Module: sage.lfunctions.dokchitser
Dokchitser's L-functions Calculator
Author Log:
TODO: - add more examples from data/extcode/pari/dokchitser that illustrate use with Eisenstein series, number fields, etc. - plug this code into number fields and modular forms code (elliptic curves are done).
Module-level Functions
D) |
Class: Dokchitser
Create a Dokchitser
-series with
Dokchitser(conductor, gammaV, weight, eps, poles, residues, init, prec)
where
conductor - integer, the conductor
gammaV - list of Gamma-factor parameters,
e.g. [0] for Riemann zeta, [0,1] for ell.curves,
(see examples).
weight - positive real number, usually an integer
e.g. 1 for Riemann zeta, 2 for
of curves/
eps - complex number; sign in functional equation
poles - (default: []) list of points where
has (simple) poles;
only poles with Re(s)>weight/2 should be included
residues - vector of residues of
in those poles
or set residues='automatic' (default value)
prec - integer (default: 53) number of *bits* of precision
RIEMANN ZETA FUNCTION:
We compute with the Riemann Zeta function.
sage: L = Dokchitser(conductor=1, gammaV=[0], weight=1, eps=1, poles=[1], residues=[-1], init='1') sage: L Dokchitser L-series of conductor 1 and weight 1 sage: L(1) Traceback (most recent call last): ... ArithmeticError: ### user error: L*(s) has a pole at s=1.000000000000000000 sage: L(2) 1.64493406684823 sage: L(2, 1.1) 1.64493406684823 sage: L.derivative(2) -0.937548254315844 sage: h = RR('0.0000000000001') sage: (zeta(2+h) - zeta(2))/h -0.937028232783632 sage: L.taylor_series(2, k=5) 1.64493406684823 - 0.937548254315844*z + 0.994640117149451*z^2 - 1.00002430047384*z^3 + 1.00006193307235*z^4 + O(z^5)
RANK 1 ELLIPTIC CURVE:
We compute with the
-series of a rank
curve.
sage: E = EllipticCurve('37a') sage: L = E.lseries().dokchitser(); L Dokchitser L-function associated to Elliptic Curve defined by y^2 + y = x^3 - x over Rational Field sage: L(1) 0 sage: L.derivative(1) 0.305999773834052 sage: L.derivative(1,2) 0.373095594536324 sage: L.num_coeffs() 48 sage: L.taylor_series(1,4) 0.305999773834052*z + 0.186547797268162*z^2 - 0.136791463097188*z^3 + O(z^4) sage: L.check_functional_equation() 6.11218974800000e-18 # 32-bit 6.04442711160669e-18 # 64-bit
RANK 2 ELLIPTIC CURVE:
We compute the leading coefficient and Taylor expansion of the
-series
of a rank
curve.
sage: E = EllipticCurve('389a') sage: L = E.lseries().dokchitser() sage: L.num_coeffs () 156 sage: L.derivative(1,E.rank()) 1.51863300057685 sage: L.taylor_series(1,4) -1.28158145691931e-23 + (7.26268290635587e-24)*z + 0.759316500288427*z^2 - 0.430302337583362*z^3 + O(z^4) # 32-bit -2.69129566562797e-23 + (1.52514901968783e-23)*z + 0.759316500288427*z^2 - 0.430302337583362*z^3 + O(z^4) # 64-bit
RAMANUJAN DELTA L-FUNCTION:
The coefficients are given by Ramanujan's tau function:
sage: L = Dokchitser(conductor=1, gammaV=[0,1], weight=12, eps=1) sage: pari_precode = 'tau(n)=(5*sigma(n,3)+7*sigma(n,5))*n/12 - 35*sum(k=1,n-1,(6*k-4*(n-k))*sigma(k,3)*sigma(n-k,5))' sage: L.init_coeffs('tau(k)', pari_precode=pari_precode)
We redefine the default bound on the coefficients: Deligne's
estimate on tau(n) is better than the default
coefgrow(n)=
(by a factor 1024), so re-defining
coefgrow() improves efficiency (slightly faster).
sage: L.num_coeffs() 12 sage: L.set_coeff_growth('2*n^(11/2)') sage: L.num_coeffs() 11
Now we're ready to evaluate, etc.
sage: L(1) 0.0374412812685155 sage: L(1, 1.1) 0.0374412812685155 sage: L.taylor_series(1,3) 0.0374412812685155 + 0.0709221123619322*z + 0.0380744761270520*z^2 + O(z^3)
self, conductor, gammaV, weight, eps, [poles=[]], [residues=automatic], [prec=53], [init=None]) |
Initialization of Dokchitser calculator
sage: L = Dokchitser(conductor=1, gammaV=[0], weight=1, eps=1, poles=[1], residues=[-1], init='1') sage: L.num_coeffs() 4
Functions: check_functional_equation,
derivative,
gp,
init_coeffs,
num_coeffs,
set_coeff_growth,
taylor_series
self, [T=1.2]) |
Verifies how well numerically the functional equation is
satisfied, and also determines the residues if
self.poles != []
and residues='automatic'.
More specifically: for
(default 1.2),
self.check_functional_equation(T)
should ideally
return 0 (to the current precision).
sage: L = Dokchitser(conductor=1, gammaV=[0], weight=1, eps=1, poles=[1], residues=[-1], init='1') sage: L.check_functional_equation () -2.71050543200000e-20 # 32-bit -2.71050543121376e-20 # 64-bit
If we choose the sign in functional equation for the
function incorrectly, the functional equation doesn't check out.
sage: L = Dokchitser(conductor=1, gammaV=[0], weight=1, eps=-11, poles=[1], residues=[-1], init='1') sage: L.check_functional_equation () -9.73967861488124
self, s, [k=1]) |
Return the
-th derivative of the
-series at
.
WARNING: If
is greater than the order of vanishing of
at
you may get nonsense.
sage: E = EllipticCurve('389a') sage: L = E.lseries().dokchitser() sage: L.derivative(1,E.rank()) 1.51863300057685
self) |
Return the gp interpreter that is used to implement this Dokchitser L-function.
sage: E = EllipticCurve('11a') sage: L = E.lseries().dokchitser() sage: L(2) 0.546048036215014 sage: L.gp() GP/PARI interpreter
self, v, [cutoff=1], [w=None], [pari_precode=], [max_imaginary_part=0], [max_asymp_coeffs=40]) |
Set the coefficients
of the
-series. If
is
not equal to its dual, pass the coefficients of the dual as
the second optional argument.
Input:
sage: L = Dokchitser(conductor=1, gammaV=[0,1], weight=12, eps=1) sage: pari_precode = 'tau(n)=(5*sigma(n,3)+7*sigma(n,5))*n/12 - 35*sum(k=1,n-1,(6*k-4*(n-k))*sigma(k,3)*sigma(n-k,5))' sage: L.init_coeffs('tau(k)', pari_precode=pari_precode)
self, [T=1]) |
Return number of coefficients
that are needed in order
to perform most relevant
-function computations to the
desired precision.
sage: E = EllipticCurve('11a') sage: L = E.lseries().dokchitser() sage: L.num_coeffs() 26 sage: E = EllipticCurve('5077a') sage: L = E.lseries().dokchitser() sage: L.num_coeffs() 568 sage: L = Dokchitser(conductor=1, gammaV=[0], weight=1, eps=1, poles=[1], residues=[-1], init='1') sage: L.num_coeffs() 4
self, coefgrow) |
You might have to redefine the coefficient growth function if
the
of the
-series are not given by the following
PARI function:
coefgrow(n) = if(length(Lpoles), 1.5*n^(vecmax(real(Lpoles))-1), sqrt(4*n)^(weight-1));
Input:
sage: L = Dokchitser(conductor=1, gammaV=[0,1], weight=12, eps=1) sage: pari_precode = 'tau(n)=(5*sigma(n,3)+7*sigma(n,5))*n/12 - 35*sum(k=1,n-1,(6*k-4*(n-k))*sigma(k,3)*sigma(n-k,5))' sage: L.init_coeffs('tau(k)', pari_precode=pari_precode) sage: L.set_coeff_growth('2*n^(11/2)') sage: L(1) 0.0374412812685155
self, [a=0], [k=6], [var=z]) |
Return the first
terms of the Taylor series expansion of the
-series about
.
This is returned as a series in var, where you should view var
as equal to
. Thus this function returns the formal
power series whose coefficients are
.
Input:
sage: L = Dokchitser(conductor=1, gammaV=[0], weight=1, eps=1, poles=[1], residues=[-1], init='1') sage: L.taylor_series(2, 3) 1.64493406684823 - 0.937548254315844*z + 0.994640117149451*z^2 + O(z^3) sage: E = EllipticCurve('37a') sage: L = E.lseries().dokchitser() sage: L.taylor_series(1) 0.305999773834052*z + 0.186547797268162*z^2 - 0.136791463097188*z^3 + 0.0161066468496401*z^4 + 0.0185955175398802*z^5 + O(z^6)
We compute a Taylor series where each coefficient is to high precision.
sage: E = EllipticCurve('389a') sage: L = E.lseries().dokchitser(200) sage: L.taylor_series(1,3) 6.2239725530250970363983975962696997888173850098274602272589e-73 + (-3.527106203544994604921190324282024612952450859320...e-73)*z + 0.75931650028842677023019260789472201907809751649492435158581*z^2 + O(z^3) # 32-bit 6.2239725530250970363983942830358807065824196704980264311105e-73 + (-3.5271062035449946049211884466825561834034352383203420991340e-73)*z + 0.75931650028842677023019260789472201907809751649492435158581*z^2 + O(z^3) # 64-bit
Special Functions: __call__,
__init__,
__reduce__,
_clear_value_cache,
_Dokchitser__check_init,
_Dokchitser__to_CC,
_gp_eval,
_repr_
self, s, [c=None]) |
Input:
NOTE: Evaluation of the function takes a long time, so each
evaluation is cached. Call self._clear_value_cache()
to clear the evaluation cache.
sage: E = EllipticCurve('5077a') sage: L = E.lseries().dokchitser(100) sage: L(1) 0 sage: L(1+I) -1.3085436607849493358323930438 + 0.81298000036784359634835412129*I
See About this document... for information on suggesting changes.