Source code for sympy.functions.special.error_functions

""" This module contains various functions that are special cases
    of incomplete gamma functions. It should probably be renamed. """

from sympy.core import EulerGamma # Must be imported from core, not core.numbers
from sympy.core.add import Add
from sympy.core.cache import cacheit
from sympy.core.function import Function, ArgumentIndexError, expand_mul
from sympy.core.logic import fuzzy_or
from sympy.core.numbers import I, pi, Rational, Integer
from sympy.core.relational import is_eq
from sympy.core.power import Pow
from sympy.core.singleton import S
from sympy.core.symbol import Dummy, uniquely_named_symbol
from sympy.core.sympify import sympify
from sympy.functions.combinatorial.factorials import factorial, factorial2, RisingFactorial
from sympy.functions.elementary.complexes import  polar_lift, re, unpolarify
from sympy.functions.elementary.integers import ceiling, floor
from sympy.functions.elementary.miscellaneous import sqrt, root
from sympy.functions.elementary.exponential import exp, log, exp_polar
from sympy.functions.elementary.hyperbolic import cosh, sinh
from sympy.functions.elementary.trigonometric import cos, sin, sinc
from sympy.functions.special.hyper import hyper, meijerg

# TODO series expansions
# TODO see the "Note:" in Ei

# Helper function
def real_to_real_as_real_imag(self, deep=True, **hints):
    if self.args[0].is_extended_real:
        if deep:
            hints['complex'] = False
            return (self.expand(deep, **hints), S.Zero)
        else:
            return (self, S.Zero)
    if deep:
        x, y = self.args[0].expand(deep, **hints).as_real_imag()
    else:
        x, y = self.args[0].as_real_imag()
    re = (self.func(x + I*y) + self.func(x - I*y))/2
    im = (self.func(x + I*y) - self.func(x - I*y))/(2*I)
    return (re, im)


###############################################################################
################################ ERROR FUNCTION ###############################
###############################################################################


[docs] class erf(Function): r""" The Gauss error function. Explanation =========== This function is defined as: .. math :: \mathrm{erf}(x) = \frac{2}{\sqrt{\pi}} \int_0^x e^{-t^2} \mathrm{d}t. Examples ======== >>> from sympy import I, oo, erf >>> from sympy.abc import z Several special values are known: >>> erf(0) 0 >>> erf(oo) 1 >>> erf(-oo) -1 >>> erf(I*oo) oo*I >>> erf(-I*oo) -oo*I In general one can pull out factors of -1 and $I$ from the argument: >>> erf(-z) -erf(z) The error function obeys the mirror symmetry: >>> from sympy import conjugate >>> conjugate(erf(z)) erf(conjugate(z)) Differentiation with respect to $z$ is supported: >>> from sympy import diff >>> diff(erf(z), z) 2*exp(-z**2)/sqrt(pi) We can numerically evaluate the error function to arbitrary precision on the whole complex plane: >>> erf(4).evalf(30) 0.999999984582742099719981147840 >>> erf(-4*I).evalf(30) -1296959.73071763923152794095062*I See Also ======== erfc: Complementary error function. erfi: Imaginary error function. erf2: Two-argument error function. erfinv: Inverse error function. erfcinv: Inverse Complementary error function. erf2inv: Inverse two-argument error function. References ========== .. [1] https://en.wikipedia.org/wiki/Error_function .. [2] https://dlmf.nist.gov/7 .. [3] https://mathworld.wolfram.com/Erf.html .. [4] https://functions.wolfram.com/GammaBetaErf/Erf """ unbranched = True def fdiff(self, argindex=1): if argindex == 1: return 2*exp(-self.args[0]**2)/sqrt(pi) else: raise ArgumentIndexError(self, argindex)
[docs] def inverse(self, argindex=1): """ Returns the inverse of this function. """ return erfinv
@classmethod def eval(cls, arg): if arg.is_Number: if arg is S.NaN: return S.NaN elif arg is S.Infinity: return S.One elif arg is S.NegativeInfinity: return S.NegativeOne elif arg.is_zero: return S.Zero if isinstance(arg, erfinv): return arg.args[0] if isinstance(arg, erfcinv): return S.One - arg.args[0] if arg.is_zero: return S.Zero # Only happens with unevaluated erf2inv if isinstance(arg, erf2inv) and arg.args[0].is_zero: return arg.args[1] # Try to pull out factors of I t = arg.extract_multiplicatively(I) if t in (S.Infinity, S.NegativeInfinity): return arg # Try to pull out factors of -1 if arg.could_extract_minus_sign(): return -cls(-arg) @staticmethod @cacheit def taylor_term(n, x, *previous_terms): if n < 0 or n % 2 == 0: return S.Zero else: x = sympify(x) k = floor((n - 1)/S(2)) if len(previous_terms) > 2: return -previous_terms[-2] * x**2 * (n - 2)/(n*k) else: return 2*S.NegativeOne**k * x**n/(n*factorial(k)*sqrt(pi)) def _eval_conjugate(self): return self.func(self.args[0].conjugate()) def _eval_is_real(self): if self.args[0].is_extended_real is True: return True # There are cases where erf(z) becomes a real number # even if z is a complex number def _eval_is_imaginary(self): if self.args[0].is_imaginary is True: return True def _eval_is_finite(self): z = self.args[0] return fuzzy_or([z.is_finite, z.is_extended_real]) def _eval_is_zero(self): if self.args[0].is_extended_real is True: return self.args[0].is_zero def _eval_is_positive(self): if self.args[0].is_extended_real is True: return self.args[0].is_extended_positive def _eval_is_negative(self): if self.args[0].is_extended_real is True: return self.args[0].is_extended_negative def _eval_rewrite_as_uppergamma(self, z, **kwargs): from sympy.functions.special.gamma_functions import uppergamma return sqrt(z**2)/z*(S.One - uppergamma(S.Half, z**2)/sqrt(pi)) def _eval_rewrite_as_fresnels(self, z, **kwargs): arg = (S.One - I)*z/sqrt(pi) return (S.One + I)*(fresnelc(arg) - I*fresnels(arg)) def _eval_rewrite_as_fresnelc(self, z, **kwargs): arg = (S.One - I)*z/sqrt(pi) return (S.One + I)*(fresnelc(arg) - I*fresnels(arg)) def _eval_rewrite_as_meijerg(self, z, **kwargs): return z/sqrt(pi)*meijerg([S.Half], [], [0], [Rational(-1, 2)], z**2) def _eval_rewrite_as_hyper(self, z, **kwargs): return 2*z/sqrt(pi)*hyper([S.Half], [3*S.Half], -z**2) def _eval_rewrite_as_expint(self, z, **kwargs): return sqrt(z**2)/z - z*expint(S.Half, z**2)/sqrt(pi) def _eval_rewrite_as_tractable(self, z, limitvar=None, **kwargs): from sympy.series.limits import limit if limitvar: lim = limit(z, limitvar, S.Infinity) if lim is S.NegativeInfinity: return S.NegativeOne + _erfs(-z)*exp(-z**2) return S.One - _erfs(z)*exp(-z**2) def _eval_rewrite_as_erfc(self, z, **kwargs): return S.One - erfc(z) def _eval_rewrite_as_erfi(self, z, **kwargs): return -I*erfi(I*z) def _eval_as_leading_term(self, x, logx, cdir): arg = self.args[0].as_leading_term(x, logx=logx, cdir=cdir) arg0 = arg.subs(x, 0) if arg0 is S.ComplexInfinity: arg0 = arg.limit(x, 0, dir='-' if cdir == -1 else '+') if x in arg.free_symbols and arg0.is_zero: return 2*arg/sqrt(pi) else: return self.func(arg0) def _eval_aseries(self, n, args0, x, logx): from sympy.series.order import Order point = args0[0] if point in [S.Infinity, S.NegativeInfinity]: z = self.args[0] try: _, ex = z.leadterm(x) except (ValueError, NotImplementedError): return self ex = -ex # as x->1/x for aseries if ex.is_positive: newn = ceiling(n/ex) s = [S.NegativeOne**k * factorial2(2*k - 1) / (z**(2*k + 1) * 2**k) for k in range(newn)] + [Order(1/z**newn, x)] return S.One - (exp(-z**2)/sqrt(pi)) * Add(*s) return super(erf, self)._eval_aseries(n, args0, x, logx) as_real_imag = real_to_real_as_real_imag
[docs] class erfc(Function): r""" Complementary Error Function. Explanation =========== The function is defined as: .. math :: \mathrm{erfc}(x) = \frac{2}{\sqrt{\pi}} \int_x^\infty e^{-t^2} \mathrm{d}t Examples ======== >>> from sympy import I, oo, erfc >>> from sympy.abc import z Several special values are known: >>> erfc(0) 1 >>> erfc(oo) 0 >>> erfc(-oo) 2 >>> erfc(I*oo) -oo*I >>> erfc(-I*oo) oo*I The error function obeys the mirror symmetry: >>> from sympy import conjugate >>> conjugate(erfc(z)) erfc(conjugate(z)) Differentiation with respect to $z$ is supported: >>> from sympy import diff >>> diff(erfc(z), z) -2*exp(-z**2)/sqrt(pi) It also follows >>> erfc(-z) 2 - erfc(z) We can numerically evaluate the complementary error function to arbitrary precision on the whole complex plane: >>> erfc(4).evalf(30) 0.0000000154172579002800188521596734869 >>> erfc(4*I).evalf(30) 1.0 - 1296959.73071763923152794095062*I See Also ======== erf: Gaussian error function. erfi: Imaginary error function. erf2: Two-argument error function. erfinv: Inverse error function. erfcinv: Inverse Complementary error function. erf2inv: Inverse two-argument error function. References ========== .. [1] https://en.wikipedia.org/wiki/Error_function .. [2] https://dlmf.nist.gov/7 .. [3] https://mathworld.wolfram.com/Erfc.html .. [4] https://functions.wolfram.com/GammaBetaErf/Erfc """ unbranched = True def fdiff(self, argindex=1): if argindex == 1: return -2*exp(-self.args[0]**2)/sqrt(pi) else: raise ArgumentIndexError(self, argindex)
[docs] def inverse(self, argindex=1): """ Returns the inverse of this function. """ return erfcinv
@classmethod def eval(cls, arg): if arg.is_Number: if arg is S.NaN: return S.NaN elif arg is S.Infinity: return S.Zero elif arg.is_zero: return S.One if isinstance(arg, erfinv): return S.One - arg.args[0] if isinstance(arg, erfcinv): return arg.args[0] if arg.is_zero: return S.One # Try to pull out factors of I t = arg.extract_multiplicatively(I) if t in (S.Infinity, S.NegativeInfinity): return -arg # Try to pull out factors of -1 if arg.could_extract_minus_sign(): return 2 - cls(-arg) @staticmethod @cacheit def taylor_term(n, x, *previous_terms): if n == 0: return S.One elif n < 0 or n % 2 == 0: return S.Zero else: x = sympify(x) k = floor((n - 1)/S(2)) if len(previous_terms) > 2: return -previous_terms[-2] * x**2 * (n - 2)/(n*k) else: return -2*S.NegativeOne**k * x**n/(n*factorial(k)*sqrt(pi)) def _eval_conjugate(self): return self.func(self.args[0].conjugate()) def _eval_is_real(self): if self.args[0].is_extended_real is True: return True if self.args[0].is_imaginary is True: return False def _eval_rewrite_as_tractable(self, z, limitvar=None, **kwargs): return self.rewrite(erf).rewrite("tractable", deep=True, limitvar=limitvar) def _eval_rewrite_as_erf(self, z, **kwargs): return S.One - erf(z) def _eval_rewrite_as_erfi(self, z, **kwargs): return S.One + I*erfi(I*z) def _eval_rewrite_as_fresnels(self, z, **kwargs): arg = (S.One - I)*z/sqrt(pi) return S.One - (S.One + I)*(fresnelc(arg) - I*fresnels(arg)) def _eval_rewrite_as_fresnelc(self, z, **kwargs): arg = (S.One-I)*z/sqrt(pi) return S.One - (S.One + I)*(fresnelc(arg) - I*fresnels(arg)) def _eval_rewrite_as_meijerg(self, z, **kwargs): return S.One - z/sqrt(pi)*meijerg([S.Half], [], [0], [Rational(-1, 2)], z**2) def _eval_rewrite_as_hyper(self, z, **kwargs): return S.One - 2*z/sqrt(pi)*hyper([S.Half], [3*S.Half], -z**2) def _eval_rewrite_as_uppergamma(self, z, **kwargs): from sympy.functions.special.gamma_functions import uppergamma return S.One - sqrt(z**2)/z*(S.One - uppergamma(S.Half, z**2)/sqrt(pi)) def _eval_rewrite_as_expint(self, z, **kwargs): return S.One - sqrt(z**2)/z + z*expint(S.Half, z**2)/sqrt(pi) def _eval_expand_func(self, **hints): return self.rewrite(erf) def _eval_as_leading_term(self, x, logx, cdir): arg = self.args[0].as_leading_term(x, logx=logx, cdir=cdir) arg0 = arg.subs(x, 0) if arg0 is S.ComplexInfinity: arg0 = arg.limit(x, 0, dir='-' if cdir == -1 else '+') if arg0.is_zero: return S.One else: return self.func(arg0) as_real_imag = real_to_real_as_real_imag def _eval_aseries(self, n, args0, x, logx): return S.One - erf(*self.args)._eval_aseries(n, args0, x, logx)
[docs] class erfi(Function): r""" Imaginary error function. Explanation =========== The function erfi is defined as: .. math :: \mathrm{erfi}(x) = \frac{2}{\sqrt{\pi}} \int_0^x e^{t^2} \mathrm{d}t Examples ======== >>> from sympy import I, oo, erfi >>> from sympy.abc import z Several special values are known: >>> erfi(0) 0 >>> erfi(oo) oo >>> erfi(-oo) -oo >>> erfi(I*oo) I >>> erfi(-I*oo) -I In general one can pull out factors of -1 and $I$ from the argument: >>> erfi(-z) -erfi(z) >>> from sympy import conjugate >>> conjugate(erfi(z)) erfi(conjugate(z)) Differentiation with respect to $z$ is supported: >>> from sympy import diff >>> diff(erfi(z), z) 2*exp(z**2)/sqrt(pi) We can numerically evaluate the imaginary error function to arbitrary precision on the whole complex plane: >>> erfi(2).evalf(30) 18.5648024145755525987042919132 >>> erfi(-2*I).evalf(30) -0.995322265018952734162069256367*I See Also ======== erf: Gaussian error function. erfc: Complementary error function. erf2: Two-argument error function. erfinv: Inverse error function. erfcinv: Inverse Complementary error function. erf2inv: Inverse two-argument error function. References ========== .. [1] https://en.wikipedia.org/wiki/Error_function .. [2] https://mathworld.wolfram.com/Erfi.html .. [3] https://functions.wolfram.com/GammaBetaErf/Erfi """ unbranched = True def fdiff(self, argindex=1): if argindex == 1: return 2*exp(self.args[0]**2)/sqrt(pi) else: raise ArgumentIndexError(self, argindex) @classmethod def eval(cls, z): if z.is_Number: if z is S.NaN: return S.NaN elif z.is_zero: return S.Zero elif z is S.Infinity: return S.Infinity if z.is_zero: return S.Zero # Try to pull out factors of -1 if z.could_extract_minus_sign(): return -cls(-z) # Try to pull out factors of I nz = z.extract_multiplicatively(I) if nz is not None: if nz is S.Infinity: return I if isinstance(nz, erfinv): return I*nz.args[0] if isinstance(nz, erfcinv): return I*(S.One - nz.args[0]) # Only happens with unevaluated erf2inv if isinstance(nz, erf2inv) and nz.args[0].is_zero: return I*nz.args[1] @staticmethod @cacheit def taylor_term(n, x, *previous_terms): if n < 0 or n % 2 == 0: return S.Zero else: x = sympify(x) k = floor((n - 1)/S(2)) if len(previous_terms) > 2: return previous_terms[-2] * x**2 * (n - 2)/(n*k) else: return 2 * x**n/(n*factorial(k)*sqrt(pi)) def _eval_conjugate(self): return self.func(self.args[0].conjugate()) def _eval_is_extended_real(self): return self.args[0].is_extended_real def _eval_is_zero(self): return self.args[0].is_zero def _eval_rewrite_as_tractable(self, z, limitvar=None, **kwargs): return self.rewrite(erf).rewrite("tractable", deep=True, limitvar=limitvar) def _eval_rewrite_as_erf(self, z, **kwargs): return -I*erf(I*z) def _eval_rewrite_as_erfc(self, z, **kwargs): return I*erfc(I*z) - I def _eval_rewrite_as_fresnels(self, z, **kwargs): arg = (S.One + I)*z/sqrt(pi) return (S.One - I)*(fresnelc(arg) - I*fresnels(arg)) def _eval_rewrite_as_fresnelc(self, z, **kwargs): arg = (S.One + I)*z/sqrt(pi) return (S.One - I)*(fresnelc(arg) - I*fresnels(arg)) def _eval_rewrite_as_meijerg(self, z, **kwargs): return z/sqrt(pi)*meijerg([S.Half], [], [0], [Rational(-1, 2)], -z**2) def _eval_rewrite_as_hyper(self, z, **kwargs): return 2*z/sqrt(pi)*hyper([S.Half], [3*S.Half], z**2) def _eval_rewrite_as_uppergamma(self, z, **kwargs): from sympy.functions.special.gamma_functions import uppergamma return sqrt(-z**2)/z*(uppergamma(S.Half, -z**2)/sqrt(pi) - S.One) def _eval_rewrite_as_expint(self, z, **kwargs): return sqrt(-z**2)/z - z*expint(S.Half, -z**2)/sqrt(pi) def _eval_expand_func(self, **hints): return self.rewrite(erf) as_real_imag = real_to_real_as_real_imag def _eval_as_leading_term(self, x, logx, cdir): arg = self.args[0].as_leading_term(x, logx=logx, cdir=cdir) arg0 = arg.subs(x, 0) if x in arg.free_symbols and arg0.is_zero: return 2*arg/sqrt(pi) elif arg0.is_finite: return self.func(arg0) return self.func(arg) def _eval_aseries(self, n, args0, x, logx): from sympy.series.order import Order point = args0[0] if point is S.Infinity: z = self.args[0] s = [factorial2(2*k - 1) / (2**k * z**(2*k + 1)) for k in range(n)] + [Order(1/z**n, x)] return -I + (exp(z**2)/sqrt(pi)) * Add(*s) return super(erfi, self)._eval_aseries(n, args0, x, logx)
[docs] class erf2(Function): r""" Two-argument error function. Explanation =========== This function is defined as: .. math :: \mathrm{erf2}(x, y) = \frac{2}{\sqrt{\pi}} \int_x^y e^{-t^2} \mathrm{d}t Examples ======== >>> from sympy import oo, erf2 >>> from sympy.abc import x, y Several special values are known: >>> erf2(0, 0) 0 >>> erf2(x, x) 0 >>> erf2(x, oo) 1 - erf(x) >>> erf2(x, -oo) -erf(x) - 1 >>> erf2(oo, y) erf(y) - 1 >>> erf2(-oo, y) erf(y) + 1 In general one can pull out factors of -1: >>> erf2(-x, -y) -erf2(x, y) The error function obeys the mirror symmetry: >>> from sympy import conjugate >>> conjugate(erf2(x, y)) erf2(conjugate(x), conjugate(y)) Differentiation with respect to $x$, $y$ is supported: >>> from sympy import diff >>> diff(erf2(x, y), x) -2*exp(-x**2)/sqrt(pi) >>> diff(erf2(x, y), y) 2*exp(-y**2)/sqrt(pi) See Also ======== erf: Gaussian error function. erfc: Complementary error function. erfi: Imaginary error function. erfinv: Inverse error function. erfcinv: Inverse Complementary error function. erf2inv: Inverse two-argument error function. References ========== .. [1] https://functions.wolfram.com/GammaBetaErf/Erf2/ """ def fdiff(self, argindex): x, y = self.args if argindex == 1: return -2*exp(-x**2)/sqrt(pi) elif argindex == 2: return 2*exp(-y**2)/sqrt(pi) else: raise ArgumentIndexError(self, argindex) @classmethod def eval(cls, x, y): chk = (S.Infinity, S.NegativeInfinity, S.Zero) if x is S.NaN or y is S.NaN: return S.NaN elif x == y: return S.Zero elif x in chk or y in chk: return erf(y) - erf(x) if isinstance(y, erf2inv) and y.args[0] == x: return y.args[1] if x.is_zero or y.is_zero or x.is_extended_real and x.is_infinite or \ y.is_extended_real and y.is_infinite: return erf(y) - erf(x) #Try to pull out -1 factor sign_x = x.could_extract_minus_sign() sign_y = y.could_extract_minus_sign() if (sign_x and sign_y): return -cls(-x, -y) elif (sign_x or sign_y): return erf(y)-erf(x) def _eval_conjugate(self): return self.func(self.args[0].conjugate(), self.args[1].conjugate()) def _eval_is_extended_real(self): return self.args[0].is_extended_real and self.args[1].is_extended_real def _eval_rewrite_as_erf(self, x, y, **kwargs): return erf(y) - erf(x) def _eval_rewrite_as_erfc(self, x, y, **kwargs): return erfc(x) - erfc(y) def _eval_rewrite_as_erfi(self, x, y, **kwargs): return I*(erfi(I*x)-erfi(I*y)) def _eval_rewrite_as_fresnels(self, x, y, **kwargs): return erf(y).rewrite(fresnels) - erf(x).rewrite(fresnels) def _eval_rewrite_as_fresnelc(self, x, y, **kwargs): return erf(y).rewrite(fresnelc) - erf(x).rewrite(fresnelc) def _eval_rewrite_as_meijerg(self, x, y, **kwargs): return erf(y).rewrite(meijerg) - erf(x).rewrite(meijerg) def _eval_rewrite_as_hyper(self, x, y, **kwargs): return erf(y).rewrite(hyper) - erf(x).rewrite(hyper) def _eval_rewrite_as_uppergamma(self, x, y, **kwargs): from sympy.functions.special.gamma_functions import uppergamma return (sqrt(y**2)/y*(S.One - uppergamma(S.Half, y**2)/sqrt(pi)) - sqrt(x**2)/x*(S.One - uppergamma(S.Half, x**2)/sqrt(pi))) def _eval_rewrite_as_expint(self, x, y, **kwargs): return erf(y).rewrite(expint) - erf(x).rewrite(expint) def _eval_expand_func(self, **hints): return self.rewrite(erf) def _eval_is_zero(self): return is_eq(*self.args)
[docs] class erfinv(Function): r""" Inverse Error Function. The erfinv function is defined as: .. math :: \mathrm{erf}(x) = y \quad \Rightarrow \quad \mathrm{erfinv}(y) = x Examples ======== >>> from sympy import erfinv >>> from sympy.abc import x Several special values are known: >>> erfinv(0) 0 >>> erfinv(1) oo Differentiation with respect to $x$ is supported: >>> from sympy import diff >>> diff(erfinv(x), x) sqrt(pi)*exp(erfinv(x)**2)/2 We can numerically evaluate the inverse error function to arbitrary precision on [-1, 1]: >>> erfinv(0.2).evalf(30) 0.179143454621291692285822705344 See Also ======== erf: Gaussian error function. erfc: Complementary error function. erfi: Imaginary error function. erf2: Two-argument error function. erfcinv: Inverse Complementary error function. erf2inv: Inverse two-argument error function. References ========== .. [1] https://en.wikipedia.org/wiki/Error_function#Inverse_functions .. [2] https://functions.wolfram.com/GammaBetaErf/InverseErf/ """ def fdiff(self, argindex =1): if argindex == 1: return sqrt(pi)*exp(self.func(self.args[0])**2)*S.Half else : raise ArgumentIndexError(self, argindex)
[docs] def inverse(self, argindex=1): """ Returns the inverse of this function. """ return erf
@classmethod def eval(cls, z): if z is S.NaN: return S.NaN elif z is S.NegativeOne: return S.NegativeInfinity elif z.is_zero: return S.Zero elif z is S.One: return S.Infinity if isinstance(z, erf) and z.args[0].is_extended_real: return z.args[0] if z.is_zero: return S.Zero # Try to pull out factors of -1 nz = z.extract_multiplicatively(-1) if nz is not None and (isinstance(nz, erf) and (nz.args[0]).is_extended_real): return -nz.args[0] def _eval_rewrite_as_erfcinv(self, z, **kwargs): return erfcinv(1-z) def _eval_is_zero(self): return self.args[0].is_zero
[docs] class erfcinv (Function): r""" Inverse Complementary Error Function. The erfcinv function is defined as: .. math :: \mathrm{erfc}(x) = y \quad \Rightarrow \quad \mathrm{erfcinv}(y) = x Examples ======== >>> from sympy import erfcinv >>> from sympy.abc import x Several special values are known: >>> erfcinv(1) 0 >>> erfcinv(0) oo Differentiation with respect to $x$ is supported: >>> from sympy import diff >>> diff(erfcinv(x), x) -sqrt(pi)*exp(erfcinv(x)**2)/2 See Also ======== erf: Gaussian error function. erfc: Complementary error function. erfi: Imaginary error function. erf2: Two-argument error function. erfinv: Inverse error function. erf2inv: Inverse two-argument error function. References ========== .. [1] https://en.wikipedia.org/wiki/Error_function#Inverse_functions .. [2] https://functions.wolfram.com/GammaBetaErf/InverseErfc/ """ def fdiff(self, argindex =1): if argindex == 1: return -sqrt(pi)*exp(self.func(self.args[0])**2)*S.Half else: raise ArgumentIndexError(self, argindex)
[docs] def inverse(self, argindex=1): """ Returns the inverse of this function. """ return erfc
@classmethod def eval(cls, z): if z is S.NaN: return S.NaN elif z.is_zero: return S.Infinity elif z is S.One: return S.Zero elif z == 2: return S.NegativeInfinity if z.is_zero: return S.Infinity def _eval_rewrite_as_erfinv(self, z, **kwargs): return erfinv(1-z) def _eval_is_zero(self): return (self.args[0] - 1).is_zero def _eval_is_infinite(self): z = self.args[0] return fuzzy_or([z.is_zero, is_eq(z, Integer(2))])
[docs] class erf2inv(Function): r""" Two-argument Inverse error function. The erf2inv function is defined as: .. math :: \mathrm{erf2}(x, w) = y \quad \Rightarrow \quad \mathrm{erf2inv}(x, y) = w Examples ======== >>> from sympy import erf2inv, oo >>> from sympy.abc import x, y Several special values are known: >>> erf2inv(0, 0) 0 >>> erf2inv(1, 0) 1 >>> erf2inv(0, 1) oo >>> erf2inv(0, y) erfinv(y) >>> erf2inv(oo, y) erfcinv(-y) Differentiation with respect to $x$ and $y$ is supported: >>> from sympy import diff >>> diff(erf2inv(x, y), x) exp(-x**2 + erf2inv(x, y)**2) >>> diff(erf2inv(x, y), y) sqrt(pi)*exp(erf2inv(x, y)**2)/2 See Also ======== erf: Gaussian error function. erfc: Complementary error function. erfi: Imaginary error function. erf2: Two-argument error function. erfinv: Inverse error function. erfcinv: Inverse complementary error function. References ========== .. [1] https://functions.wolfram.com/GammaBetaErf/InverseErf2/ """ def fdiff(self, argindex): x, y = self.args if argindex == 1: return exp(self.func(x,y)**2-x**2) elif argindex == 2: return sqrt(pi)*S.Half*exp(self.func(x,y)**2) else: raise ArgumentIndexError(self, argindex) @classmethod def eval(cls, x, y): if x is S.NaN or y is S.NaN: return S.NaN elif x.is_zero and y.is_zero: return S.Zero elif x.is_zero and y is S.One: return S.Infinity elif x is S.One and y.is_zero: return S.One elif x.is_zero: return erfinv(y) elif x is S.Infinity: return erfcinv(-y) elif y.is_zero: return x elif y is S.Infinity: return erfinv(x) if x.is_zero: if y.is_zero: return S.Zero else: return erfinv(y) if y.is_zero: return x def _eval_is_zero(self): x, y = self.args if x.is_zero and y.is_zero: return True
############################################################################### #################### EXPONENTIAL INTEGRALS #################################### ###############################################################################
[docs] class Ei(Function): r""" The classical exponential integral. Explanation =========== For use in SymPy, this function is defined as .. math:: \operatorname{Ei}(x) = \sum_{n=1}^\infty \frac{x^n}{n\, n!} + \log(x) + \gamma, where $\gamma$ is the Euler-Mascheroni constant. If $x$ is a polar number, this defines an analytic function on the Riemann surface of the logarithm. Otherwise this defines an analytic function in the cut plane $\mathbb{C} \setminus (-\infty, 0]$. **Background** The name exponential integral comes from the following statement: .. math:: \operatorname{Ei}(x) = \int_{-\infty}^x \frac{e^t}{t} \mathrm{d}t If the integral is interpreted as a Cauchy principal value, this statement holds for $x > 0$ and $\operatorname{Ei}(x)$ as defined above. Examples ======== >>> from sympy import Ei, polar_lift, exp_polar, I, pi >>> from sympy.abc import x >>> Ei(-1) Ei(-1) This yields a real value: >>> Ei(-1).n(chop=True) -0.219383934395520 On the other hand the analytic continuation is not real: >>> Ei(polar_lift(-1)).n(chop=True) -0.21938393439552 + 3.14159265358979*I The exponential integral has a logarithmic branch point at the origin: >>> Ei(x*exp_polar(2*I*pi)) Ei(x) + 2*I*pi Differentiation is supported: >>> Ei(x).diff(x) exp(x)/x The exponential integral is related to many other special functions. For example: >>> from sympy import expint, Shi >>> Ei(x).rewrite(expint) -expint(1, x*exp_polar(I*pi)) - I*pi >>> Ei(x).rewrite(Shi) Chi(x) + Shi(x) See Also ======== expint: Generalised exponential integral. E1: Special case of the generalised exponential integral. li: Logarithmic integral. Li: Offset logarithmic integral. Si: Sine integral. Ci: Cosine integral. Shi: Hyperbolic sine integral. Chi: Hyperbolic cosine integral. uppergamma: Upper incomplete gamma function. References ========== .. [1] https://dlmf.nist.gov/6.6 .. [2] https://en.wikipedia.org/wiki/Exponential_integral .. [3] Abramowitz & Stegun, section 5: https://web.archive.org/web/20201128173312/http://people.math.sfu.ca/~cbm/aands/page_228.htm """ @classmethod def eval(cls, z): if z.is_zero: return S.NegativeInfinity elif z is S.Infinity: return S.Infinity elif z is S.NegativeInfinity: return S.Zero if z.is_zero: return S.NegativeInfinity nz, n = z.extract_branch_factor() if n: return Ei(nz) + 2*I*pi*n def fdiff(self, argindex=1): arg = unpolarify(self.args[0]) if argindex == 1: return exp(arg)/arg else: raise ArgumentIndexError(self, argindex) def _eval_evalf(self, prec): if (self.args[0]/polar_lift(-1)).is_positive: return Function._eval_evalf(self, prec) + (I*pi)._eval_evalf(prec) return Function._eval_evalf(self, prec) def _eval_rewrite_as_uppergamma(self, z, **kwargs): from sympy.functions.special.gamma_functions import uppergamma # XXX this does not currently work usefully because uppergamma # immediately turns into expint return -uppergamma(0, polar_lift(-1)*z) - I*pi def _eval_rewrite_as_expint(self, z, **kwargs): return -expint(1, polar_lift(-1)*z) - I*pi def _eval_rewrite_as_li(self, z, **kwargs): if isinstance(z, log): return li(z.args[0]) # TODO: # Actually it only holds that: # Ei(z) = li(exp(z)) # for -pi < imag(z) <= pi return li(exp(z)) def _eval_rewrite_as_Si(self, z, **kwargs): if z.is_negative: return Shi(z) + Chi(z) - I*pi else: return Shi(z) + Chi(z) _eval_rewrite_as_Ci = _eval_rewrite_as_Si _eval_rewrite_as_Chi = _eval_rewrite_as_Si _eval_rewrite_as_Shi = _eval_rewrite_as_Si def _eval_rewrite_as_tractable(self, z, limitvar=None, **kwargs): return exp(z) * _eis(z) def _eval_rewrite_as_Integral(self, z, **kwargs): from sympy.integrals.integrals import Integral t = Dummy(uniquely_named_symbol('t', [z]).name) return Integral(S.Exp1**t/t, (t, S.NegativeInfinity, z)) def _eval_as_leading_term(self, x, logx, cdir): from sympy import re x0 = self.args[0].limit(x, 0) arg = self.args[0].as_leading_term(x, cdir=cdir) cdir = arg.dir(x, cdir) if x0.is_zero: c, e = arg.as_coeff_exponent(x) logx = log(x) if logx is None else logx return log(c) + e*logx + EulerGamma - ( I*pi if re(cdir).is_negative else S.Zero) return super()._eval_as_leading_term(x, logx=logx, cdir=cdir) def _eval_nseries(self, x, n, logx, cdir=0): x0 = self.args[0].limit(x, 0) if x0.is_zero: f = self._eval_rewrite_as_Si(*self.args) return f._eval_nseries(x, n, logx) return super()._eval_nseries(x, n, logx) def _eval_aseries(self, n, args0, x, logx): from sympy.series.order import Order point = args0[0] if point in (S.Infinity, S.NegativeInfinity): z = self.args[0] s = [factorial(k) / (z)**k for k in range(n)] + \ [Order(1/z**n, x)] return (exp(z)/z) * Add(*s) return super(Ei, self)._eval_aseries(n, args0, x, logx)
[docs] class expint(Function): r""" Generalized exponential integral. Explanation =========== This function is defined as .. math:: \operatorname{E}_\nu(z) = z^{\nu - 1} \Gamma(1 - \nu, z), where $\Gamma(1 - \nu, z)$ is the upper incomplete gamma function (``uppergamma``). Hence for $z$ with positive real part we have .. math:: \operatorname{E}_\nu(z) = \int_1^\infty \frac{e^{-zt}}{t^\nu} \mathrm{d}t, which explains the name. The representation as an incomplete gamma function provides an analytic continuation for $\operatorname{E}_\nu(z)$. If $\nu$ is a non-positive integer, the exponential integral is thus an unbranched function of $z$, otherwise there is a branch point at the origin. Refer to the incomplete gamma function documentation for details of the branching behavior. Examples ======== >>> from sympy import expint, S >>> from sympy.abc import nu, z Differentiation is supported. Differentiation with respect to $z$ further explains the name: for integral orders, the exponential integral is an iterated integral of the exponential function. >>> expint(nu, z).diff(z) -expint(nu - 1, z) Differentiation with respect to $\nu$ has no classical expression: >>> expint(nu, z).diff(nu) -z**(nu - 1)*meijerg(((), (1, 1)), ((0, 0, 1 - nu), ()), z) At non-postive integer orders, the exponential integral reduces to the exponential function: >>> expint(0, z) exp(-z)/z >>> expint(-1, z) exp(-z)/z + exp(-z)/z**2 At half-integers it reduces to error functions: >>> expint(S(1)/2, z) sqrt(pi)*erfc(sqrt(z))/sqrt(z) At positive integer orders it can be rewritten in terms of exponentials and ``expint(1, z)``. Use ``expand_func()`` to do this: >>> from sympy import expand_func >>> expand_func(expint(5, z)) z**4*expint(1, z)/24 + (-z**3 + z**2 - 2*z + 6)*exp(-z)/24 The generalised exponential integral is essentially equivalent to the incomplete gamma function: >>> from sympy import uppergamma >>> expint(nu, z).rewrite(uppergamma) z**(nu - 1)*uppergamma(1 - nu, z) As such it is branched at the origin: >>> from sympy import exp_polar, pi, I >>> expint(4, z*exp_polar(2*pi*I)) I*pi*z**3/3 + expint(4, z) >>> expint(nu, z*exp_polar(2*pi*I)) z**(nu - 1)*(exp(2*I*pi*nu) - 1)*gamma(1 - nu) + expint(nu, z) See Also ======== Ei: Another related function called exponential integral. E1: The classical case, returns expint(1, z). li: Logarithmic integral. Li: Offset logarithmic integral. Si: Sine integral. Ci: Cosine integral. Shi: Hyperbolic sine integral. Chi: Hyperbolic cosine integral. uppergamma References ========== .. [1] https://dlmf.nist.gov/8.19 .. [2] https://functions.wolfram.com/GammaBetaErf/ExpIntegralE/ .. [3] https://en.wikipedia.org/wiki/Exponential_integral """ @classmethod def eval(cls, nu, z): from sympy.functions.special.gamma_functions import (gamma, uppergamma) nu2 = unpolarify(nu) if nu != nu2: return expint(nu2, z) if nu.is_Integer and nu <= 0 or (not nu.is_Integer and (2*nu).is_Integer): return unpolarify(expand_mul(z**(nu - 1)*uppergamma(1 - nu, z))) # Extract branching information. This can be deduced from what is # explained in lowergamma.eval(). z, n = z.extract_branch_factor() if n is S.Zero: return if nu.is_integer: if not nu > 0: return return expint(nu, z) \ - 2*pi*I*n*S.NegativeOne**(nu - 1)/factorial(nu - 1)*unpolarify(z)**(nu - 1) else: return (exp(2*I*pi*nu*n) - 1)*z**(nu - 1)*gamma(1 - nu) + expint(nu, z) def fdiff(self, argindex): nu, z = self.args if argindex == 1: return -z**(nu - 1)*meijerg([], [1, 1], [0, 0, 1 - nu], [], z) elif argindex == 2: return -expint(nu - 1, z) else: raise ArgumentIndexError(self, argindex) def _eval_rewrite_as_uppergamma(self, nu, z, **kwargs): from sympy.functions.special.gamma_functions import uppergamma return z**(nu - 1)*uppergamma(1 - nu, z) def _eval_rewrite_as_Ei(self, nu, z, **kwargs): if nu == 1: return -Ei(z*exp_polar(-I*pi)) - I*pi elif nu.is_Integer and nu > 1: # DLMF, 8.19.7 x = -unpolarify(z) return x**(nu - 1)/factorial(nu - 1)*E1(z).rewrite(Ei) + \ exp(x)/factorial(nu - 1) * \ Add(*[factorial(nu - k - 2)*x**k for k in range(nu - 1)]) else: return self def _eval_expand_func(self, **hints): return self.rewrite(Ei).rewrite(expint, **hints) def _eval_rewrite_as_Si(self, nu, z, **kwargs): if nu != 1: return self return Shi(z) - Chi(z) _eval_rewrite_as_Ci = _eval_rewrite_as_Si _eval_rewrite_as_Chi = _eval_rewrite_as_Si _eval_rewrite_as_Shi = _eval_rewrite_as_Si def _eval_nseries(self, x, n, logx, cdir=0): if not self.args[0].has(x): nu = self.args[0] if nu == 1: f = self._eval_rewrite_as_Si(*self.args) return f._eval_nseries(x, n, logx) elif nu.is_Integer and nu > 1: f = self._eval_rewrite_as_Ei(*self.args) return f._eval_nseries(x, n, logx) return super()._eval_nseries(x, n, logx) def _eval_aseries(self, n, args0, x, logx): from sympy.series.order import Order point = args0[1] nu = self.args[0] if point is S.Infinity: z = self.args[1] s = [S.NegativeOne**k * RisingFactorial(nu, k) / z**k for k in range(n)] + [Order(1/z**n, x)] return (exp(-z)/z) * Add(*s) return super(expint, self)._eval_aseries(n, args0, x, logx) def _eval_rewrite_as_Integral(self, *args, **kwargs): from sympy.integrals.integrals import Integral n, x = self.args t = Dummy(uniquely_named_symbol('t', args).name) return Integral(t**-n * exp(-t*x), (t, 1, S.Infinity))
[docs] def E1(z): """ Classical case of the generalized exponential integral. Explanation =========== This is equivalent to ``expint(1, z)``. Examples ======== >>> from sympy import E1 >>> E1(0) expint(1, 0) >>> E1(5) expint(1, 5) See Also ======== Ei: Exponential integral. expint: Generalised exponential integral. li: Logarithmic integral. Li: Offset logarithmic integral. Si: Sine integral. Ci: Cosine integral. Shi: Hyperbolic sine integral. Chi: Hyperbolic cosine integral. """ return expint(1, z)
[docs] class li(Function): r""" The classical logarithmic integral. Explanation =========== For use in SymPy, this function is defined as .. math:: \operatorname{li}(x) = \int_0^x \frac{1}{\log(t)} \mathrm{d}t \,. Examples ======== >>> from sympy import I, oo, li >>> from sympy.abc import z Several special values are known: >>> li(0) 0 >>> li(1) -oo >>> li(oo) oo Differentiation with respect to $z$ is supported: >>> from sympy import diff >>> diff(li(z), z) 1/log(z) Defining the ``li`` function via an integral: >>> from sympy import integrate >>> integrate(li(z)) z*li(z) - Ei(2*log(z)) >>> integrate(li(z),z) z*li(z) - Ei(2*log(z)) The logarithmic integral can also be defined in terms of ``Ei``: >>> from sympy import Ei >>> li(z).rewrite(Ei) Ei(log(z)) >>> diff(li(z).rewrite(Ei), z) 1/log(z) We can numerically evaluate the logarithmic integral to arbitrary precision on the whole complex plane (except the singular points): >>> li(2).evalf(30) 1.04516378011749278484458888919 >>> li(2*I).evalf(30) 1.0652795784357498247001125598 + 3.08346052231061726610939702133*I We can even compute Soldner's constant by the help of mpmath: >>> from mpmath import findroot >>> findroot(li, 2) 1.45136923488338 Further transformations include rewriting ``li`` in terms of the trigonometric integrals ``Si``, ``Ci``, ``Shi`` and ``Chi``: >>> from sympy import Si, Ci, Shi, Chi >>> li(z).rewrite(Si) -log(I*log(z)) - log(1/log(z))/2 + log(log(z))/2 + Ci(I*log(z)) + Shi(log(z)) >>> li(z).rewrite(Ci) -log(I*log(z)) - log(1/log(z))/2 + log(log(z))/2 + Ci(I*log(z)) + Shi(log(z)) >>> li(z).rewrite(Shi) -log(1/log(z))/2 + log(log(z))/2 + Chi(log(z)) - Shi(log(z)) >>> li(z).rewrite(Chi) -log(1/log(z))/2 + log(log(z))/2 + Chi(log(z)) - Shi(log(z)) See Also ======== Li: Offset logarithmic integral. Ei: Exponential integral. expint: Generalised exponential integral. E1: Special case of the generalised exponential integral. Si: Sine integral. Ci: Cosine integral. Shi: Hyperbolic sine integral. Chi: Hyperbolic cosine integral. References ========== .. [1] https://en.wikipedia.org/wiki/Logarithmic_integral .. [2] https://mathworld.wolfram.com/LogarithmicIntegral.html .. [3] https://dlmf.nist.gov/6 .. [4] https://mathworld.wolfram.com/SoldnersConstant.html """ @classmethod def eval(cls, z): if z.is_zero: return S.Zero elif z is S.One: return S.NegativeInfinity elif z is S.Infinity: return S.Infinity if z.is_zero: return S.Zero def fdiff(self, argindex=1): arg = self.args[0] if argindex == 1: return S.One / log(arg) else: raise ArgumentIndexError(self, argindex) def _eval_conjugate(self): z = self.args[0] # Exclude values on the branch cut (-oo, 0) if not z.is_extended_negative: return self.func(z.conjugate()) def _eval_rewrite_as_Li(self, z, **kwargs): return Li(z) + li(2) def _eval_rewrite_as_Ei(self, z, **kwargs): return Ei(log(z)) def _eval_rewrite_as_uppergamma(self, z, **kwargs): from sympy.functions.special.gamma_functions import uppergamma return (-uppergamma(0, -log(z)) + S.Half*(log(log(z)) - log(S.One/log(z))) - log(-log(z))) def _eval_rewrite_as_Si(self, z, **kwargs): return (Ci(I*log(z)) - I*Si(I*log(z)) - S.Half*(log(S.One/log(z)) - log(log(z))) - log(I*log(z))) _eval_rewrite_as_Ci = _eval_rewrite_as_Si def _eval_rewrite_as_Shi(self, z, **kwargs): return (Chi(log(z)) - Shi(log(z)) - S.Half*(log(S.One/log(z)) - log(log(z)))) _eval_rewrite_as_Chi = _eval_rewrite_as_Shi def _eval_rewrite_as_hyper(self, z, **kwargs): return (log(z)*hyper((1, 1), (2, 2), log(z)) + S.Half*(log(log(z)) - log(S.One/log(z))) + EulerGamma) def _eval_rewrite_as_meijerg(self, z, **kwargs): return (-log(-log(z)) - S.Half*(log(S.One/log(z)) - log(log(z))) - meijerg(((), (1,)), ((0, 0), ()), -log(z))) def _eval_rewrite_as_tractable(self, z, limitvar=None, **kwargs): return z * _eis(log(z)) def _eval_nseries(self, x, n, logx, cdir=0): z = self.args[0] s = [(log(z))**k / (factorial(k) * k) for k in range(1, n)] return EulerGamma + log(log(z)) + Add(*s) def _eval_is_zero(self): z = self.args[0] if z.is_zero: return True
[docs] class Li(Function): r""" The offset logarithmic integral. Explanation =========== For use in SymPy, this function is defined as .. math:: \operatorname{Li}(x) = \operatorname{li}(x) - \operatorname{li}(2) Examples ======== >>> from sympy import Li >>> from sympy.abc import z The following special value is known: >>> Li(2) 0 Differentiation with respect to $z$ is supported: >>> from sympy import diff >>> diff(Li(z), z) 1/log(z) The shifted logarithmic integral can be written in terms of $li(z)$: >>> from sympy import li >>> Li(z).rewrite(li) li(z) - li(2) We can numerically evaluate the logarithmic integral to arbitrary precision on the whole complex plane (except the singular points): >>> Li(2).evalf(30) 0 >>> Li(4).evalf(30) 1.92242131492155809316615998938 See Also ======== li: Logarithmic integral. Ei: Exponential integral. expint: Generalised exponential integral. E1: Special case of the generalised exponential integral. Si: Sine integral. Ci: Cosine integral. Shi: Hyperbolic sine integral. Chi: Hyperbolic cosine integral. References ========== .. [1] https://en.wikipedia.org/wiki/Logarithmic_integral .. [2] https://mathworld.wolfram.com/LogarithmicIntegral.html .. [3] https://dlmf.nist.gov/6 """ @classmethod def eval(cls, z): if z is S.Infinity: return S.Infinity elif z == S(2): return S.Zero def fdiff(self, argindex=1): arg = self.args[0] if argindex == 1: return S.One / log(arg) else: raise ArgumentIndexError(self, argindex) def _eval_evalf(self, prec): return self.rewrite(li).evalf(prec) def _eval_rewrite_as_li(self, z, **kwargs): return li(z) - li(2) def _eval_rewrite_as_tractable(self, z, limitvar=None, **kwargs): return self.rewrite(li).rewrite("tractable", deep=True) def _eval_nseries(self, x, n, logx, cdir=0): f = self._eval_rewrite_as_li(*self.args) return f._eval_nseries(x, n, logx)
############################################################################### #################### TRIGONOMETRIC INTEGRALS ################################## ############################################################################### class TrigonometricIntegral(Function): """ Base class for trigonometric integrals. """ @classmethod def eval(cls, z): if z is S.Zero: return cls._atzero elif z is S.Infinity: return cls._atinf() elif z is S.NegativeInfinity: return cls._atneginf() if z.is_zero: return cls._atzero nz = z.extract_multiplicatively(polar_lift(I)) if nz is None and cls._trigfunc(0) == 0: nz = z.extract_multiplicatively(I) if nz is not None: return cls._Ifactor(nz, 1) nz = z.extract_multiplicatively(polar_lift(-I)) if nz is not None: return cls._Ifactor(nz, -1) nz = z.extract_multiplicatively(polar_lift(-1)) if nz is None and cls._trigfunc(0) == 0: nz = z.extract_multiplicatively(-1) if nz is not None: return cls._minusfactor(nz) nz, n = z.extract_branch_factor() if n == 0 and nz == z: return return 2*pi*I*n*cls._trigfunc(0) + cls(nz) def fdiff(self, argindex=1): arg = unpolarify(self.args[0]) if argindex == 1: return self._trigfunc(arg)/arg else: raise ArgumentIndexError(self, argindex) def _eval_rewrite_as_Ei(self, z, **kwargs): return self._eval_rewrite_as_expint(z).rewrite(Ei) def _eval_rewrite_as_uppergamma(self, z, **kwargs): from sympy.functions.special.gamma_functions import uppergamma return self._eval_rewrite_as_expint(z).rewrite(uppergamma) def _eval_nseries(self, x, n, logx, cdir=0): # NOTE this is fairly inefficient if self.args[0].subs(x, 0) != 0: return super()._eval_nseries(x, n, logx) baseseries = self._trigfunc(x)._eval_nseries(x, n, logx) if self._trigfunc(0) != 0: baseseries -= 1 baseseries = baseseries.replace(Pow, lambda t, n: t**n/n, simultaneous=False) if self._trigfunc(0) != 0: baseseries += EulerGamma + log(x) return baseseries.subs(x, self.args[0])._eval_nseries(x, n, logx)
[docs] class Si(TrigonometricIntegral): r""" Sine integral. Explanation =========== This function is defined by .. math:: \operatorname{Si}(z) = \int_0^z \frac{\sin{t}}{t} \mathrm{d}t. It is an entire function. Examples ======== >>> from sympy import Si >>> from sympy.abc import z The sine integral is an antiderivative of $sin(z)/z$: >>> Si(z).diff(z) sin(z)/z It is unbranched: >>> from sympy import exp_polar, I, pi >>> Si(z*exp_polar(2*I*pi)) Si(z) Sine integral behaves much like ordinary sine under multiplication by ``I``: >>> Si(I*z) I*Shi(z) >>> Si(-z) -Si(z) It can also be expressed in terms of exponential integrals, but beware that the latter is branched: >>> from sympy import expint >>> Si(z).rewrite(expint) -I*(-expint(1, z*exp_polar(-I*pi/2))/2 + expint(1, z*exp_polar(I*pi/2))/2) + pi/2 It can be rewritten in the form of sinc function (by definition): >>> from sympy import sinc >>> Si(z).rewrite(sinc) Integral(sinc(_t), (_t, 0, z)) See Also ======== Ci: Cosine integral. Shi: Hyperbolic sine integral. Chi: Hyperbolic cosine integral. Ei: Exponential integral. expint: Generalised exponential integral. sinc: unnormalized sinc function E1: Special case of the generalised exponential integral. li: Logarithmic integral. Li: Offset logarithmic integral. References ========== .. [1] https://en.wikipedia.org/wiki/Trigonometric_integral """ _trigfunc = sin _atzero = S.Zero @classmethod def _atinf(cls): return pi*S.Half @classmethod def _atneginf(cls): return -pi*S.Half @classmethod def _minusfactor(cls, z): return -Si(z) @classmethod def _Ifactor(cls, z, sign): return I*Shi(z)*sign def _eval_rewrite_as_expint(self, z, **kwargs): # XXX should we polarify z? return pi/2 + (E1(polar_lift(I)*z) - E1(polar_lift(-I)*z))/2/I def _eval_rewrite_as_Integral(self, z, **kwargs): from sympy.integrals.integrals import Integral t = Dummy(uniquely_named_symbol('t', [z]).name) return Integral(sinc(t), (t, 0, z)) _eval_rewrite_as_sinc = _eval_rewrite_as_Integral def _eval_as_leading_term(self, x, logx, cdir): arg = self.args[0].as_leading_term(x, logx=logx, cdir=cdir) arg0 = arg.subs(x, 0) if arg0 is S.NaN: arg0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') if arg0.is_zero: return arg elif not arg0.is_infinite: return self.func(arg0) else: return self def _eval_aseries(self, n, args0, x, logx): from sympy.series.order import Order point = args0[0] # Expansion at oo if point is S.Infinity: z = self.args[0] p = [S.NegativeOne**k * factorial(2*k) / z**(2*k + 1) for k in range(n//2 + 1)] + [Order(1/z**n, x)] q = [S.NegativeOne**k * factorial(2*k + 1) / z**(2*(k + 1)) for k in range(n//2)] + [Order(1/z**n, x)] return pi/2 - cos(z)*Add(*p) - sin(z)*Add(*q) # All other points are not handled return super(Si, self)._eval_aseries(n, args0, x, logx) def _eval_is_zero(self): z = self.args[0] if z.is_zero: return True
[docs] class Ci(TrigonometricIntegral): r""" Cosine integral. Explanation =========== This function is defined for positive $x$ by .. math:: \operatorname{Ci}(x) = \gamma + \log{x} + \int_0^x \frac{\cos{t} - 1}{t} \mathrm{d}t = -\int_x^\infty \frac{\cos{t}}{t} \mathrm{d}t, where $\gamma$ is the Euler-Mascheroni constant. We have .. math:: \operatorname{Ci}(z) = -\frac{\operatorname{E}_1\left(e^{i\pi/2} z\right) + \operatorname{E}_1\left(e^{-i \pi/2} z\right)}{2} which holds for all polar $z$ and thus provides an analytic continuation to the Riemann surface of the logarithm. The formula also holds as stated for $z \in \mathbb{C}$ with $\Re(z) > 0$. By lifting to the principal branch, we obtain an analytic function on the cut complex plane. Examples ======== >>> from sympy import Ci >>> from sympy.abc import z The cosine integral is a primitive of $\cos(z)/z$: >>> Ci(z).diff(z) cos(z)/z It has a logarithmic branch point at the origin: >>> from sympy import exp_polar, I, pi >>> Ci(z*exp_polar(2*I*pi)) Ci(z) + 2*I*pi The cosine integral behaves somewhat like ordinary $\cos$ under multiplication by $i$: >>> from sympy import polar_lift >>> Ci(polar_lift(I)*z) Chi(z) + I*pi/2 >>> Ci(polar_lift(-1)*z) Ci(z) + I*pi It can also be expressed in terms of exponential integrals: >>> from sympy import expint >>> Ci(z).rewrite(expint) -expint(1, z*exp_polar(-I*pi/2))/2 - expint(1, z*exp_polar(I*pi/2))/2 See Also ======== Si: Sine integral. Shi: Hyperbolic sine integral. Chi: Hyperbolic cosine integral. Ei: Exponential integral. expint: Generalised exponential integral. E1: Special case of the generalised exponential integral. li: Logarithmic integral. Li: Offset logarithmic integral. References ========== .. [1] https://en.wikipedia.org/wiki/Trigonometric_integral """ _trigfunc = cos _atzero = S.ComplexInfinity @classmethod def _atinf(cls): return S.Zero @classmethod def _atneginf(cls): return I*pi @classmethod def _minusfactor(cls, z): return Ci(z) + I*pi @classmethod def _Ifactor(cls, z, sign): return Chi(z) + I*pi/2*sign def _eval_rewrite_as_expint(self, z, **kwargs): return -(E1(polar_lift(I)*z) + E1(polar_lift(-I)*z))/2 def _eval_rewrite_as_Integral(self, z, **kwargs): from sympy.integrals.integrals import Integral t = Dummy(uniquely_named_symbol('t', [z]).name) return S.EulerGamma + log(z) - Integral((1-cos(t))/t, (t, 0, z)) def _eval_as_leading_term(self, x, logx, cdir): arg = self.args[0].as_leading_term(x, logx=logx, cdir=cdir) arg0 = arg.subs(x, 0) if arg0 is S.NaN: arg0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') if arg0.is_zero: c, e = arg.as_coeff_exponent(x) logx = log(x) if logx is None else logx return log(c) + e*logx + EulerGamma elif arg0.is_finite: return self.func(arg0) else: return self def _eval_aseries(self, n, args0, x, logx): from sympy.series.order import Order point = args0[0] if point in (S.Infinity, S.NegativeInfinity): z = self.args[0] p = [S.NegativeOne**k * factorial(2*k) / z**(2*k + 1) for k in range(n//2 + 1)] + [Order(1/z**n, x)] q = [S.NegativeOne**k * factorial(2*k + 1) / z**(2*(k + 1)) for k in range(n//2)] + [Order(1/z**n, x)] result = sin(z)*(Add(*p)) - cos(z)*(Add(*q)) if point is S.NegativeInfinity: result += I*pi return result return super(Ci, self)._eval_aseries(n, args0, x, logx)
[docs] class Shi(TrigonometricIntegral): r""" Sinh integral. Explanation =========== This function is defined by .. math:: \operatorname{Shi}(z) = \int_0^z \frac{\sinh{t}}{t} \mathrm{d}t. It is an entire function. Examples ======== >>> from sympy import Shi >>> from sympy.abc import z The Sinh integral is a primitive of $\sinh(z)/z$: >>> Shi(z).diff(z) sinh(z)/z It is unbranched: >>> from sympy import exp_polar, I, pi >>> Shi(z*exp_polar(2*I*pi)) Shi(z) The $\sinh$ integral behaves much like ordinary $\sinh$ under multiplication by $i$: >>> Shi(I*z) I*Si(z) >>> Shi(-z) -Shi(z) It can also be expressed in terms of exponential integrals, but beware that the latter is branched: >>> from sympy import expint >>> Shi(z).rewrite(expint) expint(1, z)/2 - expint(1, z*exp_polar(I*pi))/2 - I*pi/2 See Also ======== Si: Sine integral. Ci: Cosine integral. Chi: Hyperbolic cosine integral. Ei: Exponential integral. expint: Generalised exponential integral. E1: Special case of the generalised exponential integral. li: Logarithmic integral. Li: Offset logarithmic integral. References ========== .. [1] https://en.wikipedia.org/wiki/Trigonometric_integral """ _trigfunc = sinh _atzero = S.Zero @classmethod def _atinf(cls): return S.Infinity @classmethod def _atneginf(cls): return S.NegativeInfinity @classmethod def _minusfactor(cls, z): return -Shi(z) @classmethod def _Ifactor(cls, z, sign): return I*Si(z)*sign def _eval_rewrite_as_expint(self, z, **kwargs): # XXX should we polarify z? return (E1(z) - E1(exp_polar(I*pi)*z))/2 - I*pi/2 def _eval_is_zero(self): z = self.args[0] if z.is_zero: return True def _eval_as_leading_term(self, x, logx, cdir): arg = self.args[0].as_leading_term(x) arg0 = arg.subs(x, 0) if arg0 is S.NaN: arg0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') if arg0.is_zero: return arg elif not arg0.is_infinite: return self.func(arg0) else: return self
[docs] class Chi(TrigonometricIntegral): r""" Cosh integral. Explanation =========== This function is defined for positive $x$ by .. math:: \operatorname{Chi}(x) = \gamma + \log{x} + \int_0^x \frac{\cosh{t} - 1}{t} \mathrm{d}t, where $\gamma$ is the Euler-Mascheroni constant. We have .. math:: \operatorname{Chi}(z) = \operatorname{Ci}\left(e^{i \pi/2}z\right) - i\frac{\pi}{2}, which holds for all polar $z$ and thus provides an analytic continuation to the Riemann surface of the logarithm. By lifting to the principal branch we obtain an analytic function on the cut complex plane. Examples ======== >>> from sympy import Chi >>> from sympy.abc import z The $\cosh$ integral is a primitive of $\cosh(z)/z$: >>> Chi(z).diff(z) cosh(z)/z It has a logarithmic branch point at the origin: >>> from sympy import exp_polar, I, pi >>> Chi(z*exp_polar(2*I*pi)) Chi(z) + 2*I*pi The $\cosh$ integral behaves somewhat like ordinary $\cosh$ under multiplication by $i$: >>> from sympy import polar_lift >>> Chi(polar_lift(I)*z) Ci(z) + I*pi/2 >>> Chi(polar_lift(-1)*z) Chi(z) + I*pi It can also be expressed in terms of exponential integrals: >>> from sympy import expint >>> Chi(z).rewrite(expint) -expint(1, z)/2 - expint(1, z*exp_polar(I*pi))/2 - I*pi/2 See Also ======== Si: Sine integral. Ci: Cosine integral. Shi: Hyperbolic sine integral. Ei: Exponential integral. expint: Generalised exponential integral. E1: Special case of the generalised exponential integral. li: Logarithmic integral. Li: Offset logarithmic integral. References ========== .. [1] https://en.wikipedia.org/wiki/Trigonometric_integral """ _trigfunc = cosh _atzero = S.ComplexInfinity @classmethod def _atinf(cls): return S.Infinity @classmethod def _atneginf(cls): return S.Infinity @classmethod def _minusfactor(cls, z): return Chi(z) + I*pi @classmethod def _Ifactor(cls, z, sign): return Ci(z) + I*pi/2*sign def _eval_rewrite_as_expint(self, z, **kwargs): return -I*pi/2 - (E1(z) + E1(exp_polar(I*pi)*z))/2 def _eval_as_leading_term(self, x, logx, cdir): arg = self.args[0].as_leading_term(x, logx=logx, cdir=cdir) arg0 = arg.subs(x, 0) if arg0 is S.NaN: arg0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') if arg0.is_zero: c, e = arg.as_coeff_exponent(x) logx = log(x) if logx is None else logx return log(c) + e*logx + EulerGamma elif arg0.is_finite: return self.func(arg0) else: return self
############################################################################### #################### FRESNEL INTEGRALS ######################################## ###############################################################################
[docs] class FresnelIntegral(Function): """ Base class for the Fresnel integrals.""" unbranched = True @classmethod def eval(cls, z): # Values at positive infinities signs # if any were extracted automatically if z is S.Infinity: return S.Half # Value at zero if z.is_zero: return S.Zero # Try to pull out factors of -1 and I prefact = S.One newarg = z changed = False nz = newarg.extract_multiplicatively(-1) if nz is not None: prefact = -prefact newarg = nz changed = True nz = newarg.extract_multiplicatively(I) if nz is not None: prefact = cls._sign*I*prefact newarg = nz changed = True if changed: return prefact*cls(newarg) def fdiff(self, argindex=1): if argindex == 1: return self._trigfunc(S.Half*pi*self.args[0]**2) else: raise ArgumentIndexError(self, argindex) def _eval_is_extended_real(self): return self.args[0].is_extended_real _eval_is_finite = _eval_is_extended_real def _eval_is_zero(self): return self.args[0].is_zero def _eval_conjugate(self): return self.func(self.args[0].conjugate()) as_real_imag = real_to_real_as_real_imag
[docs] class fresnels(FresnelIntegral): r""" Fresnel integral S. Explanation =========== This function is defined by .. math:: \operatorname{S}(z) = \int_0^z \sin{\frac{\pi}{2} t^2} \mathrm{d}t. It is an entire function. Examples ======== >>> from sympy import I, oo, fresnels >>> from sympy.abc import z Several special values are known: >>> fresnels(0) 0 >>> fresnels(oo) 1/2 >>> fresnels(-oo) -1/2 >>> fresnels(I*oo) -I/2 >>> fresnels(-I*oo) I/2 In general one can pull out factors of -1 and $i$ from the argument: >>> fresnels(-z) -fresnels(z) >>> fresnels(I*z) -I*fresnels(z) The Fresnel S integral obeys the mirror symmetry $\overline{S(z)} = S(\bar{z})$: >>> from sympy import conjugate >>> conjugate(fresnels(z)) fresnels(conjugate(z)) Differentiation with respect to $z$ is supported: >>> from sympy import diff >>> diff(fresnels(z), z) sin(pi*z**2/2) Defining the Fresnel functions via an integral: >>> from sympy import integrate, pi, sin, expand_func >>> integrate(sin(pi*z**2/2), z) 3*fresnels(z)*gamma(3/4)/(4*gamma(7/4)) >>> expand_func(integrate(sin(pi*z**2/2), z)) fresnels(z) We can numerically evaluate the Fresnel integral to arbitrary precision on the whole complex plane: >>> fresnels(2).evalf(30) 0.343415678363698242195300815958 >>> fresnels(-2*I).evalf(30) 0.343415678363698242195300815958*I See Also ======== fresnelc: Fresnel cosine integral. References ========== .. [1] https://en.wikipedia.org/wiki/Fresnel_integral .. [2] https://dlmf.nist.gov/7 .. [3] https://mathworld.wolfram.com/FresnelIntegrals.html .. [4] https://functions.wolfram.com/GammaBetaErf/FresnelS .. [5] The converging factors for the fresnel integrals by John W. Wrench Jr. and Vicki Alley """ _trigfunc = sin _sign = -S.One @staticmethod @cacheit def taylor_term(n, x, *previous_terms): if n < 0: return S.Zero else: x = sympify(x) if len(previous_terms) > 1: p = previous_terms[-1] return (-pi**2*x**4*(4*n - 1)/(8*n*(2*n + 1)*(4*n + 3))) * p else: return x**3 * (-x**4)**n * (S(2)**(-2*n - 1)*pi**(2*n + 1)) / ((4*n + 3)*factorial(2*n + 1)) def _eval_rewrite_as_erf(self, z, **kwargs): return (S.One + I)/4 * (erf((S.One + I)/2*sqrt(pi)*z) - I*erf((S.One - I)/2*sqrt(pi)*z)) def _eval_rewrite_as_hyper(self, z, **kwargs): return pi*z**3/6 * hyper([Rational(3, 4)], [Rational(3, 2), Rational(7, 4)], -pi**2*z**4/16) def _eval_rewrite_as_meijerg(self, z, **kwargs): return (pi*z**Rational(9, 4) / (sqrt(2)*(z**2)**Rational(3, 4)*(-z)**Rational(3, 4)) * meijerg([], [1], [Rational(3, 4)], [Rational(1, 4), 0], -pi**2*z**4/16)) def _eval_rewrite_as_Integral(self, z, **kwargs): from sympy.integrals.integrals import Integral t = Dummy(uniquely_named_symbol('t', [z]).name) return Integral(sin(pi*t**2/2), (t, 0, z)) def _eval_as_leading_term(self, x, logx, cdir): from sympy.series.order import Order arg = self.args[0].as_leading_term(x, logx=logx, cdir=cdir) arg0 = arg.subs(x, 0) if arg0 is S.ComplexInfinity: arg0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') if arg0.is_zero: return pi*arg**3/6 elif arg0 in [S.Infinity, S.NegativeInfinity]: s = 1 if arg0 is S.Infinity else -1 return s*S.Half + Order(x, x) else: return self.func(arg0) def _eval_aseries(self, n, args0, x, logx): from sympy.series.order import Order point = args0[0] # Expansion at oo and -oo if point in [S.Infinity, -S.Infinity]: z = self.args[0] # expansion of S(x) = S1(x*sqrt(pi/2)), see reference[5] page 1-8 # as only real infinities are dealt with, sin and cos are O(1) p = [S.NegativeOne**k * factorial(4*k + 1) / (2**(2*k + 2) * z**(4*k + 3) * 2**(2*k)*factorial(2*k)) for k in range(0, n) if 4*k + 3 < n] q = [1/(2*z)] + [S.NegativeOne**k * factorial(4*k - 1) / (2**(2*k + 1) * z**(4*k + 1) * 2**(2*k - 1)*factorial(2*k - 1)) for k in range(1, n) if 4*k + 1 < n] p = [-sqrt(2/pi)*t for t in p] q = [-sqrt(2/pi)*t for t in q] s = 1 if point is S.Infinity else -1 # The expansion at oo is 1/2 + some odd powers of z # To get the expansion at -oo, replace z by -z and flip the sign # The result -1/2 + the same odd powers of z as before. return s*S.Half + (sin(z**2)*Add(*p) + cos(z**2)*Add(*q) ).subs(x, sqrt(2/pi)*x) + Order(1/z**n, x) # All other points are not handled return super()._eval_aseries(n, args0, x, logx)
[docs] class fresnelc(FresnelIntegral): r""" Fresnel integral C. Explanation =========== This function is defined by .. math:: \operatorname{C}(z) = \int_0^z \cos{\frac{\pi}{2} t^2} \mathrm{d}t. It is an entire function. Examples ======== >>> from sympy import I, oo, fresnelc >>> from sympy.abc import z Several special values are known: >>> fresnelc(0) 0 >>> fresnelc(oo) 1/2 >>> fresnelc(-oo) -1/2 >>> fresnelc(I*oo) I/2 >>> fresnelc(-I*oo) -I/2 In general one can pull out factors of -1 and $i$ from the argument: >>> fresnelc(-z) -fresnelc(z) >>> fresnelc(I*z) I*fresnelc(z) The Fresnel C integral obeys the mirror symmetry $\overline{C(z)} = C(\bar{z})$: >>> from sympy import conjugate >>> conjugate(fresnelc(z)) fresnelc(conjugate(z)) Differentiation with respect to $z$ is supported: >>> from sympy import diff >>> diff(fresnelc(z), z) cos(pi*z**2/2) Defining the Fresnel functions via an integral: >>> from sympy import integrate, pi, cos, expand_func >>> integrate(cos(pi*z**2/2), z) fresnelc(z)*gamma(1/4)/(4*gamma(5/4)) >>> expand_func(integrate(cos(pi*z**2/2), z)) fresnelc(z) We can numerically evaluate the Fresnel integral to arbitrary precision on the whole complex plane: >>> fresnelc(2).evalf(30) 0.488253406075340754500223503357 >>> fresnelc(-2*I).evalf(30) -0.488253406075340754500223503357*I See Also ======== fresnels: Fresnel sine integral. References ========== .. [1] https://en.wikipedia.org/wiki/Fresnel_integral .. [2] https://dlmf.nist.gov/7 .. [3] https://mathworld.wolfram.com/FresnelIntegrals.html .. [4] https://functions.wolfram.com/GammaBetaErf/FresnelC .. [5] The converging factors for the fresnel integrals by John W. Wrench Jr. and Vicki Alley """ _trigfunc = cos _sign = S.One @staticmethod @cacheit def taylor_term(n, x, *previous_terms): if n < 0: return S.Zero else: x = sympify(x) if len(previous_terms) > 1: p = previous_terms[-1] return (-pi**2*x**4*(4*n - 3)/(8*n*(2*n - 1)*(4*n + 1))) * p else: return x * (-x**4)**n * (S(2)**(-2*n)*pi**(2*n)) / ((4*n + 1)*factorial(2*n)) def _eval_rewrite_as_erf(self, z, **kwargs): return (S.One - I)/4 * (erf((S.One + I)/2*sqrt(pi)*z) + I*erf((S.One - I)/2*sqrt(pi)*z)) def _eval_rewrite_as_hyper(self, z, **kwargs): return z * hyper([Rational(1, 4)], [S.Half, Rational(5, 4)], -pi**2*z**4/16) def _eval_rewrite_as_meijerg(self, z, **kwargs): return (pi*z**Rational(3, 4) / (sqrt(2)*root(z**2, 4)*root(-z, 4)) * meijerg([], [1], [Rational(1, 4)], [Rational(3, 4), 0], -pi**2*z**4/16)) def _eval_rewrite_as_Integral(self, z, **kwargs): from sympy.integrals.integrals import Integral t = Dummy(uniquely_named_symbol('t', [z]).name) return Integral(cos(pi*t**2/2), (t, 0, z)) def _eval_as_leading_term(self, x, logx, cdir): from sympy.series.order import Order arg = self.args[0].as_leading_term(x, logx=logx, cdir=cdir) arg0 = arg.subs(x, 0) if arg0 is S.ComplexInfinity: arg0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') if arg0.is_zero: return arg elif arg0 in [S.Infinity, S.NegativeInfinity]: s = 1 if arg0 is S.Infinity else -1 return s*S.Half + Order(x, x) else: return self.func(arg0) def _eval_aseries(self, n, args0, x, logx): from sympy.series.order import Order point = args0[0] # Expansion at oo if point in [S.Infinity, -S.Infinity]: z = self.args[0] # expansion of C(x) = C1(x*sqrt(pi/2)), see reference[5] page 1-8 # as only real infinities are dealt with, sin and cos are O(1) p = [S.NegativeOne**k * factorial(4*k + 1) / (2**(2*k + 2) * z**(4*k + 3) * 2**(2*k)*factorial(2*k)) for k in range(n) if 4*k + 3 < n] q = [1/(2*z)] + [S.NegativeOne**k * factorial(4*k - 1) / (2**(2*k + 1) * z**(4*k + 1) * 2**(2*k - 1)*factorial(2*k - 1)) for k in range(1, n) if 4*k + 1 < n] p = [-sqrt(2/pi)*t for t in p] q = [ sqrt(2/pi)*t for t in q] s = 1 if point is S.Infinity else -1 # The expansion at oo is 1/2 + some odd powers of z # To get the expansion at -oo, replace z by -z and flip the sign # The result -1/2 + the same odd powers of z as before. return s*S.Half + (cos(z**2)*Add(*p) + sin(z**2)*Add(*q) ).subs(x, sqrt(2/pi)*x) + Order(1/z**n, x) # All other points are not handled return super()._eval_aseries(n, args0, x, logx)
############################################################################### #################### HELPER FUNCTIONS ######################################### ############################################################################### class _erfs(Function): """ Helper function to make the $\\mathrm{erf}(z)$ function tractable for the Gruntz algorithm. """ @classmethod def eval(cls, arg): if arg.is_zero: return S.One def _eval_aseries(self, n, args0, x, logx): from sympy.series.order import Order point = args0[0] # Expansion at oo if point is S.Infinity: z = self.args[0] l = [1/sqrt(pi) * factorial(2*k)*(-S( 4))**(-k)/factorial(k) * (1/z)**(2*k + 1) for k in range(n)] o = Order(1/z**(2*n + 1), x) # It is very inefficient to first add the order and then do the nseries return (Add(*l))._eval_nseries(x, n, logx) + o # Expansion at I*oo t = point.extract_multiplicatively(I) if t is S.Infinity: z = self.args[0] # TODO: is the series really correct? l = [1/sqrt(pi) * factorial(2*k)*(-S( 4))**(-k)/factorial(k) * (1/z)**(2*k + 1) for k in range(n)] o = Order(1/z**(2*n + 1), x) # It is very inefficient to first add the order and then do the nseries return (Add(*l))._eval_nseries(x, n, logx) + o # All other points are not handled return super()._eval_aseries(n, args0, x, logx) def fdiff(self, argindex=1): if argindex == 1: z = self.args[0] return -2/sqrt(pi) + 2*z*_erfs(z) else: raise ArgumentIndexError(self, argindex) def _eval_rewrite_as_intractable(self, z, **kwargs): return (S.One - erf(z))*exp(z**2) class _eis(Function): """ Helper function to make the $\\mathrm{Ei}(z)$ and $\\mathrm{li}(z)$ functions tractable for the Gruntz algorithm. """ def _eval_aseries(self, n, args0, x, logx): from sympy.series.order import Order if args0[0] not in (S.Infinity, S.NegativeInfinity): return super()._eval_aseries(n, args0, x, logx) z = self.args[0] l = [factorial(k) * (1/z)**(k + 1) for k in range(n)] o = Order(1/z**(n + 1), x) # It is very inefficient to first add the order and then do the nseries return (Add(*l))._eval_nseries(x, n, logx) + o def fdiff(self, argindex=1): if argindex == 1: z = self.args[0] return S.One / z - _eis(z) else: raise ArgumentIndexError(self, argindex) def _eval_rewrite_as_intractable(self, z, **kwargs): return exp(-z)*Ei(z) def _eval_as_leading_term(self, x, logx, cdir): x0 = self.args[0].limit(x, 0) if x0.is_zero: f = self._eval_rewrite_as_intractable(*self.args) return f._eval_as_leading_term(x, logx=logx, cdir=cdir) return super()._eval_as_leading_term(x, logx=logx, cdir=cdir) def _eval_nseries(self, x, n, logx, cdir=0): x0 = self.args[0].limit(x, 0) if x0.is_zero: f = self._eval_rewrite_as_intractable(*self.args) return f._eval_nseries(x, n, logx) return super()._eval_nseries(x, n, logx)