# Sage class for PARI’s GEN type¶

See the PariInstance class for documentation and examples.

AUTHORS:

• William Stein (2006-03-01): updated to work with PARI 2.2.12-beta
• William Stein (2006-03-06): added newtonpoly
• Justin Walker: contributed some of the function definitions
• Gonzalo Tornaria: improvements to conversions; much better error handling.
• Robert Bradshaw, Jeroen Demeyer, William Stein (2010-08-15): Upgrade to PARI 2.4.3 (#9343)
• Jeroen Demeyer (2011-11-12): rewrite various conversion routines (#11611, #11854, #11952)
• Peter Bruin (2013-11-17): move PariInstance to a separate file (#15185)
• Jeroen Demeyer (2014-02-09): upgrade to PARI 2.7 (#15767)
• Martin von Gagern (2014-12-17): Added some Galois functions (#17519)
class sage.libs.pari.gen.gen

Python extension class that models the PARI GEN type.

Col(x, n=0)

Transform the object $$x$$ into a column vector with minimal size $$|n|$$.

INPUT:

• x – gen
• n – Make the column vector of minimal length $$|n|$$. If $$n > 0$$, append zeros; if $$n < 0$$, prepend zeros.

OUTPUT:

A PARI column vector (type t_COL)

EXAMPLES:

sage: pari(1.5).Col()
[1.50000000000000]~
sage: pari([1,2,3,4]).Col()
[1, 2, 3, 4]~
sage: pari('[1,2; 3,4]').Col()
[[1, 2], [3, 4]]~
sage: pari('"Sage"').Col()
["S", "a", "g", "e"]~
sage: pari('x + 3*x^3').Col()
[3, 0, 1, 0]~
sage: pari('x + 3*x^3 + O(x^5)').Col()
[1, 0, 3, 0]~


We demonstate the $$n$$ argument:

sage: pari([1,2,3,4]).Col(2)
[1, 2, 3, 4]~
sage: pari([1,2,3,4]).Col(-2)
[1, 2, 3, 4]~
sage: pari([1,2,3,4]).Col(6)
[1, 2, 3, 4, 0, 0]~
sage: pari([1,2,3,4]).Col(-6)
[0, 0, 1, 2, 3, 4]~


See also Vec() (create a row vector) for more examples and Colrev() (create a column in reversed order).

Colrev(x, n=0)

Transform the object $$x$$ into a column vector with minimal size $$|n|$$. The order of the resulting vector is reversed compared to Col().

INPUT:

• x – gen
• n – Make the vector of minimal length $$|n|$$. If $$n > 0$$, prepend zeros; if $$n < 0$$, append zeros.

OUTPUT:

A PARI column vector (type t_COL)

EXAMPLES:

sage: pari(1.5).Colrev()
[1.50000000000000]~
sage: pari([1,2,3,4]).Colrev()
[4, 3, 2, 1]~
sage: pari('[1,2; 3,4]').Colrev()
[[3, 4], [1, 2]]~
sage: pari('x + 3*x^3').Colrev()
[0, 1, 0, 3]~


We demonstate the $$n$$ argument:

sage: pari([1,2,3,4]).Colrev(2)
[4, 3, 2, 1]~
sage: pari([1,2,3,4]).Colrev(-2)
[4, 3, 2, 1]~
sage: pari([1,2,3,4]).Colrev(6)
[0, 0, 4, 3, 2, 1]~
sage: pari([1,2,3,4]).Colrev(-6)
[4, 3, 2, 1, 0, 0]~

List(x)

List(x): transforms the PARI vector or list x into a list.

EXAMPLES:

sage: v = pari([1,2,3])
sage: v
[1, 2, 3]
sage: v.type()
't_VEC'
sage: w = v.List()
sage: w
List([1, 2, 3])
sage: w.type()
't_LIST'

Mat(x)

Mat(x): Returns the matrix defined by x.

• If x is already a matrix, a copy of x is created and returned.
• If x is not a vector or a matrix, this function returns a 1x1 matrix.
• If x is a row (resp. column) vector, this functions returns a 1-row (resp. 1-column) matrix, unless all elements are column (resp. row) vectors of the same length, in which case the vectors are concatenated sideways and the associated big matrix is returned.

INPUT:

• x - gen

OUTPUT:

• gen - a PARI matrix

EXAMPLES:

sage: x = pari(5)
sage: x.type()
't_INT'
sage: y = x.Mat()
sage: y
Mat(5)
sage: y.type()
't_MAT'
sage: x = pari('[1,2;3,4]')
sage: x.type()
't_MAT'
sage: x = pari('[1,2,3,4]')
sage: x.type()
't_VEC'
sage: y = x.Mat()
sage: y
Mat([1, 2, 3, 4])
sage: y.type()
't_MAT'

sage: v = pari('[1,2;3,4]').Vec(); v
[[1, 3]~, [2, 4]~]
sage: v.Mat()
[1, 2; 3, 4]
sage: v = pari('[1,2;3,4]').Col(); v
[[1, 2], [3, 4]]~
sage: v.Mat()
[1, 2; 3, 4]

Mod(x, y)

Mod(x, y): Returns the object x modulo y, denoted Mod(x, y).

The input y must be a an integer or a polynomial:

• If y is an INTEGER, x must also be an integer, a rational number, or a p-adic number compatible with the modulus y.
• If y is a POLYNOMIAL, x must be a scalar (which is not a polmod), a polynomial, a rational function, or a power series.

Warning

This function is not the same as x % y which is an integer or a polynomial.

INPUT:

• x - gen
• y - integer or polynomial

OUTPUT:

• gen - intmod or polmod

EXAMPLES:

sage: z = pari(3)
sage: x = z.Mod(pari(7))
sage: x
Mod(3, 7)
sage: x^2
Mod(2, 7)
sage: x^100
Mod(4, 7)
sage: x.type()
't_INTMOD'

sage: f = pari("x^2 + x + 1")
sage: g = pari("x")
sage: a = g.Mod(f)
sage: a
Mod(x, x^2 + x + 1)
sage: a*a
Mod(-x - 1, x^2 + x + 1)
sage: a.type()
't_POLMOD'

Pol(v=-1)

Pol(x, v): convert x into a polynomial with main variable v and return the result.

• If x is a scalar, returns a constant polynomial.
• If x is a power series, the effect is identical to truncate, i.e. it chops off the $$O(X^k)$$.
• If x is a vector, this function creates the polynomial whose coefficients are given in x, with x[0] being the leading coefficient (which can be zero).

Warning

This is not a substitution function. It will not transform an object containing variables of higher priority than v:

sage: pari('x+y').Pol('y')
Traceback (most recent call last):
...
PariError: incorrect priority in gtopoly: variable x < y


INPUT:

• x - gen
• v - (optional) which variable, defaults to ‘x’

OUTPUT:

• gen - a polynomial

EXAMPLES:

sage: v = pari("[1,2,3,4]")
sage: f = v.Pol()
sage: f
x^3 + 2*x^2 + 3*x + 4
sage: f*f
x^6 + 4*x^5 + 10*x^4 + 20*x^3 + 25*x^2 + 24*x + 16

sage: v = pari("[1,2;3,4]")
sage: v.Pol()
[1, 3]~*x + [2, 4]~

Polrev(v=-1)

Polrev(x, v): Convert x into a polynomial with main variable v and return the result. This is the reverse of Pol if x is a vector, otherwise it is identical to Pol. By “reverse” we mean that the coefficients are reversed.

INPUT:

• x - gen

OUTPUT:

• gen - a polynomial

EXAMPLES:

sage: v = pari("[1,2,3,4]")
sage: f = v.Polrev()
sage: f
4*x^3 + 3*x^2 + 2*x + 1
sage: v.Pol()
x^3 + 2*x^2 + 3*x + 4
sage: v.Polrev('y')
4*y^3 + 3*y^2 + 2*y + 1


Note that Polrev does not reverse the coefficients of a polynomial!

sage: f
4*x^3 + 3*x^2 + 2*x + 1
sage: f.Polrev()
4*x^3 + 3*x^2 + 2*x + 1
sage: v = pari("[1,2;3,4]")
sage: v.Polrev()
[2, 4]~*x + [1, 3]~

Qfb(a, b, c, D=0, precision=0)

Qfb(a,b,c,D=0.): Returns the binary quadratic form

$ax^2 + bxy + cy^2.$

The optional D is 0 by default and initializes Shank’s distance if $$b^2 - 4ac > 0$$. The discriminant of the quadratic form must not be a perfect square.

Note

Negative definite forms are not implemented, so use their positive definite counterparts instead. (I.e., if f is a negative definite quadratic form, then -f is positive definite.)

INPUT:

• a - gen
• b - gen
• c - gen
• D - gen (optional, defaults to 0)

OUTPUT:

• gen - binary quadratic form

EXAMPLES:

sage: pari(3).Qfb(7, 1)
Qfb(3, 7, 1, 0.E-19)
sage: pari(3).Qfb(7, 2)  # discriminant is 25
Traceback (most recent call last):
...
PariError: domain error in Qfb: issquare(disc) = 1

Ser(f, v=-1, precision=-1)

Return a power series or Laurent series in the variable $$v$$ constructed from the object $$f$$.

INPUT:

• f – PARI gen
• v – PARI variable (default: $$x$$)
• precision – the desired relative precision (default: the value returned by pari.get_series_precision()). This is the absolute precision minus the $$v$$-adic valuation.

OUTPUT:

• PARI object of type t_SER

The series is constructed from $$f$$ in the following way:

• If $$f$$ is a scalar, a constant power series is returned.
• If $$f$$ is a polynomial, it is converted into a power series in the obvious way.
• If $$f$$ is a rational function, it will be expanded in a Laurent series around $$v = 0$$.
• If $$f$$ is a vector, its coefficients become the coefficients of the power series, starting from the constant term. This is the convention used by the function Polrev(), and the reverse of that used by Pol().

Warning

This function will not transform objects containing variables of higher priority than $$v$$.

EXAMPLES:

sage: pari(2).Ser()
2 + O(x^16)
sage: pari(Mod(0, 7)).Ser()
Mod(0, 7) + O(x^16)

sage: x = pari([1, 2, 3, 4, 5])
sage: x.Ser()
1 + 2*x + 3*x^2 + 4*x^3 + 5*x^4 + O(x^16)
sage: f = x.Ser('v'); print f
1 + 2*v + 3*v^2 + 4*v^3 + 5*v^4 + O(v^16)
sage: pari(1)/f
1 - 2*v + v^2 + 6*v^5 - 17*v^6 + 16*v^7 - 5*v^8 + 36*v^10 - 132*v^11 + 181*v^12 - 110*v^13 + 25*v^14 + 216*v^15 + O(v^16)

sage: pari('x^5').Ser(precision=20)
x^5 + O(x^25)
sage: pari('1/x').Ser(precision=1)
x^-1 + O(x^0)

Set(x)

Set(x): convert x into a set, i.e. a row vector of strings in increasing lexicographic order.

INPUT:

• x - gen

OUTPUT:

• gen - a vector of strings in increasing lexicographic order.

EXAMPLES:

sage: pari([1,5,2]).Set()
[1, 2, 5]
sage: pari([]).Set()     # the empty set
[]
sage: pari([1,1,-1,-1,3,3]).Set()
[-1, 1, 3]
sage: pari(1).Set()
[1]
sage: pari('1/(x*y)').Set()
[1/(y*x)]
sage: pari('["bc","ab","bc"]').Set()
["ab", "bc"]

Str()

Str(self): Return the print representation of self as a PARI object.

INPUT:

• self - gen

OUTPUT:

• gen - a PARI gen of type t_STR, i.e., a PARI string

EXAMPLES:

sage: pari([1,2,['abc',1]]).Str()
"[1, 2, [abc, 1]]"
sage: pari([1,1, 1.54]).Str()
"[1, 1, 1.54000000000000]"
sage: pari(1).Str()       # 1 is automatically converted to string rep
"1"
sage: x = pari('x')       # PARI variable "x"
sage: x.Str()             # is converted to string rep.
"x"
sage: x.Str().type()
't_STR'

Strchr(x)

Strchr(x): converts x to a string, translating each integer into a character (in ASCII).

Note

Vecsmall() is (essentially) the inverse to Strchr().

INPUT:

• x - PARI vector of integers

OUTPUT:

• gen - a PARI string

EXAMPLES:

sage: pari([65,66,123]).Strchr()
"AB{"
sage: pari('"Sage"').Vecsmall()   # pari('"Sage"') --> PARI t_STR
Vecsmall([83, 97, 103, 101])
sage: _.Strchr()
"Sage"
sage: pari([83, 97, 103, 101]).Strchr()
"Sage"

Strexpand(x)

Concatenate the entries of the vector $$x$$ into a single string, then perform tilde expansion and environment variable expansion similar to shells.

INPUT:

• x – PARI gen. Either a vector or an element which is then treated like $$[x]$$.

OUTPUT:

• PARI string (type t_STR)

EXAMPLES:

sage: pari('"~/subdir"').Strexpand()     # random
"/home/johndoe/subdir"
sage: pari('"$SAGE_LOCAL"').Strexpand() # random "/usr/local/sage/local"  TESTS: sage: a = pari('"$HOME"')
sage: a.Strexpand() != a
True

Strtex(x)

Strtex(x): Translates the vector x of PARI gens to TeX format and returns the resulting concatenated strings as a PARI t_STR.

INPUT:

• x – PARI gen. Either a vector or an element which is then treated like $$[x]$$.

OUTPUT:

• PARI string (type t_STR)

EXAMPLES:

sage: v=pari('x^2')
sage: v.Strtex()
"x^2"
sage: v=pari(['1/x^2','x'])
sage: v.Strtex()
"\\frac{1}{x^2}x"
sage: v=pari(['1 + 1/x + 1/(y+1)','x-1'])
sage: v.Strtex()
"\\frac{ \\left(y\n + 2\\right)  x\n + \\left(y\n + 1\\right) }{ \\left(y\n + 1\\right)  x}x\n - 1"

Vec(x, n=0)

Transform the object $$x$$ into a vector with minimal size $$|n|$$.

INPUT:

• x – gen
• n – Make the vector of minimal length $$|n|$$. If $$n > 0$$, append zeros; if $$n < 0$$, prepend zeros.

OUTPUT:

A PARI vector (type t_VEC)

EXAMPLES:

sage: pari(1).Vec()
[1]
sage: pari('x^3').Vec()
[1, 0, 0, 0]
sage: pari('x^3 + 3*x - 2').Vec()
[1, 0, 3, -2]
sage: pari([1,2,3]).Vec()
[1, 2, 3]
sage: pari('[1, 2; 3, 4]').Vec()
[[1, 3]~, [2, 4]~]
sage: pari('"Sage"').Vec()
["S", "a", "g", "e"]
sage: pari('2*x^2 + 3*x^3 + O(x^5)').Vec()
[2, 3, 0]
sage: pari('2*x^-2 + 3*x^3 + O(x^5)').Vec()
[2, 0, 0, 0, 0, 3, 0]


Note the different term ordering for polynomials and series:

sage: pari('1 + x + 3*x^3 + O(x^5)').Vec()
[1, 1, 0, 3, 0]
sage: pari('1 + x + 3*x^3').Vec()
[3, 0, 1, 1]


We demonstate the $$n$$ argument:

sage: pari([1,2,3,4]).Vec(2)
[1, 2, 3, 4]
sage: pari([1,2,3,4]).Vec(-2)
[1, 2, 3, 4]
sage: pari([1,2,3,4]).Vec(6)
[1, 2, 3, 4, 0, 0]
sage: pari([1,2,3,4]).Vec(-6)
[0, 0, 1, 2, 3, 4]


See also Col() (create a column vector) and Vecrev() (create a vector in reversed order).

Vecrev(x, n=0)

Transform the object $$x$$ into a vector with minimal size $$|n|$$. The order of the resulting vector is reversed compared to Vec().

INPUT:

• x – gen
• n – Make the vector of minimal length $$|n|$$. If $$n > 0$$, prepend zeros; if $$n < 0$$, append zeros.

OUTPUT:

A PARI vector (type t_VEC)

EXAMPLES:

sage: pari(1).Vecrev()
[1]
sage: pari('x^3').Vecrev()
[0, 0, 0, 1]
sage: pari('x^3 + 3*x - 2').Vecrev()
[-2, 3, 0, 1]
sage: pari([1, 2, 3]).Vecrev()
[3, 2, 1]
sage: pari('Col([1, 2, 3])').Vecrev()
[3, 2, 1]
sage: pari('[1, 2; 3, 4]').Vecrev()
[[2, 4]~, [1, 3]~]
sage: pari('"Sage"').Vecrev()
["e", "g", "a", "S"]


We demonstate the $$n$$ argument:

sage: pari([1,2,3,4]).Vecrev(2)
[4, 3, 2, 1]
sage: pari([1,2,3,4]).Vecrev(-2)
[4, 3, 2, 1]
sage: pari([1,2,3,4]).Vecrev(6)
[0, 0, 4, 3, 2, 1]
sage: pari([1,2,3,4]).Vecrev(-6)
[4, 3, 2, 1, 0, 0]

Vecsmall(x, n=0)

Transform the object $$x$$ into a t_VECSMALL with minimal size $$|n|$$.

INPUT:

• x – gen
• n – Make the vector of minimal length $$|n|$$. If $$n > 0$$, append zeros; if $$n < 0$$, prepend zeros.

OUTPUT:

A PARI vector of small integers (type t_VECSMALL)

EXAMPLES:

sage: pari([1,2,3]).Vecsmall()
Vecsmall([1, 2, 3])
sage: pari('"Sage"').Vecsmall()
Vecsmall([83, 97, 103, 101])
sage: pari(1234).Vecsmall()
Vecsmall([1234])
sage: pari('x^2 + 2*x + 3').Vecsmall()
Vecsmall([1, 2, 3])


We demonstate the $$n$$ argument:

sage: pari([1,2,3]).Vecsmall(2)
Vecsmall([1, 2, 3])
sage: pari([1,2,3]).Vecsmall(-2)
Vecsmall([1, 2, 3])
sage: pari([1,2,3]).Vecsmall(6)
Vecsmall([1, 2, 3, 0, 0, 0])
sage: pari([1,2,3]).Vecsmall(-6)
Vecsmall([0, 0, 0, 1, 2, 3])

Zn_issquare(n)

Return True if self is a square modulo $$n$$, False if not.

INPUT:

• self – integer
• n – integer or factorisation matrix

EXAMPLES:

sage: pari(3).Zn_issquare(4)
False
sage: pari(4).Zn_issquare(30.factor())
True

Zn_sqrt(n)

Return a square root of self modulo $$n$$, if such a square root exists; otherwise, raise a ValueError.

INPUT:

• self – integer
• n – integer or factorisation matrix

EXAMPLES:

sage: pari(3).Zn_sqrt(4)
Traceback (most recent call last):
...
ValueError: 3 is not a square modulo 4
sage: pari(4).Zn_sqrt(30.factor())
22

abs(x, precision=0)

Returns the absolute value of x (its modulus, if x is complex). Rational functions are not allowed. Contrary to most transcendental functions, an exact argument is not converted to a real number before applying abs and an exact result is returned if possible.

EXAMPLES:

sage: x = pari("-27.1")
sage: x.abs()
27.1000000000000
sage: pari('1 + I').abs(precision=128).sage()
1.4142135623730950488016887242096980786


If x is a polynomial, returns -x if the leading coefficient is real and negative else returns x. For a power series, the constant coefficient is considered instead.

EXAMPLES:

sage: pari('x-1.2*x^2').abs()
1.20000000000000*x^2 - x
sage: pari('-2 + t + O(t^2)').abs()
2 - t + O(t^2)

acos(x, precision=0)

The principal branch of $$\cos^{-1}(x)$$, so that $$\RR e(\mathrm{acos}(x))$$ belongs to $$[0,Pi]$$. If $$x$$ is real and $$|x| > 1$$, then $$\mathrm{acos}(x)$$ is complex.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(0.5).acos()
1.04719755119660
sage: pari(1/2).acos()
1.04719755119660
sage: pari(1.1).acos()
0.443568254385115*I
sage: C.<i> = ComplexField()
sage: pari(1.1+i).acos()
0.849343054245252 - 1.09770986682533*I

acosh(x, precision=0)

The principal branch of $$\cosh^{-1}(x)$$, so that $$\Im(\mathrm{acosh}(x))$$ belongs to $$[0,Pi]$$. If $$x$$ is real and $$x < 1$$, then $$\mathrm{acosh}(x)$$ is complex.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).acosh()
1.31695789692482
sage: pari(0).acosh()
1.57079632679490*I
sage: C.<i> = ComplexField()
sage: pari(i).acosh()
0.881373587019543 + 1.57079632679490*I

agm(x, y, precision=0)

The arithmetic-geometric mean of x and y. In the case of complex or negative numbers, the principal square root is always chosen. p-adic or power series arguments are also allowed. Note that a p-adic AGM exists only if x/y is congruent to 1 modulo p (modulo 16 for p=2). x and y cannot both be vectors or matrices.

If any of $$x$$ or $$y$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their two precisions is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).agm(2)
2.00000000000000
sage: pari(0).agm(1)
0
sage: pari(1).agm(2)
1.45679103104691
sage: C.<i> = ComplexField()
sage: pari(1+i).agm(-3)
-0.964731722290876 + 1.15700282952632*I

algdep(n)

EXAMPLES:

sage: n = pari.set_real_precision(210)
sage: w1 = pari('z1=2-sqrt(26); (z1+I)/(z1-I)')
sage: f = w1.algdep(12); f
545*x^11 - 297*x^10 - 281*x^9 + 48*x^8 - 168*x^7 + 690*x^6 - 168*x^5 + 48*x^4 - 281*x^3 - 297*x^2 + 545*x
sage: f(w1).abs() < 1.0e-200
True
sage: f.factor()
[x, 1; x + 1, 2; x^2 + 1, 1; x^2 + x + 1, 1; 545*x^4 - 1932*x^3 + 2790*x^2 - 1932*x + 545, 1]
sage: pari.set_real_precision(n)
210

arg(x, precision=0)

arg(x): argument of x,such that $$-\pi < \arg(x) \leq \pi$$.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: C.<i> = ComplexField()
sage: pari(2+i).arg()
0.463647609000806

asin(x, precision=0)

The principal branch of $$\sin^{-1}(x)$$, so that $$\RR e(\mathrm{asin}(x))$$ belongs to $$[-\pi/2,\pi/2]$$. If $$x$$ is real and $$|x| > 1$$ then $$\mathrm{asin}(x)$$ is complex.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(pari(0.5).sin()).asin()
0.500000000000000
sage: pari(2).asin()
1.57079632679490 - 1.31695789692482*I

asinh(x, precision=0)

The principal branch of $$\sinh^{-1}(x)$$, so that $$\Im(\mathrm{asinh}(x))$$ belongs to $$[-\pi/2,\pi/2]$$.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).asinh()
1.44363547517881
sage: C.<i> = ComplexField()
sage: pari(2+i).asinh()
1.52857091948100 + 0.427078586392476*I

atan(x, precision=0)

The principal branch of $$\tan^{-1}(x)$$, so that $$\RR e(\mathrm{atan}(x))$$ belongs to $$]-\pi/2, \pi/2[$$.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(1).atan()
0.785398163397448
sage: C.<i> = ComplexField()
sage: pari(1.5+i).atan()
1.10714871779409 + 0.255412811882995*I

atanh(x, precision=0)

The principal branch of $$\tanh^{-1}(x)$$, so that $$\Im(\mathrm{atanh}(x))$$ belongs to $$]-\pi/2,\pi/2]$$. If $$x$$ is real and $$|x| > 1$$ then $$\mathrm{atanh}(x)$$ is complex.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(0).atanh()
0.E-19
sage: pari(2).atanh()
0.549306144334055 - 1.57079632679490*I

bernfrac(x)

The Bernoulli number $$B_x$$, where $$B_0 = 1$$, $$B_1 = -1/2$$, $$B_2 = 1/6,\ldots,$$ expressed as a rational number. The argument $$x$$ should be of type integer.

EXAMPLES:

sage: pari(18).bernfrac()
43867/798
sage: [pari(n).bernfrac() for n in range(10)]
[1, -1/2, 1/6, 0, -1/30, 0, 1/42, 0, -1/30, 0]

bernreal(x, precision=0)

The Bernoulli number $$B_x$$, as for the function bernfrac, but $$B_x$$ is returned as a real number (with the current precision).

EXAMPLES:

sage: pari(18).bernreal()
54.9711779448622
sage: pari(18).bernreal(precision=192).sage()
54.9711779448621553884711779448621553884711779448621553885

bernvec(x)

Creates a vector containing, as rational numbers, the Bernoulli numbers $$B_0, B_2,\ldots, B_{2x}$$. This routine is obsolete. Use bernfrac instead each time you need a Bernoulli number in exact form.

Note: this routine is implemented using repeated independent calls to bernfrac, which is faster than the standard recursion in exact arithmetic.

EXAMPLES:

sage: pari(8).bernvec()
doctest:...: DeprecationWarning: bernvec() is deprecated, use repeated calls to bernfrac() instead
See http://trac.sagemath.org/15767 for details.
[1, 1/6, -1/30, 1/42, -1/30, 5/66, -691/2730, 7/6, -3617/510]
sage: [pari(2*n).bernfrac() for n in range(9)]
[1, 1/6, -1/30, 1/42, -1/30, 5/66, -691/2730, 7/6, -3617/510]

besselh1(nu, x, precision=0)

The $$H^1$$-Bessel function of index $$\nu$$ and argument $$x$$.

If $$nu$$ or $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their precisions is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).besselh1(3)
0.486091260585891 - 0.160400393484924*I

besselh2(nu, x, precision=0)

The $$H^2$$-Bessel function of index $$\nu$$ and argument $$x$$.

If $$nu$$ or $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their precisions is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).besselh2(3)
0.486091260585891 + 0.160400393484924*I

besseli(nu, x, precision=0)

Bessel I function (Bessel function of the second kind), with index $$\nu$$ and argument $$x$$. If $$x$$ converts to a power series, the initial factor $$(x/2)^{\nu}/\Gamma(\nu+1)$$ is omitted (since it cannot be represented in PARI when $$\nu$$ is not integral).

If $$nu$$ or $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their precisions is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).besseli(3)
2.24521244092995
sage: C.<i> = ComplexField()
sage: pari(2).besseli(3+i)
1.12539407613913 + 2.08313822670661*I

besselj(nu, x, precision=0)

Bessel J function (Bessel function of the first kind), with index $$\nu$$ and argument $$x$$. If $$x$$ converts to a power series, the initial factor $$(x/2)^{\nu}/\Gamma(\nu+1)$$ is omitted (since it cannot be represented in PARI when $$\nu$$ is not integral).

If $$nu$$ or $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their precisions is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).besselj(3)
0.486091260585891

besseljh(nu, x, precision=0)

J-Bessel function of half integral index (Spherical Bessel function of the first kind). More precisely, besseljh(n,x) computes $$J_{n+1/2}(x)$$ where n must an integer, and x is any complex value. In the current implementation (PARI, version 2.2.11), this function is not very accurate when $$x$$ is small.

If $$nu$$ or $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their precisions is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).besseljh(3)
0.412710032209716

besselk(nu, x, flag=0, precision=0)

nu.besselk(x, flag=0): K-Bessel function (modified Bessel function of the second kind) of index nu, which can be complex, and argument x.

If $$nu$$ or $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their precisions is used in the computation, and the parameter precision is ignored.

INPUT:

• nu - a complex number
• x - real number (positive or negative)
• flag - default: 0 or 1: use hyperu (hyperu is much slower for small x, and doesn’t work for negative x).

EXAMPLES:

sage: C.<i> = ComplexField()
sage: pari(2+i).besselk(3)
0.0455907718407551 + 0.0289192946582081*I

sage: pari(2+i).besselk(-3)
-4.34870874986752 - 5.38744882697109*I

sage: pari(2+i).besselk(300, flag=1)
3.74224603319728 E-132 + 2.49071062641525 E-134*I

besseln(nu, x, precision=0)

nu.besseln(x): Bessel N function (Spherical Bessel function of the second kind) of index nu and argument x.

If $$nu$$ or $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their precisions is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: C.<i> = ComplexField()
sage: pari(2+i).besseln(3)
-0.280775566958244 - 0.486708533223726*I

bezout(x, y)
bid_get_cyc()

Returns the structure of the group $$(O_K/I)^*$$, where $$I$$ is the ideal represented by self.

NOTE: self must be a “big ideal” (bid) as returned by idealstar for example.

EXAMPLES:

sage: K.<i> = QuadraticField(-1)
sage: J = pari(K).idealstar(K.ideal(4*i + 2))
sage: J.bid_get_cyc()
[4, 2]

bid_get_gen()

Returns a vector of generators of the group $$(O_K/I)^*$$, where $$I$$ is the ideal represented by self.

NOTE: self must be a “big ideal” (bid) with generators, as returned by idealstar with flag = 2.

EXAMPLES:

sage: K.<i> = QuadraticField(-1)
sage: J = pari(K).idealstar(K.ideal(4*i + 2), 2)
sage: J.bid_get_gen()
[7, [-2, -1]~]


We get an exception if we do not supply flag = 2 to idealstar:

sage: J = pari(K).idealstar(K.ideal(3))
sage: J.bid_get_gen()
Traceback (most recent call last):
...
PariError: missing bid generators. Use idealstar(,,2)

binary(x)

binary(x): gives the vector formed by the binary digits of abs(x), where x is of type t_INT.

INPUT:

• x - gen of type t_INT

OUTPUT:

• gen - of type t_VEC

EXAMPLES:

sage: pari(0).binary()
[]
sage: pari(-5).binary()
[1, 0, 1]
sage: pari(5).binary()
[1, 0, 1]
sage: pari(2005).binary()
[1, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1]

sage: pari('"2"').binary()
Traceback (most recent call last):
...
TypeError: x (="2") must be of type t_INT, but is of type t_STR.

binomial(x, k)

binomial(x, k): return the binomial coefficient “x choose k”.

INPUT:

• x - any PARI object (gen)
• k - integer

EXAMPLES:

sage: pari(6).binomial(2)
15
sage: pari('x+1').binomial(3)
1/6*x^3 - 1/6*x
sage: pari('2+x+O(x^2)').binomial(3)
1/3*x + O(x^2)

bitand(x, y)

bitand(x,y): Bitwise and of two integers x and y. Negative numbers behave as if modulo some large power of 2.

INPUT:

• x - gen (of type t_INT)
• y - coercible to gen (of type t_INT)

OUTPUT:

• gen - of type type t_INT

EXAMPLES:

sage: pari(8).bitand(4)
0
sage: pari(8).bitand(8)
8
sage: pari(6).binary()
[1, 1, 0]
sage: pari(7).binary()
[1, 1, 1]
sage: pari(6).bitand(7)
6
sage: pari(19).bitand(-1)
19
sage: pari(-1).bitand(-1)
-1

bitneg(x, n=-1)

bitneg(x,n=-1): Bitwise negation of the integer x truncated to n bits. n=-1 (the default) represents an infinite sequence of the bit 1. Negative numbers behave as if modulo some large power of 2.

With n=-1, this function returns -n-1. With n = 0, it returns a number a such that $$a\cong -n-1 \pmod{2^n}$$.

INPUT:

• x - gen (t_INT)
• n - long, default = -1

OUTPUT:

• gen - t_INT

EXAMPLES:

sage: pari(10).bitneg()
-11
sage: pari(1).bitneg()
-2
sage: pari(-2).bitneg()
1
sage: pari(-1).bitneg()
0
sage: pari(569).bitneg()
-570
sage: pari(569).bitneg(10)
454
sage: 454 % 2^10
454
sage: -570 % 2^10
454

bitnegimply(x, y)

bitnegimply(x,y): Bitwise negated imply of two integers x and y, in other words, x BITAND BITNEG(y). Negative numbers behave as if modulo big power of 2.

INPUT:

• x - gen (of type t_INT)
• y - coercible to gen (of type t_INT)

OUTPUT:

• gen - of type type t_INT

EXAMPLES:

sage: pari(14).bitnegimply(0)
14
sage: pari(8).bitnegimply(8)
0
sage: pari(8+4).bitnegimply(8)
4

bitor(x, y)

bitor(x,y): Bitwise or of two integers x and y. Negative numbers behave as if modulo big power of 2.

INPUT:

• x - gen (of type t_INT)
• y - coercible to gen (of type t_INT)

OUTPUT:

• gen - of type type t_INT

EXAMPLES:

sage: pari(14).bitor(0)
14
sage: pari(8).bitor(4)
12
sage: pari(12).bitor(1)
13
sage: pari(13).bitor(1)
13

bittest(x, n)

bittest(x, long n): Returns bit number n (coefficient of $$2^n$$ in binary) of the integer x. Negative numbers behave as if modulo a big power of 2.

INPUT:

• x - gen (pari integer)

OUTPUT:

• bool - a Python bool

EXAMPLES:

sage: x = pari(6)
sage: x.bittest(0)
False
sage: x.bittest(1)
True
sage: x.bittest(2)
True
sage: x.bittest(3)
False
sage: pari(-3).bittest(0)
True
sage: pari(-3).bittest(1)
False
sage: [pari(-3).bittest(n) for n in range(10)]
[True, False, True, True, True, True, True, True, True, True]

bitxor(x, y)

bitxor(x,y): Bitwise exclusive or of two integers x and y. Negative numbers behave as if modulo big power of 2.

INPUT:

• x - gen (of type t_INT)
• y - coercible to gen (of type t_INT)

OUTPUT:

• gen - of type type t_INT

EXAMPLES:

sage: pari(6).bitxor(4)
2
sage: pari(0).bitxor(4)
4
sage: pari(6).bitxor(0)
6

bnf_get_cyc()

Returns the structure of the class group of this number field as a vector of SNF invariants.

NOTE: self must be a “big number field” (bnf).

EXAMPLES:

sage: K.<a> = QuadraticField(-65)
sage: K.pari_bnf().bnf_get_cyc()
[4, 2]

bnf_get_gen()

Returns a vector of generators of the class group of this number field.

NOTE: self must be a “big number field” (bnf).

EXAMPLES:

sage: K.<a> = QuadraticField(-65)
sage: G = K.pari_bnf().bnf_get_gen(); G
[[3, 2; 0, 1], [2, 1; 0, 1]]
sage: map(lambda J: K.ideal(J), G)
[Fractional ideal (3, a + 2), Fractional ideal (2, a + 1)]

bnf_get_no()

Returns the class number of self, a “big number field” (bnf).

EXAMPLES:

sage: K.<a> = QuadraticField(-65)
sage: K.pari_bnf().bnf_get_no()
8

bnf_get_reg()

Returns the regulator of this number field.

NOTE: self must be a “big number field” (bnf).

EXAMPLES:

sage: K.<a> = NumberField(x^4 - 4*x^2 + 1)
sage: K.pari_bnf().bnf_get_reg()
2.66089858019037...

bnfcertify()

bnf being as output by bnfinit, checks whether the result is correct, i.e. whether the calculation of the contents of self are correct without assuming the Generalized Riemann Hypothesis. If it is correct, the answer is 1. If not, the program may output some error message or loop indefinitely.

REFERENCES:

 [PariUsers] User’s Guide to PARI/GP, http://pari.math.u-bordeaux.fr/pub/pari/manuals/2.7.0/users.pdf
bnfinit(flag=0, tech=None, precision=0)
bnfisintnorm(x)
bnfisnorm(x, flag=0)
bnfisprincipal(x, flag=1)
bnfissunit(sunit_data, x)
bnfisunit(x)
bnfnarrow()
bnfsunit(S, precision=0)
bnfunit()
bnrclassno(I)

Return the order of the ray class group of self modulo I.

INPUT:

• self: a pari “BNF” object representing a number field
• I: a pari “BID” object representing an ideal of self

OUTPUT: integer

TESTS:

sage: K.<z> = QuadraticField(-23)
sage: p = K.primes_above(3)[0]
sage: K.pari_bnf().bnrclassno(p._pari_bid_())
3

ceil(x)

For real x: return the smallest integer = x. For rational functions: the quotient of numerator by denominator. For lists: apply componentwise.

INPUT:

• x - gen

OUTPUT:

• gen - depends on type of x

EXAMPLES:

sage: pari(1.4).ceil()
2
sage: pari(-1.4).ceil()
-1
sage: pari(3/4).ceil()
1
sage: pari(x).ceil()
x
sage: pari((x^2+x+1)/x).ceil()
x + 1


This may be unexpected: but it is correct, treating the argument as a rational function in RR(x).

sage: pari(x^2+5*x+2.5).ceil()
x^2 + 5*x + 2.50000000000000

centerlift(x, v=-1)

centerlift(x,v): Centered lift of x. This function returns exactly the same thing as lift, except if x is an integer mod.

INPUT:

• x - gen
• v - var (default: x)

OUTPUT: gen

EXAMPLES:

sage: x = pari(-2).Mod(5)
sage: x.centerlift()
-2
sage: x.lift()
3
sage: f = pari('x-1').Mod('x^2 + 1')
sage: f.centerlift()
x - 1
sage: f.lift()
x - 1
sage: f = pari('x-y').Mod('x^2+1')
sage: f
Mod(x - y, x^2 + 1)
sage: f.centerlift('x')
x - y
sage: f.centerlift('y')
Mod(x - y, x^2 + 1)

change_variable_name(var)

In self, which must be a t_POL or t_SER, set the variable to var. If the variable of self is already var, then return self.

Warning

You should be careful with variable priorities when applying this on a polynomial or series of which the coefficients have polynomial components. To be safe, only use this function on polynomials with integer or rational coefficients. For a safer alternative, use subst().

EXAMPLES:

sage: f = pari('x^3 + 17*x + 3')
sage: f.change_variable_name("y")
y^3 + 17*y + 3
sage: f = pari('1 + 2*y + O(y^10)')
sage: f.change_variable_name("q")
1 + 2*q + O(q^10)
sage: f.change_variable_name("y") is f
True


In PARI, I refers to the square root of -1, so it cannot be used as variable name. Note the difference with subst():

sage: f = pari('x^2 + 1')
sage: f.change_variable_name("I")
Traceback (most recent call last):
...
PariError: I already exists with incompatible valence
sage: f.subst("x", "I")
0

charpoly(var=-1, flag=0)

charpoly(A,v=x,flag=0): det(v*Id-A) = characteristic polynomial of A using the comatrix. flag is optional and may be set to 1 (use Lagrange interpolation) or 2 (use Hessenberg form), 0 being the default.

chinese(y)
component(x, n)

component(x, long n): Return n’th component of the internal representation of x. This function is 1-based instead of 0-based.

Note

For vectors or matrices, it is simpler to use x[n-1]. For list objects such as is output by nfinit, it is easier to use member functions.

INPUT:

• x - gen
• n - C long (coercible to)

OUTPUT: gen

EXAMPLES:

sage: pari([0,1,2,3,4]).component(1)
0
sage: pari([0,1,2,3,4]).component(2)
1
sage: pari([0,1,2,3,4]).component(4)
3
sage: pari('x^3 + 2').component(1)
2
sage: pari('x^3 + 2').component(2)
0
sage: pari('x^3 + 2').component(4)
1

sage: pari('x').component(0)
Traceback (most recent call last):
...
PariError: non-existent component: index < 1

concat(y)
conj(x)

conj(x): Return the algebraic conjugate of x.

INPUT:

• x - gen

OUTPUT: gen

EXAMPLES:

sage: pari('x+1').conj()
x + 1
sage: pari('x+I').conj()
x - I
sage: pari('1/(2*x+3*I)').conj()
1/(2*x - 3*I)
sage: pari([1,2,'2-I','Mod(x,x^2+1)', 'Mod(x,x^2-2)']).conj()
[1, 2, 2 + I, Mod(-x, x^2 + 1), Mod(-x, x^2 - 2)]
sage: pari('Mod(x,x^2-2)').conj()
Mod(-x, x^2 - 2)
sage: pari('Mod(x,x^3-3)').conj()
Traceback (most recent call last):
...
PariError: incorrect type in gconj (t_POLMOD)

conjvec(x, precision=0)

conjvec(x): Returns the vector of all conjugates of the algebraic number x. An algebraic number is a polynomial over Q modulo an irreducible polynomial.

INPUT:

• x - gen

OUTPUT: gen

EXAMPLES:

sage: pari('Mod(1+x,x^2-2)').conjvec()
[-0.414213562373095, 2.41421356237310]~
sage: pari('Mod(x,x^3-3)').conjvec()
[1.44224957030741, -0.721124785153704 - 1.24902476648341*I, -0.721124785153704 + 1.24902476648341*I]~
sage: pari('Mod(1+x,x^2-2)').conjvec(precision=192)[0].sage()
-0.414213562373095048801688724209698078569671875376948073177

content()

Greatest common divisor of all the components of self.

EXAMPLES:

sage: R.<x> = PolynomialRing(ZZ)
sage: pari(2*x^2 + 2).content()
2
sage: pari("4*x^3 - 2*x/3 + 2/5").content()
2/15

contfrac(x, b=0, lmax=0)

contfrac(x,b,lmax): continued fraction expansion of x (x rational, real or rational function). b and lmax are both optional, where b is the vector of numerators of the continued fraction, and lmax is a bound for the number of terms in the continued fraction expansion.

contfracpnqn(x, b=0, lmax=0)

contfracpnqn(x): [p_n,p_n-1; q_n,q_n-1] corresponding to the continued fraction x.

cos(x, precision=0)

The cosine function.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(1.5).cos()
0.0707372016677029
sage: C.<i> = ComplexField()
sage: pari(1+i).cos()
0.833730025131149 - 0.988897705762865*I
sage: pari('x+O(x^8)').cos()
1 - 1/2*x^2 + 1/24*x^4 - 1/720*x^6 + 1/40320*x^8 + O(x^9)

cosh(x, precision=0)

The hyperbolic cosine function.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(1.5).cosh()
2.35240961524325
sage: C.<i> = ComplexField()
sage: pari(1+i).cosh()
0.833730025131149 + 0.988897705762865*I
sage: pari('x+O(x^8)').cosh()
1 + 1/2*x^2 + 1/24*x^4 + 1/720*x^6 + O(x^8)

cotan(x, precision=0)

The cotangent of x.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(5).cotan()
-0.295812915532746


Computing the cotangent of $$\pi$$ doesn’t raise an error, but instead just returns a very large (positive or negative) number.

sage: x = RR(pi)
sage: pari(x).cotan()         # random
-8.17674825 E15

debug(depth=-1)

Show the internal structure of self (like the \x command in gp).

EXAMPLE:

sage: pari('[1/2, 1.0*I]').debug()  # random addresses
[&=0000000004c5f010] VEC(lg=3):2200000000000003 0000000004c5eff8 0000000004c5efb0
1st component = [&=0000000004c5eff8] FRAC(lg=3):0800000000000003 0000000004c5efe0 0000000004c5efc8
num = [&=0000000004c5efe0] INT(lg=3):0200000000000003 (+,lgefint=3):4000000000000003 0000000000000001
den = [&=0000000004c5efc8] INT(lg=3):0200000000000003 (+,lgefint=3):4000000000000003 0000000000000002
2nd component = [&=0000000004c5efb0] COMPLEX(lg=3):0c00000000000003 00007fae8a2eb840 0000000004c5ef90
real = gen_0
imag = [&=0000000004c5ef90] REAL(lg=4):0400000000000004 (+,expo=0):6000000000000000 8000000000000000 0000000000000000

denominator(x)

denominator(x): Return the denominator of x. When x is a vector, this is the least common multiple of the denominators of the components of x.

• x - gen

OUTPUT: gen

EXAMPLES:

sage: pari('5/9').denominator()
9
sage: pari('(x+1)/(x-2)').denominator()
x - 2
sage: pari('2/3 + 5/8*x + 7/3*x^2 + 1/5*y').denominator()
1
sage: pari('2/3*x').denominator()
1
sage: pari('[2/3, 5/8, 7/3, 1/5]').denominator()
120

deriv(v=-1)
dilog(x, precision=0)

The principal branch of the dilogarithm of $$x$$, i.e. the analytic continuation of the power series $$\log_2(x) = \sum_{n>=1} x^n/n^2$$.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(1).dilog()
1.64493406684823
sage: C.<i> = ComplexField()
sage: pari(1+i).dilog()
0.616850275068085 + 1.46036211675312*I

dirzetak(n)
disc()

e.disc(): return the discriminant of the elliptic curve e.

EXAMPLES:

sage: e = pari([0, -1, 1, -10, -20]).ellinit()
sage: e.disc()
-161051
sage: _.factor()
[-1, 1; 11, 5]

divrem(x, y, var=-1)

divrem(x, y, v): Euclidean division of x by y giving as a 2-dimensional column vector the quotient and the remainder, with respect to v (to main variable if v is omitted).

eint1(x, n=0, precision=0)

x.eint1(n): exponential integral E1(x):

$\int_{x}^{\infty} \frac{e^{-t}}{t} dt$

If n is present, output the vector [eint1(x), eint1(2*x), ..., eint1(n*x)]. This is faster than repeatedly calling eint1(i*x).

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

REFERENCE:

• See page 262, Prop 5.6.12, of Cohen’s book “A Course in Computational Algebraic Number Theory”.

EXAMPLES:

elementval(x, p)

e.elladd(z0, z1): return the sum of the points z0 and z1 on this elliptic curve.

INPUT:

• e - elliptic curve E
• z0 - point on E
• z1 - point on E

OUTPUT: point on E

EXAMPLES:

First we create an elliptic curve:

sage: e = pari([0, 1, 1, -2, 0]).ellinit()


Next we add two points on the elliptic curve. Notice that the Python lists are automatically converted to PARI objects so you don’t have to do that explicitly in your code.

sage: e.elladd([1,0], [-1,1])
[-3/4, -15/8]

ellak(n)

e.ellak(n): Returns the coefficient $$a_n$$ of the $$L$$-function of the elliptic curve e, i.e. the $$n$$-th Fourier coefficient of the weight 2 newform associated to e (according to Shimura-Taniyama).

The curve $$e$$ must be a medium or long vector of the type given by ellinit. For this function to work for every n and not just those prime to the conductor, e must be a minimal Weierstrass equation. If this is not the case, use the function ellminimalmodel first before using ellak (or you will get INCORRECT RESULTS!)

INPUT:

• e - a PARI elliptic curve.
• n - integer.

EXAMPLES:

sage: e = pari([0, -1, 1, -10, -20]).ellinit()
sage: e.ellak(6)
2
sage: e.ellak(2005)
2
sage: e.ellak(-1)
0
sage: e.ellak(0)
0

ellan(n, python_ints=False)

Return the first $$n$$ Fourier coefficients of the modular form attached to this elliptic curve. See ellak for more details.

INPUT:

• n - a long integer
• python_ints - bool (default is False); if True, return a list of Python ints instead of a PARI gen wrapper.

EXAMPLES:

sage: e = pari([0, -1, 1, -10, -20]).ellinit()
sage: e.ellan(3)
[1, -2, -1]
sage: e.ellan(20)
[1, -2, -1, 2, 1, 2, -2, 0, -2, -2, 1, -2, 4, 4, -1, -4, -2, 4, 0, 2]
sage: e.ellan(-1)
[]
sage: v = e.ellan(10, python_ints=True); v
[1, -2, -1, 2, 1, 2, -2, 0, -2, -2]
sage: type(v)
<type 'list'>
sage: type(v[0])
<type 'int'>

ellanalyticrank(precision=0)

Returns a 2-component vector with the order of vanishing at $$s = 1$$ of the L-function of the elliptic curve and the value of the first non-zero derivative.

EXAMPLE:

sage: E = EllipticCurve('389a1')
sage: pari(E).ellanalyticrank()
[2, 1.51863300057685]

ellap(p)

e.ellap(p): Returns the prime-indexed coefficient $$a_p$$ of the $$L$$-function of the elliptic curve $$e$$, i.e. the $$p$$-th Fourier coefficient of the newform attached to e.

The computation uses the Shanks–Mestre method, or the SEA algorithm.

Warning

For this function to work for every n and not just those prime to the conductor, e must be a minimal Weierstrass equation. If this is not the case, use the function ellminimalmodel first before using ellap (or you will get INCORRECT RESULTS!)

INPUT:

• e - a PARI elliptic curve.
• p - prime integer

EXAMPLES:

sage: e = pari([0, -1, 1, -10, -20]).ellinit()
sage: e.ellap(2)
-2
sage: e.ellap(2003)
4
sage: e.ellak(-1)
0

ellaplist(n, python_ints=False)

e.ellaplist(n): Returns a PARI list of all the prime-indexed coefficients $$a_p$$ (up to n) of the $$L$$-function of the elliptic curve $$e$$, i.e. the Fourier coefficients of the newform attached to $$e$$.

INPUT:

• self – an elliptic curve
• n – a long integer
• python_ints – bool (default is False); if True, return a list of Python ints instead of a PARI gen wrapper.

Warning

The curve e must be a medium or long vector of the type given by ellinit. For this function to work for every n and not just those prime to the conductor, e must be a minimal Weierstrass equation. If this is not the case, use the function ellminimalmodel first before using ellaplist (or you will get INCORRECT RESULTS!)

EXAMPLES:

sage: e = pari([0, -1, 1, -10, -20]).ellinit()
sage: v = e.ellaplist(10); v
[-2, -1, 1, -2]
sage: type(v)
<type 'sage.libs.pari.gen.gen'>
sage: v.type()
't_VEC'
sage: e.ellan(10)
[1, -2, -1, 2, 1, 2, -2, 0, -2, -2]
sage: v = e.ellaplist(10, python_ints=True); v
[-2, -1, 1, -2]
sage: type(v)
<type 'list'>
sage: type(v[0])
<type 'int'>


TESTS:

sage: v = e.ellaplist(1)
sage: print v, type(v)
[] <type 'sage.libs.pari.gen.gen'>
sage: v = e.ellaplist(1, python_ints=True)
sage: print v, type(v)
[] <type 'list'>

ellbil(z0, z1, precision=0)

e.ellbil(z0, z1): return the value of the canonical bilinear form on z0 and z1.

INPUT:

• e - elliptic curve (assumed integral given by a minimal model, as returned by ellminimalmodel)
• z0, z1 - rational points on e

EXAMPLES:

sage: e = pari([0,1,1,-2,0]).ellinit().ellminimalmodel()[0]
sage: e.ellbil([1, 0], [-1, 1])
0.418188984498861

ellchangecurve(ch)

e.ellchangecurve(ch): return the new model (equation) for the elliptic curve e given by the change of coordinates ch.

The change of coordinates is specified by a vector ch=[u,r,s,t]; if $$x'$$ and $$y'$$ are the new coordinates, then $$x = u^2 x' + r$$ and $$y = u^3 y' + su^2 x' + t$$.

INPUT:

• e - elliptic curve
• ch - change of coordinates vector with 4 entries

EXAMPLES:

sage: e = pari([1,2,3,4,5]).ellinit()
sage: e.ellglobalred()
[10351, [1, -1, 0, -1], 1, [11, 1; 941, 1], [[1, 5, 0, 1], [1, 5, 0, 1]]]
sage: f = e.ellchangecurve([1,-1,0,-1])
sage: f[:5]
[1, -1, 0, 4, 3]

ellchangepoint(y)

self.ellchangepoint(y): change data on point or vector of points self on an elliptic curve according to y=[u,r,s,t]

EXAMPLES:

sage: e = pari([0,1,1,-2,0]).ellinit()
sage: x = pari([1,0])
sage: e.ellisoncurve([1,4])
False
sage: e.ellisoncurve(x)
True
sage: f = e.ellchangecurve([1,2,3,-1])
sage: f[:5]   # show only first five entries
[6, -2, -1, 17, 8]
sage: x.ellchangepoint([1,2,3,-1])
[-1, 4]
sage: f.ellisoncurve([-1,4])
True

elleisnum(k, flag=0, precision=0)

om.elleisnum(k, flag=0): om=[om1,om2] being a 2-component vector giving a basis of a lattice L and k an even positive integer, computes the numerical value of the Eisenstein series of weight k. When flag is non-zero and k=4 or 6, this gives g2 or g3 with the correct normalization.

INPUT:

• om - gen, 2-component vector giving a basis of a lattice L
• k - int (even positive)
• flag - int (default 0)

OUTPUT:

• gen - numerical value of E_k

EXAMPLES:

sage: e = pari([0,1,1,-2,0]).ellinit()
sage: om = e.omega()
sage: om
[2.49021256085506, -1.97173770155165*I]
sage: om.elleisnum(2)
10.0672605281120
sage: om.elleisnum(4)
112.000000000000
sage: om.elleisnum(100)
2.15314248576078 E50

elleta(precision=0)

e.elleta(): return the vector [eta1,eta2] of quasi-periods associated with the period lattice e.omega() of the elliptic curve e.

EXAMPLES:

sage: e = pari([0,0,0,-82,0]).ellinit()
sage: e.elleta()
[3.60546360143265, 3.60546360143265*I]
sage: w1, w2 = e.omega()
sage: eta1, eta2 = e.elleta()
sage: w1*eta2 - w2*eta1
6.28318530717959*I

ellglobalred()

Return information related to the global minimal model of the elliptic curve e.

INPUT:

• e – elliptic curve (returned by ellinit)

OUTPUT: A vector [N, [u,r,s,t], c, faN, L] with

• N - the (arithmetic) conductor of $$e$$

• [u,r,s,t] - a vector giving the coordinate change over

• c - the product of the local Tamagawa numbers of $$e$$.

• faN is the factorization of $$N$$

• L[i] is elllocalred(E, faN[i,1])

EXAMPLES:

sage: e = pari([0, 5, 2, -1, 1]).ellinit()
sage: e.ellglobalred()
[20144, [1, -2, 0, -1], 1, [2, 4; 1259, 1], [[4, 2, 0, 1], [1, 5, 0, 1]]]
sage: e = pari(EllipticCurve('17a').a_invariants()).ellinit()
sage: e.ellglobalred()
[17, [1, 0, 0, 0], 4, Mat([17, 1]), [[1, 8, 0, 4]]]

ellheight(a, flag=2, precision=0)

e.ellheight(a, flag=2): return the global Neron-Tate height of the point a on the elliptic curve e.

INPUT:

• e - elliptic curve over $$\QQ$$, assumed to be in a standard minimal integral model (as given by ellminimalmodel)

• a - rational point on e

• flag (optional) - specifies which algorithm to be used for computing the archimedean local height:

• 0 - uses sigma- and theta-functions and a trick

due to J. Silverman

• 1 - uses Tate’s $$4^n$$ algorithm

• 2 - uses Mestre’s AGM algorithm (this is the default, being faster than the other two)

• precision (optional) - the precision of the result, in bits.

EXAMPLES:

sage: e = pari([0,1,1,-2,0]).ellinit().ellminimalmodel()[0]
sage: e.ellheight([1,0])
0.476711659343740
sage: e.ellheight([1,0], flag=0)
0.476711659343740
sage: e.ellheight([1,0], flag=1)
0.476711659343740
sage: e.ellheight([1,0], precision=128).sage()
0.47671165934373953737948605888465305945902294217            # 32-bit
0.476711659343739537379486058884653059459022942211150879336  # 64-bit

ellheightmatrix(x, precision=0)

e.ellheightmatrix(x): return the height matrix for the vector x of points on the elliptic curve e.

In other words, it returns the Gram matrix of x with respect to the height bilinear form on e (see ellbil).

INPUT:

• e - elliptic curve over $$\QQ$$, assumed to be in a standard minimal integral model (as given by ellminimalmodel)
• x - vector of rational points on e

EXAMPLES:

sage: e = pari([0,1,1,-2,0]).ellinit().ellminimalmodel()[0]
sage: e.ellheightmatrix([[1,0], [-1,1]])
[0.476711659343740, 0.418188984498861; 0.418188984498861, 0.686667083305587]

ellinit(flag=0, precision=0)

Return the PARI elliptic curve object with Weierstrass coefficients given by self, a list with 5 elements.

INPUT:

• self – a list of 5 coefficients

• flag – ignored (for backwards compatibility)

• precision (optional, default: 0) - the real precision to be used in the computation of the components of the PARI (s)ell structure; if 0, use the default 64 bits.

Note

The parameter precision in ellinit controls not only the real precision of the resulting (s)ell structure, but in some cases also the precision of most subsequent computations with this elliptic curve (if those rely on the precomputations done by ellinit). You should therefore set the precision from the start to the value you require.

OUTPUT:

• gen – a PARI ell structure.

EXAMPLES:

An elliptic curve with integer coefficients:

sage: e = pari([0,1,0,1,0]).ellinit(); e
[0, 1, 0, 1, 0, 4, 2, 0, -1, -32, 224, -48, 2048/3, Vecsmall([1]), [Vecsmall([64, -1])], [0, 0, 0, 0, 0, 0, 0, 0]]


The coefficients can be any ring elements that convert to PARI:

sage: pari([0,1/2,0,-3/4,0]).ellinit()
[0, 1/2, 0, -3/4, 0, 2, -3/2, 0, -9/16, 40, -116, 117/4, 256000/117, Vecsmall([1]), [Vecsmall([64, 1])], [0, 0, 0, 0, 0, 0, 0, 0]]
sage: pari([0,0.5,0,-0.75,0]).ellinit()
[0, 0.500000000000000, 0, -0.750000000000000, 0, 2.00000000000000, -1.50000000000000, 0, -0.562500000000000, 40.0000000000000, -116.000000000000, 29.2500000000000, 2188.03418803419, Vecsmall([0]), [Vecsmall([64, 1])], [0, 0, 0, 0]]
sage: pari([0,I,0,1,0]).ellinit()
[0, I, 0, 1, 0, 4*I, 2, 0, -1, -64, 352*I, -80, 16384/5, Vecsmall([0]), [Vecsmall([64, 0])], [0, 0, 0, 0]]
sage: pari([0,x,0,2*x,1]).ellinit()
[0, x, 0, 2*x, 1, 4*x, 4*x, 4, -4*x^2 + 4*x, 16*x^2 - 96*x, -64*x^3 + 576*x^2 - 864, 64*x^4 - 576*x^3 + 576*x^2 - 432, (256*x^6 - 4608*x^5 + 27648*x^4 - 55296*x^3)/(4*x^4 - 36*x^3 + 36*x^2 - 27), Vecsmall([0]), [Vecsmall([64, 0])], [0, 0, 0, 0]]

ellisoncurve(x)

e.ellisoncurve(x): return True if the point x is on the elliptic curve e, False otherwise.

If the point or the curve have inexact coefficients, an attempt is made to take this into account.

EXAMPLES:

sage: e = pari([0,1,1,-2,0]).ellinit()
sage: e.ellisoncurve([1,0])
True
sage: e.ellisoncurve([1,1])
False
sage: e.ellisoncurve([1,0.00000000000000001])
False
sage: e.ellisoncurve([1,0.000000000000000001])
True
sage: e.ellisoncurve([0])
True

ellj(precision=0)

Elliptic $$j$$-invariant of self.

EXAMPLES:

sage: pari(I).ellj()
1728.00000000000
sage: pari(3*I).ellj()
153553679.396729
0.E-54
-3375.000000000000000000000000000000000000000000000000000000000000000000000000
sage: pari(-I).ellj()
Traceback (most recent call last):
...
PariError: domain error in modular function: Im(argument) <= 0

elllocalred(p)

e.elllocalred(p): computes the data of local reduction at the prime p on the elliptic curve e

For more details on local reduction and Kodaira types, see IV.8 and IV.9 in J. Silverman’s book “Advanced topics in the arithmetic of elliptic curves”.

INPUT:

• e - elliptic curve with coefficients in $$\ZZ$$
• p - prime number

OUTPUT:

• gen - the exponent of p in the arithmetic conductor of e
• gen - the Kodaira type of e at p, encoded as an integer:
• 1 - type $$I_0$$: good reduction, nonsingular curve of genus 1
• 2 - type $$II$$: rational curve with a cusp
• 3 - type $$III$$: two nonsingular rational curves intersecting tangentially at one point
• 4 - type $$IV$$: three nonsingular rational curves intersecting at one point
• 5 - type $$I_1$$: rational curve with a node
• 6 or larger - think of it as $$4+v$$, then it is type $$I_v$$: $$v$$ nonsingular rational curves arranged as a $$v$$-gon
• -1 - type $$I_0^*$$: nonsingular rational curve of multiplicity two with four nonsingular rational curves of multiplicity one attached
• -2 - type $$II^*$$: nine nonsingular rational curves in a special configuration
• -3 - type $$III^*$$: eight nonsingular rational curves in a special configuration
• -4 - type $$IV^*$$: seven nonsingular rational curves in a special configuration
• -5 or smaller - think of it as $$-4-v$$, then it is type $$I_v^*$$: chain of $$v+1$$ nonsingular rational curves of multiplicity two, with two nonsingular rational curves of multiplicity one attached at either end
• gen - a vector with 4 components, giving the coordinate changes done during the local reduction; if the first component is 1, then the equation for e was already minimal at p
• gen - the local Tamagawa number $$c_p$$

EXAMPLES:

Type $$I_0$$:

sage: e = pari([0,0,0,0,1]).ellinit()
sage: e.elllocalred(7)
[0, 1, [1, 0, 0, 0], 1]


Type $$II$$:

sage: e = pari(EllipticCurve('27a3').a_invariants()).ellinit()
sage: e.elllocalred(3)
[3, 2, [1, -1, 0, 1], 1]


Type $$III$$:

sage: e = pari(EllipticCurve('24a4').a_invariants()).ellinit()
sage: e.elllocalred(2)
[3, 3, [1, 1, 0, 1], 2]


Type $$IV$$:

sage: e = pari(EllipticCurve('20a2').a_invariants()).ellinit()
sage: e.elllocalred(2)
[2, 4, [1, 1, 0, 1], 3]


Type $$I_1$$:

sage: e = pari(EllipticCurve('11a2').a_invariants()).ellinit()
sage: e.elllocalred(11)
[1, 5, [1, 0, 0, 0], 1]


Type $$I_2$$:

sage: e = pari(EllipticCurve('14a4').a_invariants()).ellinit()
sage: e.elllocalred(2)
[1, 6, [1, 0, 0, 0], 2]


Type $$I_6$$:

sage: e = pari(EllipticCurve('14a1').a_invariants()).ellinit()
sage: e.elllocalred(2)
[1, 10, [1, 0, 0, 0], 2]


Type $$I_0^*$$:

sage: e = pari(EllipticCurve('32a3').a_invariants()).ellinit()
sage: e.elllocalred(2)
[5, -1, [1, 1, 1, 0], 1]


Type $$II^*$$:

sage: e = pari(EllipticCurve('24a5').a_invariants()).ellinit()
sage: e.elllocalred(2)
[3, -2, [1, 2, 1, 4], 1]


Type $$III^*$$:

sage: e = pari(EllipticCurve('24a2').a_invariants()).ellinit()
sage: e.elllocalred(2)
[3, -3, [1, 2, 1, 4], 2]


Type $$IV^*$$:

sage: e = pari(EllipticCurve('20a1').a_invariants()).ellinit()
sage: e.elllocalred(2)
[2, -4, [1, 0, 1, 2], 3]


Type $$I_1^*$$:

sage: e = pari(EllipticCurve('24a1').a_invariants()).ellinit()
sage: e.elllocalred(2)
[3, -5, [1, 0, 1, 2], 4]


Type $$I_6^*$$:

sage: e = pari(EllipticCurve('90c2').a_invariants()).ellinit()
sage: e.elllocalred(3)
[2, -10, [1, 96, 1, 316], 4]

elllseries(s, A=1, precision=0)

e.elllseries(s, A=1): return the value of the $$L$$-series of the elliptic curve e at the complex number s.

This uses an $$O(N^{1/2})$$ algorithm in the conductor N of e, so it is impractical for large conductors (say greater than $$10^{12}$$).

INPUT:

• e - elliptic curve defined over $$\QQ$$
• s - complex number
• A (optional) - cutoff point for the integral, which must be chosen close to 1 for best speed.

EXAMPLES:

sage: e = pari([0,1,1,-2,0]).ellinit()
sage: e.elllseries(2.1)
0.402838047956645
sage: e.elllseries(1, precision=128)
2.98766720445395 E-38
sage: e.elllseries(1, precision=256)
5.48956813891054 E-77
sage: e.elllseries(-2)
0
sage: e.elllseries(2.1, A=1.1)
0.402838047956645

ellminimalmodel()

ellminimalmodel(e): return the standard minimal integral model of the rational elliptic curve e and the corresponding change of variables. INPUT:

• e - gen (that defines an elliptic curve)

OUTPUT:

• gen - minimal model
• gen - change of coordinates

EXAMPLES:

sage: e = pari([1,2,3,4,5]).ellinit()
sage: F, ch = e.ellminimalmodel()
sage: F[:5]
[1, -1, 0, 4, 3]
sage: ch
[1, -1, 0, -1]
sage: e.ellchangecurve(ch)[:5]
[1, -1, 0, 4, 3]

ellorder(x)

e.ellorder(x): return the order of the point x on the elliptic curve e (return 0 if x is not a torsion point)

INPUT:

• e - elliptic curve defined over $$\QQ$$
• x - point on e

EXAMPLES:

sage: e = pari(EllipticCurve('65a1').a_invariants()).ellinit()


A point of order two:

sage: e.ellorder([0,0])
2


And a point of infinite order:

sage: e.ellorder([1,0])
0

ellordinate(x, precision=0)

e.ellordinate(x): return the $$y$$-coordinates of the points on the elliptic curve e having x as $$x$$-coordinate.

INPUT:

• e - elliptic curve
• x - x-coordinate (can be a complex or p-adic number, or a more complicated object like a power series)

EXAMPLES:

sage: e = pari([0,1,1,-2,0]).ellinit()
sage: e.ellordinate(0)
[0, -1]
sage: e.ellordinate(I)
[0.582203589721741 - 1.38606082464177*I, -1.58220358972174 + 1.38606082464177*I]
sage: e.ellordinate(I, precision=128)[0].sage()
0.58220358972174117723338947874993600727 - 1.3860608246417697185311834209833653345*I
sage: e.ellordinate(1+3*5^1+O(5^3))
[4*5 + 5^2 + O(5^3), 4 + 3*5^2 + O(5^3)]
sage: e.ellordinate('z+2*z^2+O(z^4)')
[-2*z - 7*z^2 - 23*z^3 + O(z^4), -1 + 2*z + 7*z^2 + 23*z^3 + O(z^4)]


The field in which PARI looks for the point depends on the input field:

sage: e.ellordinate(5)
[]
sage: e.ellordinate(5.0)
[11.3427192823270, -12.3427192823270]

ellpointtoz(pt, precision=0)

e.ellpointtoz(pt): return the complex number (in the fundamental parallelogram) corresponding to the point pt on the elliptic curve e, under the complex uniformization of e given by the Weierstrass p-function.

The complex number z returned by this function lies in the parallelogram formed by the real and complex periods of e, as given by e.omega().

EXAMPLES:

sage: e = pari([0,0,0,1,0]).ellinit()
sage: e.ellpointtoz([0,0])
1.85407467730137


The point at infinity is sent to the complex number 0:

sage: e.ellpointtoz([0])
0

ellpow(z, n)

e.ellpow(z, n): return $$n$$ times the point $$z$$ on the elliptic curve $$e$$.

INPUT:

• e - elliptic curve
• z - point on $$e$$
• n - integer, or a complex quadratic integer of complex multiplication for $$e$$. Complex multiplication currently only works if $$e$$ is defined over $$Q$$.

EXAMPLES: We consider a curve with CM by $$Z[i]$$:

sage: e = pari([0,0,0,3,0]).ellinit()
sage: p = [1,2]  # Point of infinite order


Multiplication by two:

sage: e.ellpow([0,0], 2)
[0]
sage: e.ellpow(p, 2)
[1/4, -7/8]


Complex multiplication:

sage: q = e.ellpow(p, 1+I); q
[-2*I, 1 + I]
sage: e.ellpow(q, 1-I)
[1/4, -7/8]


TESTS:

sage: for D in [-7, -8, -11, -12, -16, -19, -27, -28]:  # long time (1s)
....:     hcpol = hilbert_class_polynomial(D)
....:     j = hcpol.roots(multiplicities=False)[0]
....:     t = (1728-j)/(27*j)
....:     E = EllipticCurve([4*t,16*t^2])
....:     P = E.point([0, 4*t])
....:     assert(E.j_invariant() == j)
....:     #
....:     # Compute some CM number and its minimal polynomial
....:     #
....:     cm = pari('cm = (3*quadgen(%s)+2)'%D)
....:     cm_minpoly = pari('minpoly(cm)')
....:     #
....:     # Evaluate cm_minpoly(cm)(P), which should be zero
....:     #
....:     e = pari(E)  # Convert E to PARI
....:     P2 = e.ellpow(P, cm_minpoly[2]*cm + cm_minpoly[1])
....:     P0 = e.elladd(e.ellpow(P, cm_minpoly[0]), e.ellpow(P2, cm))
....:     assert(P0 == E(0))

ellrootno(p=None)

Return the root number for the L-function of the elliptic curve E/Q at a prime p (including 0, for the infinite place); return the global root number if p is omitted.

INPUT:

• e - elliptic curve over $$\QQ$$
• p - a prime number or None.

OUTPUT: 1 or -1

EXAMPLES: Here is a curve of rank 3:

sage: e = pari([0,0,0,-82,0]).ellinit()
sage: e.ellrootno()
-1
sage: e.ellrootno(2)
1
sage: e.ellrootno(1009)
1

ellsigma(z, flag=0, precision=0)

e.ellsigma(z, flag=0): return the value at the complex point z of the Weierstrass $$\sigma$$ function associated to the elliptic curve e.

EXAMPLES:

sage: e = pari([0,0,0,1,0]).ellinit()
sage: C.<i> = ComplexField()
sage: e.ellsigma(2+i)
1.43490215804166 + 1.80307856719256*I

ellsub(z0, z1)

e.ellsub(z0, z1): return z0-z1 on this elliptic curve.

INPUT:

• e - elliptic curve E
• z0 - point on E
• z1 - point on E

OUTPUT: point on E

EXAMPLES:

sage: e = pari([0, 1, 1, -2, 0]).ellinit()
sage: e.ellsub([1,0], [-1,1])
[0, 0]

elltaniyama(n=-1)
elltors(flag=0)

e.elltors(flag = 0): return information about the torsion subgroup of the elliptic curve e

INPUT:

• e - elliptic curve over $$\QQ$$
• flag (optional) - specify which algorithm to use:
• 0 (default) - use Doud’s algorithm: bound torsion by computing the cardinality of e(GF(p)) for small primes of good reduction, then look for torsion points using Weierstrass parametrization and Mazur’s classification
• 1 - use algorithm given by the Nagell-Lutz theorem (this is much slower)

OUTPUT:

• gen - the order of the torsion subgroup, a.k.a. the number of points of finite order
• gen - vector giving the structure of the torsion subgroup as a product of cyclic groups, sorted in non-increasing order
• gen - vector giving points on e generating these cyclic groups

EXAMPLES:

sage: e = pari([1,0,1,-19,26]).ellinit()
sage: e.elltors()
[12, [6, 2], [[1, 2], [3, -2]]]

ellwp(z='z', n=20, flag=0, precision=0)

Return the value or the series expansion of the Weierstrass $$P$$-function at $$z$$ on the lattice $$self$$ (or the lattice defined by the elliptic curve $$self$$).

INPUT:

• self – an elliptic curve created using ellinit or a list [om1, om2] representing generators for a lattice.
• z – (default: ‘z’) a complex number or a variable name (as string or PARI variable).
• n – (default: 20) if ‘z’ is a variable, compute the series expansion up to at least $$O(z^n)$$.
• flag – (default = 0): If flag is 0, compute only $$P(z)$$. If flag is 1, compute $$[P(z), P'(z)]$$.

OUTPUT:

• $$P(z)$$ (if flag is 0) or $$[P(z), P'(z)]$$ (if flag is 1).

numbers

EXAMPLES:

We first define the elliptic curve X_0(11):

sage: E = pari([0,-1,1,-10,-20]).ellinit()


Compute P(1):

sage: E.ellwp(1)
13.9658695257485


Compute P(1+i), where i = sqrt(-1):

sage: C.<i> = ComplexField()
sage: E.ellwp(pari(1+i))
-1.11510682565555 + 2.33419052307470*I
sage: E.ellwp(1+i)
-1.11510682565555 + 2.33419052307470*I


The series expansion, to the default $$O(z^20)$$ precision:

sage: E.ellwp()
z^-2 + 31/15*z^2 + 2501/756*z^4 + 961/675*z^6 + 77531/41580*z^8 + 1202285717/928746000*z^10 + 2403461/2806650*z^12 + 30211462703/43418875500*z^14 + 3539374016033/7723451736000*z^16 + 413306031683977/1289540602350000*z^18 + O(z^20)


Compute the series for wp to lower precision:

sage: E.ellwp(n=4)
z^-2 + 31/15*z^2 + O(z^4)


Next we use the version where the input is generators for a lattice:

sage: pari([1.2692, 0.63 + 1.45*i]).ellwp(1)
13.9656146936689 + 0.000644829272810...*I


With flag=1, compute the pair P(z) and P’(z):

sage: E.ellwp(1, flag=1)
[13.9658695257485, 50.5619893875144]

ellzeta(z, precision=0)

e.ellzeta(z): return the value at the complex point z of the Weierstrass $$\zeta$$ function associated with the elliptic curve e.

Note

This function has infinitely many poles (one of which is at z=0); attempting to evaluate it too close to one of the poles will result in a PariError.

INPUT:

• e - elliptic curve
• z - complex number

EXAMPLES:

sage: e = pari([0,0,0,1,0]).ellinit()
sage: e.ellzeta(1)
1.06479841295883
sage: C.<i> = ComplexField()
sage: e.ellzeta(i-1)
-0.350122658523049 - 0.350122658523049*I

ellztopoint(z, precision=0)

e.ellztopoint(z): return the point on the elliptic curve e corresponding to the complex number z, under the usual complex uniformization of e by the Weierstrass p-function.

INPUT:

• e - elliptic curve
• z - complex number

OUTPUT point on e

EXAMPLES:

sage: e = pari([0,0,0,1,0]).ellinit()
sage: C.<i> = ComplexField()
sage: e.ellztopoint(1+i)
[0.E-19 - 1.02152286795670*I, -0.149072813701096 - 0.149072813701096*I]


Complex numbers belonging to the period lattice of e are of course sent to the point at infinity on e:

sage: e.ellztopoint(0)
[0]

erfc(x, precision=0)

Return the complementary error function:

$(2/\sqrt{\pi}) \int_{x}^{\infty} e^{-t^2} dt.$

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(1).erfc()
0.157299207050285

eta(x, flag=0, precision=0)

x.eta(flag=0): if flag=0, $$\eta$$ function without the $$q^{1/24}$$; otherwise $$\eta$$ of the complex number $$x$$ in the upper half plane intelligently computed using $$\mathrm{SL}(2,\ZZ)$$ transformations.

DETAILS: This functions computes the following. If the input $$x$$ is a complex number with positive imaginary part, the result is $$\prod_{n=1}^{\infty} (q-1^n)$$, where $$q=e^{2 i \pi x}$$. If $$x$$ is a power series (or can be converted to a power series) with positive valuation, the result is $$\prod_{n=1}^{\infty} (1-x^n)$$.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: C.<i> = ComplexField()
sage: pari(i).eta()
0.998129069925959

eval(*args, **kwds)

Evaluate self with the given arguments.

This is currently implemented in 3 cases:

• univariate polynomials, rational functions, power series and Laurent series (using a single unnamed argument or keyword arguments),
• any PARI object supporting the PARI function substvec (in particular, multivariate polynomials) using keyword arguments,
• objects of type t_CLOSURE (functions in GP bytecode form) using unnamed arguments.

In no case is mixing unnamed and keyword arguments allowed.

EXAMPLES:

sage: f = pari('x^2 + 1')
sage: f.type()
't_POL'
sage: f.eval(I)
0
sage: f.eval(x=2)
5
sage: (1/f).eval(x=1)
1/2


The notation f(x) is an alternative for f.eval(x):

sage: f(3) == f.eval(3)
True


Evaluating power series:

sage: f = pari('1 + x + x^3 + O(x^7)')
sage: f(2*pari('y')^2)
1 + 2*y^2 + 8*y^6 + O(y^14)


Substituting zero is sometimes possible, and trying to do so in illegal cases can raise various errors:

sage: pari('1 + O(x^3)').eval(0)
1
sage: pari('1/x').eval(0)
Traceback (most recent call last):
...
PariError: impossible inverse in gdiv: 0
sage: pari('1/x + O(x^2)').eval(0)
Traceback (most recent call last):
...
ZeroDivisionError: substituting 0 in Laurent series with negative valuation
sage: pari('1/x + O(x^2)').eval(pari('O(x^3)'))
Traceback (most recent call last):
...
PariError: impossible inverse in gdiv: O(x^3)
sage: pari('O(x^0)').eval(0)
Traceback (most recent call last):
...
PariError: domain error in polcoeff: t_SER = O(x^0)


Evaluating multivariate polynomials:

sage: f = pari('y^2 + x^3')
sage: f(1)    # Dangerous, depends on PARI variable ordering
y^2 + 1
sage: f(x=1)  # Safe
y^2 + 1
sage: f(y=1)
x^3 + 1
sage: f(1, 2)
Traceback (most recent call last):
...
TypeError: evaluating PARI t_POL takes exactly 1 argument (2 given)
sage: f(y='x', x='2*y')
x^2 + 8*y^3
sage: f()
x^3 + y^2


It’s not an error to substitute variables which do not appear:

sage: f.eval(z=37)
x^3 + y^2
sage: pari(42).eval(t=0)
42


We can define and evaluate closures as follows:

sage: T = pari('n -> n + 2')
sage: T.type()
't_CLOSURE'
sage: T.eval(3)
5

sage: T = pari('() -> 42')
sage: T()
42

sage: pr = pari('s -> print(s)')
sage: pr.eval('"hello world"')
hello world

sage: f = pari('myfunc(x,y) = x*y')
sage: f.eval(5, 6)
30


Default arguments work, missing arguments are treated as zero (like in GP):

sage: f = pari("(x, y, z=1.0) -> [x, y, z]")
sage: f(1, 2, 3)
[1, 2, 3]
sage: f(1, 2)
[1, 2, 1.00000000000000]
sage: f(1)
[1, 0, 1.00000000000000]
sage: f()
[0, 0, 1.00000000000000]


Using keyword arguments, we can substitute in more complicated objects, for example a number field:

sage: K.<a> = NumberField(x^2 + 1)
sage: nf = K._pari_()
sage: nf
[y^2 + 1, [0, 1], -4, 1, [Mat([1, 0.E-19 + 1.00000000000000*I]), [1, 1.00000000000000; 1, -1.00000000000000], [1, 1; 1, -1], [2, 0; 0, -2], [2, 0; 0, 2], [1, 0; 0, -1], [1, [0, -1; 1, 0]], []], [0.E-19 + 1.00000000000000*I], [1, y], [1, 0; 0, 1], [1, 0, 0, -1; 0, 1, 1, 0]]
sage: nf(y='x')
[x^2 + 1, [0, 1], -4, 1, [Mat([1, 0.E-19 + 1.00000000000000*I]), [1, 1.00000000000000; 1, -1.00000000000000], [1, 1; 1, -1], [2, 0; 0, -2], [2, 0; 0, 2], [1, 0; 0, -1], [1, [0, -1; 1, 0]], []], [0.E-19 + 1.00000000000000*I], [1, x], [1, 0; 0, 1], [1, 0, 0, -1; 0, 1, 1, 0]]

exp(precision=0)

x.exp(): exponential of x.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(0).exp()
1.00000000000000
sage: pari(1).exp()
2.71828182845905
sage: pari('x+O(x^8)').exp()
1 + x + 1/2*x^2 + 1/6*x^3 + 1/24*x^4 + 1/120*x^5 + 1/720*x^6 + 1/5040*x^7 + O(x^8)

factor(limit=-1, proof=1)

Return the factorization of x.

INPUT:

• limit – (default: -1) is optional and can be set whenever x is of (possibly recursive) rational type. If limit is set return partial factorization, using primes up to limit (up to primelimit if limit=0).
• proof – (default: True) optional. If False (not the default), returned factors larger than $$2^{64}$$ may only be pseudoprimes.

Note

In the standard PARI/GP interpreter and C-library the factor command always has proof=False, so beware!

EXAMPLES:

sage: pari('x^10-1').factor()
[x - 1, 1; x + 1, 1; x^4 - x^3 + x^2 - x + 1, 1; x^4 + x^3 + x^2 + x + 1, 1]
sage: pari(2^100-1).factor()
[3, 1; 5, 3; 11, 1; 31, 1; 41, 1; 101, 1; 251, 1; 601, 1; 1801, 1; 4051, 1; 8101, 1; 268501, 1]
sage: pari(2^100-1).factor(proof=False)
[3, 1; 5, 3; 11, 1; 31, 1; 41, 1; 101, 1; 251, 1; 601, 1; 1801, 1; 4051, 1; 8101, 1; 268501, 1]


We illustrate setting a limit:

sage: pari(next_prime(10^50)*next_prime(10^60)*next_prime(10^4)).factor(10^5)
[10007, 1; 100000000000000000000000000000000000000000000000151000000000700000000000000000000000000000000000000000000001057, 1]


PARI doesn’t have an algorithm for factoring multivariate polynomials:

sage: pari('x^3 - y^3').factor()
Traceback (most recent call last):
...
PariError: sorry, factor for general polynomials is not yet implemented

factormod(p, flag=0)

x.factormod(p,flag=0): factorization mod p of the polynomial x using Berlekamp. flag is optional, and can be 0: default or 1: simple factormod, same except that only the degrees of the irreducible factors are given.

factornf(t)

Factorization of the polynomial self over the number field defined by the polynomial t. This does not require that $$t$$ is integral, nor that the discriminant of the number field can be factored.

EXAMPLES:

sage: x = polygen(QQ)
sage: K.<a> = NumberField(x^2 - 1/8)
sage: pari(x^2 - 2).factornf(K.pari_polynomial("a"))
[x + Mod(-4*a, 8*a^2 - 1), 1; x + Mod(4*a, 8*a^2 - 1), 1]


self.factorpadic(p,r=20,flag=0): p-adic factorization of the polynomial x to precision r. flag is optional and may be set to 0 (use round 4) or 1 (use Buchmann-Lenstra)

ffgen(T, v=-1)

Return the generator $$g=x \bmod T$$ of the finite field defined by the polynomial $$T$$.

INPUT:

• T – a gen of type t_POL with coefficients of type t_INTMOD:

a polynomial over a prime finite field

• v – string: a variable name or -1 (optional)

If $$v$$ is a string, then $$g$$ will be a polynomial in $$v$$, else the variable of the polynomial $$T$$ is used.

EXAMPLES:

sage: x = GF(2)['x'].gen()
sage: pari(x^2+x+2).ffgen()
x
sage: pari(x^2+x+1).ffgen('a')
a

ffinit(p, n, v=-1)

Return a monic irreducible polynomial $$g$$ of degree $$n$$ over the finite field of $$p$$ elements.

INPUT:

• p – a gen of type t_INT: a prime number
• n – integer: the degree of the polynomial
• v – string: a variable name or -1 (optional)

If $$v \geq 0', then g$$ will be a polynomial in $$v$$, else the variable $$x$$ is used.

EXAMPLES:

sage: pari(7).ffinit(11)
Mod(1, 7)*x^11 + Mod(1, 7)*x^10 + Mod(4, 7)*x^9 + Mod(5, 7)*x^8 + Mod(1, 7)*x^7 + Mod(1, 7)*x^2 + Mod(1, 7)*x + Mod(6, 7)
sage: pari(2003).ffinit(3)
Mod(1, 2003)*x^3 + Mod(1, 2003)*x^2 + Mod(1993, 2003)*x + Mod(1995, 2003)

fflog(g, o=None)

Return the discrete logarithm of the finite field element self in base $$g$$.

INPUT:

• self – a PARI finite field element (FFELT) in the multiplicative group generated by $$g$$.
• g – the base of the logarithm as a PARI finite field element (FFELT). If $$o$$ is None, this must be a generator of the parent finite field.
• o – either None (then $$g$$ must a primitive root) or the order of $$g$$ or a tuple (o, o.factor()).

OUTPUT:

• An integer $$n$$ such that self = g^n.

EXAMPLES:

sage: k.<a> = GF(2^12)
sage: g = pari(a).ffprimroot()
sage: (g^1234).fflog(g)
1234
sage: pari(k(1)).fflog(g)
0


This element does not generate the full multiplicative group:

sage: b = g^5
sage: ord = b.fforder(); ord
819
sage: (b^555).fflog(b, ord)
555
sage: (b^555).fflog(b, (ord, ord.factor()) )
555

fforder(o=None)

Return the multiplicative order of the finite field element self.

INPUT:

• self – a PARI finite field element (FFELT).
• o – either None or a multiple of the order of $$o$$ or a tuple (o, o.factor()).

OUTPUT:

• The smallest positive integer $$n$$ such that self^n = 1.

EXAMPLES:

sage: k.<a> = GF(5^80)
sage: g = pari(a).ffprimroot()
sage: g.fforder()
82718061255302767487140869206996285356581211090087890624
sage: g.fforder( (5^80-1, factor(5^80-1)) )
82718061255302767487140869206996285356581211090087890624
sage: k(2)._pari_().fforder(o=4)
4

ffprimroot()

Return a primitive root of the multiplicative group of the definition field of the given finite field element.

INPUT:

• self – a PARI finite field element (FFELT)

OUTPUT:

• A generator of the multiplicative group of the finite field generated by self.

EXAMPLES:

sage: x = polygen(GF(3))
sage: k.<a> = GF(9, modulus=x^2+1)
sage: b = pari(a).ffprimroot()
sage: b  # random
a + 1
sage: b.fforder()
8

fibonacci(x)

Return the Fibonacci number of index x.

EXAMPLES:

sage: pari(18).fibonacci()
2584
sage: [pari(n).fibonacci() for n in range(10)]
[0, 1, 1, 2, 3, 5, 8, 13, 21, 34]

floor(x)

For real x: return the largest integer = x. For rational functions: the quotient of numerator by denominator. For lists: apply componentwise.

INPUT:

• x - gen

OUTPUT: gen

EXAMPLES:

sage: pari(5/9).floor()
0
sage: pari(11/9).floor()
1
sage: pari(1.17).floor()
1
sage: pari([1.5,2.3,4.99]).floor()
[1, 2, 4]
sage: pari([[1.1,2.2],[3.3,4.4]]).floor()
[[1, 2], [3, 4]]
sage: pari(x).floor()
x
sage: pari((x^2+x+1)/x).floor()
x + 1
sage: pari(x^2+5*x+2.5).floor()
x^2 + 5*x + 2.50000000000000

sage: pari('"hello world"').floor()
Traceback (most recent call last):
...
PariError: incorrect type in gfloor (t_STR)

frac(x)

frac(x): Return the fractional part of x, which is x - floor(x).

INPUT:

• x - gen

OUTPUT: gen

EXAMPLES:

sage: pari(1.75).frac()
0.750000000000000
sage: pari(sqrt(2)).frac()
0.414213562373095
sage: pari('sqrt(-2)').frac()
Traceback (most recent call last):
...
PariError: incorrect type in gfloor (t_COMPLEX)

galoisapply(aut, x)
galoisfixedfield(perm, flag=0, v=-1)

Compute the fixed field of the Galois group self.

This wraps the galoisfixedfield function from PARI.

INPUT:

• self – A Galois group as generated by galoisinit().
• perm – An element of a Galois group, a vector of such elements, or a subgroup generated by galoissubgroups().
• flag – Amount of data to include in output (see below).
• v – Name of the second variable to use (default: 'y').

OUTPUT:

This depends on the value of flag:

• flag = 0 – A two-element tuple consisting of the defining polynomial of the fixed field and a description of its roots modulo the primes used in the group.
• flag = 1 – Just the polynomial.
• flag = 2 – A third tuple element will describe the factorization of the original polynomial, using the variable indicated by v to stand for a root of the polynomial from the first tuple element.

EXAMPLES:

sage: G = pari(x^4 + 1).galoisinit()
sage: G.galoisfixedfield(G[5][1], flag=2)
[x^2 + 4, Mod(2*x^2, x^4 + 1), [x^2 - 1/2*y, x^2 + 1/2*y]]
sage: G.galoisfixedfield(G[5][5:7])
[x^4 + 1, Mod(x, x^4 + 1)]
sage: L = G.galoissubgroups()
sage: G.galoisfixedfield(L[2], flag=2, v='z')
[x^2 + 2, Mod(x^3 + x, x^4 + 1), [x^2 - z*x - 1, x^2 + z*x - 1]]

galoisinit(den=None)

Calculate the Galois group of self.

This wraps the galoisinit function from PARI.

INPUT:

• self – A number field or a polynomial.
• den – If set, this must be a multiple of the least common denominator of the automorphisms, expressed as polynomials in a root of the defining polynomial.

OUTPUT:

An eight-tuple, represented as a GEN object, with details about the Galois group of the number field. For details see the PARI manual. Note that the element indices in Sage and PARI are 0-based and 1-based, respectively.

EXAMPLES:

sage: P = pari(x^6 + 108)
sage: G = P.galoisinit()
sage: G[0] == P
True
sage: len(G[5]) == prod(G[7])
True

galoisisabelian(flag=0)

Decide whether self is an abelian group.

This wraps the galoisisabelian function from PARI.

INPUT:

• self – A Galois group as generated by galoisinit(), or a subgroup thereof as returned by galoissubgroups().
• flag – Controls the details contained in the returned result.

OUTPUT:

This returns 0 if self is not an abelian group. If it is, then the output depends on flag:

• flag = 0 – The HNF matrix of self over its generators is returned.
• flag = 1 – The return value is simply 1.

EXAMPLES:

sage: G = pari(x^6 + 108).galoisinit()
sage: G.galoisisabelian()
0
sage: H = G.galoissubgroups()[2]
sage: H.galoisisabelian()
Mat(2)
sage: H.galoisisabelian(flag=1)
1

galoisisnormal(subgrp)

Decide whether subgrp is a normal subgroup of self.

This wraps the galoisisnormal function from PARI.

INPUT:

OUTPUT:

One if subgrp is a subgroup of self, zero otherwise.

EXAMPLES:

sage: G = pari(x^6 + 108).galoisinit()
sage: L = G.galoissubgroups()
sage: G.galoisisnormal(L[0])
1
sage: G.galoisisnormal(L[2])
0

galoispermtopol(perm)

Return the polynomial defining the Galois automorphism perm.

This wraps the galoispermtopol function from PARI.

INPUT:

• self – A Galois group as generated by galoisinit().
• perm – A permutation from that group, or a vector or matrix of such permutations.

OUTPUT:

The defining polynomial of the specified automorphism.

EXAMPLES:

sage: G = pari(x^6 + 108).galoisinit()
sage: G.galoispermtopol(G[5])
[x, 1/12*x^4 - 1/2*x, -1/12*x^4 - 1/2*x, 1/12*x^4 + 1/2*x, -1/12*x^4 + 1/2*x, -x]
sage: G.galoispermtopol(G[5][1])
1/12*x^4 - 1/2*x
sage: G.galoispermtopol(G[5][1:4])
[1/12*x^4 - 1/2*x, -1/12*x^4 - 1/2*x, 1/12*x^4 + 1/2*x]

galoissubfields(flag=0, v=-1)

List all subfields of the Galois group self.

This wraps the galoissubfields function from PARI.

This method is essentially the same as applying galoisfixedfield() to each group returned by galoissubgroups().

INPUT:

OUTPUT:

A vector of all subfields of this group. Each entry is as described in the galoisfixedfield() method.

EXAMPLES:

sage: G = pari(x^6 + 108).galoisinit()
sage: G.galoissubfields(flag=1)
[x, x^2 + 972, x^3 + 54, x^3 + 864, x^3 - 54, x^6 + 108]
sage: G = pari(x^4 + 1).galoisinit()
sage: G.galoissubfields(flag=2, v='z')[2]
[x^2 + 2, Mod(x^3 + x, x^4 + 1), [x^2 - z*x - 1, x^2 + z*x - 1]]

galoissubgroups()

List all subgroups of the Galois group self.

This wraps the galoissubgroups function from PARI.

INPUT:

OUTPUT:

A vector of all subgroups of this group. Each subgroup is described as a two-tuple, with the subgroup generators as first element and the orders of these generators as second element.

EXAMPLES:

sage: G = pari(x^6 + 108).galoisinit()
sage: L = G.galoissubgroups()
sage: list(L[0][1])
[3, 2]

gamma(s, precision=0)

s.gamma(precision): Gamma function at s.

If $$s$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$s$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).gamma()
1.00000000000000
sage: pari(5).gamma()
24.0000000000000
sage: C.<i> = ComplexField()
sage: pari(1+i).gamma()
0.498015668118356 - 0.154949828301811*I


TESTS:

sage: pari(-1).gamma()
Traceback (most recent call last):
...
PariError: domain error in gamma: argument = non-positive integer

gammah(s, precision=0)

s.gammah(): Gamma function evaluated at the argument x+1/2.

If $$s$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$s$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).gammah()
1.32934038817914
sage: pari(5).gammah()
52.3427777845535
sage: C.<i> = ComplexField()
sage: pari(1+i).gammah()
0.575315188063452 + 0.0882106775440939*I

gcd(x, y=None)

Return the greatest common divisor of $$x$$ and $$y$$.

If $$y$$ is None, then $$x$$ must be a list or tuple, and the greatest common divisor of its components is returned.

EXAMPLES:

sage: pari(10).gcd(15)
5
sage: pari([5, 'y']).gcd()
1
sage: pari(['x', x^2]).gcd()
x

gequal(a, b)

Check whether $$a$$ and $$b$$ are equal using PARI’s gequal.

EXAMPLES:

sage: a = pari(1); b = pari(1.0); c = pari('"some_string"')
sage: a.gequal(a)
True
sage: b.gequal(b)
True
sage: c.gequal(c)
True
sage: a.gequal(b)
True
sage: a.gequal(c)
False


WARNING: this relation is not transitive:

sage: a = pari('[0]'); b = pari(0); c = pari('[0,0]')
sage: a.gequal(b)
True
sage: b.gequal(c)
True
sage: a.gequal(c)
False

gequal0(a)

Check whether $$a$$ is equal to zero.

EXAMPLES:

sage: pari(0).gequal0()
True
sage: pari(1).gequal0()
False
sage: pari(1e-100).gequal0()
False
sage: pari("0.0 + 0.0*I").gequal0()
True
sage: pari(GF(3^20,'t')(0)).gequal0()
True

gequal_long(a, b)

Check whether $$a$$ is equal to the long int $$b$$ using PARI’s gequalsg.

EXAMPLES:

sage: a = pari(1); b = pari(2.0); c = pari('3*matid(3)')
sage: a.gequal_long(1)
True
sage: a.gequal_long(-1)
False
sage: a.gequal_long(0)
False
sage: b.gequal_long(2)
True
sage: b.gequal_long(-2)
False
sage: c.gequal_long(3)
True
sage: c.gequal_long(-3)
False

getattr(attr)

Return the PARI attribute with the given name.

EXAMPLES:

sage: K = pari("nfinit(x^2 - x - 1)")
sage: K.getattr("pol")
x^2 - x - 1
sage: K.getattr("disc")
5

sage: K.getattr("reg")
Traceback (most recent call last):
...
PariError: _.reg: incorrect type in reg (t_VEC)
sage: K.getattr("zzz")
Traceback (most recent call last):
...
PariError: no function named "_.zzz"

hilbert(x, y, p)
hyperu(a, b, x, precision=0)

a.hyperu(b,x): U-confluent hypergeometric function.

If $$a$$, $$b$$, or $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their precisions is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(1).hyperu(2,3)
0.333333333333333

idealappr(x, flag=0)
idealchinese(x, y)

Chinese Remainder Theorem over number fields.

INPUT:

• x – prime ideal factorization
• y – vector of elements

OUTPUT:

An element b in the ambient number field self such that $$v_p(b-y_p) \ge v_p(x)$$ for all prime ideals $$p$$ dividing $$x$$, and $$v_p(b) \ge 0$$ for all other $$p$$.

EXAMPLES:

sage: F = QuadraticField(5, 'alpha')
sage: nf = F._pari_()
sage: P = F.ideal(F.gen())
sage: Q = F.ideal(2)
sage: moduli = pari.matrix(2,2,[P.pari_prime(),4,Q.pari_prime(),4])
sage: residues = pari.vector(2,[0,1])
sage: b = F(nf.idealchinese(moduli,residues))
sage: b.valuation(P) >= 4
True
sage: (b-1).valuation(Q) >= 2
True

idealcoprime(x, y)

Given two integral ideals x and y of a pari number field self, return an element a of the field (expressed in the integral basis of self) such that a*x is an integral ideal coprime to y.

EXAMPLES:

sage: F = NumberField(x^3-2, 'alpha')
sage: nf = F._pari_()
sage: x = pari('[1, -1, 2]~')
sage: y = pari('[1, -1, 3]~')
sage: nf.idealcoprime(x, y)
[1, 0, 0]~

sage: y = pari('[2, -2, 4]~')
sage: nf.idealcoprime(x, y)
[5/43, 9/43, -1/43]~

idealdiv(x, y, flag=0)
idealfactor(x)
idealhnf(a, b=None)
idealintersection(x, y)
ideallist(bound, flag=4)

Vector of vectors $$L$$ of all idealstar of all ideals of $$norm <= bound$$.

The binary digits of flag mean:

• 1: give generators;
• 4: (default) give only the ideals and not the bid.

EXAMPLES:

sage: R.<x> = PolynomialRing(QQ)
sage: K.<a> = NumberField(x^2 + 1)
sage: L = K.pari_nf().ideallist(100)


Now we have our list $$L$$. Entry $$L[n-1]$$ contains all ideals of norm $$n$$:

sage: L[0]   # One ideal of norm 1.
[[1, 0; 0, 1]]
sage: L[64]  # 4 ideals of norm 65.
[[65, 8; 0, 1], [65, 47; 0, 1], [65, 18; 0, 1], [65, 57; 0, 1]]

ideallog(x, bid)

Return the discrete logarithm of the unit x in (ring of integers)/bid.

INPUT:

• self - a pari number field
• bid - a big ideal structure (corresponding to an ideal I of self) output by idealstar
• x - an element of self with valuation zero at all primes dividing I

OUTPUT:

• the discrete logarithm of x on the generators given in bid[2]

EXAMPLE:

sage: F = NumberField(x^3-2, 'alpha')
sage: nf = F._pari_()
sage: I = pari('[1, -1, 2]~')
sage: bid = nf.idealstar(I)
sage: x = pari('5')
sage: nf.ideallog(x, bid)
[25]~

idealmul(x, y, flag=0)
idealnorm(x)
idealprimedec(nf, p)

Prime ideal decomposition of the prime number $$p$$ in the number field $$nf$$ as a vector of 5 component vectors $$[p,a,e,f,b]$$ representing the prime ideals $$p O_K + a O_K$$, $$e$$ ,f as usual, $$a$$ as vector of components on the integral basis, $$b$$ Lenstra’s constant.

EXAMPLES:

sage: K.<i> = QuadraticField(-1)
sage: F = pari(K).idealprimedec(5); F
[[5, [-2, 1]~, 1, 1, [2, -1; 1, 2]], [5, [2, 1]~, 1, 1, [-2, -1; 1, -2]]]
sage: F[0].pr_get_p()
5

idealred(I, vdir=0)
idealstar(I, flag=1)

Return the big ideal (bid) structure of modulus I.

INPUT:

• self - a pari number field
• I – an ideal of self, or a row vector whose first component is an ideal and whose second component is a row vector of r_1 0 or 1.
• flag - determines the amount of computation and the shape of the output:
• 1 (default): return a bid structure without generators
• 2: return a bid structure with generators (slower)
• 0 (deprecated): only outputs units of (ring of integers/I) as an abelian group, i.e as a 3-component vector [h,d,g]: h is the order, d is the vector of SNF cyclic components and g the corresponding generators. This flag is deprecated: it is in fact slightly faster to compute a true bid structure, which contains much more information.

EXAMPLE:

sage: F = NumberField(x^3-2, 'alpha')
sage: nf = F._pari_()
sage: I = pari('[1, -1, 2]~')
sage: nf.idealstar(I)
[[[43, 9, 5; 0, 1, 0; 0, 0, 1], [0]], [42, [42]], Mat([[43, [9, 1, 0]~, 1, 1, [-5, 2, -18; -9, -5, 2; 1, -9, -5]], 1]), [[[[42], [3], [3], [Vecsmall([])], 1]], [[], [], []]], Mat(1)]

idealtwoelt(x, a=None)
idealval(x, p)
imag(x)

imag(x): Return the imaginary part of x. This function also works component-wise.

INPUT:

• x - gen

OUTPUT: gen

EXAMPLES:

sage: pari('1+2*I').imag()
2
sage: pari(sqrt(-2)).imag()
1.41421356237310
sage: pari('x+I').imag()
1
sage: pari('x+2*I').imag()
2
sage: pari('(1+I)*x^2+2*I').imag()
x^2 + 2
sage: pari('[1,2,3] + [4*I,5,6]').imag()
[4, 0, 0]

incgam(s, x, y=None, precision=0)

s.incgam(x, y, precision): incomplete gamma function. y is optional and is the precomputed value of gamma(s).

If $$s$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$s$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: C.<i> = ComplexField()
sage: pari(1+i).incgam(3-i)
-0.0458297859919946 + 0.0433696818726677*I

incgamc(s, x, precision=0)

s.incgamc(x): complementary incomplete gamma function.

The arguments $$x$$ and $$s$$ are complex numbers such that $$s$$ is not a pole of $$\Gamma$$ and $$|x|/(|s|+1)$$ is not much larger than $$1$$ (otherwise, the convergence is very slow). The function returns the value of the integral $$\int_{0}^{x} e^{-t} t^{s-1} dt.$$

If $$s$$ or $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their precisions is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(1).incgamc(2)
0.864664716763387

intformal(y=-1)

x.intformal(y): formal integration of x with respect to the main variable of y, or to the main variable of x if y is omitted

ispower(k=None)

Determine whether or not self is a perfect k-th power. If k is not specified, find the largest k so that self is a k-th power.

INPUT:

• k - int (optional)

OUTPUT:

• power - int, what power it is
• g - what it is a power of

EXAMPLES:

sage: pari(9).ispower()
(2, 3)
sage: pari(17).ispower()
(1, 17)
sage: pari(17).ispower(2)
(False, None)
sage: pari(17).ispower(1)
(1, 17)
sage: pari(2).ispower()
(1, 2)

isprime(flag=0)

isprime(x, flag=0): Returns True if x is a PROVEN prime number, and False otherwise.

INPUT:

• flag - int 0 (default): use a combination of algorithms. 1: certify primality using the Pocklington-Lehmer Test. 2: certify primality using the APRCL test.

OUTPUT:

• bool - True or False

EXAMPLES:

sage: pari(9).isprime()
False
sage: pari(17).isprime()
True
sage: n = pari(561)    # smallest Carmichael number
sage: n.isprime()      # not just a pseudo-primality test!
False
sage: n.isprime(1)
False
sage: n.isprime(2)
False
sage: n = pari(2^31-1)
sage: n.isprime(1)
(True, [2, 3, 1; 3, 5, 1; 7, 3, 1; 11, 3, 1; 31, 2, 1; 151, 3, 1; 331, 3, 1])

ispseudoprime(flag=0)

ispseudoprime(x, flag=0): Returns True if x is a pseudo-prime number, and False otherwise.

INPUT:

• flag - int 0 (default): checks whether x is a Baillie-Pomerance-Selfridge-Wagstaff pseudo prime (strong Rabin-Miller pseudo prime for base 2, followed by strong Lucas test for the sequence (P,-1), P smallest positive integer such that $$P^2 - 4$$ is not a square mod x). 0: checks whether x is a strong Miller-Rabin pseudo prime for flag randomly chosen bases (with end-matching to catch square roots of -1).

OUTPUT:

• bool - True or False, or when flag=1, either False or a tuple (True, cert) where cert is a primality certificate.

EXAMPLES:

sage: pari(9).ispseudoprime()
False
sage: pari(17).ispseudoprime()
True
sage: n = pari(561)     # smallest Carmichael number
sage: n.ispseudoprime(2)
False

issquare(x, find_root=False)

issquare(x,n): True if x is a square, False if not. If find_root is given, also returns the exact square root.

issquarefree()

EXAMPLES:

sage: pari(10).issquarefree()
True
sage: pari(20).issquarefree()
False

j()

e.j(): return the j-invariant of the elliptic curve e.

EXAMPLES:

sage: e = pari([0, -1, 1, -10, -20]).ellinit()
sage: e.j()
-122023936/161051
sage: _.factor()
[-1, 1; 2, 12; 11, -5; 31, 3]

kronecker(y)
lcm(x, y=None)

Return the least common multiple of $$x$$ and $$y$$.

If $$y$$ is None, then $$x$$ must be a list or tuple, and the least common multiple of its components is returned.

EXAMPLES:

sage: pari(10).lcm(15)
30
sage: pari([5, 'y']).lcm()
5*y
sage: pari([10, 'x', x^2]).lcm()
10*x^2

length()
lex(x, y)

lex(x,y): Compare x and y lexicographically (1 if xy, 0 if x==y, -1 if xy)

lift(x, v=-1)

lift(x,v): Returns the lift of an element of Z/nZ to Z or R[x]/(P) to R[x] for a type R if v is omitted. If v is given, lift only polymods with main variable v. If v does not occur in x, lift only intmods.

INPUT:

• x - gen
• v - (optional) variable

OUTPUT: gen

EXAMPLES:

sage: x = pari("x")
sage: a = x.Mod('x^3 + 17*x + 3')
sage: a
Mod(x, x^3 + 17*x + 3)
sage: b = a^4; b
Mod(-17*x^2 - 3*x, x^3 + 17*x + 3)
sage: b.lift()
-17*x^2 - 3*x


??? more examples

lindep(flag=0)
list()

Convert self to a list of PARI gens.

EXAMPLES:

A PARI vector becomes a Sage list:

sage: L = pari("vector(10,i,i^2)").list()
sage: L
[1, 4, 9, 16, 25, 36, 49, 64, 81, 100]
sage: type(L)
<type 'list'>
sage: type(L[0])
<type 'sage.libs.pari.gen.gen'>


For polynomials, list() behaves as for ordinary Sage polynomials:

sage: pol = pari("x^3 + 5/3*x"); pol.list()
[0, 5/3, 0, 1]


For power series or Laurent series, we get all coefficients starting from the lowest degree term. This includes trailing zeros:

sage: R.<x> = LaurentSeriesRing(QQ)
sage: s = x^2 + O(x^8)
sage: s.list()
[1]
sage: pari(s).list()
[1, 0, 0, 0, 0, 0]
sage: s = x^-2 + O(x^0)
sage: s.list()
[1]
sage: pari(s).list()
[1, 0]


For matrices, we get a list of columns:

sage: M = matrix(ZZ,3,2,[1,4,2,5,3,6]); M
[1 4]
[2 5]
[3 6]
sage: pari(M).list()
[[1, 2, 3]~, [4, 5, 6]~]


For “scalar” types, we get a 1-element list containing self:

sage: pari("42").list()
[42]

list_str()

Return str that might correctly evaluate to a Python-list.

listinsert(obj, n)
listput(obj, n)
lllgram()
lllgramint()
lngamma(x, precision=0)

Alias for log_gamma().

EXAMPLES:

sage: pari(100).lngamma()
359.134205369575

log(x, precision=0)

x.log(): natural logarithm of x.

This function returns the principal branch of the natural logarithm of $$x$$, i.e., the branch such that $$\Im(\log(x)) \in ]-\pi, \pi].$$ The result is complex (with imaginary part equal to $$\pi$$) if $$x\in \RR$$ and $$x<0$$. In general, the algorithm uses the formula

$\log(x) \simeq \frac{\pi}{2{\rm agm}(1,4/s)} - m\log(2),$

if $$s=x 2^m$$ is large enough. (The result is exact to $$B$$ bits provided that $$s>2^{B/2}$$.) At low accuracies, this function computes $$\log$$ using the series expansion near $$1$$.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

Note that $$p$$-adic arguments can also be given as input, with the convention that $$\log(p)=0$$. Hence, in particular, $$\exp(\log(x))/x$$ is not in general equal to $$1$$ but instead to a $$(p-1)$$-st root of unity (or $$\pm 1$$ if $$p=2$$) times a power of $$p$$.

EXAMPLES:

sage: pari(5).log()
1.60943791243410
sage: C.<i> = ComplexField()
sage: pari(i).log()
0.E-19 + 1.57079632679490*I

log_gamma(x, precision=0)

Logarithm of the gamma function of x.

This function returns the principal branch of the logarithm of the gamma function of $$x$$. The function $$\log(\Gamma(x))$$ is analytic on the complex plane with non-positive integers removed. This function can have much larger inputs than $$\Gamma$$ itself.

The $$p$$-adic analogue of this function is unfortunately not implemented.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(100).log_gamma()
359.134205369575


EXAMPLES:

sage: pari('[1,2,3; 4,5,6;  7,8,9]').matadjoint()
[-3, 6, -3; 6, -12, 6; -3, 6, -3]
[(i*e - h*f), (-i*b + h*c), (f*b - e*c); (-i*d + g*f), i*a - g*c, -f*a + d*c; (h*d - g*e), -h*a + g*b, e*a - d*b]

matdet(flag=0)

Return the determinant of this matrix.

INPUT:

• flag - (optional) flag 0: using Gauss-Bareiss. 1: use classical Gaussian elimination (slightly better for integer entries)

EXAMPLES:

sage: pari('[1,2; 3,4]').matdet(0)
-2
sage: pari('[1,2; 3,4]').matdet(1)
-2

matfrobenius(flag=0)

M.matfrobenius(flag=0): Return the Frobenius form of the square matrix M. If flag is 1, return only the elementary divisors (a list of polynomials). If flag is 2, return a two-components vector [F,B] where F is the Frobenius form and B is the basis change so that $$M=B^{-1} F B$$.

EXAMPLES:

sage: a = pari('[1,2;3,4]')
sage: a.matfrobenius()
[0, 2; 1, 5]
sage: a.matfrobenius(flag=1)
[x^2 - 5*x - 2]
sage: a.matfrobenius(2)
[[0, 2; 1, 5], [1, -1/3; 0, 1/3]]
sage: v = a.matfrobenius(2)
sage: v[0]
[0, 2; 1, 5]
sage: v[1]^(-1)*v[0]*v[1]
[1, 2; 3, 4]


We let t be the matrix of $$T_2$$ acting on modular symbols of level 43, which was computed using ModularSymbols(43,sign=1).T(2).matrix():

sage: t = pari('[3, -2, 0, 0; 0, -2, 0, 1; 0, -1, -2, 2; 0, -2, 0, 2]')
sage: t.matfrobenius()
[0, 0, 0, -12; 1, 0, 0, -2; 0, 1, 0, 8; 0, 0, 1, 1]
sage: t.charpoly('x')
x^4 - x^3 - 8*x^2 + 2*x + 12
sage: t.matfrobenius(1)
[x^4 - x^3 - 8*x^2 + 2*x + 12]


AUTHORS:

• Martin Albrect (2006-04-02)
mathnf(flag=0)

A.mathnf(flag=0): (upper triangular) Hermite normal form of A, basis for the lattice formed by the columns of A.

INPUT:

• flag - optional, value range from 0 to 4 (0 if omitted), meaning : 0: naive algorithm
• 1: Use Batut's algorithm - output 2-component vector [H,U] such that H is the HNF of A, and U is a unimodular matrix such that xU=H. 3: Use Batut’s algorithm. Output [H,U,P] where P is a permutation matrix such that P A U = H. 4: As 1, using a heuristic variant of LLL reduction along the way.

EXAMPLES:

sage: pari('[1,2,3; 4,5,6;  7,8,9]').mathnf()
[6, 1; 3, 1; 0, 1]

mathnfmod(d)

Returns the Hermite normal form if d is a multiple of the determinant

Beware that PARI’s concept of a Hermite normal form is an upper triangular matrix with the same column space as the input matrix.

INPUT:

• d - multiple of the determinant of self

EXAMPLES:

       sage: M=matrix([[1,2,3],[4,5,6],[7,8,11]])
sage: d=M.det()
sage: pari(M).mathnfmod(d)
[6, 4, 3; 0, 1, 0; 0, 0, 1]


Note that d really needs to be a multiple of the discriminant, not just of the exponent of the cokernel:

       sage: M=matrix([[1,0,0],[0,2,0],[0,0,6]])
sage: pari(M).mathnfmod(6)
[1, 0, 0; 0, 1, 0; 0, 0, 6]
sage: pari(M).mathnfmod(12)
[1, 0, 0; 0, 2, 0; 0, 0, 6]

mathnfmodid(d)

Returns the Hermite Normal Form of M concatenated with d*Identity

Beware that PARI’s concept of a Hermite normal form is a maximal rank upper triangular matrix with the same column space as the input matrix.

INPUT:

• d - Determines

EXAMPLES:

       sage: M=matrix([[1,0,0],[0,2,0],[0,0,6]])
sage: pari(M).mathnfmodid(6)
[1, 0, 0; 0, 2, 0; 0, 0, 6]


This routine is not completely equivalent to mathnfmod:

sage: pari(M).mathnfmod(6)
[1, 0, 0; 0, 1, 0; 0, 0, 6]

matker(flag=0)

Return a basis of the kernel of this matrix.

INPUT:

• flag - optional; may be set to 0: default; non-zero: x is known to have integral entries.

EXAMPLES:

sage: pari('[1,2,3;4,5,6;7,8,9]').matker()
[1; -2; 1]


With algorithm 1, even if the matrix has integer entries the kernel need not be saturated (which is weird):

sage: pari('[1,2,3;4,5,6;7,8,9]').matker(1)
[3; -6; 3]
sage: pari('matrix(3,3,i,j,i)').matker()
[-1, -1; 1, 0; 0, 1]
sage: pari('[1,2,3;4,5,6;7,8,9]*Mod(1,2)').matker()
[Mod(1, 2); Mod(0, 2); Mod(1, 2)]

matkerint(flag=0)

Return the integer kernel of a matrix.

This is the LLL-reduced Z-basis of the kernel of the matrix x with integral entries.

INPUT:

• flag - optional, and may be set to 0: default, uses a modified LLL, 1: uses matrixqz.

EXAMPLES:

sage: pari('[2,1;2,1]').matker()
[-1/2; 1]
sage: pari('[2,1;2,1]').matkerint()
[1; -2]
sage: pari('[2,1;2,1]').matkerint(1)
[1; -2]

matsnf(flag=0)

x.matsnf(flag=0): Smith normal form (i.e. elementary divisors) of the matrix x, expressed as a vector d. Binary digits of flag mean 1: returns [u,v,d] where d=u*x*v, otherwise only the diagonal d is returned, 2: allow polynomial entries, otherwise assume x is integral, 4: removes all information corresponding to entries equal to 1 in d.

EXAMPLES:

sage: pari('[1,2,3; 4,5,6;  7,8,9]').matsnf()
[0, 3, 1]

matsolve(B)

matsolve(B): Solve the linear system Mx=B for an invertible matrix M

matsolve(B) uses Gaussian elimination to solve Mx=B, where M is invertible and B is a column vector.

The corresponding pari library routine is gauss. The gp-interface name matsolve has been given preference here.

INPUT:

• B - a column vector of the same dimension as the square matrix self

EXAMPLES:

sage: pari('[1,1;1,-1]').matsolve(pari('[1;0]'))
[1/2; 1/2]

matsolvemod(D, B, flag=0)

For column vectors $$D=(d_i)$$ and $$B=(b_i)$$, find a small integer solution to the system of linear congruences

$R_ix=b_i\text{ (mod }d_i),$

where $$R_i$$ is the ith row of self. If $$d_i=0$$, the equation is considered over the integers. The entries of self, D, and B should all be integers (those of D should also be non-negative).

If flag is 1, the output is a two-component row vector whose first component is a solution and whose second component is a matrix whose columns form a basis of the solution set of the homogeneous system.

For either value of flag, the output is 0 if there is no solution.

Note that if D or B is an integer, then it will be considered as a vector all of whose entries are that integer.

EXAMPLES:

sage: D = pari('[3,4]~')
sage: B = pari('[1,2]~')
sage: M = pari('[1,2;3,4]')
sage: M.matsolvemod(D, B)
[-2, 0]~
sage: M.matsolvemod(3, 1)
[-1, 1]~
sage: M.matsolvemod(pari('[3,0]~'), pari('[1,2]~'))
[6, -4]~
sage: M2 = pari('[1,10;9,18]')
sage: M2.matsolvemod(3, pari('[2,3]~'), 1)
[[0, -1]~, [-1, -2; 1, -1]]
sage: M2.matsolvemod(9, pari('[2,3]~'))
0
sage: M2.matsolvemod(9, pari('[2,45]~'), 1)
[[1, 1]~, [-1, -4; 1, -5]]

mattranspose()

Transpose of the matrix self.

EXAMPLES:

sage: pari('[1,2,3; 4,5,6; 7,8,9]').mattranspose()
[1, 4, 7; 2, 5, 8; 3, 6, 9]

max(x, y)

max(x,y): Return the maximum of x and y.

min(x, y)

min(x,y): Return the minimum of x and y.

mod()

Given an INTMOD or POLMOD Mod(a,m), return the modulus $$m$$.

EXAMPLES:

sage: pari(4).Mod(5).mod()
5
sage: pari("Mod(x, x*y)").mod()
y*x
sage: pari("[Mod(4,5)]").mod()
Traceback (most recent call last):
...
TypeError: Not an INTMOD or POLMOD in mod()

modreverse()

modreverse(x): reverse polymod of the polymod x, if it exists.

EXAMPLES:

moebius(x)

moebius(x): Moebius function of x.

ncols()

Return the number of columns of self.

EXAMPLES:

sage: pari('matrix(19,8)').ncols()
8

newtonpoly(p)

x.newtonpoly(p): Newton polygon of polynomial x with respect to the prime p.

EXAMPLES:

sage: x = pari('y^8+6*y^6-27*y^5+1/9*y^2-y+1')
sage: x.newtonpoly(3)
[1, 1, -1/3, -1/3, -1/3, -1/3, -1/3, -1/3]


nextprime(x): smallest pseudoprime greater than or equal to $$x$$. If add_one is non-zero, return the smallest pseudoprime strictly greater than $$x$$.

EXAMPLES:

sage: pari(1).nextprime()
2
sage: pari(2).nextprime()
2
3
sage: pari(2^100).nextprime()
1267650600228229401496703205653

nf_get_diff()

Returns the different of this number field as a PARI ideal.

INPUT:

• self – A PARI number field being the output of nfinit(),

bnfinit() or bnrinit().

EXAMPLES:

sage: K.<a> = NumberField(x^4 - 4*x^2 + 1)
sage: pari(K).nf_get_diff()
[12, 0, 0, 0; 0, 12, 8, 0; 0, 0, 4, 0; 0, 0, 0, 4]

nf_get_pol()

Returns the defining polynomial of this number field.

INPUT:

• self – A PARI number field being the output of nfinit(),

bnfinit() or bnrinit().

EXAMPLES:

sage: K.<a> = NumberField(x^4 - 4*x^2 + 1)
sage: pari(K).nf_get_pol()
y^4 - 4*y^2 + 1
sage: bnr = pari("K = bnfinit(x^4 - 4*x^2 + 1); bnrinit(K, 2*x)")
sage: bnr.nf_get_pol()
x^4 - 4*x^2 + 1


For relative number fields, this returns the relative polynomial. However, beware that pari(L) returns an absolute number field:

sage: L.<b> = K.extension(x^2 - 5)
sage: pari(L).nf_get_pol()        # Absolute
y^8 - 28*y^6 + 208*y^4 - 408*y^2 + 36
sage: L.pari_rnf().nf_get_pol()   # Relative
x^2 - 5


TESTS:

sage: x = polygen(QQ)
sage: K.<a> = NumberField(x^4 - 4*x^2 + 1)
sage: K.pari_nf().nf_get_pol()
y^4 - 4*y^2 + 1
sage: K.pari_bnf().nf_get_pol()
y^4 - 4*y^2 + 1


An error is raised for invalid input:

sage: pari("[0]").nf_get_pol()
Traceback (most recent call last):
...
PariError: incorrect type in pol (t_VEC)

nf_get_sign()

Returns a Python list [r1, r2], where r1 and r2 are Python ints representing the number of real embeddings and pairs of complex embeddings of this number field, respectively.

INPUT:

• self – A PARI number field being the output of nfinit(),

bnfinit() or bnrinit().

EXAMPLES:

sage: K.<a> = NumberField(x^4 - 4*x^2 + 1)
sage: s = K.pari_nf().nf_get_sign(); s
[4, 0]
sage: type(s); type(s[0])
<type 'list'>
<type 'int'>
sage: CyclotomicField(15).pari_nf().nf_get_sign()
[0, 4]

nf_get_zk()

Returns a vector with a $$\ZZ$$-basis for the ring of integers of this number field. The first element is always $$1$$.

INPUT:

• self – A PARI number field being the output of nfinit(),

bnfinit() or bnrinit().

EXAMPLES:

sage: K.<a> = NumberField(x^4 - 4*x^2 + 1)
sage: pari(K).nf_get_zk()
[1, y, y^3 - 4*y, y^2 - 2]

nf_subst(z)

Given a PARI number field self, return the same PARI number field but in the variable z.

INPUT:

• self – A PARI number field being the output of nfinit(),

bnfinit() or bnrinit().

EXAMPLES:

sage: x = polygen(QQ)
sage: K = NumberField(x^2 + 5, 'a')


We can substitute in a PARI nf structure:

sage: Kpari = K.pari_nf()
sage: Kpari.nf_get_pol()
y^2 + 5
sage: Lpari = Kpari.nf_subst('a')
sage: Lpari.nf_get_pol()
a^2 + 5


We can also substitute in a PARI bnf structure:

sage: Kpari = K.pari_bnf()
sage: Kpari.nf_get_pol()
y^2 + 5
sage: Kpari.bnf_get_cyc()  # Structure of class group
[2]
sage: Lpari = Kpari.nf_subst('a')
sage: Lpari.nf_get_pol()
a^2 + 5
sage: Lpari.bnf_get_cyc()  # We still have a bnf after substituting
[2]

nfbasis(flag=0, fa=None)

Integral basis of the field $$\QQ[a]$$, where a is a root of the polynomial x.

INPUT:

• flag: if set to 1 and fa is not given: assume that no square of a prime > 500000 divides the discriminant of x.

• fa: If present, encodes a subset of primes at which to check for maximality. This must be one of the three following things:

• an integer: check all primes up to fa using trial division.
• a vector: a list of primes to check.
• a matrix: a partial factorization of the discriminant of x.

Note

In earlier versions of Sage, other bits in flag were defined but these are now simply ignored.

EXAMPLES:

sage: pari('x^3 - 17').nfbasis()
[1, x, 1/3*x^2 - 1/3*x + 1/3]


We test flag = 1, noting it gives a wrong result when the discriminant (-4 * $$p^2 * q$$ in the example below) has a big square factor:

sage: p = next_prime(10^10); q = next_prime(p)
sage: x = polygen(QQ); f = x^2 + p^2*q
sage: pari(f).nfbasis(1)   # Wrong result
[1, x]
sage: pari(f).nfbasis()    # Correct result
[1, 1/10000000019*x]
sage: pari(f).nfbasis(fa=10^6)   # Check primes up to 10^6: wrong result
[1, x]
sage: pari(f).nfbasis(fa="[2,2; %s,2]"%p)    # Correct result and faster
[1, 1/10000000019*x]
sage: pari(f).nfbasis(fa=[2,p])              # Equivalent with the above
[1, 1/10000000019*x]

nfbasis_d(flag=0, fa=None)

Like nfbasis(), but return a tuple (B, D) where $$B$$ is the integral basis and $$D$$ the discriminant.

EXAMPLES:

sage: F = NumberField(x^3-2,'alpha')
sage: F._pari_()[0].nfbasis_d()
([1, y, y^2], -108)

sage: G = NumberField(x^5-11,'beta')
sage: G._pari_()[0].nfbasis_d()
([1, y, y^2, y^3, y^4], 45753125)

sage: pari([-2,0,0,1]).Polrev().nfbasis_d()
([1, x, x^2], -108)

nfbasistoalg(nf, x)

Transforms the column vector x on the integral basis into an algebraic number.

INPUT:

• nf – a number field
• x – a column of rational numbers of length equal to the degree of nf or a single rational number

OUTPUT:

• A POLMOD representing the element of nf whose coordinates are x in the Z-basis of nf.

EXAMPLES:

sage: x = polygen(QQ)
sage: K.<a> = NumberField(x^3 - 17)
sage: Kpari = K.pari_nf()
sage: Kpari.getattr('zk')
[1, 1/3*y^2 - 1/3*y + 1/3, y]
sage: Kpari.nfbasistoalg(42)
Mod(42, y^3 - 17)
sage: Kpari.nfbasistoalg("[3/2, -5, 0]~")
Mod(-5/3*y^2 + 5/3*y - 1/6, y^3 - 17)
sage: Kpari.getattr('zk') * pari("[3/2, -5, 0]~")
-5/3*y^2 + 5/3*y - 1/6

nfbasistoalg_lift(nf, x)

Transforms the column vector x on the integral basis into a polynomial representing the algebraic number.

INPUT:

• nf – a number field
• x – a column of rational numbers of length equal to the degree of nf or a single rational number

OUTPUT:

• nf.nfbasistoalg(x).lift()

EXAMPLES:

sage: x = polygen(QQ)
sage: K.<a> = NumberField(x^3 - 17)
sage: Kpari = K.pari_nf()
sage: Kpari.getattr('zk')
[1, 1/3*y^2 - 1/3*y + 1/3, y]
sage: Kpari.nfbasistoalg_lift(42)
42
sage: Kpari.nfbasistoalg_lift("[3/2, -5, 0]~")
-5/3*y^2 + 5/3*y - 1/6
sage: Kpari.getattr('zk') * pari("[3/2, -5, 0]~")
-5/3*y^2 + 5/3*y - 1/6

nfdisc(flag=0, p=0)

nfdisc(x): Return the discriminant of the number field defined over QQ by x.

EXAMPLES:

sage: F = NumberField(x^3-2,'alpha')
sage: F._pari_()[0].nfdisc()
-108

sage: G = NumberField(x^5-11,'beta')
sage: G._pari_()[0].nfdisc()
45753125

sage: f = x^3-2
sage: f._pari_()
x^3 - 2
sage: f._pari_().nfdisc()
-108

nfeltdiveuc(x, y)

Given $$x$$ and $$y$$ in the number field self, return $$q$$ such that $$x - q y$$ is “small”.

EXAMPLES:

sage: k.<a> = NumberField(x^2 + 5)
sage: x = 10
sage: y = a + 1
sage: pari(k).nfeltdiveuc(pari(x), pari(y))
[2, -2]~

nfeltreduce(x, I)

Given an ideal I in Hermite normal form and an element x of the pari number field self, finds an element r in self such that x-r belongs to the ideal and r is small.

EXAMPLES:

sage: k.<a> = NumberField(x^2 + 5)
sage: I = k.ideal(a)
sage: kp = pari(k)
sage: kp.nfeltreduce(12, I.pari_hnf())
[2, 0]~
sage: 12 - k(kp.nfeltreduce(12, I.pari_hnf())) in I
True

nffactor(x)
nfgaloisconj(flag=0, denom=None, precision=0)

Edited from the pari documentation:

nfgaloisconj(nf): list of conjugates of a root of the polynomial x=nf.pol in the same number field.

Uses a combination of Allombert’s algorithm and nfroots.

EXAMPLES:

sage: x = QQ['x'].0; nf = pari(x^2 + 2).nfinit()
sage: nf.nfgaloisconj()
[-x, x]~
sage: nf = pari(x^3 + 2).nfinit()
sage: nf.nfgaloisconj()
[x]~
sage: nf = pari(x^4 + 2).nfinit()
sage: nf.nfgaloisconj()
[-x, x]~

nfgenerator()
nfhilbert(a, b, p=None)

nfhilbert(nf,a,b,{p}): if p is omitted, global Hilbert symbol (a,b) in nf, that is 1 if X^2-aY^2-bZ^2 has a non-trivial solution (X,Y,Z) in nf, -1 otherwise. Otherwise compute the local symbol modulo the prime ideal p.

EXAMPLES:

sage: x = polygen(QQ)
sage: K.<t> = NumberField(x^3 - x + 1)
sage: pari(K).nfhilbert(t, t + 2)
-1
sage: P = K.ideal(t^2 + t - 2)   # Prime ideal above 5
sage: pari(K).nfhilbert(t, t + 2, P.pari_prime())
-1
sage: P = K.ideal(t^2 + 3*t - 1) # Prime ideal above 23, ramified
sage: pari(K).nfhilbert(t, t + 2, P.pari_prime())
1

nfhnf(x)

nfhnf(nf,x) : given a pseudo-matrix (A, I) or an integral pseudo-matrix (A,I,J), finds a pseudo-basis in Hermite normal form of the module it generates.

A pseudo-matrix is a 2-component row vector (A, I) where A is a relative m x n matrix and I an ideal list of length n. An integral pseudo-matrix is a 3-component row vector (A, I, J).

Note

The definition of a pseudo-basis ([Cohen]): Let M be a finitely generated, torsion-free R-module, and set V = KM. If $$\mathfrak{a}_i$$ are fractional ideals of R and $$w_i$$ are elements of V, we say that $$(w_i, \mathfrak{a}_k)_{1 \leq i \leq k}$$ is a pseudo-basis of M if $$M = \mathfrak{a}_1 w_1 \oplus \cdots \oplus \mathfrak{a}_k w_k.$$

REFERENCES:

 [Cohen] Cohen, “Advanced Topics in Computational Number Theory”

EXAMPLES:

sage: F.<a> = NumberField(x^2-x-1)
sage: Fp = pari(F)
sage: A = matrix(F,[[1,2,a,3],[3,0,a+2,0],[0,0,a,2],[3+a,a,0,1]])
sage: I = [F.ideal(-2*a+1),F.ideal(7), F.ideal(3),F.ideal(1)]
sage: Fp.nfhnf([pari(A),[pari(P) for P in I]])
[[1, [-969/5, -1/15]~, [15, -2]~, [-1938, -3]~; 0, 1, 0, 0; 0, 0, 1, 0; 0, 0, 0, 1], [[3997, 1911; 0, 7], [15, 6; 0, 3], 1, 1]]
sage: K.<b> = NumberField(x^3-2)
sage: Kp = pari(K)
sage: A = matrix(K,[[1,0,0,5*b],[1,2*b^2,b,57],[0,2,1,b^2-3],[2,0,0,b]])
sage: I = [K.ideal(2),K.ideal(3+b^2),K.ideal(1),K.ideal(1)]
sage: Kp.nfhnf([pari(A),[pari(P) for P in I]])
[[1, -225, 72, -31; 0, 1, [0, -1, 0]~, [0, 0, -1/2]~; 0, 0, 1, [0, 0, -1/2]~; 0, 0, 0, 1], [[1116, 756, 612; 0, 18, 0; 0, 0, 18], 2, 1, [2, 0, 0; 0, 1, 0; 0, 0, 1]]]


An example where the ring of integers of the number field is not a PID:

sage: K.<b> = NumberField(x^2+5)
sage: Kp = pari(K)
sage: A = matrix(K,[[1,0,0,5*b],[1,2*b^2,b,57],[0,2,1,b^2-3],[2,0,0,b]])
sage: I = [K.ideal(2),K.ideal(3+b^2),K.ideal(1),K.ideal(1)]
sage: Kp.nfhnf([pari(A),[pari(P) for P in I]])
[[1, [15, 6]~, [0, -54]~, [113, 72]~; 0, 1, [-4, -1]~, [0, -1]~; 0, 0, 1, 0; 0, 0, 0, 1], [[360, 180; 0, 180], [6, 4; 0, 2], 1, 1]]
sage: A = matrix(K,[[1,0,0,5*b],[1,2*b,b,57],[0,2,1,b-3],[2,0,b,b]])
sage: I = [K.ideal(2).factor()[0][0],K.ideal(3+b),K.ideal(1),K.ideal(1)]
sage: Kp.nfhnf([pari(A),[pari(P) for P in I]])
[[1, [7605, 4]~, [5610, 5]~, [7913, -6]~; 0, 1, 0, -1; 0, 0, 1, 0; 0, 0, 0, 1], [[19320, 13720; 0, 56], [2, 1; 0, 1], 1, 1]]


AUTHORS:

• Aly Deines (2012-09-19)
nfinit(flag=0, precision=0)

nfinit(pol, {flag=0}): pol being a nonconstant irreducible polynomial, gives a vector containing all the data necessary for PARI to compute in this number field.

flag is optional and can be set to:
• 0: default

• 1: do not compute different

• 2: first use polred to find a simpler polynomial

• 3: outputs a two-element vector [nf,Mod(a,P)], where nf is as in 2

and Mod(a,P) is a polmod equal to Mod(x,pol) and P=nf.pol

EXAMPLES:

sage: pari('x^3 - 17').nfinit()
[x^3 - 17, [1, 1], -867, 3, [[1, 1.68006914259990, 2.57128159065824; 1, -0.340034571299952 - 2.65083754153991*I, -1.28564079532912 + 2.22679517779329*I], [1, 1.68006914259990, 2.57128159065824; 1, -2.99087211283986, 0.941154382464174; 1, 2.31080297023995, -3.51243597312241], [1, 2, 3; 1, -3, 1; 1, 2, -4], [3, 1, 0; 1, -11, 17; 0, 17, 0], [51, 0, 16; 0, 17, 3; 0, 0, 1], [17, 0, -1; 0, 0, 3; -1, 3, 2], [51, [-17, 6, -1; 0, -18, 3; 1, 0, -16]], [3, 17]], [2.57128159065824, -1.28564079532912 + 2.22679517779329*I], [1, 1/3*x^2 - 1/3*x + 1/3, x], [1, 0, -1; 0, 0, 3; 0, 1, 1], [1, 0, 0, 0, -4, 6, 0, 6, -1; 0, 1, 0, 1, 1, -1, 0, -1, 3; 0, 0, 1, 0, 2, 0, 1, 0, 1]]


TESTS:

sage: pari('x^2 + 10^100 + 1').nfinit()
[...]
sage: pari('1.0').nfinit()
Traceback (most recent call last):
...
PariError: incorrect type in checknf [please apply nfinit()] (t_REAL)

nfisisom(other)

nfisisom(x, y): Determine if the number fields defined by x and y are isomorphic. According to the PARI documentation, this is much faster if at least one of x or y is a number field. If they are isomorphic, it returns an embedding for the generators. If not, returns 0.

EXAMPLES:

sage: F = NumberField(x^3-2,'alpha')
sage: G = NumberField(x^3-2,'beta')
sage: F._pari_().nfisisom(G._pari_())
[y]

sage: GG = NumberField(x^3-4,'gamma')
sage: F._pari_().nfisisom(GG._pari_())
[1/2*y^2]

sage: F._pari_().nfisisom(GG.pari_nf())
[1/2*y^2]

sage: F.pari_nf().nfisisom(GG._pari_()[0])
[y^2]

sage: H = NumberField(x^2-2,'alpha')
sage: F._pari_().nfisisom(H._pari_())
0

nfroots(poly)

Return the roots of $$poly$$ in the number field self without multiplicity.

EXAMPLES:

sage: y = QQ['yy'].0; _ = pari(y) # pari has variable ordering rules
sage: x = QQ['zz'].0; nf = pari(x^2 + 2).nfinit()
sage: nf.nfroots(y^2 + 2)
[Mod(-zz, zz^2 + 2), Mod(zz, zz^2 + 2)]
sage: nf = pari(x^3 + 2).nfinit()
sage: nf.nfroots(y^3 + 2)
[Mod(zz, zz^3 + 2)]
sage: nf = pari(x^4 + 2).nfinit()
sage: nf.nfroots(y^4 + 2)
[Mod(-zz, zz^4 + 2), Mod(zz, zz^4 + 2)]

nfrootsof1()

nf.nfrootsof1()

number of roots of unity and primitive root of unity in the number field nf.

EXAMPLES:

sage: nf = pari('x^2 + 1').nfinit()
sage: nf.nfrootsof1()
[4, x]

nfsubfields(d=0)

Find all subfields of degree d of number field nf (all subfields if d is null or omitted). Result is a vector of subfields, each being given by [g,h], where g is an absolute equation and h expresses one of the roots of g in terms of the root x of the polynomial defining nf.

INPUT:

• self - nf number field
• d - C long integer
norm()
nrows()

Return the number of rows of self.

EXAMPLES:

sage: pari('matrix(19,8)').nrows()
19

numbpart(x)

numbpart(x): returns the number of partitions of x.

EXAMPLES:

sage: pari(20).numbpart()
627
sage: pari(100).numbpart()
190569292

numdiv(n)

Return the number of divisors of the integer n.

EXAMPLES:

sage: pari(10).numdiv()
4

numerator(x)

numerator(x): Returns the numerator of x.

INPUT:

• x - gen

OUTPUT: gen

EXAMPLES:

numtoperm(k, n)

numtoperm(k, n): Return the permutation number k (mod n!) of n letters, where n is an integer.

INPUT:

• k - gen, integer
• n - int

OUTPUT:

• gen - vector (permutation of 1,...,n)

EXAMPLES:

omega(precision=0)

e.omega(): return basis for the period lattice of the elliptic curve e.

EXAMPLES:

sage: e = pari([0, -1, 1, -10, -20]).ellinit()
sage: e.omega()
[1.26920930427955, 0.634604652139777 - 1.45881661693850*I]

order()

x.padicappr(a): p-adic roots of the polynomial x congruent to a mod p

INPUT:

• x - gen

OUTPUT: int

EXAMPLES:

sage: K = Qp(11,5)
sage: x = K(11^-10 + 5*11^-7 + 11^-6)
sage: y = pari(x)
-5
Traceback (most recent call last):
...
PariError: inconsistent moduli in padicprec: 11 != 17


This works for polynomials too:

sage: R.<t> = PolynomialRing(Zp(3))
sage: pol = R([O(3^4), O(3^6), O(3^5)])
4


The uniformizer of the p-adic ring this element lies in, as a t_INT.

INPUT:

• x - gen, of type t_PADIC

OUTPUT:

• p - gen, of type t_INT

EXAMPLES:

sage: K = Qp(11,5)
sage: x = K(11^-10 + 5*11^-7 + 11^-6)
sage: y = pari(x)
11
't_INT'

permtonum(x)

permtonum(x): Return the ordinal (between 1 and n!) of permutation vector x. ??? Huh ??? say more. what is a perm vector. 0 to n-1 or 1-n.

INPUT:

• x - gen (vector of integers)

OUTPUT:

• gen - integer

EXAMPLES:

phi(n)

Return the Euler phi function of n.

EXAMPLES:

sage: pari(10).phi()
4

polcoeff(n, var=-1)

EXAMPLES:

sage: f = pari("x^2 + y^3 + x*y")
sage: f
x^2 + y*x + y^3
sage: f.polcoeff(1)
y
sage: f.polcoeff(3)
0
sage: f.polcoeff(3, "y")
1
sage: f.polcoeff(1, "y")
x

polcompositum(pol2, flag=0)
poldegree(var=-1)

f.poldegree(var=x): Return the degree of this polynomial.

poldisc(var=-1)

Return the discriminant of this polynomial.

EXAMPLES:

sage: pari("x^2 + 1").poldisc()
-4


Before trac ticket #15654, this used to take a very long time. Now it takes much less than a second:

sage: pari.allocatemem(200000)
PARI stack size set to 200000 bytes
sage: x = polygen(ZpFM(3,10))
sage: pol = ((x-1)^50 + x)
sage: pari(pol).poldisc()
2*3 + 3^4 + 2*3^6 + 3^7 + 2*3^8 + 2*3^9 + O(3^10)

poldiscreduced()
polgalois(precision=0)

f.polgalois(): Galois group of the polynomial f

polhensellift(y, p, e)

self.polhensellift(y, p, e): lift the factorization y of self modulo p to a factorization modulo $$p^e$$ using Hensel lift. The factors in y must be pairwise relatively prime modulo p.

polinterpolate(ya, x)

self.polinterpolate(ya,x,e): polynomial interpolation at x according to data vectors self, ya (i.e. return P such that P(self[i]) = ya[i] for all i). Also return an error estimate on the returned value.

polisirreducible()

f.polisirreducible(): Returns True if f is an irreducible non-constant polynomial, or False if f is reducible or constant.

self.pollead(v): leading coefficient of polynomial or series self, or self itself if self is a scalar. Error otherwise. With respect to the main variable of self if v is omitted, with respect to the variable v otherwise

polrecip()
polred(flag=0, fa=None)
polredabs(flag=0)
polredbest(flag=0)
polresultant(y, var=-1, flag=0)
polroots(flag=0, precision=0)

Complex roots of the given polynomial using Schonhage’s method, as modified by Gourdon.

polrootsmod(p, flag=0)
polsturm(a, b)
polsturm_full()
polsylvestermatrix(g)
polsym(n)
polylog(x, m, flag=0, precision=0)

x.polylog(m,flag=0): m-th polylogarithm of x. flag is optional, and can be 0: default, 1: D_m -modified m-th polylog of x, 2: D_m-modified m-th polylog of x, 3: P_m-modified m-th polylog of x.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

TODO: Add more explanation, copied from the PARI manual.

EXAMPLES:

sage: pari(10).polylog(3)
5.64181141475134 - 8.32820207698027*I
sage: pari(10).polylog(3,0)
5.64181141475134 - 8.32820207698027*I
sage: pari(10).polylog(3,1)
0.523778453502411
sage: pari(10).polylog(3,2)
-0.400459056163451

pr_get_e()

Returns the ramification index (over $$\QQ$$) of this prime ideal.

NOTE: self must be a PARI prime ideal (as returned by idealfactor for example).

EXAMPLES:

sage: K.<i> = QuadraticField(-1)
sage: pari(K).idealfactor(K.ideal(2))[0,0].pr_get_e()
2
sage: pari(K).idealfactor(K.ideal(3))[0,0].pr_get_e()
1
sage: pari(K).idealfactor(K.ideal(5))[0,0].pr_get_e()
1

pr_get_f()

Returns the residue class degree (over $$\QQ$$) of this prime ideal.

NOTE: self must be a PARI prime ideal (as returned by idealfactor for example).

EXAMPLES:

sage: K.<i> = QuadraticField(-1)
sage: pari(K).idealfactor(K.ideal(2))[0,0].pr_get_f()
1
sage: pari(K).idealfactor(K.ideal(3))[0,0].pr_get_f()
2
sage: pari(K).idealfactor(K.ideal(5))[0,0].pr_get_f()
1

pr_get_gen()

Returns the second generator of this PARI prime ideal, where the first generator is self.pr_get_p().

NOTE: self must be a PARI prime ideal (as returned by idealfactor for example).

EXAMPLES:

sage: K.<i> = QuadraticField(-1)
sage: g = pari(K).idealfactor(K.ideal(2))[0,0].pr_get_gen(); g; K(g)
[1, 1]~
i + 1
sage: g = pari(K).idealfactor(K.ideal(3))[0,0].pr_get_gen(); g; K(g)
[3, 0]~
3
sage: g = pari(K).idealfactor(K.ideal(5))[0,0].pr_get_gen(); g; K(g)
[-2, 1]~
i - 2

pr_get_p()

Returns the prime of $$\ZZ$$ lying below this prime ideal.

NOTE: self must be a PARI prime ideal (as returned by idealfactor for example).

EXAMPLES:

sage: K.<i> = QuadraticField(-1)
sage: F = pari(K).idealfactor(K.ideal(5)); F
[[5, [-2, 1]~, 1, 1, [2, -1; 1, 2]], 1; [5, [2, 1]~, 1, 1, [-2, -1; 1, -2]], 1]
sage: F[0,0].pr_get_p()
5

precision(x, n=-1)

precision(x,n): Change the precision of x to be n, where n is a C-integer). If n is omitted, output the real precision of x.

INPUT:

• x - gen
• n - (optional) int

OUTPUT: nothing or gen if n is omitted

EXAMPLES:

primepi()

Return the number of primes less than or equal to self.

EXAMPLES:

sage: pari(7).primepi()
4
sage: pari(100).primepi()
25
sage: pari(1000).primepi()
168
sage: pari(100000).primepi()
9592
sage: pari(0).primepi()
0
sage: pari(-15).primepi()
0
sage: pari(500509).primepi()
41581

printtex(x)
psi(x, precision=0)

x.psi(): psi-function at x.

Return the $$\psi$$-function of $$x$$, i.e., the logarithmic derivative $$\Gamma'(x)/\Gamma(x)$$.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(1).psi()
-0.577215664901533

python(locals=None)

Return Python eval of self.

Note: is self is a real (type t_REAL) the result will be a RealField element of the equivalent precision; if self is a complex (type t_COMPLEX) the result will be a ComplexField element of precision the minimum precision of the real and imaginary parts.

EXAMPLES:

sage: pari('389/17').python()
389/17
sage: f = pari('(2/3)*x^3 + x - 5/7 + y'); f
2/3*x^3 + x + (y - 5/7)
sage: var('x,y')
(x, y)
sage: f.python({'x':x, 'y':y})
2/3*x^3 + x + y - 5/7


You can also use .sage, which is a psynonym:

sage: f.sage({'x':x, 'y':y})
2/3*x^3 + x + y - 5/7

python_list()

Return a Python list of the PARI gens. This object must be of type t_VEC.

INPUT: None

OUTPUT:

• list - Python list whose elements are the elements of the input gen.

EXAMPLES:

sage: v=pari([1,2,3,10,102,10])
sage: w = v.python_list()
sage: w
[1, 2, 3, 10, 102, 10]
sage: type(w[0])
<type 'sage.libs.pari.gen.gen'>
sage: pari("[1,2,3]").python_list()
[1, 2, 3]

python_list_small()

Return a Python list of the PARI gens. This object must be of type t_VECSMALL, and the resulting list contains python ‘int’s.

EXAMPLES:

sage: v=pari([1,2,3,10,102,10]).Vecsmall()
sage: w = v.python_list_small()
sage: w
[1, 2, 3, 10, 102, 10]
sage: type(w[0])
<type 'int'>

qfbclassno(d, flag=0)

Computes the class number of the quadratic order of discriminant $$d$$.

INPUT:

• $$d$$ (gen) – a quadratic discriminant, which is an integer congruent to $$0$$ or $$1$$mod4, not a square.
• flag (long int) – if 0 (default), uses Euler product and the functional equation for $$d>0$$ or Shanks’s method for $$d<0$$; if 1, uses Euler products and the functional equation in both cases.

OUTPUT:

The class number of the quadratic order with discriminant $$d$$.

Warning

Using Euler products and the functional equation is reliable but has complexity $$O(|d|^{1/2})$$. Using Shanks’s method for $$d<0$$ is $$O(|d|^{1/4})$$ but this function may give incorrect results when the class group has many cyclic factors, because implementing Shanks’s method in full generality slows it down immensely. It is therefore strongly recommended to double-check results using either the version with flag = 1 or the function quadclassunit. The result is unconditionally correct for $$-d < 2e10$$.

EXAMPLES:

sage: pari(-4).qfbclassno()
1
sage: pari(-23).qfbclassno()
3
sage: pari(-104).qfbclassno()
6

sage: pari(109).qfbclassno()
1
sage: pari(10001).qfbclassno()
16
sage: pari(10001).qfbclassno(flag=1)
16


TESTS:

The input must be congruent to $$0$$ or $$1\mod4$$ and not a square:

sage: pari(3).qfbclassno()
Traceback (most recent call last):
...
PariError: domain error in classno2: disc % 4 > 1
sage: pari(4).qfbclassno()
Traceback (most recent call last):
...
PariError: domain error in classno2: issquare(disc) = 1

qfbhclassno(n)

Computes the Hurwitz-Kronecker class number of $$n$$.

INPUT:

• $$n$$ (gen) – a non-negative integer

OUTPUT:

0 if $$n<0$$, otherwise the Hurwitz-Kronecker class number of $$n$$. This is $$0$$ if $$n\equiv1,2\mod4$$, $$-1/12$$ when $$n=0$$, and otherwise it is the number of classes of positive definite binary quadratic forms with discriminant $$-n$$, each weighted by the number of its automorphisms.

Note

If $$n$$ is large (more than $$5*10^5$$), the result is conditional upon GRH.

EXAMPLES:

The Hurwitz class number is 0 if n is congruent to 1 or 2 modulo 4:

sage: pari(10009).qfbhclassno()
0
sage: pari(2).qfbhclassno()
0


It is -1/12 for n=0:

sage: pari(0).qfbhclassno()
-1/12


Otherwise it is the number of classes of positive definite binary quadratic forms with discriminant $$-n$$, weighted by $$1/m$$ where $$m$$ is the number of automorphisms of the form:

sage: pari(4).qfbhclassno()
1/2
sage: pari(3).qfbhclassno()
1/3
sage: pari(23).qfbhclassno()
3

qflll(flag=0)

qflll(x,flag=0): LLL reduction of the vectors forming the matrix x (gives the unimodular transformation matrix). The columns of x must be linearly independent, unless specified otherwise below. flag is optional, and can be 0: default, 1: assumes x is integral, columns may be dependent, 2: assumes x is integral, returns a partially reduced basis, 4: assumes x is integral, returns [K,I] where K is the integer kernel of x and I the LLL reduced image, 5: same as 4 but x may have polynomial coefficients, 8: same as 0 but x may have polynomial coefficients.

qflllgram(flag=0)

qflllgram(x,flag=0): LLL reduction of the lattice whose gram matrix is x (gives the unimodular transformation matrix). flag is optional and can be 0: default,1: lllgramint algorithm for integer matrices, 4: lllgramkerim giving the kernel and the LLL reduced image, 5: lllgramkerimgen same when the matrix has polynomial coefficients, 8: lllgramgen, same as qflllgram when the coefficients are polynomials.

qfminim(b=None, m=None, flag=0, precision=0)

Return vectors with bounded norm for this quadratic form.

INPUT:

• self – a quadratic form

• b – a bound on vector norm (finds minimal non-zero vectors if b=0)

• m – maximum number of vectors to return. If None (default), return all vectors of norm at most B

• flag (optional) –

• 0: default;
• 1: return only the first minimal vector found (ignore max);
• 2: as 0 but uses a more robust, slower implementation, valid for non integral quadratic forms.

OUTPUT:

A triple consisting of

• the number of vectors of norm <= b,
• the actual maximum norm of vectors listed
• a matrix whose columns are vectors with norm less than or equal to b for the definite quadratic form. Only one of $$v$$ and $$-v$$ is returned and the zero vector is never returned.

Note

If max is specified then only max vectors will be output, but all vectors withing the given norm bound will be computed.

EXAMPLES:

sage: A = Matrix(3,3,[1,2,3,2,5,5,3,5,11])
sage: A.is_positive_definite()
True


The first 5 vectors of norm at most 10:

sage: pari(A).qfminim(10, 5).python()
[
[-17 -14 -15 -16 -13]
[  4   3   3   3   2]
146, 10, [  3   3   3   3   3]
]


All vectors of minimal norm:

sage: pari(A).qfminim(0).python()
[
[-5 -2  1]
[ 1  1  0]
6, 1, [ 1  0  0]
]


Use flag=2 for non-integral input:

sage: pari(A.change_ring(RR)).qfminim(5, m=5, flag=2).python()
[
[ -5 -10  -2  -7   3]
[  1   2   1   2   0]
10, 5.00000000000000000, [  1   2   0   1  -1]
]

qfrep(B, flag=0)

Vector of (half) the number of vectors of norms from 1 to $$B$$ for the integral and definite quadratic form self. Binary digits of flag mean 1: count vectors of even norm from 1 to $$2B$$, 2: return a t_VECSMALL instead of a t_VEC (which is faster).

EXAMPLES:

sage: M = pari("[5,1,1;1,3,1;1,1,1]")
sage: M.qfrep(20)
[1, 1, 2, 2, 2, 4, 4, 3, 3, 4, 2, 4, 6, 0, 4, 6, 4, 5, 6, 4]
sage: M.qfrep(20, flag=1)
[1, 2, 4, 3, 4, 4, 0, 6, 5, 4, 12, 4, 4, 8, 0, 3, 8, 6, 12, 12]
sage: M.qfrep(20, flag=2)
Vecsmall([1, 1, 2, 2, 2, 4, 4, 3, 3, 4, 2, 4, 6, 0, 4, 6, 4, 5, 6, 4])


Returns the class group of a quadratic order of discriminant $$d$$.

INPUT:

• $$d$$ (gen) – a quadratic discriminant, which is an integer congruent to $$0$$ or $$1$$mod4, not a square.

OUTPUT:

(h,cyc,gen,reg) where:

• h is the class number
• cyc is the class group structure (list of invariants)
• gen is the class group generators (list of quadratic forms)
• reg is the regulator

ALGORITHM:

Buchmann-McCurley’s sub-exponential algorithm

EXAMPLES:

sage: pari(-4).quadclassunit()
[1, [], [], 1]
[3, [3], [Qfb(2, 1, 3)], 1]
[6, [6], [Qfb(5, -4, 6)], 1]

[1, [], [], 5.56453508676047]
[16, [16], [Qfb(10, 99, -5, 0.E-38)], 5.29834236561059]
16
[16]
5.29834236561059


TESTS:

The input must be congruent to $$0$$ or $$1\mod4$$ and not a square:

sage: pari(3).quadclassunit()
Traceback (most recent call last):
...
PariError: domain error in Buchquad: disc % 4 > 1
Traceback (most recent call last):
...
PariError: domain error in Buchquad: issquare(disc) = 1


Returns a polynomial over $$\QQ$$ whose roots generate the Hilbert class field of the quadratic field of discriminant self (which must be fundamental).

EXAMPLES:

sage: pari(-23).quadhilbert()
x^3 - x^2 + 1
x^4 - 6*x^2 - 5*x - 1
Traceback (most recent call last):
...
PariError: domain error in quadray: isfundamental(D) = 0

random(N)

random(N=2^31): Return a pseudo-random integer between 0 and $$N-1$$.

INPUT:

-N - gen, integer

OUTPUT:

• gen - integer

EXAMPLES:

real(x)

real(x): Return the real part of x.

INPUT:

• x - gen

OUTPUT: gen

EXAMPLES:

reverse()

Return the polynomial obtained by reversing the coefficients of this polynomial.

rnfcharpoly(T, a, v='x')
rnfdisc(x)
rnfeltabstorel(x)
rnfeltreltoabs(x)
rnfequation(poly, flag=0)
rnfidealabstorel(x)
rnfidealdown(x)

rnfidealdown(rnf,x): finds the intersection of the ideal x with the base field.

EXAMPLES:
sage: x = ZZ[‘xx1’].0; pari(x) xx1 sage: y = ZZ[‘yy1’].0; pari(y) yy1 sage: nf = pari(y^2 - 6*y + 24).nfinit() sage: rnf = nf.rnfinit(x^2 - pari(y))

This is the relative HNF of the inert ideal (2) in rnf:

sage: P = pari('[[[1, 0]~, [0, 0]~; [0, 0]~, [1, 0]~], [[2, 0; 0, 2], [2, 0; 0, 1/2]]]')


And this is the inert ideal (2) in nf:

sage: rnf.rnfidealdown(P) 2
rnfidealhnf(x)
rnfidealnormrel(x)
rnfidealreltoabs(x)
rnfidealtwoelt(x)
rnfinit(poly)

EXAMPLES: We construct a relative number field.

sage: f = pari('y^3+y+1')
sage: K = f.nfinit()
sage: x = pari('x'); y = pari('y')
sage: g = x^5 - x^2 + y
sage: L = K.rnfinit(g)

rnfisfree(poly)
rnfisnorm(T, flag=0)
rnfisnorminit(polrel, flag=2)
round(x, estimate=False)

round(x,estimate=False): If x is a real number, returns x rounded to the nearest integer (rounding up). If the optional argument estimate is True, also returns the binary exponent e of the difference between the original and the rounded value (the “fractional part”) (this is the integer ceiling of log_2(error)).

When x is a general PARI object, this function returns the result of rounding every coefficient at every level of PARI object. Note that this is different than what the truncate function does (see the example below).

One use of round is to get exact results after a long approximate computation, when theory tells you that the coefficients must be integers.

INPUT:

• x - gen
• estimate - (optional) bool, False by default

OUTPUT:

• if estimate is False, return a single gen.
• if estimate is True, return rounded version of x and error estimate in bits, both as gens.

EXAMPLES:

sage: pari('1.5').round()
2
sage: pari('1.5').round(True)
(2, -1)
sage: pari('1.5 + 2.1*I').round()
2 + 2*I
sage: pari('1.0001').round(True)
(1, -14)
sage: pari('(2.4*x^2 - 1.7)/x').round()
(2*x^2 - 2)/x
sage: pari('(2.4*x^2 - 1.7)/x').truncate()
2.40000000000000*x

sage(locals=None)

Return Python eval of self.

Note: is self is a real (type t_REAL) the result will be a RealField element of the equivalent precision; if self is a complex (type t_COMPLEX) the result will be a ComplexField element of precision the minimum precision of the real and imaginary parts.

EXAMPLES:

sage: pari('389/17').python()
389/17
sage: f = pari('(2/3)*x^3 + x - 5/7 + y'); f
2/3*x^3 + x + (y - 5/7)
sage: var('x,y')
(x, y)
sage: f.python({'x':x, 'y':y})
2/3*x^3 + x + y - 5/7


You can also use .sage, which is a psynonym:

sage: f.sage({'x':x, 'y':y})
2/3*x^3 + x + y - 5/7

serconvol(g)
serlaplace()
serreverse()

serreverse(f): reversion of the power series f.

If f(t) is a series in t with valuation 1, find the series g(t) such that g(f(t)) = t.

EXAMPLES:

sage: f = pari('x+x^2+x^3+O(x^4)'); f
x + x^2 + x^3 + O(x^4)
sage: g = f.serreverse(); g
x - x^2 + x^3 + O(x^4)
sage: f.subst('x',g)
x + O(x^4)
sage: g.subst('x',f)
x + O(x^4)

shift(x, n)

shift(x,n): shift x left n bits if n=0, right -n bits if n0.

shiftmul(x, n)

shiftmul(x,n): Return the product of x by $$2^n$$.

sign(x)

Return the sign of x, where x is of type integer, real or fraction.

EXAMPLES:

sage: pari(pi).sign()
1
sage: pari(0).sign()
0
sage: pari(-1/2).sign()
-1


PARI throws an error if you attempt to take the sign of a complex number:

sage: pari(I).sign()
Traceback (most recent call last):
...
PariError: incorrect type in gsigne (t_COMPLEX)

simplify(x)

simplify(x): Simplify the object x as much as possible, and return the result.

A complex or quadratic number whose imaginary part is an exact 0 (i.e., not an approximate one such as O(3) or 0.E-28) is converted to its real part, and a a polynomial of degree 0 is converted to its constant term. Simplification occurs recursively.

This function is useful before using arithmetic functions, which expect integer arguments:

EXAMPLES:

sage: y = pari('y')
sage: x = pari('9') + y - y
sage: x
9
sage: x.type()
't_POL'
sage: x.factor()
matrix(0,2)
sage: pari('9').factor()
Mat([3, 2])
sage: x.simplify()
9
sage: x.simplify().factor()
Mat([3, 2])
sage: x = pari('1.5 + 0*I')
sage: x.type()
't_REAL'
sage: x.simplify()
1.50000000000000
sage: y = x.simplify()
sage: y.type()
't_REAL'

sin(x, precision=0)

x.sin(): The sine of x.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(1).sin()
0.841470984807897
sage: C.<i> = ComplexField()
sage: pari(1+i).sin()
1.29845758141598 + 0.634963914784736*I

sinh(x, precision=0)

The hyperbolic sine function.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(0).sinh()
0.E-19
sage: C.<i> = ComplexField()
sage: pari(1+i).sinh()
0.634963914784736 + 1.29845758141598*I

sizebyte(x)

Return the total number of bytes occupied by the complete tree of the object x. Note that this number depends on whether the computer is 32-bit or 64-bit.

INPUT:

• x - gen

OUTPUT: int (a Python int)

EXAMPLE:

sage: pari('1').sizebyte()
12           # 32-bit
24           # 64-bit

sizedigit(x)

sizedigit(x): Return a quick estimate for the maximal number of decimal digits before the decimal point of any component of x.

INPUT:

• x - gen

OUTPUT:

• int - Python integer

EXAMPLES:

sage: x = pari('10^100')
sage: x.Str().length()
101
sage: x.sizedigit()
101


Note that digits after the decimal point are ignored.

sage: x = pari('1.234')
sage: x
1.23400000000000
sage: x.sizedigit()
1


The estimate can be one too big:

sage: pari('7234.1').sizedigit()
4
sage: pari('9234.1').sizedigit()
5

sizeword(x)

Return the total number of machine words occupied by the complete tree of the object x. A machine word is 32 or 64 bits, depending on the computer.

INPUT:

• x - gen

OUTPUT: int (a Python int)

EXAMPLES:

sage: pari('0').sizeword()
2
sage: pari('1').sizeword()
3
sage: pari('1000000').sizeword()
3
sage: pari('10^100').sizeword()
13      # 32-bit
8       # 64-bit
sage: pari(RDF(1.0)).sizeword()
4       # 32-bit
3       # 64-bit
sage: pari('x').sizeword()
9
sage: pari('x^20').sizeword()
66
sage: pari('[x, I]').sizeword()
20

sqr(x)

x.sqr(): square of x. Faster than, and most of the time (but not always - see the examples) identical to x*x.

EXAMPLES:

sage: pari(2).sqr()
4


For $$2$$-adic numbers, x.sqr() may not be identical to x*x (squaring a $$2$$-adic number increases its precision):

sage: pari("1+O(2^5)").sqr()
1 + O(2^6)
sage: pari("1+O(2^5)")*pari("1+O(2^5)")
1 + O(2^5)


However:

sage: x = pari("1+O(2^5)"); x*x
1 + O(2^6)

sqrt(x, precision=0)

x.sqrt(precision): The square root of x.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).sqrt()
1.41421356237310

sqrtint(x)

Return the integer square root of the integer $$x$$, rounded towards zero.

EXAMPLES:

sage: pari(8).sqrtint()
2
sage: pari(10^100).sqrtint()
100000000000000000000000000000000000000000000000000

sqrtn(x, n, precision=0)

x.sqrtn(n): return the principal branch of the n-th root of x, i.e., the one such that $$\arg(\sqrt(x)) \in ]-\pi/n, \pi/n]$$. Also returns a second argument which is a suitable root of unity allowing one to recover all the other roots. If it was not possible to find such a number, then this second return value is 0. If the argument is present and no square root exists, return 0 instead of raising an error.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

Note

intmods (modulo a prime) and $$p$$-adic numbers are allowed as arguments.

INPUT:

• x - gen
• n - integer

OUTPUT:

• gen - principal n-th root of x
• gen - root of unity z that gives the other roots

EXAMPLES:

sage: s, z = pari(2).sqrtn(5)
sage: z
0.309016994374947 + 0.951056516295154*I
sage: s
1.14869835499704
sage: s^5
2.00000000000000
sage: z^5
1.00000000000000 - 2.710505431 E-20*I       # 32-bit
1.00000000000000 - 2.71050543121376 E-20*I  # 64-bit
sage: (s*z)^5
2.00000000000000 + 0.E-19*I                 # 32-bit
2.00000000000000 - 1.08420217248550 E-19*I  # 64-bit

subst(var, z)

In self, replace the variable var by the expression $$z$$.

EXAMPLES:

sage: x = pari("x"); y = pari("y")
sage: f = pari('x^3 + 17*x + 3')
sage: f.subst(x, y)
y^3 + 17*y + 3
sage: f.subst(x, "z")
z^3 + 17*z + 3
sage: f.subst(x, "z")^2
z^6 + 34*z^4 + 6*z^3 + 289*z^2 + 102*z + 9
sage: f.subst(x, "x+1")
x^3 + 3*x^2 + 20*x + 21
sage: f.subst(x, "xyz")
xyz^3 + 17*xyz + 3
sage: f.subst(x, "xyz")^2
xyz^6 + 34*xyz^4 + 6*xyz^3 + 289*xyz^2 + 102*xyz + 9

substpol(y, z)
sumdiv(n)

Return the sum of the divisors of $$n$$.

EXAMPLES:

sage: pari(10).sumdiv()
18

sumdivk(n, k)

Return the sum of the k-th powers of the divisors of n.

EXAMPLES:

sage: pari(10).sumdivk(2)
130

tan(x, precision=0)

x.tan() - tangent of x

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(2).tan()
-2.18503986326152
sage: C.<i> = ComplexField()
sage: pari(i).tan()
0.761594155955765*I

tanh(x, precision=0)

x.tanh() - hyperbolic tangent of x

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(1).tanh()
0.761594155955765
sage: C.<i> = ComplexField()
sage: z = pari(i); z
1.00000000000000*I
sage: result = z.tanh()
sage: result.real() <= 1e-18
True
sage: result.imag()
1.55740772465490

taylor(v=-1)
teichmuller(x)

teichmuller(x): teichmuller character of p-adic number x.

This is the unique $$(p-1)$$-st root of unity congruent to $$x/p^{v_p(x)}$$ modulo $$p$$.

EXAMPLES:

sage: pari('2+O(7^5)').teichmuller()
2 + 4*7 + 6*7^2 + 3*7^3 + O(7^5)

theta(q, z, precision=0)

q.theta(z): Jacobi sine theta-function.

If $$q$$ or $$z$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If the arguments are inexact (e.g. real), the smallest of their precisions is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(0.5).theta(2)
1.63202590295260

thetanullk(q, k, precision=0)

q.thetanullk(k): return the k-th derivative at z=0 of theta(q,z).

If $$q$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$q$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

EXAMPLES:

sage: pari(0.5).thetanullk(1)
0.548978532560341

thue(rhs, ne)
thueinit(flag=0, precision=0)
trace()

Return the trace of this PARI object.

EXAMPLES:

sage: pari('[1,2; 3,4]').trace()
5

truncate(x, estimate=False)

truncate(x,estimate=False): Return the truncation of x. If estimate is True, also return the number of error bits.

When x is in the real numbers, this means that the part after the decimal point is chopped away, e is the binary exponent of the difference between the original and truncated value (the “fractional part”). If x is a rational function, the result is the integer part (Euclidean quotient of numerator by denominator) and if requested the error estimate is 0.

When truncate is applied to a power series (in X), it transforms it into a polynomial or a rational function with denominator a power of X, by chopping away the $$O(X^k)$$. Similarly, when applied to a p-adic number, it transforms it into an integer or a rational number by chopping away the $$O(p^k)$$.

INPUT:

• x - gen
• estimate - (optional) bool, which is False by default

OUTPUT:

• if estimate is False, return a single gen.
• if estimate is True, return rounded version of x and error estimate in bits, both as gens.

EXAMPLES:

sage: pari('(x^2+1)/x').round()
(x^2 + 1)/x
sage: pari('(x^2+1)/x').truncate()
x
sage: pari('1.043').truncate()
1
sage: pari('1.043').truncate(True)
(1, -5)
sage: pari('1.6').truncate()
1
sage: pari('1.6').round()
2
sage: pari('1/3 + 2 + 3^2 + O(3^3)').truncate()
34/3
sage: pari('sin(x+O(x^10))').truncate()
1/362880*x^9 - 1/5040*x^7 + 1/120*x^5 - 1/6*x^3 + x
sage: pari('sin(x+O(x^10))').round()   # each coefficient has abs < 1
x + O(x^10)

type()

Return the PARI type of self as a string.

Note

In Cython, it is much faster to simply use typ(self.g) for checking PARI types.

EXAMPLES:

sage: pari(7).type()
't_INT'
sage: pari('x').type()
't_POL'

valuation(x, p)

valuation(x,p): Return the valuation of x with respect to p.

The valuation is the highest exponent of p dividing x.

• If p is an integer, x must be an integer, an intmod whose modulus is divisible by p, a rational number, a p-adic number, or a polynomial or power series in which case the valuation is the minimum of the valuations of the coefficients.
• If p is a polynomial, x must be a polynomial or a rational function. If p is a monomial then x may also be a power series.
• If x is a vector, complex or quadratic number, then the valuation is the minimum of the component valuations.
• If x = 0, the result is $$2^31-1$$ on 32-bit machines or $$2^63-1$$ on 64-bit machines if x is an exact object. If x is a p-adic number or power series, the result is the exponent of the zero.

INPUT:

• x - gen
• p - coercible to gen

OUTPUT:

• gen - integer

EXAMPLES:

sage: pari(9).valuation(3)
2
sage: pari(9).valuation(9)
1
sage: x = pari(9).Mod(27); x.valuation(3)
2
sage: pari('5/3').valuation(3)
-1
sage: pari('9 + 3*x + 15*x^2').valuation(3)
1
sage: pari([9,3,15]).valuation(3)
1
sage: pari('9 + 3*x + 15*x^2 + O(x^5)').valuation(3)
1

sage: pari('x^2*(x+1)^3').valuation(pari('x+1'))
3
sage: pari('x + O(x^5)').valuation('x')
1
sage: pari('2*x^2 + O(x^5)').valuation('x')
2

sage: pari(0).valuation(3)
2147483647            # 32-bit
9223372036854775807   # 64-bit

variable(x)

variable(x): Return the main variable of the object x, or p if x is a p-adic number.

This function raises a TypeError exception on scalars, i.e., on objects with no variable associated to them.

INPUT:

• x - gen

OUTPUT: gen

EXAMPLES:

sage: pari('x^2 + x -2').variable()
x
sage: pari('1+2^3 + O(2^5)').variable()
2
sage: pari('x+y0').variable()
x
sage: pari('y0+z0').variable()
y0

vecextract(y, z=None)

self.vecextract(y,z): extraction of the components of the matrix or vector x according to y and z. If z is omitted, y designates columns, otherwise y corresponds to rows and z to columns. y and z can be vectors (of indices), strings (indicating ranges as in”1..10”) or masks (integers whose binary representation indicates the indices to extract, from left to right 1, 2, 4, 8, etc.)

Note

This function uses the PARI row and column indexing, so the first row or column is indexed by 1 instead of 0.

vecmax(x)

vecmax(x): Return the maximum of the elements of the vector/matrix x.

vecmin(x)

vecmin(x): Return the maximum of the elements of the vector/matrix x.

weber(x, flag=0, precision=0)

x.weber(flag=0): One of Weber’s f functions of x. flag is optional, and can be 0: default, function f(x)=exp(-i*Pi/24)*eta((x+1)/2)/eta(x) such that $$j=(f^{24}-16)^3/f^{24}$$, 1: function f1(x)=eta(x/2)/eta(x) such that $$j=(f1^24+16)^3/f2^{24}$$, 2: function f2(x)=sqrt(2)*eta(2*x)/eta(x) such that $$j=(f2^{24}+16)^3/f2^{24}$$.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

TODO: Add further explanation from PARI manual.

EXAMPLES:

sage: C.<i> = ComplexField()
sage: pari(i).weber()
1.18920711500272
sage: pari(i).weber(1)
1.09050773266526
sage: pari(i).weber(2)
1.09050773266526

xgcd(x, y)

Returns u,v,d such that d=gcd(x,y) and u*x+v*y=d.

EXAMPLES:

sage: pari(10).xgcd(15)
(5, -1, 1)

zeta(s, precision=0)

zeta(s): zeta function at s with s a complex or a p-adic number.

If $$s$$ is a complex number, this is the Riemann zeta function $$\zeta(s)=\sum_{n\geq 1} n^{-s}$$, computed either using the Euler-Maclaurin summation formula (if $$s$$ is not an integer), or using Bernoulli numbers (if $$s$$ is a negative integer or an even nonnegative integer), or using modular forms (if $$s$$ is an odd nonnegative integer).

If $$s$$ is a $$p$$-adic number, this is the Kubota-Leopoldt zeta function, i.e. the unique continuous $$p$$-adic function on the $$p$$-adic integers that interpolates the values of $$(1-p^{-k})\zeta(k)$$ at negative integers $$k$$ such that $$k\equiv 1\pmod{p-1}$$ if $$p$$ is odd, and at odd $$k$$ if $$p=2$$.

If $$x$$ is an exact argument, it is first converted to a real or complex number using the optional parameter precision (in bits). If $$x$$ is inexact (e.g. real), its own precision is used in the computation, and the parameter precision is ignored.

INPUT:

• s - gen (real, complex, or p-adic number)

OUTPUT:

• gen - value of zeta at s.

EXAMPLES:

sage: pari(2).zeta()
1.64493406684823
sage: x = RR(pi)^2/6
sage: pari(x)
1.64493406684823
sage: pari(3).zeta()
1.20205690315959
sage: pari('1+5*7+2*7^2+O(7^3)').zeta()
4*7^-2 + 5*7^-1 + O(7^0)

znprimroot()

Return a primitive root modulo self, whenever it exists.

INPUT:

• self – an integer $$n$$ such that $$|n|$$ is equal to 1, 2, 4, a power of an odd prime, or twice a power of an odd prime

OUTPUT:

A generator (type t_INTMOD) of $$(\ZZ/n\ZZ)^*$$. Note that this group is cyclic if and only if $$n$$ is of the above form.

EXAMPLES:

sage: pari(4).znprimroot()
Mod(3, 4)
sage: pari(10007^3).znprimroot()
Mod(5, 1002101470343)
sage: pari(2*109^10).znprimroot()
Mod(236736367459211723407, 473472734918423446802)

znstar()

Return the structure of the group $$(\ZZ/n\ZZ)^*$$.

INPUT:

• self – any integer $$n$$ (type t_INT)

OUTPUT:

A triple $$[\phi(n), [d_1, \ldots, d_k], [x_1, \ldots, x_k]]$$, where

• $$\phi(n)$$ is the order of $$(\ZZ/n\ZZ)^*$$;
• $$d_1, \ldots, d_k$$ are the unique integers greater than 1 with $$d_k \mid d_{k-1} \mid \ldots \mid d_1$$ such that $$(\ZZ/n\ZZ)^*$$ is isomorphic to $$\prod_{i=1}^k \ZZ/d_i\ZZ$$;
• $$x_1, \ldots, x_k$$ are the images of the standard generators under some isomorphism from $$\prod_{i=1}^k \ZZ/d_i\ZZ$$ to $$(\ZZ/n\ZZ)^*$$.

EXAMPLES:

sage: pari(0).znstar()
[2, [2], [-1]]
sage: pari(96).znstar()
[32, [8, 2, 2], [Mod(37, 96), Mod(79, 96), Mod(65, 96)]]
sage: pari(-5).znstar()
[4, [4], [Mod(2, 5)]]

sage.libs.pari.gen.init_pari_stack(s=8000000)

EXAMPLES:

sage: from sage.libs.pari.gen import init_pari_stack
sage: init_pari_stack()
doctest:...: DeprecationWarning: init_pari_stack() is deprecated; use pari.allocatemem() instead.
See http://trac.sagemath.org/10018 for details.
sage: pari.stacksize()
8000000
`
sage.libs.pari.gen.objtogen(s)

Convert any Sage/Python object to a PARI gen.

#### Previous topic

PARI C-library interface

#### Next topic

Cython wrapper for the Parma Polyhedra Library (PPL)