage: from sage.rings.qqbar import qq_generator
sage: qq_generator.pari_field()
Traceback (most recent call last):
...
ValueError: No PARI field attached to trivial generator

sage: from sage.rings.qqbar import ANRoot, AlgebraicGenerator, qq_generator
sage: y = polygen(QQ)
sage: x = polygen(QQbar)
sage: nf = NumberField(y^2 - y - 1, name='a', check=False)
sage: root = ANRoot(x^2 - x - 1, RIF(1, 2))
sage: gen = AlgebraicGenerator(nf, root)
sage: gen.pari_field()
[[y^2 - y - 1, [2, 0], ...]

root_as_algebraic()#

Return the root attached to self as an algebraic number.

EXAMPLES:

sage: t = sage.rings.qqbar.qq_generator.root_as_algebraic(); t
1
sage: t.parent()
Algebraic Real Field

super_poly(super, checked=None)#

Given a generator gen and another generator super, where super is the result of a tree of union() operations where one of the leaves is gen, gen.super_poly(super) returns a polynomial expressing the value of gen in terms of the value of super (except that if gen is qq_generator, super_poly() always returns None.)

EXAMPLES:

sage: from sage.rings.qqbar import AlgebraicGenerator, ANRoot, qq_generator
sage: _.<y> = QQ['y']
sage: x = polygen(QQbar)
sage: nf2 = NumberField(y^2 - 2, name='a', check=False)
sage: root2 = ANRoot(x^2 - 2, RIF(1, 2))
sage: gen2 = AlgebraicGenerator(nf2, root2)
sage: gen2
Number Field in a with defining polynomial y^2 - 2 with a in 1.414213562373095?
sage: nf3 = NumberField(y^2 - 3, name='a', check=False)
sage: root3 = ANRoot(x^2 - 3, RIF(1, 2))
sage: gen3 = AlgebraicGenerator(nf3, root3)
sage: gen3
Number Field in a with defining polynomial y^2 - 3 with a in 1.732050807568878?
sage: gen2_3 = gen2.union(gen3)
sage: gen2_3
Number Field in a with defining polynomial y^4 - 4*y^2 + 1 with a in -1.931851652578137?
sage: qq_generator.super_poly(gen2) is None
True
sage: gen2.super_poly(gen2_3)
-a^3 + 3*a
sage: gen3.super_poly(gen2_3)
a^2 - 2

union(other)#

Given generators alpha and beta, alpha.union(beta) gives a generator for the number field $$\QQ[\alpha][\beta]$$.

EXAMPLES:

sage: from sage.rings.qqbar import ANRoot, AlgebraicGenerator, qq_generator
sage: _.<y> = QQ['y']
sage: x = polygen(QQbar)
sage: nf2 = NumberField(y^2 - 2, name='a', check=False)
sage: root2 = ANRoot(x^2 - 2, RIF(1, 2))
sage: gen2 = AlgebraicGenerator(nf2, root2)
sage: gen2
Number Field in a with defining polynomial y^2 - 2 with a in 1.414213562373095?
sage: nf3 = NumberField(y^2 - 3, name='a', check=False)
sage: root3 = ANRoot(x^2 - 3, RIF(1, 2))
sage: gen3 = AlgebraicGenerator(nf3, root3)
sage: gen3
Number Field in a with defining polynomial y^2 - 3 with a in 1.732050807568878?
sage: gen2.union(qq_generator) is gen2
True
sage: qq_generator.union(gen3) is gen3
True
sage: gen2.union(gen3)
Number Field in a with defining polynomial y^4 - 4*y^2 + 1 with a in -1.931851652578137?

class sage.rings.qqbar.AlgebraicGeneratorRelation(child1, child1_poly, child2, child2_poly, parent)#

Bases: SageObject

A simple class for maintaining relations in the lattice of algebraic extensions.

class sage.rings.qqbar.AlgebraicNumber(x)#

The class for algebraic numbers (complex numbers which are the roots of a polynomial with integer coefficients). Much of its functionality is inherited from AlgebraicNumber_base.

_richcmp_(other, op)#

Compare two algebraic numbers, lexicographically. (That is, first compare the real components; if the real components are equal, compare the imaginary components.)

EXAMPLES:

sage: x = QQbar.zeta(3); x
-0.500000000000000? + 0.866025403784439?*I
sage: QQbar(-1) < x
True
sage: QQbar(-1/2) < x
True
sage: QQbar(0) > x
True


One problem with this lexicographic ordering is the fact that if two algebraic numbers have the same real component, that real component has to be compared for exact equality, which can be a costly operation. For the special case where both numbers have the same minimal polynomial, that cost can be avoided, though (see github issue #16964):

sage: x = polygen(ZZ)
sage: p = 69721504*x^8 + 251777664*x^6 + 329532012*x^4 + 184429548*x^2 + 37344321
sage: sorted(p.roots(QQbar,False))
[-0.0221204634374361? - 1.090991904211621?*I,
-0.0221204634374361? + 1.090991904211621?*I,
-0.8088604911480535?*I,
0.?e-215 - 0.7598602580415435?*I,
0.?e-229 + 0.7598602580415435?*I,
0.8088604911480535?*I,
0.0221204634374361? - 1.090991904211621?*I,
0.0221204634374361? + 1.090991904211621?*I]


It also works for comparison of conjugate roots even in a degenerate situation where many roots have the same real part. In the following example, the polynomial p2 is irreducible and all its roots have real part equal to $$1$$:

sage: p1 = x^8 + 74*x^7 + 2300*x^6 + 38928*x^5 + \
....: 388193*x^4 + 2295312*x^3 + 7613898*x^2 + \
....: 12066806*x + 5477001
sage: p2 = p1((x-1)^2)
sage: sum(1 for r in p2.roots(CC,False) if abs(r.real() - 1) < 0.0001)
16
sage: r1 = QQbar.polynomial_root(p2, CIF(1, (-4.1,-4.0)))
sage: r2 = QQbar.polynomial_root(p2, CIF(1, (4.0, 4.1)))
sage: all([r1<r2, r1==r1, r2==r2, r2>r1])
True


Though, comparing roots which are not equal or conjugate is much slower because the algorithm needs to check the equality of the real parts:

sage: sorted(p2.roots(QQbar,False))   # long time - 3 secs
[1.000000000000000? - 4.016778562562223?*I,
1.000000000000000? - 3.850538755978243?*I,
1.000000000000000? - 3.390564396412898?*I,
...
1.000000000000000? + 3.390564396412898?*I,
1.000000000000000? + 3.850538755978243?*I,
1.000000000000000? + 4.016778562562223?*I]

complex_exact(field)#

Given a ComplexField, return the best possible approximation of this number in that field. Note that if either component is sufficiently close to the halfway point between two floating-point numbers in the corresponding RealField, then this will trigger exact computation, which may be very slow.

EXAMPLES:

sage: a = QQbar.zeta(9) + QQbar(I) + QQbar.zeta(9).conjugate(); a
1.532088886237957? + 1.000000000000000?*I
sage: a.complex_exact(CIF)
1.532088886237957? + 1*I

complex_number(field)#

Given the complex field field, compute an accurate approximation of this element in that field.

The approximation will be off by at most two ulp’s in each component, except for components which are very close to zero, which will have an absolute error at most $$2^{-prec+1}$$ where prec is the precision of the field.

EXAMPLES:

sage: a = QQbar.zeta(5)
sage: a.complex_number(CC)
0.309016994374947 + 0.951056516295154*I

sage: b = QQbar(2).sqrt() + QQbar(3).sqrt() * QQbar.gen()
sage: b.complex_number(ComplexField(128))
1.4142135623730950488016887242096980786 + 1.7320508075688772935274463415058723669*I

conjugate()#

Return the complex conjugate of self.

EXAMPLES:

sage: QQbar(3 + 4*I).conjugate()
3 - 4*I
sage: QQbar.zeta(7).conjugate()
0.6234898018587335? - 0.7818314824680299?*I
sage: QQbar.zeta(7) + QQbar.zeta(7).conjugate()
1.246979603717467? + 0.?e-18*I

imag()#

Return the imaginary part of self.

EXAMPLES:

sage: QQbar.zeta(7).imag()
0.7818314824680299?

interval_exact(field)#

Given a ComplexIntervalField, compute the best possible approximation of this number in that field. Note that if either the real or imaginary parts of this number are sufficiently close to some floating-point number (and, in particular, if either is exactly representable in floating-point), then this will trigger exact computation, which may be very slow.

EXAMPLES:

sage: a = QQbar(I).sqrt(); a
0.7071067811865475? + 0.7071067811865475?*I
sage: a.interval_exact(CIF)
0.7071067811865475? + 0.7071067811865475?*I
sage: b = QQbar((1+I)*sqrt(2)/2)                                            # optional - sage.symbolic
sage: (a - b).interval(CIF)                                                 # optional - sage.symbolic
0.?e-19 + 0.?e-18*I
sage: (a - b).interval_exact(CIF)                                           # optional - sage.symbolic
0

multiplicative_order()#

Compute the multiplicative order of this algebraic number.

That is, find the smallest positive integer $$n$$ such that $$x^n = 1$$. If there is no such $$n$$, returns +Infinity.

We first check that abs(x) is very close to 1. If so, we compute $$x$$ exactly and examine its argument.

EXAMPLES:

sage: QQbar(-sqrt(3)/2 - I/2).multiplicative_order()                        # optional - sage.symbolic
12
sage: QQbar(1).multiplicative_order()
1
sage: QQbar(-I).multiplicative_order()
4
sage: QQbar(707/1000 + 707/1000*I).multiplicative_order()
+Infinity
sage: QQbar(3/5 + 4/5*I).multiplicative_order()
+Infinity

norm()#

Return self * self.conjugate().

This is the algebraic definition of norm, if we view QQbar as AA[I].

EXAMPLES:

sage: QQbar(3 + 4*I).norm()
25
sage: type(QQbar(I).norm())
<class 'sage.rings.qqbar.AlgebraicReal'>
sage: QQbar.zeta(1007).norm()
1.000000000000000?

rational_argument()#

Return the argument of self, divided by $$2\pi$$, as long as this result is rational. Otherwise returns None. Always triggers exact computation.

EXAMPLES:

sage: QQbar((1+I)*(sqrt(2)+sqrt(5))).rational_argument()                    # optional - sage.symbolic
1/8
sage: QQbar(-1 + I*sqrt(3)).rational_argument()                             # optional - sage.symbolic
1/3
sage: QQbar(-1 - I*sqrt(3)).rational_argument()                             # optional - sage.symbolic
-1/3
sage: QQbar(3+4*I).rational_argument() is None
True
sage: (QQbar(2)**(1/5) * QQbar.zeta(7)**2).rational_argument()  # long time
2/7
sage: (QQbar.zeta(73)**5).rational_argument()
5/73
sage: (QQbar.zeta(3)^65536).rational_argument()
1/3

real()#

Return the real part of self.

EXAMPLES:

sage: QQbar.zeta(5).real()
0.3090169943749474?

class sage.rings.qqbar.AlgebraicNumberPowQQAction(G, S)#

Bases: Action

Implement powering of an algebraic number (an element of QQbar or AA) by a rational.

This is always a right action.

INPUT:

• G – must be QQ

• S – the parent on which to act, either AA or QQbar.

Note

To compute x ^ (a/b), we take the $$b$$’th root of $$x$$; then we take that to the $$a$$’th power. If $$x$$ is a negative algebraic real and $$b$$ is odd, take the real $$b$$’th root; otherwise take the principal $$b$$’th root.

EXAMPLES:

In QQbar:

sage: QQbar(2)^(1/2)
1.414213562373095?
sage: QQbar(8)^(2/3)
4
sage: QQbar(8)^(2/3) == 4
True
sage: x = polygen(QQbar)
sage: phi = QQbar.polynomial_root(x^2 - x - 1, RIF(1, 2))
sage: tau = QQbar.polynomial_root(x^2 - x - 1, RIF(-1, 0))
sage: rt5 = QQbar(5)^(1/2)
sage: phi^10 / rt5
55.00363612324742?
sage: tau^10 / rt5
0.003636123247413266?
sage: (phi^10 - tau^10) / rt5
55.00000000000000?
sage: (phi^10 - tau^10) / rt5 == fibonacci(10)
True
sage: (phi^50 - tau^50) / rt5 == fibonacci(50)
True
sage: QQbar(-8)^(1/3)
1.000000000000000? + 1.732050807568878?*I
sage: (QQbar(-8)^(1/3))^3
-8
sage: QQbar(32)^(1/5)
2
sage: a = QQbar.zeta(7)^(1/3); a
0.9555728057861407? + 0.2947551744109043?*I
sage: a == QQbar.zeta(21)
True
sage: QQbar.zeta(7)^6
0.6234898018587335? - 0.7818314824680299?*I
sage: (QQbar.zeta(7)^6)^(1/3) * QQbar.zeta(21)
1.000000000000000? + 0.?e-17*I


In AA:

sage: AA(2)^(1/2)
1.414213562373095?
sage: AA(8)^(2/3)
4
sage: AA(8)^(2/3) == 4
True
sage: x = polygen(AA)
sage: phi = AA.polynomial_root(x^2 - x - 1, RIF(0, 2))
sage: tau = AA.polynomial_root(x^2 - x - 1, RIF(-2, 0))
sage: rt5 = AA(5)^(1/2)
sage: phi^10 / rt5
55.00363612324742?
sage: tau^10 / rt5
0.003636123247413266?
sage: (phi^10 - tau^10) / rt5
55.00000000000000?
sage: (phi^10 - tau^10) / rt5 == fibonacci(10)
True
sage: (phi^50 - tau^50) / rt5 == fibonacci(50)
True

class sage.rings.qqbar.AlgebraicNumber_base(parent, x)#

Bases: FieldElement

This is the common base class for algebraic numbers (complex numbers which are the zero of a polynomial in $$\ZZ[x]$$) and algebraic reals (algebraic numbers which happen to be real).

AlgebraicNumber objects can be created using QQbar (== AlgebraicNumberField()), and AlgebraicReal objects can be created using AA (== AlgebraicRealField()). They can be created either by coercing a rational or a symbolic expression, or by using the QQbar.polynomial_root() or AA.polynomial_root() method to construct a particular root of a polynomial with algebraic coefficients. Also, AlgebraicNumber and AlgebraicReal are closed under addition, subtraction, multiplication, division (except by 0), and rational powers (including roots), except that for a negative AlgebraicReal, taking a power with an even denominator returns an AlgebraicNumber instead of an AlgebraicReal.

AlgebraicNumber and AlgebraicReal objects can be approximated to any desired precision. They can be compared exactly; if the two numbers are very close, or are equal, this may require exact computation, which can be extremely slow.

As long as exact computation is not triggered, computation with algebraic numbers should not be too much slower than computation with intervals. As mentioned above, exact computation is triggered when comparing two algebraic numbers which are very close together. This can be an explicit comparison in user code, but the following list of actions (not necessarily complete) can also trigger exact computation:

• Dividing by an algebraic number which is very close to 0.

• Using an algebraic number which is very close to 0 as the leading coefficient in a polynomial.

• Taking a root of an algebraic number which is very close to 0.

The exact definition of “very close” is subject to change; currently, we compute our best approximation of the two numbers using 128-bit arithmetic, and see if that’s sufficient to decide the comparison. Note that comparing two algebraic numbers which are actually equal will always trigger exact computation, unless they are actually the same object.

EXAMPLES:

sage: sqrt(QQbar(2))
1.414213562373095?
sage: sqrt(QQbar(2))^2 == 2
True
sage: x = polygen(QQbar)
sage: phi = QQbar.polynomial_root(x^2 - x - 1, RIF(1, 2))
sage: phi
1.618033988749895?
sage: phi^2 == phi+1
True
sage: AA(sqrt(65537))
256.0019531175495?

as_number_field_element(minimal=False, embedded=False, prec=53)#

Return a number field containing this value, a representation of this value as an element of that number field, and a homomorphism from the number field back to AA or QQbar.

INPUT:

• minimal – Boolean (default: False). Whether to minimize the degree of the extension.

• embedded – Boolean (default: False). Whether to make the NumberField embedded.

• prec – integer (default: 53). The number of bit of precision to guarantee finding real roots.

This may not return the smallest such number field, unless minimal=True is specified.

To compute a single number field containing multiple algebraic numbers, use the function number_field_elements_from_algebraics instead.

EXAMPLES:

sage: QQbar(sqrt(8)).as_number_field_element()
(Number Field in a with defining polynomial y^2 - 2, 2*a, Ring morphism:
From: Number Field in a with defining polynomial y^2 - 2
To:   Algebraic Real Field
Defn: a |--> 1.414213562373095?)

sage: x = polygen(ZZ)
sage: p = x^3 + x^2 + x + 17
sage: (rt,) = p.roots(ring=AA, multiplicities=False); rt
-2.804642726932742?

sage: (nf, elt, hom) = rt.as_number_field_element()
sage: nf, elt, hom
(Number Field in a with defining polynomial y^3 - 2*y^2 - 31*y - 50,
a^2 - 5*a - 19,
Ring morphism:
From: Number Field in a with defining polynomial y^3 - 2*y^2 - 31*y - 50
To:   Algebraic Real Field
Defn: a |--> 7.237653139801104?)
sage: elt == rt
False
sage: AA(elt)
Traceback (most recent call last):
...
ValueError: need a real or complex embedding to convert a non rational element of a number field into an algebraic number
sage: hom(elt) == rt
True


Creating an element of an embedded number field:

sage: (nf, elt, hom) = rt.as_number_field_element(embedded=True)
sage: nf.coerce_embedding()
Generic morphism:
From: Number Field in a with defining polynomial y^3 - 2*y^2 - 31*y - 50 with a = 7.237653139801104?
To:   Algebraic Real Field
Defn: a -> 7.237653139801104?
sage: elt
a^2 - 5*a - 19
sage: elt.parent() == nf
True
sage: hom(elt).parent()
Algebraic Real Field
sage: hom(elt) == rt
True
sage: elt == rt
True
sage: AA(elt)
-2.804642726932742?
sage: RR(elt)
-2.80464272693274


A complex algebraic number as an element of an embedded number field:

sage: num = QQbar(sqrt(2) + 3^(1/3)*I)
sage: nf, elt, hom = num.as_number_field_element(embedded=True)
sage: hom(elt).parent() is QQbar
True
sage: nf.coerce_embedding() is not None
True
sage: QQbar(elt) == num == hom(elt)
True


We see an example where we do not get the minimal number field unless we specify minimal=True:

sage: rt2 = AA(sqrt(2))
sage: rt3 = AA(sqrt(3))
sage: rt3b = rt2 + rt3 - rt2
sage: rt3b.as_number_field_element()
(Number Field in a with defining polynomial y^4 - 4*y^2 + 1, a^2 - 2, Ring morphism:
From: Number Field in a with defining polynomial y^4 - 4*y^2 + 1
To:   Algebraic Real Field
Defn: a |--> -1.931851652578137?)
sage: rt3b.as_number_field_element(minimal=True)
(Number Field in a with defining polynomial y^2 - 3, a, Ring morphism:
From: Number Field in a with defining polynomial y^2 - 3
To:   Algebraic Real Field
Defn: a |--> 1.732050807568878?)

degree()#

Return the degree of this algebraic number (the degree of its minimal polynomial, or equivalently, the degree of the smallest algebraic extension of the rationals containing this number).

EXAMPLES:

sage: QQbar(5/3).degree()
1
sage: sqrt(QQbar(2)).degree()
2
sage: QQbar(17).nth_root(5).degree()
5
sage: sqrt(3+sqrt(QQbar(8))).degree()
2

exactify()#

Compute an exact representation for this number.

EXAMPLES:

sage: two = QQbar(4).nth_root(4)^2
sage: two
2.000000000000000?
sage: two.exactify()
sage: two
2

interval(field)#

Given an interval (or ball) field (real or complex, as appropriate) of precision $$p$$, compute an interval representation of self with diameter() at most $$2^{-p}$$; then round that representation into the given field. Here diameter() is relative diameter for intervals not containing 0, and absolute diameter for intervals that do contain 0; thus, if the returned interval does not contain 0, it has at least $$p-1$$ good bits.

EXAMPLES:

sage: RIF64 = RealIntervalField(64)
sage: x = AA(2).sqrt()
sage: y = x*x
sage: y = 1000 * y - 999 * y
sage: y.interval_fast(RIF64)
2.000000000000000?
sage: y.interval(RIF64)
2.000000000000000000?
sage: CIF64 = ComplexIntervalField(64)
sage: x = QQbar.zeta(11)
sage: x.interval_fast(CIF64)
0.8412535328311811689? + 0.5406408174555975821?*I
sage: x.interval(CIF64)
0.8412535328311811689? + 0.5406408174555975822?*I
sage: x.interval(CBF) # abs tol 1e-16
[0.8412535328311812 +/- 3.12e-17] + [0.5406408174555976 +/- 1.79e-17]*I


The following implicitly use this method:

sage: RIF(AA(5).sqrt())
2.236067977499790?
sage: AA(-5).sqrt().interval(RIF)
Traceback (most recent call last):
...
TypeError: unable to convert 2.236067977499790?*I to real interval

interval_diameter(diam)#

Compute an interval representation of self with diameter() at most diam. The precision of the returned value is unpredictable.

EXAMPLES:

sage: AA(2).sqrt().interval_diameter(1e-10)
1.4142135623730950488?
sage: AA(2).sqrt().interval_diameter(1e-30)
1.41421356237309504880168872420969807857?
sage: QQbar(2).sqrt().interval_diameter(1e-10)
1.4142135623730950488?
sage: QQbar(2).sqrt().interval_diameter(1e-30)
1.41421356237309504880168872420969807857?

interval_fast(field)#

Given a RealIntervalField or ComplexIntervalField, compute the value of this number using interval arithmetic of at least the precision of the field, and return the value in that field. (More precision may be used in the computation.) The returned interval may be arbitrarily imprecise, if this number is the result of a sufficiently long computation chain.

EXAMPLES:

sage: x = AA(2).sqrt()
sage: x.interval_fast(RIF)
1.414213562373095?
sage: x.interval_fast(RealIntervalField(200))
1.414213562373095048801688724209698078569671875376948073176680?
sage: x = QQbar(I).sqrt()
sage: x.interval_fast(CIF)
0.7071067811865475? + 0.7071067811865475?*I
sage: x.interval_fast(RIF)
Traceback (most recent call last):
...
TypeError: unable to convert complex interval 0.7071067811865475244? + 0.7071067811865475244?*I to real interval

is_integer()#

Return True if this number is a integer.

EXAMPLES:

sage: QQbar(2).is_integer()
True
sage: QQbar(1/2).is_integer()
False

is_square()#

Return whether or not this number is square.

OUTPUT:

(boolean) True in all cases for elements of QQbar; True for non-negative elements of AA; otherwise False

EXAMPLES:

sage: AA(2).is_square()
True
sage: AA(-2).is_square()
False
sage: QQbar(-2).is_square()
True
sage: QQbar(I).is_square()
True

minpoly()#

Compute the minimal polynomial of this algebraic number. The minimal polynomial is the monic polynomial of least degree having this number as a root; it is unique.

EXAMPLES:

sage: QQbar(4).sqrt().minpoly()
x - 2
sage: ((QQbar(2).nth_root(4))^2).minpoly()
x^2 - 2
sage: v = sqrt(QQbar(2)) + sqrt(QQbar(3)); v
3.146264369941973?
sage: p = v.minpoly(); p
x^4 - 10*x^2 + 1
sage: p(RR(v.real()))
1.31006316905768e-14

nth_root(n, all=False)#

Return the n-th root of this number.

INPUT:

• all - bool (default: False). If True, return a list of all $$n$$-th roots as complex algebraic numbers.

Warning

Note that for odd $$n$$, all=False and negative real numbers, AlgebraicReal and AlgebraicNumber values give different answers: AlgebraicReal values prefer real results, and AlgebraicNumber values return the principal root.

EXAMPLES:

sage: AA(-8).nth_root(3)
-2
sage: QQbar(-8).nth_root(3)
1.000000000000000? + 1.732050807568878?*I
sage: QQbar.zeta(12).nth_root(15)
0.9993908270190957? + 0.03489949670250097?*I


You can get all n-th roots of algebraic numbers:

sage: AA(-8).nth_root(3, all=True)
[1.000000000000000? + 1.732050807568878?*I,
-2.000000000000000? + 0.?e-18*I,
1.000000000000000? - 1.732050807568878?*I]

sage: QQbar(1+I).nth_root(4, all=True)
[1.069553932363986? + 0.2127475047267431?*I,
-0.2127475047267431? + 1.069553932363986?*I,
-1.069553932363986? - 0.2127475047267431?*I,
0.2127475047267431? - 1.069553932363986?*I]


Attempt to obtain a symbolic expression using radicals. If no exact symbolic expression can be found, the algebraic number will be returned without modification.

EXAMPLES:

sage: AA(1/sqrt(5)).radical_expression()                                    # optional - sage.symbolic
sqrt(1/5)
sage: AA(sqrt(5 + sqrt(5))).radical_expression()                            # optional - sage.symbolic
sqrt(sqrt(5) + 5)
sage: QQbar.zeta(5).radical_expression()                                    # optional - sage.symbolic
1/4*sqrt(5) + 1/2*sqrt(-1/2*sqrt(5) - 5/2) - 1/4
sage: x = polygen(QQ, 'x')
sage: a = (x^7 - x - 1).roots(AA, False)
sage: a.radical_expression()                                                # optional - sage.symbolic
1.112775684278706?
sage: a.radical_expression().parent() == SR                                 # optional - sage.symbolic
False
sage: a = sorted((x^7-x-1).roots(QQbar, False), key=imag)
sage: a.radical_expression()                                                # optional - sage.symbolic
-0.3636235193291805? - 0.9525611952610331?*I
sage: QQbar.zeta(5).imag().radical_expression()                             # optional - sage.symbolic
1/2*sqrt(1/2*sqrt(5) + 5/2)
sage: AA(5/3).radical_expression()                                          # optional - sage.symbolic
5/3
sage: AA(5/3).radical_expression().parent() == SR                           # optional - sage.symbolic
True
sage: QQbar(0).radical_expression()                                         # optional - sage.symbolic
0

simplify()#

Compute an exact representation for this number, in the smallest possible number field.

EXAMPLES:

sage: rt2 = AA(sqrt(2))
sage: rt3 = AA(sqrt(3))
sage: rt2b = rt3 + rt2 - rt3
sage: rt2b.exactify()
sage: rt2b._exact_value()
a^3 - 3*a where a^4 - 4*a^2 + 1 = 0 and a in -0.5176380902050415?
sage: rt2b.simplify()
sage: rt2b._exact_value()
a where a^2 - 2 = 0 and a in 1.414213562373095?

sqrt(all=False, extend=True)#

Return the square root(s) of this number.

INPUT:

• extend - bool (default: True); ignored if self is in QQbar, or positive in AA. If self is negative in AA, do the following: if True, return a square root of self in QQbar, otherwise raise a ValueError.

• all - bool (default: False); if True, return a list of all square roots. If False, return just one square root, or raise an ValueError if self is a negative element of AA and extend=False.

OUTPUT:

Either the principal square root of self, or a list of its square roots (with the principal one first).

EXAMPLES:

sage: AA(2).sqrt()
1.414213562373095?

sage: QQbar(I).sqrt()
0.7071067811865475? + 0.7071067811865475?*I
sage: QQbar(I).sqrt(all=True)
[0.7071067811865475? + 0.7071067811865475?*I, -0.7071067811865475? - 0.7071067811865475?*I]

sage: a = QQbar(0)
sage: a.sqrt()
0
sage: a.sqrt(all=True)


sage: a = AA(0)
sage: a.sqrt()
0
sage: a.sqrt(all=True)



This second example just shows that the program does not care where 0 is defined, it gives the same answer regardless. After all, how many ways can you square-root zero?

sage: AA(-2).sqrt()
1.414213562373095?*I

sage: AA(-2).sqrt(all=True)
[1.414213562373095?*I, -1.414213562373095?*I]

sage: AA(-2).sqrt(extend=False)
Traceback (most recent call last):
...
ValueError: -2 is not a square in AA, being negative. Use extend = True for a square root in QQbar.

class sage.rings.qqbar.AlgebraicPolynomialTracker(poly)#

Bases: SageObject

Keeps track of a polynomial used for algebraic numbers.

If multiple algebraic numbers are created as roots of a single polynomial, this allows the polynomial and information about the polynomial to be shared. This reduces work if the polynomial must be recomputed at higher precision, or if it must be factored.

This class is private, and should only be constructed by AA.common_polynomial() or QQbar.common_polynomial(), and should only be used as an argument to AA.polynomial_root() or QQbar.polynomial_root(). (It does not matter whether you create the common polynomial with AA.common_polynomial() or QQbar.common_polynomial().)

EXAMPLES:

sage: x = polygen(QQbar)
sage: P = QQbar.common_polynomial(x^2 - x - 1)
sage: P
x^2 - x - 1
sage: QQbar.polynomial_root(P, RIF(1, 2))
1.618033988749895?

complex_roots(prec, multiplicity)#

Find the roots of self in the complex field to precision prec.

EXAMPLES:

sage: x = polygen(ZZ)
sage: cp = AA.common_polynomial(x^4 - 2)


Note that the precision is not guaranteed to find the tightest possible interval since complex_roots() depends on the underlying BLAS implementation.

sage: cp.complex_roots(30, 1)
[-1.18920711500272...?,
1.189207115002721?,
-1.189207115002721?*I,
1.189207115002721?*I]

exactify()#

Compute a common field that holds all of the algebraic coefficients of this polynomial, then factor the polynomial over that field. Store the factors for later use (ignoring multiplicity).

EXAMPLES:

sage: x = polygen(AA)
sage: p = sqrt(AA(2)) * x^2 - sqrt(AA(3))
sage: cp = AA.common_polynomial(p)
sage: cp._exact
False
sage: cp.exactify()
sage: cp._exact
True

factors()#

EXAMPLES:

sage: x = polygen(QQ)
sage: f = QQbar.common_polynomial(x^4 + 4)
sage: f.factors()
[y^2 - 2*y + 2, y^2 + 2*y + 2]

generator()#

Return an AlgebraicGenerator for a number field containing all the coefficients of self.

EXAMPLES:

sage: x = polygen(AA)
sage: p = sqrt(AA(2)) * x^2 - sqrt(AA(3))
sage: cp = AA.common_polynomial(p)
sage: cp.generator()
Number Field in a with defining polynomial y^4 - 4*y^2 + 1
with a in -0.5176380902050415?

is_complex()#

Return True if the coefficients of this polynomial are non-real.

EXAMPLES:

sage: x = polygen(QQ); f = x^3 - 7
sage: g = AA.common_polynomial(f)
sage: g.is_complex()
False
sage: QQbar.common_polynomial(x^3 - QQbar(I)).is_complex()
True

poly()#

Return the underlying polynomial of self.

EXAMPLES:

sage: x = polygen(QQ)
sage: f = x^3 - 7
sage: g = AA.common_polynomial(f)
sage: g.poly()
y^3 - 7

class sage.rings.qqbar.AlgebraicReal(x)#

A real algebraic number.

_richcmp_(other, op)#

Compare two algebraic reals.

EXAMPLES:

sage: AA(2).sqrt() < AA(3).sqrt()
True
sage: ((5+AA(5).sqrt())/2).sqrt() == 2*QQbar.zeta(5).imag()
True
sage: AA(3).sqrt() + AA(2).sqrt() < 3
False

ceil()#

Return the smallest integer not smaller than self.

EXAMPLES:

sage: AA(sqrt(2)).ceil()                                                    # optional - sage.symbolic
2
sage: AA(-sqrt(2)).ceil()                                                   # optional - sage.symbolic
-1
sage: AA(42).ceil()
42

conjugate()#

Return the complex conjugate of self, i.e. returns itself.

EXAMPLES:

sage: a = AA(sqrt(2) + sqrt(3))                                             # optional - sage.symbolic
sage: a.conjugate()                                                         # optional - sage.symbolic
3.146264369941973?
sage: a.conjugate() is a                                                    # optional - sage.symbolic
True

floor()#

Return the largest integer not greater than self.

EXAMPLES:

sage: AA(sqrt(2)).floor()                                                   # optional - sage.symbolic
1
sage: AA(-sqrt(2)).floor()                                                  # optional - sage.symbolic
-2
sage: AA(42).floor()
42

imag()#

Return the imaginary part of this algebraic real.

It always returns 0.

EXAMPLES:

sage: a = AA(sqrt(2) + sqrt(3))                                             # optional - sage.symbolic
sage: a.imag()                                                              # optional - sage.symbolic
0
sage: parent(a.imag())                                                      # optional - sage.symbolic
Algebraic Real Field

interval_exact(field)#

Given a RealIntervalField, compute the best possible approximation of this number in that field. Note that if this number is sufficiently close to some floating-point number (and, in particular, if this number is exactly representable in floating-point), then this will trigger exact computation, which may be very slow.

EXAMPLES:

sage: x = AA(2).sqrt()
sage: y = x*x
sage: x.interval(RIF)
1.414213562373095?
sage: x.interval_exact(RIF)
1.414213562373095?
sage: y.interval(RIF)
2.000000000000000?
sage: y.interval_exact(RIF)
2
sage: z = 1 + AA(2).sqrt() / 2^200
sage: z.interval(RIF)
1.000000000000001?
sage: z.interval_exact(RIF)
1.000000000000001?

multiplicative_order()#

Compute the multiplicative order of this real algebraic number.

That is, find the smallest positive integer $$n$$ such that $$x^n = 1$$. If there is no such $$n$$, returns +Infinity.

We first check that abs(x) is very close to 1. If so, we compute $$x$$ exactly and compare it to $$1$$ and $$-1$$.

EXAMPLES:

sage: AA(1).multiplicative_order()
1
sage: AA(-1).multiplicative_order()
2
sage: AA(5).sqrt().multiplicative_order()
+Infinity

real()#

Return the real part of this algebraic real.

It always returns self.

EXAMPLES:

sage: a = AA(sqrt(2) + sqrt(3))                                             # optional - sage.symbolic
sage: a.real()                                                              # optional - sage.symbolic
3.146264369941973?
sage: a.real() is a                                                         # optional - sage.symbolic
True

real_exact(field)#

Given a RealField, compute the best possible approximation of this number in that field. Note that if this number is sufficiently close to the halfway point between two floating-point numbers in the field (for the default round-to-nearest mode) or if the number is sufficiently close to a floating-point number in the field (for directed rounding modes), then this will trigger exact computation, which may be very slow.

The rounding mode of the field is respected.

EXAMPLES:

sage: x = AA(2).sqrt()^2
sage: x.real_exact(RR)
2.00000000000000
sage: x.real_exact(RealField(53, rnd='RNDD'))
2.00000000000000
sage: x.real_exact(RealField(53, rnd='RNDU'))
2.00000000000000
sage: x.real_exact(RealField(53, rnd='RNDZ'))
2.00000000000000
sage: (-x).real_exact(RR)
-2.00000000000000
sage: (-x).real_exact(RealField(53, rnd='RNDD'))
-2.00000000000000
sage: (-x).real_exact(RealField(53, rnd='RNDU'))
-2.00000000000000
sage: (-x).real_exact(RealField(53, rnd='RNDZ'))
-2.00000000000000
sage: y = (x-2).real_exact(RR).abs()
sage: y == 0.0 or y == -0.0 # the sign of 0.0 is not significant in MPFI
True
sage: y = (x-2).real_exact(RealField(53, rnd='RNDD'))
sage: y == 0.0 or y == -0.0 # same as above
True
sage: y = (x-2).real_exact(RealField(53, rnd='RNDU'))
sage: y == 0.0 or y == -0.0 # idem
True
sage: y = (x-2).real_exact(RealField(53, rnd='RNDZ'))
sage: y == 0.0 or y == -0.0 # ibidem
True
sage: y = AA(2).sqrt()
sage: y.real_exact(RR)
1.41421356237310
sage: y.real_exact(RealField(53, rnd='RNDD'))
1.41421356237309
sage: y.real_exact(RealField(53, rnd='RNDU'))
1.41421356237310
sage: y.real_exact(RealField(53, rnd='RNDZ'))
1.41421356237309

real_number(field)#

Given a RealField, compute a good approximation to self in that field. The approximation will be off by at most two ulp’s, except for numbers which are very close to 0, which will have an absolute error at most 2**(-(field.prec()-1)). Also, the rounding mode of the field is respected.

EXAMPLES:

sage: x = AA(2).sqrt()^2
sage: x.real_number(RR)
2.00000000000000
sage: x.real_number(RealField(53, rnd='RNDD'))
1.99999999999999
sage: x.real_number(RealField(53, rnd='RNDU'))
2.00000000000001
sage: x.real_number(RealField(53, rnd='RNDZ'))
1.99999999999999
sage: (-x).real_number(RR)
-2.00000000000000
sage: (-x).real_number(RealField(53, rnd='RNDD'))
-2.00000000000001
sage: (-x).real_number(RealField(53, rnd='RNDU'))
-1.99999999999999
sage: (-x).real_number(RealField(53, rnd='RNDZ'))
-1.99999999999999
sage: (x-2).real_number(RR)
5.42101086242752e-20
sage: (x-2).real_number(RealField(53, rnd='RNDD'))
-1.08420217248551e-19
sage: (x-2).real_number(RealField(53, rnd='RNDU'))
2.16840434497101e-19
sage: (x-2).real_number(RealField(53, rnd='RNDZ'))
0.000000000000000
sage: y = AA(2).sqrt()
sage: y.real_number(RR)
1.41421356237309
sage: y.real_number(RealField(53, rnd='RNDD'))
1.41421356237309
sage: y.real_number(RealField(53, rnd='RNDU'))
1.41421356237310
sage: y.real_number(RealField(53, rnd='RNDZ'))
1.41421356237309

round()#

Round self to the nearest integer.

EXAMPLES:

sage: AA(sqrt(2)).round()                                                   # optional - sage.symbolic
1
sage: AA(1/2).round()
1
sage: AA(-1/2).round()
-1

sign()#

Compute the sign of this algebraic number (return $$-1$$ if negative, $$0$$ if zero, or $$1$$ if positive).

This computes an interval enclosing this number using 128-bit interval arithmetic; if this interval includes 0, then fall back to exact computation (which can be very slow).

EXAMPLES:

sage: AA(-5).nth_root(7).sign()
-1
sage: (AA(2).sqrt() - AA(2).sqrt()).sign()
0

sage: a = AA(2).sqrt() + AA(3).sqrt() - 58114382797550084497/18470915334626475921
sage: a.sign()
1
sage: b = AA(2).sqrt() + AA(3).sqrt() - 2602510228533039296408/827174681630786895911
sage: b.sign()
-1

sage: c = AA(5)**(1/3) - 1437624125539676934786/840727688792155114277
sage: c.sign()
1

sage: (((a+b)*(a+c)*(b+c))**9 / (a*b*c)).sign()
1
sage: (a-b).sign()
1
sage: (b-a).sign()
-1
sage: (a*b).sign()
-1
sage: ((a*b).abs() + a).sign()
1
sage: (a*b - b*a).sign()
0

sage: a = AA(sqrt(1/2))                                                     # optional - sage.symbolic
sage: b = AA(-sqrt(1/2))                                                    # optional - sage.symbolic
sage: (a + b).sign()                                                        # optional - sage.symbolic
0

trunc()#

Round self to the nearest integer toward zero.

EXAMPLES:

sage: AA(sqrt(2)).trunc()                                                   # optional - sage.symbolic
1
sage: AA(-sqrt(2)).trunc()                                                  # optional - sage.symbolic
-1
sage: AA(1).trunc()
1
sage: AA(-1).trunc()
-1

class sage.rings.qqbar.AlgebraicRealField#

The field of algebraic reals.

algebraic_closure()#

Return the algebraic closure of this field, which is the field $$\overline{\QQ}$$ of algebraic numbers.

EXAMPLES:

sage: AA.algebraic_closure()
Algebraic Field

completion(p, prec, extras={})#

Return the completion of self at the place $$p$$.

Only implemented for $$p = \infty$$ at present.

INPUT:

• p – either a prime (not implemented at present) or Infinity

• prec – precision of approximate field to return

• extras – (optional) a dict of extra keyword arguments for the RealField constructor

EXAMPLES:

sage: AA.completion(infinity, 500)
Real Field with 500 bits of precision
sage: AA.completion(infinity, prec=53, extras={'type':'RDF'})
Real Double Field
sage: AA.completion(infinity, 53) is RR
True
sage: AA.completion(7, 10)
Traceback (most recent call last):
...
NotImplementedError

gen(n=0)#

Return the $$n$$-th element of the tuple returned by gens().

EXAMPLES:

sage: AA.gen(0)
1
sage: AA.gen(1)
Traceback (most recent call last):
...
IndexError: n must be 0

gens()#

Return a set of generators for this field.

As this field is not finitely generated, we opt for just returning 1.

EXAMPLES:

sage: AA.gens()
(1,)

ngens()#

Return the size of the tuple returned by gens().

EXAMPLES:

sage: AA.ngens()
1

polynomial_root(poly, interval, multiplicity=1)#

Given a polynomial with algebraic coefficients and an interval enclosing exactly one root of the polynomial, constructs an algebraic real representation of that root.

The polynomial need not be irreducible, or even squarefree; but if the given root is a multiple root, its multiplicity must be specified. (IMPORTANT NOTE: Currently, multiplicity-$$k$$ roots are handled by taking the $$(k-1)$$-st derivative of the polynomial. This means that the interval must enclose exactly one root of this derivative.)

The conditions on the arguments (that the interval encloses exactly one root, and that multiple roots match the given multiplicity) are not checked; if they are not satisfied, an error may be thrown (possibly later, when the algebraic number is used), or wrong answers may result.

Note that if you are constructing multiple roots of a single polynomial, it is better to use AA.common_polynomial (or QQbar.common_polynomial; the two are equivalent) to get a shared polynomial.

EXAMPLES:

sage: x = polygen(AA)
sage: phi = AA.polynomial_root(x^2 - x - 1, RIF(1, 2)); phi
1.618033988749895?
sage: p = (x-1)^7 * (x-2)
sage: r = AA.polynomial_root(p, RIF(9/10, 11/10), multiplicity=7)
sage: r; r == 1
1.000000000000000?
True
sage: p = (x-phi)*(x-sqrt(AA(2)))
sage: r = AA.polynomial_root(p, RIF(1, 3/2))
sage: r; r == sqrt(AA(2))
1.414213562373095?
True


We allow complex polynomials, as long as the particular root in question is real.

sage: K.<im> = QQ[I]
sage: x = polygen(K)
sage: p = (im + 1) * (x^3 - 2); p
(I + 1)*x^3 - 2*I - 2
sage: r = AA.polynomial_root(p, RIF(1, 2)); r^3
2.000000000000000?

random_element(poly_degree=2, *args, **kwds)#

Return a random algebraic real number.

INPUT:

• poly_degree - default: 2 - degree of the random polynomial over the integers of which the returned algebraic real number is a (real part of a) root. This is not necessarily the degree of the minimal polynomial of the number. Increase this parameter to achieve a greater diversity of algebraic numbers, at a cost of greater computation time. You can also vary the distribution of the coefficients but that will not vary the degree of the extension containing the element.

• args, kwds - arguments and keywords passed to the random number generator for elements of ZZ, the integers. See random_element() for details, or see example below.

OUTPUT:

An element of AA, the field of algebraic real numbers (see sage.rings.qqbar).

ALGORITHM:

We pass all arguments to AlgebraicField.random_element(), and then take the real part of the result.

EXAMPLES:

sage: a = AA.random_element()
sage: a in AA
True

sage: b = AA.random_element(poly_degree=5)
sage: b in AA
True


Parameters for the distribution of the integer coefficients of the polynomials can be passed on to the random element method for integers. For example, we can rule out zero as a coefficient (and therefore as a root) by requesting coefficients between 1 and 10:

sage: z = [AA.random_element(x=1, y=10) for _ in range(5)]
sage: AA(0) in z
False

zeta(n=2)#

Return an $$n$$-th root of unity in this field. This will raise a ValueError if $$n \ne \{1, 2\}$$ since no such root exists.

INPUT:

• n (integer) – default 2

EXAMPLES:

sage: AA.zeta(1)
1
sage: AA.zeta(2)
-1
sage: AA.zeta()
-1
sage: AA.zeta(3)
Traceback (most recent call last):
...
ValueError: no n-th root of unity in algebraic reals


Some silly inputs:

sage: AA.zeta(Mod(-5, 7))
-1
sage: AA.zeta(0)
Traceback (most recent call last):
...
ValueError: no n-th root of unity in algebraic reals

sage.rings.qqbar.an_binop_element(a, b, op)#

Add, subtract, multiply or divide two elements represented as elements of number fields.

EXAMPLES:

sage: sqrt2 = QQbar(2).sqrt()
sage: sqrt3 = QQbar(3).sqrt()
sage: sqrt5 = QQbar(5).sqrt()

sage: a = sqrt2 + sqrt3; a.exactify()
sage: b = sqrt3 + sqrt5; b.exactify()
sage: type(a._descr)
<class 'sage.rings.qqbar.ANExtensionElement'>
sage: from sage.rings.qqbar import an_binop_element
<sage.rings.qqbar.ANBinaryExpr object at ...>
sage: an_binop_element(a, b, operator.sub)
<sage.rings.qqbar.ANBinaryExpr object at ...>
sage: an_binop_element(a, b, operator.mul)
<sage.rings.qqbar.ANBinaryExpr object at ...>
sage: an_binop_element(a, b, operator.truediv)
<sage.rings.qqbar.ANBinaryExpr object at ...>


The code tries to use existing unions of number fields:

sage: sqrt17 = QQbar(17).sqrt()
sage: sqrt19 = QQbar(19).sqrt()
sage: a = sqrt17 + sqrt19
sage: b = sqrt17 * sqrt19 - sqrt17 + sqrt19 * (sqrt17 + 2)
sage: b, type(b._descr)
(40.53909377268655?, <class 'sage.rings.qqbar.ANBinaryExpr'>)
sage: a.exactify()
sage: b = sqrt17 * sqrt19 - sqrt17 + sqrt19 * (sqrt17 + 2)
sage: b, type(b._descr)
(40.53909377268655?, <class 'sage.rings.qqbar.ANExtensionElement'>)

sage.rings.qqbar.an_binop_expr(a, b, op)#

Add, subtract, multiply or divide algebraic numbers represented as binary expressions.

INPUT:

• a, b – two elements

• op – an operator

EXAMPLES:

sage: a = QQbar(sqrt(2)) + QQbar(sqrt(3))                                       # optional - sage.symbolic
sage: b = QQbar(sqrt(3)) + QQbar(sqrt(5))                                       # optional - sage.symbolic
sage: type(a._descr); type(b._descr)                                            # optional - sage.symbolic
<class 'sage.rings.qqbar.ANBinaryExpr'>
<class 'sage.rings.qqbar.ANBinaryExpr'>
sage: from sage.rings.qqbar import an_binop_expr
sage: x = an_binop_expr(a, b, operator.add); x                                  # optional - sage.symbolic
<sage.rings.qqbar.ANBinaryExpr object at ...>
sage: x.exactify()                                                              # optional - sage.symbolic
6/7*a^7 - 2/7*a^6 - 71/7*a^5 + 26/7*a^4 + 125/7*a^3 - 72/7*a^2 - 43/7*a + 47/7
where a^8 - 12*a^6 + 23*a^4 - 12*a^2 + 1 = 0 and a in -0.3199179336182997?

sage: a = QQbar(sqrt(2)) + QQbar(sqrt(3))                                       # optional - sage.symbolic
sage: b = QQbar(sqrt(3)) + QQbar(sqrt(5))                                       # optional - sage.symbolic
sage: type(a._descr)                                                            # optional - sage.symbolic
<class 'sage.rings.qqbar.ANBinaryExpr'>
sage: x = an_binop_expr(a, b, operator.mul); x                                  # optional - sage.symbolic
<sage.rings.qqbar.ANBinaryExpr object at ...>
sage: x.exactify()                                                              # optional - sage.symbolic
2*a^7 - a^6 - 24*a^5 + 12*a^4 + 46*a^3 - 22*a^2 - 22*a + 9
where a^8 - 12*a^6 + 23*a^4 - 12*a^2 + 1 = 0 and a in -0.3199179336182997?

sage.rings.qqbar.an_binop_rational(a, b, op)#

Used to add, subtract, multiply or divide algebraic numbers.

Used when both are actually rational.

EXAMPLES:

sage: from sage.rings.qqbar import an_binop_rational
sage: f = an_binop_rational(QQbar(2), QQbar(3/7), operator.add)
sage: f
17/7
sage: type(f)
<class 'sage.rings.qqbar.ANRational'>

sage: f = an_binop_rational(QQbar(2), QQbar(3/7), operator.mul)
sage: f
6/7
sage: type(f)
<class 'sage.rings.qqbar.ANRational'>

sage.rings.qqbar.clear_denominators(poly)#

Take a monic polynomial and rescale the variable to get a monic polynomial with “integral” coefficients.

This works on any univariate polynomial whose base ring has a denominator() method that returns integers; for example, the base ring might be $$\QQ$$ or a number field.

Returns the scale factor and the new polynomial.

(Inspired by pari:primitive_pol_to_monic .)

We assume that coefficient denominators are “small”; the algorithm factors the denominators, to give the smallest possible scale factor.

EXAMPLES:

sage: from sage.rings.qqbar import clear_denominators

sage: _.<x> = QQ['x']
sage: clear_denominators(x + 3/2)
(2, x + 3)
sage: clear_denominators(x^2 + x/2 + 1/4)
(2, x^2 + x + 1)

sage.rings.qqbar.cmp_elements_with_same_minpoly(a, b, p)#

Compare the algebraic elements a and b knowing that they have the same minimal polynomial p.

This is an helper function for comparison of algebraic elements (i.e. the methods AlgebraicNumber._richcmp_() and AlgebraicReal._richcmp_()).

INPUT:

• a and b – elements of the algebraic or the real algebraic field with same minimal polynomial

• p – the minimal polynomial

OUTPUT:

$$-1$$, $$0$$, $$1$$, $$None$$ depending on whether $$a < b$$, $$a = b$$ or $$a > b$$ or the function did not succeed with the given precision of $$a$$ and $$b$$.

EXAMPLES:

sage: from sage.rings.qqbar import cmp_elements_with_same_minpoly
sage: x = polygen(ZZ)
sage: p = x^2 - 2
sage: a = AA.polynomial_root(p, RIF(1,2))
sage: b = AA.polynomial_root(p, RIF(-2,-1))
sage: cmp_elements_with_same_minpoly(a, b, p)
1
sage: cmp_elements_with_same_minpoly(-a, b, p)
0

sage.rings.qqbar.conjugate_expand(v)#

If the interval v (which may be real or complex) includes some purely real numbers, return v' containing v such that v' == v'.conjugate(). Otherwise return v unchanged. (Note that if v' == v'.conjugate(), and v' includes one non-real root of a real polynomial, then v' also includes the conjugate of that root. Also note that the diameter of the return value is at most twice the diameter of the input.)

EXAMPLES:

sage: from sage.rings.qqbar import conjugate_expand
sage: conjugate_expand(CIF(RIF(0, 1), RIF(1, 2))).str(style='brackets')
'[0.0000000000000000 .. 1.0000000000000000] + [1.0000000000000000 .. 2.0000000000000000]*I'
sage: conjugate_expand(CIF(RIF(0, 1), RIF(0, 1))).str(style='brackets')
'[0.0000000000000000 .. 1.0000000000000000] + [-1.0000000000000000 .. 1.0000000000000000]*I'
sage: conjugate_expand(CIF(RIF(0, 1), RIF(-2, 1))).str(style='brackets')
'[0.0000000000000000 .. 1.0000000000000000] + [-2.0000000000000000 .. 2.0000000000000000]*I'
sage: conjugate_expand(RIF(1, 2)).str(style='brackets')
'[1.0000000000000000 .. 2.0000000000000000]'

sage.rings.qqbar.conjugate_shrink(v)#

If the interval v includes some purely real numbers, return a real interval containing only those real numbers. Otherwise return v unchanged.

If v includes exactly one root of a real polynomial, and v was returned by conjugate_expand(), then conjugate_shrink(v) still includes that root, and is a RealIntervalFieldElement iff the root in question is real.

EXAMPLES:

sage: from sage.rings.qqbar import conjugate_shrink
sage: conjugate_shrink(RIF(3, 4)).str(style='brackets')
'[3.0000000000000000 .. 4.0000000000000000]'
sage: conjugate_shrink(CIF(RIF(1, 2), RIF(1, 2))).str(style='brackets')
'[1.0000000000000000 .. 2.0000000000000000] + [1.0000000000000000 .. 2.0000000000000000]*I'
sage: conjugate_shrink(CIF(RIF(1, 2), RIF(0, 1))).str(style='brackets')
'[1.0000000000000000 .. 2.0000000000000000]'
sage: conjugate_shrink(CIF(RIF(1, 2), RIF(-1, 2))).str(style='brackets')
'[1.0000000000000000 .. 2.0000000000000000]'

sage.rings.qqbar.do_polred(poly, threshold=32)#

Find a polynomial of reasonably small discriminant that generates the same number field as poly, using the PARI polredbest function.

INPUT:

• poly - a monic irreducible polynomial with integer coefficients

• threshold - an integer used to decide whether to run polredbest

OUTPUT:

A triple (elt_fwd, elt_back, new_poly), where:

• new_poly is the new polynomial defining the same number field,

• elt_fwd is a polynomial expression for a root of the new polynomial in terms of a root of the original polynomial,

• elt_back is a polynomial expression for a root of the original polynomial in terms of a root of the new polynomial.

EXAMPLES:

sage: from sage.rings.qqbar import do_polred
sage: R.<x> = QQ['x']
sage: oldpol = x^2 - 5
sage: fwd, back, newpol = do_polred(oldpol)
sage: newpol
x^2 - x - 1
sage: Kold.<a> = NumberField(oldpol)
sage: Knew.<b> = NumberField(newpol)
sage: newpol(fwd(a))
0
sage: oldpol(back(b))
0
sage: do_polred(x^2 - x - 11)
(1/3*x + 1/3, 3*x - 1, x^2 - x - 1)
sage: do_polred(x^3 + 123456)
(-1/4*x, -4*x, x^3 - 1929)


This shows that github issue #13054 has been fixed:

sage: do_polred(x^4 - 4294967296*x^2 + 54265257667816538374400)
(1/4*x, 4*x, x^4 - 268435456*x^2 + 211973662764908353025)

sage.rings.qqbar.find_zero_result(fn, l)#

l is a list of some sort. fn is a function which maps an element of l and a precision into an interval (either real or complex) of that precision, such that for sufficient precision, exactly one element of l results in an interval containing 0. Returns that one element of l.

EXAMPLES:

sage: from sage.rings.qqbar import find_zero_result
sage: _.<x> = QQ['x']
sage: delta = 10^(-70)
sage: p1 = x - 1
sage: p2 = x - 1 - delta
sage: p3 = x - 1 + delta
sage: p2 == find_zero_result(lambda p, prec: p(RealIntervalField(prec)(1 + delta)), [p1, p2, p3])
True

sage.rings.qqbar.get_AA_golden_ratio()#

Return the golden ratio as an element of the algebraic real field. Used by sage.symbolic.constants.golden_ratio._algebraic_().

EXAMPLES:

sage: AA(golden_ratio)  # indirect doctest                                      # optional - sage.symbolic
1.618033988749895?

sage.rings.qqbar.is_AlgebraicField(F)#

Check whether F is an AlgebraicField instance.

This function is deprecated. Use isinstance() with AlgebraicField instead.

EXAMPLES:

sage: from sage.rings.qqbar import is_AlgebraicField
sage: [is_AlgebraicField(x) for x in [AA, QQbar, None, 0, "spam"]]
doctest:warning...
DeprecationWarning: is_AlgebraicField is deprecated;
See https://github.com/sagemath/sage/issues/32660 for details.
[False, True, False, False, False]

sage.rings.qqbar.is_AlgebraicField_common(F)#

Check whether F is an AlgebraicField_common instance.

This function is deprecated. Use isinstance() with AlgebraicField_common instead.

EXAMPLES:

sage: from sage.rings.qqbar import is_AlgebraicField_common
sage: [is_AlgebraicField_common(x) for x in [AA, QQbar, None, 0, "spam"]]
doctest:warning...
DeprecationWarning: is_AlgebraicField_common is deprecated;
See https://github.com/sagemath/sage/issues/32610 for details.
[True, True, False, False, False]

sage.rings.qqbar.is_AlgebraicNumber(x)#

Test if x is an instance of AlgebraicNumber. For internal use.

EXAMPLES:

sage: from sage.rings.qqbar import is_AlgebraicNumber
sage: is_AlgebraicNumber(AA(sqrt(2)))
False
sage: is_AlgebraicNumber(QQbar(sqrt(2)))
True
sage: is_AlgebraicNumber("spam")
False

sage.rings.qqbar.is_AlgebraicReal(x)#

Test if x is an instance of AlgebraicReal. For internal use.

EXAMPLES:

sage: from sage.rings.qqbar import is_AlgebraicReal
sage: is_AlgebraicReal(AA(sqrt(2)))
True
sage: is_AlgebraicReal(QQbar(sqrt(2)))
False
sage: is_AlgebraicReal("spam")
False

sage.rings.qqbar.is_AlgebraicRealField(F)#

Check whether F is an AlgebraicRealField instance. For internal use.

This function is deprecated. Use isinstance() with AlgebraicRealField instead.

EXAMPLES:

sage: from sage.rings.qqbar import is_AlgebraicRealField
sage: [is_AlgebraicRealField(x) for x in [AA, QQbar, None, 0, "spam"]]
doctest:warning...
DeprecationWarning: is_AlgebraicRealField is deprecated;
See https://github.com/sagemath/sage/issues/32660 for details.
[True, False, False, False, False]

sage.rings.qqbar.isolating_interval(intv_fn, pol)#

intv_fn is a function that takes a precision and returns an interval of that precision containing some particular root of pol. (It must return better approximations as the precision increases.) pol is an irreducible polynomial with rational coefficients.

Returns an interval containing at most one root of pol.

EXAMPLES:

sage: from sage.rings.qqbar import isolating_interval

sage: _.<x> = QQ['x']
sage: isolating_interval(lambda prec: sqrt(RealIntervalField(prec)(2)), x^2 - 2)
1.4142135623730950488?


And an example that requires more precision:

sage: delta = 10^(-70)
sage: p = (x - 1) * (x - 1 - delta) * (x - 1 + delta)
sage: isolating_interval(lambda prec: RealIntervalField(prec)(1 + delta), p)
1.000000000000000000000000000000000000000000000000000000000000000000000100000000000000000000000000000000000000000000000000000000000000000000000000000000000?


The function also works with complex intervals and complex roots:

sage: p = x^2 - x + 13/36
sage: isolating_interval(lambda prec: ComplexIntervalField(prec)(1/2, 1/3), p)
0.500000000000000000000? + 0.3333333333333333334?*I

sage.rings.qqbar.number_field_elements_from_algebraics(numbers, minimal=False, same_field=False, embedded=False, prec=53)#

Given a sequence of elements of either AA or QQbar (or a mixture), computes a number field containing all of these elements, these elements as members of that number field, and a homomorphism from the number field back to AA or QQbar.

INPUT:

• numbers – a number or list of numbers.

• minimal – Boolean (default: False). Whether to minimize the degree of the extension.

• same_field – Boolean (default: False). See below.

• embedded – Boolean (default: False). Whether to make the NumberField embedded.

• prec – integer (default: 53). The number of bit of precision to guarantee finding real roots.

OUTPUT:

A tuple with the NumberField, the numbers inside the NumberField, and a homomorphism from the number field back to AA or QQbar.

This may not return the smallest such number field, unless minimal=True is specified.

If same_field=True is specified, and all of the elements are from the same field (either AA or QQbar), the generated homomorphism will map back to that field. Otherwise, if all specified elements are real, the homomorphism might map back to AA (and will, if minimal=True is specified), even if the elements were in QQbar.

Also, a single number can be passed, rather than a sequence; and any values which are not elements of AA or QQbar will automatically be coerced to QQbar

This function may be useful for efficiency reasons: doing exact computations in the corresponding number field will be faster than doing exact computations directly in AA or QQbar.

EXAMPLES:

We can use this to compute the splitting field of a polynomial. (Unfortunately this takes an unreasonably long time for non-toy examples.):

sage: x = polygen(QQ)
sage: p = x^3 + x^2 + x + 17
sage: rts = p.roots(ring=QQbar, multiplicities=False)
sage: splitting = number_field_elements_from_algebraics(rts); splitting
Number Field in a with defining polynomial y^6 - 40*y^4 - 22*y^3 + 873*y^2 + 1386*y + 594
sage: p.roots(ring=splitting)
[(361/29286*a^5 - 19/3254*a^4 - 14359/29286*a^3 + 401/29286*a^2 + 18183/1627*a + 15930/1627, 1), (49/117144*a^5 - 179/39048*a^4 - 3247/117144*a^3 + 22553/117144*a^2 + 1744/4881*a - 17195/6508, 1), (-1493/117144*a^5 + 407/39048*a^4 + 60683/117144*a^3 - 24157/117144*a^2 - 56293/4881*a - 53033/6508, 1)]
sage: rt2 = AA(sqrt(2)); rt2
1.414213562373095?
sage: rt3 = AA(sqrt(3)); rt3
1.732050807568878?
sage: rt3a = QQbar(sqrt(3)); rt3a
1.732050807568878?
sage: qqI = QQbar.zeta(4); qqI
I
sage: z3 = QQbar.zeta(3); z3
-0.500000000000000? + 0.866025403784439?*I
sage: rt2b = rt3 + rt2 - rt3; rt2b
1.414213562373095?
sage: rt2c = z3 + rt2 - z3; rt2c
1.414213562373095? + 0.?e-19*I

sage: number_field_elements_from_algebraics(rt2)
(Number Field in a with defining polynomial y^2 - 2, a, Ring morphism:
From: Number Field in a with defining polynomial y^2 - 2
To:   Algebraic Real Field
Defn: a |--> 1.414213562373095?)

sage: number_field_elements_from_algebraics((rt2,rt3))
(Number Field in a with defining polynomial y^4 - 4*y^2 + 1, [-a^3 + 3*a, a^2 - 2], Ring morphism:
From: Number Field in a with defining polynomial y^4 - 4*y^2 + 1
To:   Algebraic Real Field
Defn: a |--> -1.931851652578137?)


rt3a is a real number in QQbar. Ordinarily, we’d get a homomorphism to AA (because all elements are real), but if we specify same_field=True, we’ll get a homomorphism back to QQbar:

sage: number_field_elements_from_algebraics(rt3a)
(Number Field in a with defining polynomial y^2 - 3, a, Ring morphism:
From: Number Field in a with defining polynomial y^2 - 3
To:   Algebraic Real Field
Defn: a |--> 1.732050807568878?)

sage: number_field_elements_from_algebraics(rt3a, same_field=True)
(Number Field in a with defining polynomial y^2 - 3, a, Ring morphism:
From: Number Field in a with defining polynomial y^2 - 3
To:   Algebraic Field
Defn: a |--> 1.732050807568878?)


We’ve created rt2b in such a way that sage does not initially know that it’s in a degree-2 extension of $$\QQ$$:

sage: number_field_elements_from_algebraics(rt2b)
(Number Field in a with defining polynomial y^4 - 4*y^2 + 1, -a^3 + 3*a, Ring morphism:
From: Number Field in a with defining polynomial y^4 - 4*y^2 + 1
To:   Algebraic Real Field
Defn: a |--> -1.931851652578137?)


We can specify minimal=True if we want the smallest number field:

sage: number_field_elements_from_algebraics(rt2b, minimal=True)
(Number Field in a with defining polynomial y^2 - 2, a, Ring morphism:
From: Number Field in a with defining polynomial y^2 - 2
To:   Algebraic Real Field
Defn: a |--> 1.414213562373095?)


Things work fine with rational numbers, too:

sage: number_field_elements_from_algebraics((QQbar(1/2), AA(17)))
(Rational Field, [1/2, 17], Ring morphism:
From: Rational Field
To:   Algebraic Real Field
Defn: 1 |--> 1)


Or we can just pass in symbolic expressions, as long as they can be coerced into QQbar:

sage: number_field_elements_from_algebraics((sqrt(7), sqrt(9), sqrt(11)))
(Number Field in a with defining polynomial y^4 - 9*y^2 + 1, [-a^3 + 8*a, 3, -a^3 + 10*a], Ring morphism:
From: Number Field in a with defining polynomial y^4 - 9*y^2 + 1
To:   Algebraic Real Field
Defn: a |--> 0.3354367396454047?)


Here we see an example of doing some computations with number field elements, and then mapping them back into QQbar:

sage: (fld,nums,hom) = number_field_elements_from_algebraics((rt2, rt3, qqI, z3))
sage: fld,nums,hom  # random
(Number Field in a with defining polynomial y^8 - y^4 + 1, [-a^5 + a^3 + a, a^6 - 2*a^2, a^6, -a^4], Ring morphism:
From: Number Field in a with defining polynomial y^8 - y^4 + 1
To:   Algebraic Field
Defn: a |--> -0.2588190451025208? - 0.9659258262890683?*I)
sage: (nfrt2, nfrt3, nfI, nfz3) = nums
sage: hom(nfrt2)
1.414213562373095? + 0.?e-18*I
sage: nfrt2^2
2
sage: nfrt3^2
3
sage: nfz3 + nfz3^2
-1
sage: nfI^2
-1
sage: sum = nfrt2 + nfrt3 + nfI + nfz3; sum
a^5 + a^4 - a^3 + 2*a^2 - a - 1
sage: hom(sum)
2.646264369941973? + 1.866025403784439?*I
sage: hom(sum) == rt2 + rt3 + qqI + z3
True
sage: [hom(n) for n in nums] == [rt2, rt3, qqI, z3]
True


It is also possible to have an embedded Number Field:

sage: x = polygen(ZZ)
sage: my_num = AA.polynomial_root(x^3-2, RIF(0,3))
sage: res = number_field_elements_from_algebraics(my_num,embedded=True)
sage: res.gen_embedding()
1.259921049894873?
sage: res
Ring morphism:
From: Number Field in a with defining polynomial y^3 - 2 with a = 1.259921049894873?
To:   Algebraic Real Field
Defn: a |--> 1.259921049894873?

sage: nf,nums,hom = number_field_elements_from_algebraics([2^(1/3),3^(1/5)],embedded=True)
sage: nf
Number Field in a with defining polynomial y^15 - 9*y^10 + 21*y^5 - 3 with a = 0.6866813218928813?
sage: nums
[a^10 - 5*a^5 + 2, -a^8 + 4*a^3]
sage: hom
Ring morphism:
From: Number Field in a with defining polynomial y^15 - 9*y^10 + 21*y^5 - 3 with a = 0.6866813218928813?
To:   Algebraic Real Field
Defn: a |--> 0.6866813218928813?


Complex embeddings are possible as well:

sage: elems = [sqrt(5), 2^(1/3)+sqrt(3)*I, 3/4]
sage: nf, nums, hom = number_field_elements_from_algebraics(elems, embedded=True)
sage: nf # random (polynomial and root not unique)
Number Field in a with defining polynomial y^24 - 6*y^23 ...- 9*y^2 + 1
with a = 0.2598679? + 0.0572892?*I
sage: nf.is_isomorphic(NumberField(x^24 - 9*x^22 + 135*x^20 - 720*x^18 + 1821*x^16 - 3015*x^14 + 3974*x^12 - 3015*x^10 + 1821*x^8 - 720*x^6 + 135*x^4 - 9*x^2 + 1, 'a'))
True
sage: list(map(QQbar, nums)) == elems == list(map(hom, nums))
True

sage.rings.qqbar.prec_seq()#

Return a generator object which iterates over an infinite increasing sequence of precisions to be tried in various numerical computations.

Currently just returns powers of 2 starting at 64.

EXAMPLES:

sage: g = sage.rings.qqbar.prec_seq()
sage: [next(g), next(g), next(g)]
[64, 128, 256]

sage.rings.qqbar.rational_exact_root(r, d)#

Check whether the rational $$r$$ is an exact $$d$$’th power.

If so, this returns the $$d$$’th root of $$r$$; otherwise, this returns None.

EXAMPLES:

sage: from sage.rings.qqbar import rational_exact_root
sage: rational_exact_root(16/81, 4)
2/3
sage: rational_exact_root(8/81, 3) is None
True

sage.rings.qqbar.short_prec_seq()#

Return a sequence of precisions to try in cases when an infinite-precision computation is possible: returns a couple of small powers of 2 and then None.

EXAMPLES:

sage: from sage.rings.qqbar import short_prec_seq
sage: short_prec_seq()
(64, 128, None)

sage.rings.qqbar.t1#

alias of ANBinaryExpr

sage.rings.qqbar.t2#

alias of ANRoot

sage.rings.qqbar.tail_prec_seq()#

A generator over precisions larger than those in short_prec_seq().

EXAMPLES:

sage: from sage.rings.qqbar import tail_prec_seq
sage: g = tail_prec_seq()
sage: [next(g), next(g), next(g)]
[256, 512, 1024]