diff --git "a/venv/lib/python3.10/site-packages/sympy/functions/elementary/trigonometric.py" "b/venv/lib/python3.10/site-packages/sympy/functions/elementary/trigonometric.py" new file mode 100644--- /dev/null +++ "b/venv/lib/python3.10/site-packages/sympy/functions/elementary/trigonometric.py" @@ -0,0 +1,3574 @@ +from typing import Tuple as tTuple, Union as tUnion +from sympy.core.add import Add +from sympy.core.cache import cacheit +from sympy.core.expr import Expr +from sympy.core.function import Function, ArgumentIndexError, PoleError, expand_mul +from sympy.core.logic import fuzzy_not, fuzzy_or, FuzzyBool, fuzzy_and +from sympy.core.mod import Mod +from sympy.core.numbers import Rational, pi, Integer, Float, equal_valued +from sympy.core.relational import Ne, Eq +from sympy.core.singleton import S +from sympy.core.symbol import Symbol, Dummy +from sympy.core.sympify import sympify +from sympy.functions.combinatorial.factorials import factorial, RisingFactorial +from sympy.functions.combinatorial.numbers import bernoulli, euler +from sympy.functions.elementary.complexes import arg as arg_f, im, re +from sympy.functions.elementary.exponential import log, exp +from sympy.functions.elementary.integers import floor +from sympy.functions.elementary.miscellaneous import sqrt, Min, Max +from sympy.functions.elementary.piecewise import Piecewise +from sympy.functions.elementary._trigonometric_special import ( + cos_table, ipartfrac, fermat_coords) +from sympy.logic.boolalg import And +from sympy.ntheory import factorint +from sympy.polys.specialpolys import symmetric_poly +from sympy.utilities.iterables import numbered_symbols + + +############################################################################### +########################## UTILITIES ########################################## +############################################################################### + + +def _imaginary_unit_as_coefficient(arg): + """ Helper to extract symbolic coefficient for imaginary unit """ + if isinstance(arg, Float): + return None + else: + return arg.as_coefficient(S.ImaginaryUnit) + +############################################################################### +########################## TRIGONOMETRIC FUNCTIONS ############################ +############################################################################### + + +class TrigonometricFunction(Function): + """Base class for trigonometric functions. """ + + unbranched = True + _singularities = (S.ComplexInfinity,) + + def _eval_is_rational(self): + s = self.func(*self.args) + if s.func == self.func: + if s.args[0].is_rational and fuzzy_not(s.args[0].is_zero): + return False + else: + return s.is_rational + + def _eval_is_algebraic(self): + s = self.func(*self.args) + if s.func == self.func: + if fuzzy_not(self.args[0].is_zero) and self.args[0].is_algebraic: + return False + pi_coeff = _pi_coeff(self.args[0]) + if pi_coeff is not None and pi_coeff.is_rational: + return True + else: + return s.is_algebraic + + def _eval_expand_complex(self, deep=True, **hints): + re_part, im_part = self.as_real_imag(deep=deep, **hints) + return re_part + im_part*S.ImaginaryUnit + + def _as_real_imag(self, deep=True, **hints): + if self.args[0].is_extended_real: + if deep: + hints['complex'] = False + return (self.args[0].expand(deep, **hints), S.Zero) + else: + return (self.args[0], S.Zero) + if deep: + re, im = self.args[0].expand(deep, **hints).as_real_imag() + else: + re, im = self.args[0].as_real_imag() + return (re, im) + + def _period(self, general_period, symbol=None): + f = expand_mul(self.args[0]) + if symbol is None: + symbol = tuple(f.free_symbols)[0] + + if not f.has(symbol): + return S.Zero + + if f == symbol: + return general_period + + if symbol in f.free_symbols: + if f.is_Mul: + g, h = f.as_independent(symbol) + if h == symbol: + return general_period/abs(g) + + if f.is_Add: + a, h = f.as_independent(symbol) + g, h = h.as_independent(symbol, as_Add=False) + if h == symbol: + return general_period/abs(g) + + raise NotImplementedError("Use the periodicity function instead.") + + +@cacheit +def _table2(): + # If nested sqrt's are worse than un-evaluation + # you can require q to be in (1, 2, 3, 4, 6, 12) + # q <= 12, q=15, q=20, q=24, q=30, q=40, q=60, q=120 return + # expressions with 2 or fewer sqrt nestings. + return { + 12: (3, 4), + 20: (4, 5), + 30: (5, 6), + 15: (6, 10), + 24: (6, 8), + 40: (8, 10), + 60: (20, 30), + 120: (40, 60) + } + + +def _peeloff_pi(arg): + r""" + Split ARG into two parts, a "rest" and a multiple of $\pi$. + This assumes ARG to be an Add. + The multiple of $\pi$ returned in the second position is always a Rational. + + Examples + ======== + + >>> from sympy.functions.elementary.trigonometric import _peeloff_pi + >>> from sympy import pi + >>> from sympy.abc import x, y + >>> _peeloff_pi(x + pi/2) + (x, 1/2) + >>> _peeloff_pi(x + 2*pi/3 + pi*y) + (x + pi*y + pi/6, 1/2) + + """ + pi_coeff = S.Zero + rest_terms = [] + for a in Add.make_args(arg): + K = a.coeff(pi) + if K and K.is_rational: + pi_coeff += K + else: + rest_terms.append(a) + + if pi_coeff is S.Zero: + return arg, S.Zero + + m1 = (pi_coeff % S.Half) + m2 = pi_coeff - m1 + if m2.is_integer or ((2*m2).is_integer and m2.is_even is False): + return Add(*(rest_terms + [m1*pi])), m2 + return arg, S.Zero + + +def _pi_coeff(arg: Expr, cycles: int = 1) -> tUnion[Expr, None]: + r""" + When arg is a Number times $\pi$ (e.g. $3\pi/2$) then return the Number + normalized to be in the range $[0, 2]$, else `None`. + + When an even multiple of $\pi$ is encountered, if it is multiplying + something with known parity then the multiple is returned as 0 otherwise + as 2. + + Examples + ======== + + >>> from sympy.functions.elementary.trigonometric import _pi_coeff + >>> from sympy import pi, Dummy + >>> from sympy.abc import x + >>> _pi_coeff(3*x*pi) + 3*x + >>> _pi_coeff(11*pi/7) + 11/7 + >>> _pi_coeff(-11*pi/7) + 3/7 + >>> _pi_coeff(4*pi) + 0 + >>> _pi_coeff(5*pi) + 1 + >>> _pi_coeff(5.0*pi) + 1 + >>> _pi_coeff(5.5*pi) + 3/2 + >>> _pi_coeff(2 + pi) + + >>> _pi_coeff(2*Dummy(integer=True)*pi) + 2 + >>> _pi_coeff(2*Dummy(even=True)*pi) + 0 + + """ + if arg is pi: + return S.One + elif not arg: + return S.Zero + elif arg.is_Mul: + cx = arg.coeff(pi) + if cx: + c, x = cx.as_coeff_Mul() # pi is not included as coeff + if c.is_Float: + # recast exact binary fractions to Rationals + f = abs(c) % 1 + if f != 0: + p = -int(round(log(f, 2).evalf())) + m = 2**p + cm = c*m + i = int(cm) + if equal_valued(i, cm): + c = Rational(i, m) + cx = c*x + else: + c = Rational(int(c)) + cx = c*x + if x.is_integer: + c2 = c % 2 + if c2 == 1: + return x + elif not c2: + if x.is_even is not None: # known parity + return S.Zero + return Integer(2) + else: + return c2*x + return cx + elif arg.is_zero: + return S.Zero + return None + + +class sin(TrigonometricFunction): + r""" + The sine function. + + Returns the sine of x (measured in radians). + + Explanation + =========== + + This function will evaluate automatically in the + case $x/\pi$ is some rational number [4]_. For example, + if $x$ is a multiple of $\pi$, $\pi/2$, $\pi/3$, $\pi/4$, and $\pi/6$. + + Examples + ======== + + >>> from sympy import sin, pi + >>> from sympy.abc import x + >>> sin(x**2).diff(x) + 2*x*cos(x**2) + >>> sin(1).diff(x) + 0 + >>> sin(pi) + 0 + >>> sin(pi/2) + 1 + >>> sin(pi/6) + 1/2 + >>> sin(pi/12) + -sqrt(2)/4 + sqrt(6)/4 + + + See Also + ======== + + csc, cos, sec, tan, cot + asin, acsc, acos, asec, atan, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Trigonometric_functions + .. [2] https://dlmf.nist.gov/4.14 + .. [3] https://functions.wolfram.com/ElementaryFunctions/Sin + .. [4] https://mathworld.wolfram.com/TrigonometryAngles.html + + """ + + def period(self, symbol=None): + return self._period(2*pi, symbol) + + def fdiff(self, argindex=1): + if argindex == 1: + return cos(self.args[0]) + else: + raise ArgumentIndexError(self, argindex) + + @classmethod + def eval(cls, arg): + from sympy.calculus.accumulationbounds import AccumBounds + from sympy.sets.setexpr import SetExpr + if arg.is_Number: + if arg is S.NaN: + return S.NaN + elif arg.is_zero: + return S.Zero + elif arg in (S.Infinity, S.NegativeInfinity): + return AccumBounds(-1, 1) + + if arg is S.ComplexInfinity: + return S.NaN + + if isinstance(arg, AccumBounds): + from sympy.sets.sets import FiniteSet + min, max = arg.min, arg.max + d = floor(min/(2*pi)) + if min is not S.NegativeInfinity: + min = min - d*2*pi + if max is not S.Infinity: + max = max - d*2*pi + if AccumBounds(min, max).intersection(FiniteSet(pi/2, pi*Rational(5, 2))) \ + is not S.EmptySet and \ + AccumBounds(min, max).intersection(FiniteSet(pi*Rational(3, 2), + pi*Rational(7, 2))) is not S.EmptySet: + return AccumBounds(-1, 1) + elif AccumBounds(min, max).intersection(FiniteSet(pi/2, pi*Rational(5, 2))) \ + is not S.EmptySet: + return AccumBounds(Min(sin(min), sin(max)), 1) + elif AccumBounds(min, max).intersection(FiniteSet(pi*Rational(3, 2), pi*Rational(8, 2))) \ + is not S.EmptySet: + return AccumBounds(-1, Max(sin(min), sin(max))) + else: + return AccumBounds(Min(sin(min), sin(max)), + Max(sin(min), sin(max))) + elif isinstance(arg, SetExpr): + return arg._eval_func(cls) + + if arg.could_extract_minus_sign(): + return -cls(-arg) + + i_coeff = _imaginary_unit_as_coefficient(arg) + if i_coeff is not None: + from sympy.functions.elementary.hyperbolic import sinh + return S.ImaginaryUnit*sinh(i_coeff) + + pi_coeff = _pi_coeff(arg) + if pi_coeff is not None: + if pi_coeff.is_integer: + return S.Zero + + if (2*pi_coeff).is_integer: + # is_even-case handled above as then pi_coeff.is_integer, + # so check if known to be not even + if pi_coeff.is_even is False: + return S.NegativeOne**(pi_coeff - S.Half) + + if not pi_coeff.is_Rational: + narg = pi_coeff*pi + if narg != arg: + return cls(narg) + return None + + # https://github.com/sympy/sympy/issues/6048 + # transform a sine to a cosine, to avoid redundant code + if pi_coeff.is_Rational: + x = pi_coeff % 2 + if x > 1: + return -cls((x % 1)*pi) + if 2*x > 1: + return cls((1 - x)*pi) + narg = ((pi_coeff + Rational(3, 2)) % 2)*pi + result = cos(narg) + if not isinstance(result, cos): + return result + if pi_coeff*pi != arg: + return cls(pi_coeff*pi) + return None + + if arg.is_Add: + x, m = _peeloff_pi(arg) + if m: + m = m*pi + return sin(m)*cos(x) + cos(m)*sin(x) + + if arg.is_zero: + return S.Zero + + if isinstance(arg, asin): + return arg.args[0] + + if isinstance(arg, atan): + x = arg.args[0] + return x/sqrt(1 + x**2) + + if isinstance(arg, atan2): + y, x = arg.args + return y/sqrt(x**2 + y**2) + + if isinstance(arg, acos): + x = arg.args[0] + return sqrt(1 - x**2) + + if isinstance(arg, acot): + x = arg.args[0] + return 1/(sqrt(1 + 1/x**2)*x) + + if isinstance(arg, acsc): + x = arg.args[0] + return 1/x + + if isinstance(arg, asec): + x = arg.args[0] + return sqrt(1 - 1/x**2) + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n < 0 or n % 2 == 0: + return S.Zero + else: + x = sympify(x) + + if len(previous_terms) > 2: + p = previous_terms[-2] + return -p*x**2/(n*(n - 1)) + else: + return S.NegativeOne**(n//2)*x**n/factorial(n) + + def _eval_nseries(self, x, n, logx, cdir=0): + arg = self.args[0] + if logx is not None: + arg = arg.subs(log(x), logx) + if arg.subs(x, 0).has(S.NaN, S.ComplexInfinity): + raise PoleError("Cannot expand %s around 0" % (self)) + return Function._eval_nseries(self, x, n=n, logx=logx, cdir=cdir) + + def _eval_rewrite_as_exp(self, arg, **kwargs): + from sympy.functions.elementary.hyperbolic import HyperbolicFunction + I = S.ImaginaryUnit + if isinstance(arg, (TrigonometricFunction, HyperbolicFunction)): + arg = arg.func(arg.args[0]).rewrite(exp) + return (exp(arg*I) - exp(-arg*I))/(2*I) + + def _eval_rewrite_as_Pow(self, arg, **kwargs): + if isinstance(arg, log): + I = S.ImaginaryUnit + x = arg.args[0] + return I*x**-I/2 - I*x**I /2 + + def _eval_rewrite_as_cos(self, arg, **kwargs): + return cos(arg - pi/2, evaluate=False) + + def _eval_rewrite_as_tan(self, arg, **kwargs): + tan_half = tan(S.Half*arg) + return 2*tan_half/(1 + tan_half**2) + + def _eval_rewrite_as_sincos(self, arg, **kwargs): + return sin(arg)*cos(arg)/cos(arg) + + def _eval_rewrite_as_cot(self, arg, **kwargs): + cot_half = cot(S.Half*arg) + return Piecewise((0, And(Eq(im(arg), 0), Eq(Mod(arg, pi), 0))), + (2*cot_half/(1 + cot_half**2), True)) + + def _eval_rewrite_as_pow(self, arg, **kwargs): + return self.rewrite(cos).rewrite(pow) + + def _eval_rewrite_as_sqrt(self, arg, **kwargs): + return self.rewrite(cos).rewrite(sqrt) + + def _eval_rewrite_as_csc(self, arg, **kwargs): + return 1/csc(arg) + + def _eval_rewrite_as_sec(self, arg, **kwargs): + return 1/sec(arg - pi/2, evaluate=False) + + def _eval_rewrite_as_sinc(self, arg, **kwargs): + return arg*sinc(arg) + + def _eval_conjugate(self): + return self.func(self.args[0].conjugate()) + + def as_real_imag(self, deep=True, **hints): + from sympy.functions.elementary.hyperbolic import cosh, sinh + re, im = self._as_real_imag(deep=deep, **hints) + return (sin(re)*cosh(im), cos(re)*sinh(im)) + + def _eval_expand_trig(self, **hints): + from sympy.functions.special.polynomials import chebyshevt, chebyshevu + arg = self.args[0] + x = None + if arg.is_Add: # TODO, implement more if deep stuff here + # TODO: Do this more efficiently for more than two terms + x, y = arg.as_two_terms() + sx = sin(x, evaluate=False)._eval_expand_trig() + sy = sin(y, evaluate=False)._eval_expand_trig() + cx = cos(x, evaluate=False)._eval_expand_trig() + cy = cos(y, evaluate=False)._eval_expand_trig() + return sx*cy + sy*cx + elif arg.is_Mul: + n, x = arg.as_coeff_Mul(rational=True) + if n.is_Integer: # n will be positive because of .eval + # canonicalization + + # See https://mathworld.wolfram.com/Multiple-AngleFormulas.html + if n.is_odd: + return S.NegativeOne**((n - 1)/2)*chebyshevt(n, sin(x)) + else: + return expand_mul(S.NegativeOne**(n/2 - 1)*cos(x)* + chebyshevu(n - 1, sin(x)), deep=False) + pi_coeff = _pi_coeff(arg) + if pi_coeff is not None: + if pi_coeff.is_Rational: + return self.rewrite(sqrt) + return sin(arg) + + def _eval_as_leading_term(self, x, logx=None, cdir=0): + from sympy.calculus.accumulationbounds import AccumBounds + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + n = x0/pi + if n.is_integer: + lt = (arg - n*pi).as_leading_term(x) + return (S.NegativeOne**n)*lt + if x0 is S.ComplexInfinity: + x0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') + if x0 in [S.Infinity, S.NegativeInfinity]: + return AccumBounds(-1, 1) + return self.func(x0) if x0.is_finite else self + + def _eval_is_extended_real(self): + if self.args[0].is_extended_real: + return True + + def _eval_is_finite(self): + arg = self.args[0] + if arg.is_extended_real: + return True + + def _eval_is_zero(self): + rest, pi_mult = _peeloff_pi(self.args[0]) + if rest.is_zero: + return pi_mult.is_integer + + def _eval_is_complex(self): + if self.args[0].is_extended_real \ + or self.args[0].is_complex: + return True + + +class cos(TrigonometricFunction): + """ + The cosine function. + + Returns the cosine of x (measured in radians). + + Explanation + =========== + + See :func:`sin` for notes about automatic evaluation. + + Examples + ======== + + >>> from sympy import cos, pi + >>> from sympy.abc import x + >>> cos(x**2).diff(x) + -2*x*sin(x**2) + >>> cos(1).diff(x) + 0 + >>> cos(pi) + -1 + >>> cos(pi/2) + 0 + >>> cos(2*pi/3) + -1/2 + >>> cos(pi/12) + sqrt(2)/4 + sqrt(6)/4 + + See Also + ======== + + sin, csc, sec, tan, cot + asin, acsc, acos, asec, atan, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Trigonometric_functions + .. [2] https://dlmf.nist.gov/4.14 + .. [3] https://functions.wolfram.com/ElementaryFunctions/Cos + + """ + + def period(self, symbol=None): + return self._period(2*pi, symbol) + + def fdiff(self, argindex=1): + if argindex == 1: + return -sin(self.args[0]) + else: + raise ArgumentIndexError(self, argindex) + + @classmethod + def eval(cls, arg): + from sympy.functions.special.polynomials import chebyshevt + from sympy.calculus.accumulationbounds import AccumBounds + from sympy.sets.setexpr import SetExpr + if arg.is_Number: + if arg is S.NaN: + return S.NaN + elif arg.is_zero: + return S.One + elif arg in (S.Infinity, S.NegativeInfinity): + # In this case it is better to return AccumBounds(-1, 1) + # rather than returning S.NaN, since AccumBounds(-1, 1) + # preserves the information that sin(oo) is between + # -1 and 1, where S.NaN does not do that. + return AccumBounds(-1, 1) + + if arg is S.ComplexInfinity: + return S.NaN + + if isinstance(arg, AccumBounds): + return sin(arg + pi/2) + elif isinstance(arg, SetExpr): + return arg._eval_func(cls) + + if arg.is_extended_real and arg.is_finite is False: + return AccumBounds(-1, 1) + + if arg.could_extract_minus_sign(): + return cls(-arg) + + i_coeff = _imaginary_unit_as_coefficient(arg) + if i_coeff is not None: + from sympy.functions.elementary.hyperbolic import cosh + return cosh(i_coeff) + + pi_coeff = _pi_coeff(arg) + if pi_coeff is not None: + if pi_coeff.is_integer: + return (S.NegativeOne)**pi_coeff + + if (2*pi_coeff).is_integer: + # is_even-case handled above as then pi_coeff.is_integer, + # so check if known to be not even + if pi_coeff.is_even is False: + return S.Zero + + if not pi_coeff.is_Rational: + narg = pi_coeff*pi + if narg != arg: + return cls(narg) + return None + + # cosine formula ##################### + # https://github.com/sympy/sympy/issues/6048 + # explicit calculations are performed for + # cos(k pi/n) for n = 8,10,12,15,20,24,30,40,60,120 + # Some other exact values like cos(k pi/240) can be + # calculated using a partial-fraction decomposition + # by calling cos( X ).rewrite(sqrt) + if pi_coeff.is_Rational: + q = pi_coeff.q + p = pi_coeff.p % (2*q) + if p > q: + narg = (pi_coeff - 1)*pi + return -cls(narg) + if 2*p > q: + narg = (1 - pi_coeff)*pi + return -cls(narg) + + # If nested sqrt's are worse than un-evaluation + # you can require q to be in (1, 2, 3, 4, 6, 12) + # q <= 12, q=15, q=20, q=24, q=30, q=40, q=60, q=120 return + # expressions with 2 or fewer sqrt nestings. + table2 = _table2() + if q in table2: + a, b = table2[q] + a, b = p*pi/a, p*pi/b + nvala, nvalb = cls(a), cls(b) + if None in (nvala, nvalb): + return None + return nvala*nvalb + cls(pi/2 - a)*cls(pi/2 - b) + + if q > 12: + return None + + cst_table_some = { + 3: S.Half, + 5: (sqrt(5) + 1) / 4, + } + if q in cst_table_some: + cts = cst_table_some[pi_coeff.q] + return chebyshevt(pi_coeff.p, cts).expand() + + if 0 == q % 2: + narg = (pi_coeff*2)*pi + nval = cls(narg) + if None == nval: + return None + x = (2*pi_coeff + 1)/2 + sign_cos = (-1)**((-1 if x < 0 else 1)*int(abs(x))) + return sign_cos*sqrt( (1 + nval)/2 ) + return None + + if arg.is_Add: + x, m = _peeloff_pi(arg) + if m: + m = m*pi + return cos(m)*cos(x) - sin(m)*sin(x) + + if arg.is_zero: + return S.One + + if isinstance(arg, acos): + return arg.args[0] + + if isinstance(arg, atan): + x = arg.args[0] + return 1/sqrt(1 + x**2) + + if isinstance(arg, atan2): + y, x = arg.args + return x/sqrt(x**2 + y**2) + + if isinstance(arg, asin): + x = arg.args[0] + return sqrt(1 - x ** 2) + + if isinstance(arg, acot): + x = arg.args[0] + return 1/sqrt(1 + 1/x**2) + + if isinstance(arg, acsc): + x = arg.args[0] + return sqrt(1 - 1/x**2) + + if isinstance(arg, asec): + x = arg.args[0] + return 1/x + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n < 0 or n % 2 == 1: + return S.Zero + else: + x = sympify(x) + + if len(previous_terms) > 2: + p = previous_terms[-2] + return -p*x**2/(n*(n - 1)) + else: + return S.NegativeOne**(n//2)*x**n/factorial(n) + + def _eval_nseries(self, x, n, logx, cdir=0): + arg = self.args[0] + if logx is not None: + arg = arg.subs(log(x), logx) + if arg.subs(x, 0).has(S.NaN, S.ComplexInfinity): + raise PoleError("Cannot expand %s around 0" % (self)) + return Function._eval_nseries(self, x, n=n, logx=logx, cdir=cdir) + + def _eval_rewrite_as_exp(self, arg, **kwargs): + I = S.ImaginaryUnit + from sympy.functions.elementary.hyperbolic import HyperbolicFunction + if isinstance(arg, (TrigonometricFunction, HyperbolicFunction)): + arg = arg.func(arg.args[0]).rewrite(exp) + return (exp(arg*I) + exp(-arg*I))/2 + + def _eval_rewrite_as_Pow(self, arg, **kwargs): + if isinstance(arg, log): + I = S.ImaginaryUnit + x = arg.args[0] + return x**I/2 + x**-I/2 + + def _eval_rewrite_as_sin(self, arg, **kwargs): + return sin(arg + pi/2, evaluate=False) + + def _eval_rewrite_as_tan(self, arg, **kwargs): + tan_half = tan(S.Half*arg)**2 + return (1 - tan_half)/(1 + tan_half) + + def _eval_rewrite_as_sincos(self, arg, **kwargs): + return sin(arg)*cos(arg)/sin(arg) + + def _eval_rewrite_as_cot(self, arg, **kwargs): + cot_half = cot(S.Half*arg)**2 + return Piecewise((1, And(Eq(im(arg), 0), Eq(Mod(arg, 2*pi), 0))), + ((cot_half - 1)/(cot_half + 1), True)) + + def _eval_rewrite_as_pow(self, arg, **kwargs): + return self._eval_rewrite_as_sqrt(arg) + + def _eval_rewrite_as_sqrt(self, arg: Expr, **kwargs): + from sympy.functions.special.polynomials import chebyshevt + + pi_coeff = _pi_coeff(arg) + if pi_coeff is None: + return None + + if isinstance(pi_coeff, Integer): + return None + + if not isinstance(pi_coeff, Rational): + return None + + cst_table_some = cos_table() + + if pi_coeff.q in cst_table_some: + rv = chebyshevt(pi_coeff.p, cst_table_some[pi_coeff.q]()) + if pi_coeff.q < 257: + rv = rv.expand() + return rv + + if not pi_coeff.q % 2: # recursively remove factors of 2 + pico2 = pi_coeff * 2 + nval = cos(pico2 * pi).rewrite(sqrt) + x = (pico2 + 1) / 2 + sign_cos = -1 if int(x) % 2 else 1 + return sign_cos * sqrt((1 + nval) / 2) + + FC = fermat_coords(pi_coeff.q) + if FC: + denoms = FC + else: + denoms = [b**e for b, e in factorint(pi_coeff.q).items()] + + apart = ipartfrac(*denoms) + decomp = (pi_coeff.p * Rational(n, d) for n, d in zip(apart, denoms)) + X = [(x[1], x[0]*pi) for x in zip(decomp, numbered_symbols('z'))] + pcls = cos(sum(x[0] for x in X))._eval_expand_trig().subs(X) + + if not FC or len(FC) == 1: + return pcls + return pcls.rewrite(sqrt) + + def _eval_rewrite_as_sec(self, arg, **kwargs): + return 1/sec(arg) + + def _eval_rewrite_as_csc(self, arg, **kwargs): + return 1/sec(arg).rewrite(csc) + + def _eval_conjugate(self): + return self.func(self.args[0].conjugate()) + + def as_real_imag(self, deep=True, **hints): + from sympy.functions.elementary.hyperbolic import cosh, sinh + re, im = self._as_real_imag(deep=deep, **hints) + return (cos(re)*cosh(im), -sin(re)*sinh(im)) + + def _eval_expand_trig(self, **hints): + from sympy.functions.special.polynomials import chebyshevt + arg = self.args[0] + x = None + if arg.is_Add: # TODO: Do this more efficiently for more than two terms + x, y = arg.as_two_terms() + sx = sin(x, evaluate=False)._eval_expand_trig() + sy = sin(y, evaluate=False)._eval_expand_trig() + cx = cos(x, evaluate=False)._eval_expand_trig() + cy = cos(y, evaluate=False)._eval_expand_trig() + return cx*cy - sx*sy + elif arg.is_Mul: + coeff, terms = arg.as_coeff_Mul(rational=True) + if coeff.is_Integer: + return chebyshevt(coeff, cos(terms)) + pi_coeff = _pi_coeff(arg) + if pi_coeff is not None: + if pi_coeff.is_Rational: + return self.rewrite(sqrt) + return cos(arg) + + def _eval_as_leading_term(self, x, logx=None, cdir=0): + from sympy.calculus.accumulationbounds import AccumBounds + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + n = (x0 + pi/2)/pi + if n.is_integer: + lt = (arg - n*pi + pi/2).as_leading_term(x) + return (S.NegativeOne**n)*lt + if x0 is S.ComplexInfinity: + x0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') + if x0 in [S.Infinity, S.NegativeInfinity]: + return AccumBounds(-1, 1) + return self.func(x0) if x0.is_finite else self + + def _eval_is_extended_real(self): + if self.args[0].is_extended_real: + return True + + def _eval_is_finite(self): + arg = self.args[0] + + if arg.is_extended_real: + return True + + def _eval_is_complex(self): + if self.args[0].is_extended_real \ + or self.args[0].is_complex: + return True + + def _eval_is_zero(self): + rest, pi_mult = _peeloff_pi(self.args[0]) + if rest.is_zero and pi_mult: + return (pi_mult - S.Half).is_integer + + +class tan(TrigonometricFunction): + """ + The tangent function. + + Returns the tangent of x (measured in radians). + + Explanation + =========== + + See :class:`sin` for notes about automatic evaluation. + + Examples + ======== + + >>> from sympy import tan, pi + >>> from sympy.abc import x + >>> tan(x**2).diff(x) + 2*x*(tan(x**2)**2 + 1) + >>> tan(1).diff(x) + 0 + >>> tan(pi/8).expand() + -1 + sqrt(2) + + See Also + ======== + + sin, csc, cos, sec, cot + asin, acsc, acos, asec, atan, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Trigonometric_functions + .. [2] https://dlmf.nist.gov/4.14 + .. [3] https://functions.wolfram.com/ElementaryFunctions/Tan + + """ + + def period(self, symbol=None): + return self._period(pi, symbol) + + def fdiff(self, argindex=1): + if argindex == 1: + return S.One + self**2 + else: + raise ArgumentIndexError(self, argindex) + + def inverse(self, argindex=1): + """ + Returns the inverse of this function. + """ + return atan + + @classmethod + def eval(cls, arg): + from sympy.calculus.accumulationbounds import AccumBounds + if arg.is_Number: + if arg is S.NaN: + return S.NaN + elif arg.is_zero: + return S.Zero + elif arg in (S.Infinity, S.NegativeInfinity): + return AccumBounds(S.NegativeInfinity, S.Infinity) + + if arg is S.ComplexInfinity: + return S.NaN + + if isinstance(arg, AccumBounds): + min, max = arg.min, arg.max + d = floor(min/pi) + if min is not S.NegativeInfinity: + min = min - d*pi + if max is not S.Infinity: + max = max - d*pi + from sympy.sets.sets import FiniteSet + if AccumBounds(min, max).intersection(FiniteSet(pi/2, pi*Rational(3, 2))): + return AccumBounds(S.NegativeInfinity, S.Infinity) + else: + return AccumBounds(tan(min), tan(max)) + + if arg.could_extract_minus_sign(): + return -cls(-arg) + + i_coeff = _imaginary_unit_as_coefficient(arg) + if i_coeff is not None: + from sympy.functions.elementary.hyperbolic import tanh + return S.ImaginaryUnit*tanh(i_coeff) + + pi_coeff = _pi_coeff(arg, 2) + if pi_coeff is not None: + if pi_coeff.is_integer: + return S.Zero + + if not pi_coeff.is_Rational: + narg = pi_coeff*pi + if narg != arg: + return cls(narg) + return None + + if pi_coeff.is_Rational: + q = pi_coeff.q + p = pi_coeff.p % q + # ensure simplified results are returned for n*pi/5, n*pi/10 + table10 = { + 1: sqrt(1 - 2*sqrt(5)/5), + 2: sqrt(5 - 2*sqrt(5)), + 3: sqrt(1 + 2*sqrt(5)/5), + 4: sqrt(5 + 2*sqrt(5)) + } + if q in (5, 10): + n = 10*p/q + if n > 5: + n = 10 - n + return -table10[n] + else: + return table10[n] + if not pi_coeff.q % 2: + narg = pi_coeff*pi*2 + cresult, sresult = cos(narg), cos(narg - pi/2) + if not isinstance(cresult, cos) \ + and not isinstance(sresult, cos): + if sresult == 0: + return S.ComplexInfinity + return 1/sresult - cresult/sresult + + table2 = _table2() + if q in table2: + a, b = table2[q] + nvala, nvalb = cls(p*pi/a), cls(p*pi/b) + if None in (nvala, nvalb): + return None + return (nvala - nvalb)/(1 + nvala*nvalb) + narg = ((pi_coeff + S.Half) % 1 - S.Half)*pi + # see cos() to specify which expressions should be + # expanded automatically in terms of radicals + cresult, sresult = cos(narg), cos(narg - pi/2) + if not isinstance(cresult, cos) \ + and not isinstance(sresult, cos): + if cresult == 0: + return S.ComplexInfinity + return (sresult/cresult) + if narg != arg: + return cls(narg) + + if arg.is_Add: + x, m = _peeloff_pi(arg) + if m: + tanm = tan(m*pi) + if tanm is S.ComplexInfinity: + return -cot(x) + else: # tanm == 0 + return tan(x) + + if arg.is_zero: + return S.Zero + + if isinstance(arg, atan): + return arg.args[0] + + if isinstance(arg, atan2): + y, x = arg.args + return y/x + + if isinstance(arg, asin): + x = arg.args[0] + return x/sqrt(1 - x**2) + + if isinstance(arg, acos): + x = arg.args[0] + return sqrt(1 - x**2)/x + + if isinstance(arg, acot): + x = arg.args[0] + return 1/x + + if isinstance(arg, acsc): + x = arg.args[0] + return 1/(sqrt(1 - 1/x**2)*x) + + if isinstance(arg, asec): + x = arg.args[0] + return sqrt(1 - 1/x**2)*x + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n < 0 or n % 2 == 0: + return S.Zero + else: + x = sympify(x) + + a, b = ((n - 1)//2), 2**(n + 1) + + B = bernoulli(n + 1) + F = factorial(n + 1) + + return S.NegativeOne**a*b*(b - 1)*B/F*x**n + + def _eval_nseries(self, x, n, logx, cdir=0): + i = self.args[0].limit(x, 0)*2/pi + if i and i.is_Integer: + return self.rewrite(cos)._eval_nseries(x, n=n, logx=logx) + return Function._eval_nseries(self, x, n=n, logx=logx) + + def _eval_rewrite_as_Pow(self, arg, **kwargs): + if isinstance(arg, log): + I = S.ImaginaryUnit + x = arg.args[0] + return I*(x**-I - x**I)/(x**-I + x**I) + + def _eval_conjugate(self): + return self.func(self.args[0].conjugate()) + + def as_real_imag(self, deep=True, **hints): + re, im = self._as_real_imag(deep=deep, **hints) + if im: + from sympy.functions.elementary.hyperbolic import cosh, sinh + denom = cos(2*re) + cosh(2*im) + return (sin(2*re)/denom, sinh(2*im)/denom) + else: + return (self.func(re), S.Zero) + + def _eval_expand_trig(self, **hints): + arg = self.args[0] + x = None + if arg.is_Add: + n = len(arg.args) + TX = [] + for x in arg.args: + tx = tan(x, evaluate=False)._eval_expand_trig() + TX.append(tx) + + Yg = numbered_symbols('Y') + Y = [ next(Yg) for i in range(n) ] + + p = [0, 0] + for i in range(n + 1): + p[1 - i % 2] += symmetric_poly(i, Y)*(-1)**((i % 4)//2) + return (p[0]/p[1]).subs(list(zip(Y, TX))) + + elif arg.is_Mul: + coeff, terms = arg.as_coeff_Mul(rational=True) + if coeff.is_Integer and coeff > 1: + I = S.ImaginaryUnit + z = Symbol('dummy', real=True) + P = ((1 + I*z)**coeff).expand() + return (im(P)/re(P)).subs([(z, tan(terms))]) + return tan(arg) + + def _eval_rewrite_as_exp(self, arg, **kwargs): + I = S.ImaginaryUnit + from sympy.functions.elementary.hyperbolic import HyperbolicFunction + if isinstance(arg, (TrigonometricFunction, HyperbolicFunction)): + arg = arg.func(arg.args[0]).rewrite(exp) + neg_exp, pos_exp = exp(-arg*I), exp(arg*I) + return I*(neg_exp - pos_exp)/(neg_exp + pos_exp) + + def _eval_rewrite_as_sin(self, x, **kwargs): + return 2*sin(x)**2/sin(2*x) + + def _eval_rewrite_as_cos(self, x, **kwargs): + return cos(x - pi/2, evaluate=False)/cos(x) + + def _eval_rewrite_as_sincos(self, arg, **kwargs): + return sin(arg)/cos(arg) + + def _eval_rewrite_as_cot(self, arg, **kwargs): + return 1/cot(arg) + + def _eval_rewrite_as_sec(self, arg, **kwargs): + sin_in_sec_form = sin(arg).rewrite(sec) + cos_in_sec_form = cos(arg).rewrite(sec) + return sin_in_sec_form/cos_in_sec_form + + def _eval_rewrite_as_csc(self, arg, **kwargs): + sin_in_csc_form = sin(arg).rewrite(csc) + cos_in_csc_form = cos(arg).rewrite(csc) + return sin_in_csc_form/cos_in_csc_form + + def _eval_rewrite_as_pow(self, arg, **kwargs): + y = self.rewrite(cos).rewrite(pow) + if y.has(cos): + return None + return y + + def _eval_rewrite_as_sqrt(self, arg, **kwargs): + y = self.rewrite(cos).rewrite(sqrt) + if y.has(cos): + return None + return y + + def _eval_as_leading_term(self, x, logx=None, cdir=0): + from sympy.calculus.accumulationbounds import AccumBounds + from sympy.functions.elementary.complexes import re + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + n = 2*x0/pi + if n.is_integer: + lt = (arg - n*pi/2).as_leading_term(x) + return lt if n.is_even else -1/lt + if x0 is S.ComplexInfinity: + x0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') + if x0 in (S.Infinity, S.NegativeInfinity): + return AccumBounds(S.NegativeInfinity, S.Infinity) + return self.func(x0) if x0.is_finite else self + + def _eval_is_extended_real(self): + # FIXME: currently tan(pi/2) return zoo + return self.args[0].is_extended_real + + def _eval_is_real(self): + arg = self.args[0] + if arg.is_real and (arg/pi - S.Half).is_integer is False: + return True + + def _eval_is_finite(self): + arg = self.args[0] + + if arg.is_real and (arg/pi - S.Half).is_integer is False: + return True + + if arg.is_imaginary: + return True + + def _eval_is_zero(self): + rest, pi_mult = _peeloff_pi(self.args[0]) + if rest.is_zero: + return pi_mult.is_integer + + def _eval_is_complex(self): + arg = self.args[0] + + if arg.is_real and (arg/pi - S.Half).is_integer is False: + return True + + +class cot(TrigonometricFunction): + """ + The cotangent function. + + Returns the cotangent of x (measured in radians). + + Explanation + =========== + + See :class:`sin` for notes about automatic evaluation. + + Examples + ======== + + >>> from sympy import cot, pi + >>> from sympy.abc import x + >>> cot(x**2).diff(x) + 2*x*(-cot(x**2)**2 - 1) + >>> cot(1).diff(x) + 0 + >>> cot(pi/12) + sqrt(3) + 2 + + See Also + ======== + + sin, csc, cos, sec, tan + asin, acsc, acos, asec, atan, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Trigonometric_functions + .. [2] https://dlmf.nist.gov/4.14 + .. [3] https://functions.wolfram.com/ElementaryFunctions/Cot + + """ + + def period(self, symbol=None): + return self._period(pi, symbol) + + def fdiff(self, argindex=1): + if argindex == 1: + return S.NegativeOne - self**2 + else: + raise ArgumentIndexError(self, argindex) + + def inverse(self, argindex=1): + """ + Returns the inverse of this function. + """ + return acot + + @classmethod + def eval(cls, arg): + from sympy.calculus.accumulationbounds import AccumBounds + if arg.is_Number: + if arg is S.NaN: + return S.NaN + if arg.is_zero: + return S.ComplexInfinity + elif arg in (S.Infinity, S.NegativeInfinity): + return AccumBounds(S.NegativeInfinity, S.Infinity) + + if arg is S.ComplexInfinity: + return S.NaN + + if isinstance(arg, AccumBounds): + return -tan(arg + pi/2) + + if arg.could_extract_minus_sign(): + return -cls(-arg) + + i_coeff = _imaginary_unit_as_coefficient(arg) + if i_coeff is not None: + from sympy.functions.elementary.hyperbolic import coth + return -S.ImaginaryUnit*coth(i_coeff) + + pi_coeff = _pi_coeff(arg, 2) + if pi_coeff is not None: + if pi_coeff.is_integer: + return S.ComplexInfinity + + if not pi_coeff.is_Rational: + narg = pi_coeff*pi + if narg != arg: + return cls(narg) + return None + + if pi_coeff.is_Rational: + if pi_coeff.q in (5, 10): + return tan(pi/2 - arg) + if pi_coeff.q > 2 and not pi_coeff.q % 2: + narg = pi_coeff*pi*2 + cresult, sresult = cos(narg), cos(narg - pi/2) + if not isinstance(cresult, cos) \ + and not isinstance(sresult, cos): + return 1/sresult + cresult/sresult + q = pi_coeff.q + p = pi_coeff.p % q + table2 = _table2() + if q in table2: + a, b = table2[q] + nvala, nvalb = cls(p*pi/a), cls(p*pi/b) + if None in (nvala, nvalb): + return None + return (1 + nvala*nvalb)/(nvalb - nvala) + narg = (((pi_coeff + S.Half) % 1) - S.Half)*pi + # see cos() to specify which expressions should be + # expanded automatically in terms of radicals + cresult, sresult = cos(narg), cos(narg - pi/2) + if not isinstance(cresult, cos) \ + and not isinstance(sresult, cos): + if sresult == 0: + return S.ComplexInfinity + return cresult/sresult + if narg != arg: + return cls(narg) + + if arg.is_Add: + x, m = _peeloff_pi(arg) + if m: + cotm = cot(m*pi) + if cotm is S.ComplexInfinity: + return cot(x) + else: # cotm == 0 + return -tan(x) + + if arg.is_zero: + return S.ComplexInfinity + + if isinstance(arg, acot): + return arg.args[0] + + if isinstance(arg, atan): + x = arg.args[0] + return 1/x + + if isinstance(arg, atan2): + y, x = arg.args + return x/y + + if isinstance(arg, asin): + x = arg.args[0] + return sqrt(1 - x**2)/x + + if isinstance(arg, acos): + x = arg.args[0] + return x/sqrt(1 - x**2) + + if isinstance(arg, acsc): + x = arg.args[0] + return sqrt(1 - 1/x**2)*x + + if isinstance(arg, asec): + x = arg.args[0] + return 1/(sqrt(1 - 1/x**2)*x) + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n == 0: + return 1/sympify(x) + elif n < 0 or n % 2 == 0: + return S.Zero + else: + x = sympify(x) + + B = bernoulli(n + 1) + F = factorial(n + 1) + + return S.NegativeOne**((n + 1)//2)*2**(n + 1)*B/F*x**n + + def _eval_nseries(self, x, n, logx, cdir=0): + i = self.args[0].limit(x, 0)/pi + if i and i.is_Integer: + return self.rewrite(cos)._eval_nseries(x, n=n, logx=logx) + return self.rewrite(tan)._eval_nseries(x, n=n, logx=logx) + + def _eval_conjugate(self): + return self.func(self.args[0].conjugate()) + + def as_real_imag(self, deep=True, **hints): + re, im = self._as_real_imag(deep=deep, **hints) + if im: + from sympy.functions.elementary.hyperbolic import cosh, sinh + denom = cos(2*re) - cosh(2*im) + return (-sin(2*re)/denom, sinh(2*im)/denom) + else: + return (self.func(re), S.Zero) + + def _eval_rewrite_as_exp(self, arg, **kwargs): + from sympy.functions.elementary.hyperbolic import HyperbolicFunction + I = S.ImaginaryUnit + if isinstance(arg, (TrigonometricFunction, HyperbolicFunction)): + arg = arg.func(arg.args[0]).rewrite(exp) + neg_exp, pos_exp = exp(-arg*I), exp(arg*I) + return I*(pos_exp + neg_exp)/(pos_exp - neg_exp) + + def _eval_rewrite_as_Pow(self, arg, **kwargs): + if isinstance(arg, log): + I = S.ImaginaryUnit + x = arg.args[0] + return -I*(x**-I + x**I)/(x**-I - x**I) + + def _eval_rewrite_as_sin(self, x, **kwargs): + return sin(2*x)/(2*(sin(x)**2)) + + def _eval_rewrite_as_cos(self, x, **kwargs): + return cos(x)/cos(x - pi/2, evaluate=False) + + def _eval_rewrite_as_sincos(self, arg, **kwargs): + return cos(arg)/sin(arg) + + def _eval_rewrite_as_tan(self, arg, **kwargs): + return 1/tan(arg) + + def _eval_rewrite_as_sec(self, arg, **kwargs): + cos_in_sec_form = cos(arg).rewrite(sec) + sin_in_sec_form = sin(arg).rewrite(sec) + return cos_in_sec_form/sin_in_sec_form + + def _eval_rewrite_as_csc(self, arg, **kwargs): + cos_in_csc_form = cos(arg).rewrite(csc) + sin_in_csc_form = sin(arg).rewrite(csc) + return cos_in_csc_form/sin_in_csc_form + + def _eval_rewrite_as_pow(self, arg, **kwargs): + y = self.rewrite(cos).rewrite(pow) + if y.has(cos): + return None + return y + + def _eval_rewrite_as_sqrt(self, arg, **kwargs): + y = self.rewrite(cos).rewrite(sqrt) + if y.has(cos): + return None + return y + + def _eval_as_leading_term(self, x, logx=None, cdir=0): + from sympy.calculus.accumulationbounds import AccumBounds + from sympy.functions.elementary.complexes import re + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + n = 2*x0/pi + if n.is_integer: + lt = (arg - n*pi/2).as_leading_term(x) + return 1/lt if n.is_even else -lt + if x0 is S.ComplexInfinity: + x0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') + if x0 in (S.Infinity, S.NegativeInfinity): + return AccumBounds(S.NegativeInfinity, S.Infinity) + return self.func(x0) if x0.is_finite else self + + def _eval_is_extended_real(self): + return self.args[0].is_extended_real + + def _eval_expand_trig(self, **hints): + arg = self.args[0] + x = None + if arg.is_Add: + n = len(arg.args) + CX = [] + for x in arg.args: + cx = cot(x, evaluate=False)._eval_expand_trig() + CX.append(cx) + + Yg = numbered_symbols('Y') + Y = [ next(Yg) for i in range(n) ] + + p = [0, 0] + for i in range(n, -1, -1): + p[(n - i) % 2] += symmetric_poly(i, Y)*(-1)**(((n - i) % 4)//2) + return (p[0]/p[1]).subs(list(zip(Y, CX))) + elif arg.is_Mul: + coeff, terms = arg.as_coeff_Mul(rational=True) + if coeff.is_Integer and coeff > 1: + I = S.ImaginaryUnit + z = Symbol('dummy', real=True) + P = ((z + I)**coeff).expand() + return (re(P)/im(P)).subs([(z, cot(terms))]) + return cot(arg) # XXX sec and csc return 1/cos and 1/sin + + def _eval_is_finite(self): + arg = self.args[0] + if arg.is_real and (arg/pi).is_integer is False: + return True + if arg.is_imaginary: + return True + + def _eval_is_real(self): + arg = self.args[0] + if arg.is_real and (arg/pi).is_integer is False: + return True + + def _eval_is_complex(self): + arg = self.args[0] + if arg.is_real and (arg/pi).is_integer is False: + return True + + def _eval_is_zero(self): + rest, pimult = _peeloff_pi(self.args[0]) + if pimult and rest.is_zero: + return (pimult - S.Half).is_integer + + def _eval_subs(self, old, new): + arg = self.args[0] + argnew = arg.subs(old, new) + if arg != argnew and (argnew/pi).is_integer: + return S.ComplexInfinity + return cot(argnew) + + +class ReciprocalTrigonometricFunction(TrigonometricFunction): + """Base class for reciprocal functions of trigonometric functions. """ + + _reciprocal_of = None # mandatory, to be defined in subclass + _singularities = (S.ComplexInfinity,) + + # _is_even and _is_odd are used for correct evaluation of csc(-x), sec(-x) + # TODO refactor into TrigonometricFunction common parts of + # trigonometric functions eval() like even/odd, func(x+2*k*pi), etc. + + # optional, to be defined in subclasses: + _is_even: FuzzyBool = None + _is_odd: FuzzyBool = None + + @classmethod + def eval(cls, arg): + if arg.could_extract_minus_sign(): + if cls._is_even: + return cls(-arg) + if cls._is_odd: + return -cls(-arg) + + pi_coeff = _pi_coeff(arg) + if (pi_coeff is not None + and not (2*pi_coeff).is_integer + and pi_coeff.is_Rational): + q = pi_coeff.q + p = pi_coeff.p % (2*q) + if p > q: + narg = (pi_coeff - 1)*pi + return -cls(narg) + if 2*p > q: + narg = (1 - pi_coeff)*pi + if cls._is_odd: + return cls(narg) + elif cls._is_even: + return -cls(narg) + + if hasattr(arg, 'inverse') and arg.inverse() == cls: + return arg.args[0] + + t = cls._reciprocal_of.eval(arg) + if t is None: + return t + elif any(isinstance(i, cos) for i in (t, -t)): + return (1/t).rewrite(sec) + elif any(isinstance(i, sin) for i in (t, -t)): + return (1/t).rewrite(csc) + else: + return 1/t + + def _call_reciprocal(self, method_name, *args, **kwargs): + # Calls method_name on _reciprocal_of + o = self._reciprocal_of(self.args[0]) + return getattr(o, method_name)(*args, **kwargs) + + def _calculate_reciprocal(self, method_name, *args, **kwargs): + # If calling method_name on _reciprocal_of returns a value != None + # then return the reciprocal of that value + t = self._call_reciprocal(method_name, *args, **kwargs) + return 1/t if t is not None else t + + def _rewrite_reciprocal(self, method_name, arg): + # Special handling for rewrite functions. If reciprocal rewrite returns + # unmodified expression, then return None + t = self._call_reciprocal(method_name, arg) + if t is not None and t != self._reciprocal_of(arg): + return 1/t + + def _period(self, symbol): + f = expand_mul(self.args[0]) + return self._reciprocal_of(f).period(symbol) + + def fdiff(self, argindex=1): + return -self._calculate_reciprocal("fdiff", argindex)/self**2 + + def _eval_rewrite_as_exp(self, arg, **kwargs): + return self._rewrite_reciprocal("_eval_rewrite_as_exp", arg) + + def _eval_rewrite_as_Pow(self, arg, **kwargs): + return self._rewrite_reciprocal("_eval_rewrite_as_Pow", arg) + + def _eval_rewrite_as_sin(self, arg, **kwargs): + return self._rewrite_reciprocal("_eval_rewrite_as_sin", arg) + + def _eval_rewrite_as_cos(self, arg, **kwargs): + return self._rewrite_reciprocal("_eval_rewrite_as_cos", arg) + + def _eval_rewrite_as_tan(self, arg, **kwargs): + return self._rewrite_reciprocal("_eval_rewrite_as_tan", arg) + + def _eval_rewrite_as_pow(self, arg, **kwargs): + return self._rewrite_reciprocal("_eval_rewrite_as_pow", arg) + + def _eval_rewrite_as_sqrt(self, arg, **kwargs): + return self._rewrite_reciprocal("_eval_rewrite_as_sqrt", arg) + + def _eval_conjugate(self): + return self.func(self.args[0].conjugate()) + + def as_real_imag(self, deep=True, **hints): + return (1/self._reciprocal_of(self.args[0])).as_real_imag(deep, + **hints) + + def _eval_expand_trig(self, **hints): + return self._calculate_reciprocal("_eval_expand_trig", **hints) + + def _eval_is_extended_real(self): + return self._reciprocal_of(self.args[0])._eval_is_extended_real() + + def _eval_as_leading_term(self, x, logx=None, cdir=0): + return (1/self._reciprocal_of(self.args[0]))._eval_as_leading_term(x) + + def _eval_is_finite(self): + return (1/self._reciprocal_of(self.args[0])).is_finite + + def _eval_nseries(self, x, n, logx, cdir=0): + return (1/self._reciprocal_of(self.args[0]))._eval_nseries(x, n, logx) + + +class sec(ReciprocalTrigonometricFunction): + """ + The secant function. + + Returns the secant of x (measured in radians). + + Explanation + =========== + + See :class:`sin` for notes about automatic evaluation. + + Examples + ======== + + >>> from sympy import sec + >>> from sympy.abc import x + >>> sec(x**2).diff(x) + 2*x*tan(x**2)*sec(x**2) + >>> sec(1).diff(x) + 0 + + See Also + ======== + + sin, csc, cos, tan, cot + asin, acsc, acos, asec, atan, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Trigonometric_functions + .. [2] https://dlmf.nist.gov/4.14 + .. [3] https://functions.wolfram.com/ElementaryFunctions/Sec + + """ + + _reciprocal_of = cos + _is_even = True + + def period(self, symbol=None): + return self._period(symbol) + + def _eval_rewrite_as_cot(self, arg, **kwargs): + cot_half_sq = cot(arg/2)**2 + return (cot_half_sq + 1)/(cot_half_sq - 1) + + def _eval_rewrite_as_cos(self, arg, **kwargs): + return (1/cos(arg)) + + def _eval_rewrite_as_sincos(self, arg, **kwargs): + return sin(arg)/(cos(arg)*sin(arg)) + + def _eval_rewrite_as_sin(self, arg, **kwargs): + return (1/cos(arg).rewrite(sin)) + + def _eval_rewrite_as_tan(self, arg, **kwargs): + return (1/cos(arg).rewrite(tan)) + + def _eval_rewrite_as_csc(self, arg, **kwargs): + return csc(pi/2 - arg, evaluate=False) + + def fdiff(self, argindex=1): + if argindex == 1: + return tan(self.args[0])*sec(self.args[0]) + else: + raise ArgumentIndexError(self, argindex) + + def _eval_is_complex(self): + arg = self.args[0] + + if arg.is_complex and (arg/pi - S.Half).is_integer is False: + return True + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + # Reference Formula: + # https://functions.wolfram.com/ElementaryFunctions/Sec/06/01/02/01/ + if n < 0 or n % 2 == 1: + return S.Zero + else: + x = sympify(x) + k = n//2 + return S.NegativeOne**k*euler(2*k)/factorial(2*k)*x**(2*k) + + def _eval_as_leading_term(self, x, logx=None, cdir=0): + from sympy.calculus.accumulationbounds import AccumBounds + from sympy.functions.elementary.complexes import re + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + n = (x0 + pi/2)/pi + if n.is_integer: + lt = (arg - n*pi + pi/2).as_leading_term(x) + return (S.NegativeOne**n)/lt + if x0 is S.ComplexInfinity: + x0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') + if x0 in (S.Infinity, S.NegativeInfinity): + return AccumBounds(S.NegativeInfinity, S.Infinity) + return self.func(x0) if x0.is_finite else self + + +class csc(ReciprocalTrigonometricFunction): + """ + The cosecant function. + + Returns the cosecant of x (measured in radians). + + Explanation + =========== + + See :func:`sin` for notes about automatic evaluation. + + Examples + ======== + + >>> from sympy import csc + >>> from sympy.abc import x + >>> csc(x**2).diff(x) + -2*x*cot(x**2)*csc(x**2) + >>> csc(1).diff(x) + 0 + + See Also + ======== + + sin, cos, sec, tan, cot + asin, acsc, acos, asec, atan, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Trigonometric_functions + .. [2] https://dlmf.nist.gov/4.14 + .. [3] https://functions.wolfram.com/ElementaryFunctions/Csc + + """ + + _reciprocal_of = sin + _is_odd = True + + def period(self, symbol=None): + return self._period(symbol) + + def _eval_rewrite_as_sin(self, arg, **kwargs): + return (1/sin(arg)) + + def _eval_rewrite_as_sincos(self, arg, **kwargs): + return cos(arg)/(sin(arg)*cos(arg)) + + def _eval_rewrite_as_cot(self, arg, **kwargs): + cot_half = cot(arg/2) + return (1 + cot_half**2)/(2*cot_half) + + def _eval_rewrite_as_cos(self, arg, **kwargs): + return 1/sin(arg).rewrite(cos) + + def _eval_rewrite_as_sec(self, arg, **kwargs): + return sec(pi/2 - arg, evaluate=False) + + def _eval_rewrite_as_tan(self, arg, **kwargs): + return (1/sin(arg).rewrite(tan)) + + def fdiff(self, argindex=1): + if argindex == 1: + return -cot(self.args[0])*csc(self.args[0]) + else: + raise ArgumentIndexError(self, argindex) + + def _eval_is_complex(self): + arg = self.args[0] + if arg.is_real and (arg/pi).is_integer is False: + return True + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n == 0: + return 1/sympify(x) + elif n < 0 or n % 2 == 0: + return S.Zero + else: + x = sympify(x) + k = n//2 + 1 + return (S.NegativeOne**(k - 1)*2*(2**(2*k - 1) - 1)* + bernoulli(2*k)*x**(2*k - 1)/factorial(2*k)) + + def _eval_as_leading_term(self, x, logx=None, cdir=0): + from sympy.calculus.accumulationbounds import AccumBounds + from sympy.functions.elementary.complexes import re + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + n = x0/pi + if n.is_integer: + lt = (arg - n*pi).as_leading_term(x) + return (S.NegativeOne**n)/lt + if x0 is S.ComplexInfinity: + x0 = arg.limit(x, 0, dir='-' if re(cdir).is_negative else '+') + if x0 in (S.Infinity, S.NegativeInfinity): + return AccumBounds(S.NegativeInfinity, S.Infinity) + return self.func(x0) if x0.is_finite else self + + +class sinc(Function): + r""" + Represents an unnormalized sinc function: + + .. math:: + + \operatorname{sinc}(x) = + \begin{cases} + \frac{\sin x}{x} & \qquad x \neq 0 \\ + 1 & \qquad x = 0 + \end{cases} + + Examples + ======== + + >>> from sympy import sinc, oo, jn + >>> from sympy.abc import x + >>> sinc(x) + sinc(x) + + * Automated Evaluation + + >>> sinc(0) + 1 + >>> sinc(oo) + 0 + + * Differentiation + + >>> sinc(x).diff() + cos(x)/x - sin(x)/x**2 + + * Series Expansion + + >>> sinc(x).series() + 1 - x**2/6 + x**4/120 + O(x**6) + + * As zero'th order spherical Bessel Function + + >>> sinc(x).rewrite(jn) + jn(0, x) + + See also + ======== + + sin + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Sinc_function + + """ + _singularities = (S.ComplexInfinity,) + + def fdiff(self, argindex=1): + x = self.args[0] + if argindex == 1: + # We would like to return the Piecewise here, but Piecewise.diff + # currently can't handle removable singularities, meaning things + # like sinc(x).diff(x, 2) give the wrong answer at x = 0. See + # https://github.com/sympy/sympy/issues/11402. + # + # return Piecewise(((x*cos(x) - sin(x))/x**2, Ne(x, S.Zero)), (S.Zero, S.true)) + return cos(x)/x - sin(x)/x**2 + else: + raise ArgumentIndexError(self, argindex) + + @classmethod + def eval(cls, arg): + if arg.is_zero: + return S.One + if arg.is_Number: + if arg in [S.Infinity, S.NegativeInfinity]: + return S.Zero + elif arg is S.NaN: + return S.NaN + + if arg is S.ComplexInfinity: + return S.NaN + + if arg.could_extract_minus_sign(): + return cls(-arg) + + pi_coeff = _pi_coeff(arg) + if pi_coeff is not None: + if pi_coeff.is_integer: + if fuzzy_not(arg.is_zero): + return S.Zero + elif (2*pi_coeff).is_integer: + return S.NegativeOne**(pi_coeff - S.Half)/arg + + def _eval_nseries(self, x, n, logx, cdir=0): + x = self.args[0] + return (sin(x)/x)._eval_nseries(x, n, logx) + + def _eval_rewrite_as_jn(self, arg, **kwargs): + from sympy.functions.special.bessel import jn + return jn(0, arg) + + def _eval_rewrite_as_sin(self, arg, **kwargs): + return Piecewise((sin(arg)/arg, Ne(arg, S.Zero)), (S.One, S.true)) + + def _eval_is_zero(self): + if self.args[0].is_infinite: + return True + rest, pi_mult = _peeloff_pi(self.args[0]) + if rest.is_zero: + return fuzzy_and([pi_mult.is_integer, pi_mult.is_nonzero]) + if rest.is_Number and pi_mult.is_integer: + return False + + def _eval_is_real(self): + if self.args[0].is_extended_real or self.args[0].is_imaginary: + return True + + _eval_is_finite = _eval_is_real + + +############################################################################### +########################### TRIGONOMETRIC INVERSES ############################ +############################################################################### + + +class InverseTrigonometricFunction(Function): + """Base class for inverse trigonometric functions.""" + _singularities = (S.One, S.NegativeOne, S.Zero, S.ComplexInfinity) # type: tTuple[Expr, ...] + + @staticmethod + @cacheit + def _asin_table(): + # Only keys with could_extract_minus_sign() == False + # are actually needed. + return { + sqrt(3)/2: pi/3, + sqrt(2)/2: pi/4, + 1/sqrt(2): pi/4, + sqrt((5 - sqrt(5))/8): pi/5, + sqrt(2)*sqrt(5 - sqrt(5))/4: pi/5, + sqrt((5 + sqrt(5))/8): pi*Rational(2, 5), + sqrt(2)*sqrt(5 + sqrt(5))/4: pi*Rational(2, 5), + S.Half: pi/6, + sqrt(2 - sqrt(2))/2: pi/8, + sqrt(S.Half - sqrt(2)/4): pi/8, + sqrt(2 + sqrt(2))/2: pi*Rational(3, 8), + sqrt(S.Half + sqrt(2)/4): pi*Rational(3, 8), + (sqrt(5) - 1)/4: pi/10, + (1 - sqrt(5))/4: -pi/10, + (sqrt(5) + 1)/4: pi*Rational(3, 10), + sqrt(6)/4 - sqrt(2)/4: pi/12, + -sqrt(6)/4 + sqrt(2)/4: -pi/12, + (sqrt(3) - 1)/sqrt(8): pi/12, + (1 - sqrt(3))/sqrt(8): -pi/12, + sqrt(6)/4 + sqrt(2)/4: pi*Rational(5, 12), + (1 + sqrt(3))/sqrt(8): pi*Rational(5, 12) + } + + + @staticmethod + @cacheit + def _atan_table(): + # Only keys with could_extract_minus_sign() == False + # are actually needed. + return { + sqrt(3)/3: pi/6, + 1/sqrt(3): pi/6, + sqrt(3): pi/3, + sqrt(2) - 1: pi/8, + 1 - sqrt(2): -pi/8, + 1 + sqrt(2): pi*Rational(3, 8), + sqrt(5 - 2*sqrt(5)): pi/5, + sqrt(5 + 2*sqrt(5)): pi*Rational(2, 5), + sqrt(1 - 2*sqrt(5)/5): pi/10, + sqrt(1 + 2*sqrt(5)/5): pi*Rational(3, 10), + 2 - sqrt(3): pi/12, + -2 + sqrt(3): -pi/12, + 2 + sqrt(3): pi*Rational(5, 12) + } + + @staticmethod + @cacheit + def _acsc_table(): + # Keys for which could_extract_minus_sign() + # will obviously return True are omitted. + return { + 2*sqrt(3)/3: pi/3, + sqrt(2): pi/4, + sqrt(2 + 2*sqrt(5)/5): pi/5, + 1/sqrt(Rational(5, 8) - sqrt(5)/8): pi/5, + sqrt(2 - 2*sqrt(5)/5): pi*Rational(2, 5), + 1/sqrt(Rational(5, 8) + sqrt(5)/8): pi*Rational(2, 5), + 2: pi/6, + sqrt(4 + 2*sqrt(2)): pi/8, + 2/sqrt(2 - sqrt(2)): pi/8, + sqrt(4 - 2*sqrt(2)): pi*Rational(3, 8), + 2/sqrt(2 + sqrt(2)): pi*Rational(3, 8), + 1 + sqrt(5): pi/10, + sqrt(5) - 1: pi*Rational(3, 10), + -(sqrt(5) - 1): pi*Rational(-3, 10), + sqrt(6) + sqrt(2): pi/12, + sqrt(6) - sqrt(2): pi*Rational(5, 12), + -(sqrt(6) - sqrt(2)): pi*Rational(-5, 12) + } + + +class asin(InverseTrigonometricFunction): + r""" + The inverse sine function. + + Returns the arcsine of x in radians. + + Explanation + =========== + + ``asin(x)`` will evaluate automatically in the cases + $x \in \{\infty, -\infty, 0, 1, -1\}$ and for some instances when the + result is a rational multiple of $\pi$ (see the ``eval`` class method). + + A purely imaginary argument will lead to an asinh expression. + + Examples + ======== + + >>> from sympy import asin, oo + >>> asin(1) + pi/2 + >>> asin(-1) + -pi/2 + >>> asin(-oo) + oo*I + >>> asin(oo) + -oo*I + + See Also + ======== + + sin, csc, cos, sec, tan, cot + acsc, acos, asec, atan, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Inverse_trigonometric_functions + .. [2] https://dlmf.nist.gov/4.23 + .. [3] https://functions.wolfram.com/ElementaryFunctions/ArcSin + + """ + + def fdiff(self, argindex=1): + if argindex == 1: + return 1/sqrt(1 - self.args[0]**2) + else: + raise ArgumentIndexError(self, argindex) + + def _eval_is_rational(self): + s = self.func(*self.args) + if s.func == self.func: + if s.args[0].is_rational: + return False + else: + return s.is_rational + + def _eval_is_positive(self): + return self._eval_is_extended_real() and self.args[0].is_positive + + def _eval_is_negative(self): + return self._eval_is_extended_real() and self.args[0].is_negative + + @classmethod + def eval(cls, arg): + if arg.is_Number: + if arg is S.NaN: + return S.NaN + elif arg is S.Infinity: + return S.NegativeInfinity*S.ImaginaryUnit + elif arg is S.NegativeInfinity: + return S.Infinity*S.ImaginaryUnit + elif arg.is_zero: + return S.Zero + elif arg is S.One: + return pi/2 + elif arg is S.NegativeOne: + return -pi/2 + + if arg is S.ComplexInfinity: + return S.ComplexInfinity + + if arg.could_extract_minus_sign(): + return -cls(-arg) + + if arg.is_number: + asin_table = cls._asin_table() + if arg in asin_table: + return asin_table[arg] + + i_coeff = _imaginary_unit_as_coefficient(arg) + if i_coeff is not None: + from sympy.functions.elementary.hyperbolic import asinh + return S.ImaginaryUnit*asinh(i_coeff) + + if arg.is_zero: + return S.Zero + + if isinstance(arg, sin): + ang = arg.args[0] + if ang.is_comparable: + ang %= 2*pi # restrict to [0,2*pi) + if ang > pi: # restrict to (-pi,pi] + ang = pi - ang + + # restrict to [-pi/2,pi/2] + if ang > pi/2: + ang = pi - ang + if ang < -pi/2: + ang = -pi - ang + + return ang + + if isinstance(arg, cos): # acos(x) + asin(x) = pi/2 + ang = arg.args[0] + if ang.is_comparable: + return pi/2 - acos(arg) + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n < 0 or n % 2 == 0: + return S.Zero + else: + x = sympify(x) + if len(previous_terms) >= 2 and n > 2: + p = previous_terms[-2] + return p*(n - 2)**2/(n*(n - 1))*x**2 + else: + k = (n - 1) // 2 + R = RisingFactorial(S.Half, k) + F = factorial(k) + return R/F*x**n/n + + def _eval_as_leading_term(self, x, logx=None, cdir=0): # asin + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + if x0.is_zero: + return arg.as_leading_term(x) + # Handling branch points + if x0 in (-S.One, S.One, S.ComplexInfinity): + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir).expand() + # Handling points lying on branch cuts (-oo, -1) U (1, oo) + if (1 - x0**2).is_negative: + ndir = arg.dir(x, cdir if cdir else 1) + if im(ndir).is_negative: + if x0.is_negative: + return -pi - self.func(x0) + elif im(ndir).is_positive: + if x0.is_positive: + return pi - self.func(x0) + else: + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir).expand() + return self.func(x0) + + def _eval_nseries(self, x, n, logx, cdir=0): # asin + from sympy.series.order import O + arg0 = self.args[0].subs(x, 0) + # Handling branch points + if arg0 is S.One: + t = Dummy('t', positive=True) + ser = asin(S.One - t**2).rewrite(log).nseries(t, 0, 2*n) + arg1 = S.One - self.args[0] + f = arg1.as_leading_term(x) + g = (arg1 - f)/ f + if not g.is_meromorphic(x, 0): # cannot be expanded + return O(1) if n == 0 else pi/2 + O(sqrt(x)) + res1 = sqrt(S.One + g)._eval_nseries(x, n=n, logx=logx) + res = (res1.removeO()*sqrt(f)).expand() + return ser.removeO().subs(t, res).expand().powsimp() + O(x**n, x) + + if arg0 is S.NegativeOne: + t = Dummy('t', positive=True) + ser = asin(S.NegativeOne + t**2).rewrite(log).nseries(t, 0, 2*n) + arg1 = S.One + self.args[0] + f = arg1.as_leading_term(x) + g = (arg1 - f)/ f + if not g.is_meromorphic(x, 0): # cannot be expanded + return O(1) if n == 0 else -pi/2 + O(sqrt(x)) + res1 = sqrt(S.One + g)._eval_nseries(x, n=n, logx=logx) + res = (res1.removeO()*sqrt(f)).expand() + return ser.removeO().subs(t, res).expand().powsimp() + O(x**n, x) + + res = Function._eval_nseries(self, x, n=n, logx=logx) + if arg0 is S.ComplexInfinity: + return res + # Handling points lying on branch cuts (-oo, -1) U (1, oo) + if (1 - arg0**2).is_negative: + ndir = self.args[0].dir(x, cdir if cdir else 1) + if im(ndir).is_negative: + if arg0.is_negative: + return -pi - res + elif im(ndir).is_positive: + if arg0.is_positive: + return pi - res + else: + return self.rewrite(log)._eval_nseries(x, n, logx=logx, cdir=cdir) + return res + + def _eval_rewrite_as_acos(self, x, **kwargs): + return pi/2 - acos(x) + + def _eval_rewrite_as_atan(self, x, **kwargs): + return 2*atan(x/(1 + sqrt(1 - x**2))) + + def _eval_rewrite_as_log(self, x, **kwargs): + return -S.ImaginaryUnit*log(S.ImaginaryUnit*x + sqrt(1 - x**2)) + + _eval_rewrite_as_tractable = _eval_rewrite_as_log + + def _eval_rewrite_as_acot(self, arg, **kwargs): + return 2*acot((1 + sqrt(1 - arg**2))/arg) + + def _eval_rewrite_as_asec(self, arg, **kwargs): + return pi/2 - asec(1/arg) + + def _eval_rewrite_as_acsc(self, arg, **kwargs): + return acsc(1/arg) + + def _eval_is_extended_real(self): + x = self.args[0] + return x.is_extended_real and (1 - abs(x)).is_nonnegative + + def inverse(self, argindex=1): + """ + Returns the inverse of this function. + """ + return sin + + +class acos(InverseTrigonometricFunction): + r""" + The inverse cosine function. + + Explanation + =========== + + Returns the arc cosine of x (measured in radians). + + ``acos(x)`` will evaluate automatically in the cases + $x \in \{\infty, -\infty, 0, 1, -1\}$ and for some instances when + the result is a rational multiple of $\pi$ (see the eval class method). + + ``acos(zoo)`` evaluates to ``zoo`` + (see note in :class:`sympy.functions.elementary.trigonometric.asec`) + + A purely imaginary argument will be rewritten to asinh. + + Examples + ======== + + >>> from sympy import acos, oo + >>> acos(1) + 0 + >>> acos(0) + pi/2 + >>> acos(oo) + oo*I + + See Also + ======== + + sin, csc, cos, sec, tan, cot + asin, acsc, asec, atan, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Inverse_trigonometric_functions + .. [2] https://dlmf.nist.gov/4.23 + .. [3] https://functions.wolfram.com/ElementaryFunctions/ArcCos + + """ + + def fdiff(self, argindex=1): + if argindex == 1: + return -1/sqrt(1 - self.args[0]**2) + else: + raise ArgumentIndexError(self, argindex) + + def _eval_is_rational(self): + s = self.func(*self.args) + if s.func == self.func: + if s.args[0].is_rational: + return False + else: + return s.is_rational + + @classmethod + def eval(cls, arg): + if arg.is_Number: + if arg is S.NaN: + return S.NaN + elif arg is S.Infinity: + return S.Infinity*S.ImaginaryUnit + elif arg is S.NegativeInfinity: + return S.NegativeInfinity*S.ImaginaryUnit + elif arg.is_zero: + return pi/2 + elif arg is S.One: + return S.Zero + elif arg is S.NegativeOne: + return pi + + if arg is S.ComplexInfinity: + return S.ComplexInfinity + + if arg.is_number: + asin_table = cls._asin_table() + if arg in asin_table: + return pi/2 - asin_table[arg] + elif -arg in asin_table: + return pi/2 + asin_table[-arg] + + i_coeff = _imaginary_unit_as_coefficient(arg) + if i_coeff is not None: + return pi/2 - asin(arg) + + if isinstance(arg, cos): + ang = arg.args[0] + if ang.is_comparable: + ang %= 2*pi # restrict to [0,2*pi) + if ang > pi: # restrict to [0,pi] + ang = 2*pi - ang + + return ang + + if isinstance(arg, sin): # acos(x) + asin(x) = pi/2 + ang = arg.args[0] + if ang.is_comparable: + return pi/2 - asin(arg) + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n == 0: + return pi/2 + elif n < 0 or n % 2 == 0: + return S.Zero + else: + x = sympify(x) + if len(previous_terms) >= 2 and n > 2: + p = previous_terms[-2] + return p*(n - 2)**2/(n*(n - 1))*x**2 + else: + k = (n - 1) // 2 + R = RisingFactorial(S.Half, k) + F = factorial(k) + return -R/F*x**n/n + + def _eval_as_leading_term(self, x, logx=None, cdir=0): # acos + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + # Handling branch points + if x0 == 1: + return sqrt(2)*sqrt((S.One - arg).as_leading_term(x)) + if x0 in (-S.One, S.ComplexInfinity): + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir) + # Handling points lying on branch cuts (-oo, -1) U (1, oo) + if (1 - x0**2).is_negative: + ndir = arg.dir(x, cdir if cdir else 1) + if im(ndir).is_negative: + if x0.is_negative: + return 2*pi - self.func(x0) + elif im(ndir).is_positive: + if x0.is_positive: + return -self.func(x0) + else: + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir).expand() + return self.func(x0) + + def _eval_is_extended_real(self): + x = self.args[0] + return x.is_extended_real and (1 - abs(x)).is_nonnegative + + def _eval_is_nonnegative(self): + return self._eval_is_extended_real() + + def _eval_nseries(self, x, n, logx, cdir=0): # acos + from sympy.series.order import O + arg0 = self.args[0].subs(x, 0) + # Handling branch points + if arg0 is S.One: + t = Dummy('t', positive=True) + ser = acos(S.One - t**2).rewrite(log).nseries(t, 0, 2*n) + arg1 = S.One - self.args[0] + f = arg1.as_leading_term(x) + g = (arg1 - f)/ f + if not g.is_meromorphic(x, 0): # cannot be expanded + return O(1) if n == 0 else O(sqrt(x)) + res1 = sqrt(S.One + g)._eval_nseries(x, n=n, logx=logx) + res = (res1.removeO()*sqrt(f)).expand() + return ser.removeO().subs(t, res).expand().powsimp() + O(x**n, x) + + if arg0 is S.NegativeOne: + t = Dummy('t', positive=True) + ser = acos(S.NegativeOne + t**2).rewrite(log).nseries(t, 0, 2*n) + arg1 = S.One + self.args[0] + f = arg1.as_leading_term(x) + g = (arg1 - f)/ f + if not g.is_meromorphic(x, 0): # cannot be expanded + return O(1) if n == 0 else pi + O(sqrt(x)) + res1 = sqrt(S.One + g)._eval_nseries(x, n=n, logx=logx) + res = (res1.removeO()*sqrt(f)).expand() + return ser.removeO().subs(t, res).expand().powsimp() + O(x**n, x) + + res = Function._eval_nseries(self, x, n=n, logx=logx) + if arg0 is S.ComplexInfinity: + return res + # Handling points lying on branch cuts (-oo, -1) U (1, oo) + if (1 - arg0**2).is_negative: + ndir = self.args[0].dir(x, cdir if cdir else 1) + if im(ndir).is_negative: + if arg0.is_negative: + return 2*pi - res + elif im(ndir).is_positive: + if arg0.is_positive: + return -res + else: + return self.rewrite(log)._eval_nseries(x, n, logx=logx, cdir=cdir) + return res + + def _eval_rewrite_as_log(self, x, **kwargs): + return pi/2 + S.ImaginaryUnit*\ + log(S.ImaginaryUnit*x + sqrt(1 - x**2)) + + _eval_rewrite_as_tractable = _eval_rewrite_as_log + + def _eval_rewrite_as_asin(self, x, **kwargs): + return pi/2 - asin(x) + + def _eval_rewrite_as_atan(self, x, **kwargs): + return atan(sqrt(1 - x**2)/x) + (pi/2)*(1 - x*sqrt(1/x**2)) + + def inverse(self, argindex=1): + """ + Returns the inverse of this function. + """ + return cos + + def _eval_rewrite_as_acot(self, arg, **kwargs): + return pi/2 - 2*acot((1 + sqrt(1 - arg**2))/arg) + + def _eval_rewrite_as_asec(self, arg, **kwargs): + return asec(1/arg) + + def _eval_rewrite_as_acsc(self, arg, **kwargs): + return pi/2 - acsc(1/arg) + + def _eval_conjugate(self): + z = self.args[0] + r = self.func(self.args[0].conjugate()) + if z.is_extended_real is False: + return r + elif z.is_extended_real and (z + 1).is_nonnegative and (z - 1).is_nonpositive: + return r + + +class atan(InverseTrigonometricFunction): + r""" + The inverse tangent function. + + Returns the arc tangent of x (measured in radians). + + Explanation + =========== + + ``atan(x)`` will evaluate automatically in the cases + $x \in \{\infty, -\infty, 0, 1, -1\}$ and for some instances when the + result is a rational multiple of $\pi$ (see the eval class method). + + Examples + ======== + + >>> from sympy import atan, oo + >>> atan(0) + 0 + >>> atan(1) + pi/4 + >>> atan(oo) + pi/2 + + See Also + ======== + + sin, csc, cos, sec, tan, cot + asin, acsc, acos, asec, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Inverse_trigonometric_functions + .. [2] https://dlmf.nist.gov/4.23 + .. [3] https://functions.wolfram.com/ElementaryFunctions/ArcTan + + """ + + args: tTuple[Expr] + + _singularities = (S.ImaginaryUnit, -S.ImaginaryUnit) + + def fdiff(self, argindex=1): + if argindex == 1: + return 1/(1 + self.args[0]**2) + else: + raise ArgumentIndexError(self, argindex) + + def _eval_is_rational(self): + s = self.func(*self.args) + if s.func == self.func: + if s.args[0].is_rational: + return False + else: + return s.is_rational + + def _eval_is_positive(self): + return self.args[0].is_extended_positive + + def _eval_is_nonnegative(self): + return self.args[0].is_extended_nonnegative + + def _eval_is_zero(self): + return self.args[0].is_zero + + def _eval_is_real(self): + return self.args[0].is_extended_real + + @classmethod + def eval(cls, arg): + if arg.is_Number: + if arg is S.NaN: + return S.NaN + elif arg is S.Infinity: + return pi/2 + elif arg is S.NegativeInfinity: + return -pi/2 + elif arg.is_zero: + return S.Zero + elif arg is S.One: + return pi/4 + elif arg is S.NegativeOne: + return -pi/4 + + if arg is S.ComplexInfinity: + from sympy.calculus.accumulationbounds import AccumBounds + return AccumBounds(-pi/2, pi/2) + + if arg.could_extract_minus_sign(): + return -cls(-arg) + + if arg.is_number: + atan_table = cls._atan_table() + if arg in atan_table: + return atan_table[arg] + + i_coeff = _imaginary_unit_as_coefficient(arg) + if i_coeff is not None: + from sympy.functions.elementary.hyperbolic import atanh + return S.ImaginaryUnit*atanh(i_coeff) + + if arg.is_zero: + return S.Zero + + if isinstance(arg, tan): + ang = arg.args[0] + if ang.is_comparable: + ang %= pi # restrict to [0,pi) + if ang > pi/2: # restrict to [-pi/2,pi/2] + ang -= pi + + return ang + + if isinstance(arg, cot): # atan(x) + acot(x) = pi/2 + ang = arg.args[0] + if ang.is_comparable: + ang = pi/2 - acot(arg) + if ang > pi/2: # restrict to [-pi/2,pi/2] + ang -= pi + return ang + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n < 0 or n % 2 == 0: + return S.Zero + else: + x = sympify(x) + return S.NegativeOne**((n - 1)//2)*x**n/n + + def _eval_as_leading_term(self, x, logx=None, cdir=0): # atan + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + if x0.is_zero: + return arg.as_leading_term(x) + # Handling branch points + if x0 in (-S.ImaginaryUnit, S.ImaginaryUnit, S.ComplexInfinity): + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir).expand() + # Handling points lying on branch cuts (-I*oo, -I) U (I, I*oo) + if (1 + x0**2).is_negative: + ndir = arg.dir(x, cdir if cdir else 1) + if re(ndir).is_negative: + if im(x0).is_positive: + return self.func(x0) - pi + elif re(ndir).is_positive: + if im(x0).is_negative: + return self.func(x0) + pi + else: + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir).expand() + return self.func(x0) + + def _eval_nseries(self, x, n, logx, cdir=0): # atan + arg0 = self.args[0].subs(x, 0) + + # Handling branch points + if arg0 in (S.ImaginaryUnit, S.NegativeOne*S.ImaginaryUnit): + return self.rewrite(log)._eval_nseries(x, n, logx=logx, cdir=cdir) + + res = Function._eval_nseries(self, x, n=n, logx=logx) + ndir = self.args[0].dir(x, cdir if cdir else 1) + if arg0 is S.ComplexInfinity: + if re(ndir) > 0: + return res - pi + return res + # Handling points lying on branch cuts (-I*oo, -I) U (I, I*oo) + if (1 + arg0**2).is_negative: + if re(ndir).is_negative: + if im(arg0).is_positive: + return res - pi + elif re(ndir).is_positive: + if im(arg0).is_negative: + return res + pi + else: + return self.rewrite(log)._eval_nseries(x, n, logx=logx, cdir=cdir) + return res + + def _eval_rewrite_as_log(self, x, **kwargs): + return S.ImaginaryUnit/2*(log(S.One - S.ImaginaryUnit*x) + - log(S.One + S.ImaginaryUnit*x)) + + _eval_rewrite_as_tractable = _eval_rewrite_as_log + + def _eval_aseries(self, n, args0, x, logx): + if args0[0] is S.Infinity: + return (pi/2 - atan(1/self.args[0]))._eval_nseries(x, n, logx) + elif args0[0] is S.NegativeInfinity: + return (-pi/2 - atan(1/self.args[0]))._eval_nseries(x, n, logx) + else: + return super()._eval_aseries(n, args0, x, logx) + + def inverse(self, argindex=1): + """ + Returns the inverse of this function. + """ + return tan + + def _eval_rewrite_as_asin(self, arg, **kwargs): + return sqrt(arg**2)/arg*(pi/2 - asin(1/sqrt(1 + arg**2))) + + def _eval_rewrite_as_acos(self, arg, **kwargs): + return sqrt(arg**2)/arg*acos(1/sqrt(1 + arg**2)) + + def _eval_rewrite_as_acot(self, arg, **kwargs): + return acot(1/arg) + + def _eval_rewrite_as_asec(self, arg, **kwargs): + return sqrt(arg**2)/arg*asec(sqrt(1 + arg**2)) + + def _eval_rewrite_as_acsc(self, arg, **kwargs): + return sqrt(arg**2)/arg*(pi/2 - acsc(sqrt(1 + arg**2))) + + +class acot(InverseTrigonometricFunction): + r""" + The inverse cotangent function. + + Returns the arc cotangent of x (measured in radians). + + Explanation + =========== + + ``acot(x)`` will evaluate automatically in the cases + $x \in \{\infty, -\infty, \tilde{\infty}, 0, 1, -1\}$ + and for some instances when the result is a rational multiple of $\pi$ + (see the eval class method). + + A purely imaginary argument will lead to an ``acoth`` expression. + + ``acot(x)`` has a branch cut along $(-i, i)$, hence it is discontinuous + at 0. Its range for real $x$ is $(-\frac{\pi}{2}, \frac{\pi}{2}]$. + + Examples + ======== + + >>> from sympy import acot, sqrt + >>> acot(0) + pi/2 + >>> acot(1) + pi/4 + >>> acot(sqrt(3) - 2) + -5*pi/12 + + See Also + ======== + + sin, csc, cos, sec, tan, cot + asin, acsc, acos, asec, atan, atan2 + + References + ========== + + .. [1] https://dlmf.nist.gov/4.23 + .. [2] https://functions.wolfram.com/ElementaryFunctions/ArcCot + + """ + _singularities = (S.ImaginaryUnit, -S.ImaginaryUnit) + + def fdiff(self, argindex=1): + if argindex == 1: + return -1/(1 + self.args[0]**2) + else: + raise ArgumentIndexError(self, argindex) + + def _eval_is_rational(self): + s = self.func(*self.args) + if s.func == self.func: + if s.args[0].is_rational: + return False + else: + return s.is_rational + + def _eval_is_positive(self): + return self.args[0].is_nonnegative + + def _eval_is_negative(self): + return self.args[0].is_negative + + def _eval_is_extended_real(self): + return self.args[0].is_extended_real + + @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 S.NegativeInfinity: + return S.Zero + elif arg.is_zero: + return pi/ 2 + elif arg is S.One: + return pi/4 + elif arg is S.NegativeOne: + return -pi/4 + + if arg is S.ComplexInfinity: + return S.Zero + + if arg.could_extract_minus_sign(): + return -cls(-arg) + + if arg.is_number: + atan_table = cls._atan_table() + if arg in atan_table: + ang = pi/2 - atan_table[arg] + if ang > pi/2: # restrict to (-pi/2,pi/2] + ang -= pi + return ang + + i_coeff = _imaginary_unit_as_coefficient(arg) + if i_coeff is not None: + from sympy.functions.elementary.hyperbolic import acoth + return -S.ImaginaryUnit*acoth(i_coeff) + + if arg.is_zero: + return pi*S.Half + + if isinstance(arg, cot): + ang = arg.args[0] + if ang.is_comparable: + ang %= pi # restrict to [0,pi) + if ang > pi/2: # restrict to (-pi/2,pi/2] + ang -= pi; + return ang + + if isinstance(arg, tan): # atan(x) + acot(x) = pi/2 + ang = arg.args[0] + if ang.is_comparable: + ang = pi/2 - atan(arg) + if ang > pi/2: # restrict to (-pi/2,pi/2] + ang -= pi + return ang + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n == 0: + return pi/2 # FIX THIS + elif n < 0 or n % 2 == 0: + return S.Zero + else: + x = sympify(x) + return S.NegativeOne**((n + 1)//2)*x**n/n + + def _eval_as_leading_term(self, x, logx=None, cdir=0): # acot + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + if x0 is S.ComplexInfinity: + return (1/arg).as_leading_term(x) + # Handling branch points + if x0 in (-S.ImaginaryUnit, S.ImaginaryUnit, S.Zero): + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir).expand() + # Handling points lying on branch cuts [-I, I] + if x0.is_imaginary and (1 + x0**2).is_positive: + ndir = arg.dir(x, cdir if cdir else 1) + if re(ndir).is_positive: + if im(x0).is_positive: + return self.func(x0) + pi + elif re(ndir).is_negative: + if im(x0).is_negative: + return self.func(x0) - pi + else: + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir).expand() + return self.func(x0) + + def _eval_nseries(self, x, n, logx, cdir=0): # acot + arg0 = self.args[0].subs(x, 0) + + # Handling branch points + if arg0 in (S.ImaginaryUnit, S.NegativeOne*S.ImaginaryUnit): + return self.rewrite(log)._eval_nseries(x, n, logx=logx, cdir=cdir) + + res = Function._eval_nseries(self, x, n=n, logx=logx) + if arg0 is S.ComplexInfinity: + return res + ndir = self.args[0].dir(x, cdir if cdir else 1) + if arg0.is_zero: + if re(ndir) < 0: + return res - pi + return res + # Handling points lying on branch cuts [-I, I] + if arg0.is_imaginary and (1 + arg0**2).is_positive: + if re(ndir).is_positive: + if im(arg0).is_positive: + return res + pi + elif re(ndir).is_negative: + if im(arg0).is_negative: + return res - pi + else: + return self.rewrite(log)._eval_nseries(x, n, logx=logx, cdir=cdir) + return res + + def _eval_aseries(self, n, args0, x, logx): + if args0[0] is S.Infinity: + return (pi/2 - acot(1/self.args[0]))._eval_nseries(x, n, logx) + elif args0[0] is S.NegativeInfinity: + return (pi*Rational(3, 2) - acot(1/self.args[0]))._eval_nseries(x, n, logx) + else: + return super(atan, self)._eval_aseries(n, args0, x, logx) + + def _eval_rewrite_as_log(self, x, **kwargs): + return S.ImaginaryUnit/2*(log(1 - S.ImaginaryUnit/x) + - log(1 + S.ImaginaryUnit/x)) + + _eval_rewrite_as_tractable = _eval_rewrite_as_log + + def inverse(self, argindex=1): + """ + Returns the inverse of this function. + """ + return cot + + def _eval_rewrite_as_asin(self, arg, **kwargs): + return (arg*sqrt(1/arg**2)* + (pi/2 - asin(sqrt(-arg**2)/sqrt(-arg**2 - 1)))) + + def _eval_rewrite_as_acos(self, arg, **kwargs): + return arg*sqrt(1/arg**2)*acos(sqrt(-arg**2)/sqrt(-arg**2 - 1)) + + def _eval_rewrite_as_atan(self, arg, **kwargs): + return atan(1/arg) + + def _eval_rewrite_as_asec(self, arg, **kwargs): + return arg*sqrt(1/arg**2)*asec(sqrt((1 + arg**2)/arg**2)) + + def _eval_rewrite_as_acsc(self, arg, **kwargs): + return arg*sqrt(1/arg**2)*(pi/2 - acsc(sqrt((1 + arg**2)/arg**2))) + + +class asec(InverseTrigonometricFunction): + r""" + The inverse secant function. + + Returns the arc secant of x (measured in radians). + + Explanation + =========== + + ``asec(x)`` will evaluate automatically in the cases + $x \in \{\infty, -\infty, 0, 1, -1\}$ and for some instances when the + result is a rational multiple of $\pi$ (see the eval class method). + + ``asec(x)`` has branch cut in the interval $[-1, 1]$. For complex arguments, + it can be defined [4]_ as + + .. math:: + \operatorname{sec^{-1}}(z) = -i\frac{\log\left(\sqrt{1 - z^2} + 1\right)}{z} + + At ``x = 0``, for positive branch cut, the limit evaluates to ``zoo``. For + negative branch cut, the limit + + .. math:: + \lim_{z \to 0}-i\frac{\log\left(-\sqrt{1 - z^2} + 1\right)}{z} + + simplifies to :math:`-i\log\left(z/2 + O\left(z^3\right)\right)` which + ultimately evaluates to ``zoo``. + + As ``acos(x) = asec(1/x)``, a similar argument can be given for + ``acos(x)``. + + Examples + ======== + + >>> from sympy import asec, oo + >>> asec(1) + 0 + >>> asec(-1) + pi + >>> asec(0) + zoo + >>> asec(-oo) + pi/2 + + See Also + ======== + + sin, csc, cos, sec, tan, cot + asin, acsc, acos, atan, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Inverse_trigonometric_functions + .. [2] https://dlmf.nist.gov/4.23 + .. [3] https://functions.wolfram.com/ElementaryFunctions/ArcSec + .. [4] https://reference.wolfram.com/language/ref/ArcSec.html + + """ + + @classmethod + def eval(cls, arg): + if arg.is_zero: + return S.ComplexInfinity + if arg.is_Number: + if arg is S.NaN: + return S.NaN + elif arg is S.One: + return S.Zero + elif arg is S.NegativeOne: + return pi + if arg in [S.Infinity, S.NegativeInfinity, S.ComplexInfinity]: + return pi/2 + + if arg.is_number: + acsc_table = cls._acsc_table() + if arg in acsc_table: + return pi/2 - acsc_table[arg] + elif -arg in acsc_table: + return pi/2 + acsc_table[-arg] + + if arg.is_infinite: + return pi/2 + + if isinstance(arg, sec): + ang = arg.args[0] + if ang.is_comparable: + ang %= 2*pi # restrict to [0,2*pi) + if ang > pi: # restrict to [0,pi] + ang = 2*pi - ang + + return ang + + if isinstance(arg, csc): # asec(x) + acsc(x) = pi/2 + ang = arg.args[0] + if ang.is_comparable: + return pi/2 - acsc(arg) + + def fdiff(self, argindex=1): + if argindex == 1: + return 1/(self.args[0]**2*sqrt(1 - 1/self.args[0]**2)) + else: + raise ArgumentIndexError(self, argindex) + + def inverse(self, argindex=1): + """ + Returns the inverse of this function. + """ + return sec + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n == 0: + return S.ImaginaryUnit*log(2 / x) + elif n < 0 or n % 2 == 1: + return S.Zero + else: + x = sympify(x) + if len(previous_terms) > 2 and n > 2: + p = previous_terms[-2] + return p * ((n - 1)*(n-2)) * x**2/(4 * (n//2)**2) + else: + k = n // 2 + R = RisingFactorial(S.Half, k) * n + F = factorial(k) * n // 2 * n // 2 + return -S.ImaginaryUnit * R / F * x**n / 4 + + def _eval_as_leading_term(self, x, logx=None, cdir=0): # asec + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + # Handling branch points + if x0 == 1: + return sqrt(2)*sqrt((arg - S.One).as_leading_term(x)) + if x0 in (-S.One, S.Zero): + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir) + # Handling points lying on branch cuts (-1, 1) + if x0.is_real and (1 - x0**2).is_positive: + ndir = arg.dir(x, cdir if cdir else 1) + if im(ndir).is_negative: + if x0.is_positive: + return -self.func(x0) + elif im(ndir).is_positive: + if x0.is_negative: + return 2*pi - self.func(x0) + else: + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir).expand() + return self.func(x0) + + def _eval_nseries(self, x, n, logx, cdir=0): # asec + from sympy.series.order import O + arg0 = self.args[0].subs(x, 0) + # Handling branch points + if arg0 is S.One: + t = Dummy('t', positive=True) + ser = asec(S.One + t**2).rewrite(log).nseries(t, 0, 2*n) + arg1 = S.NegativeOne + self.args[0] + f = arg1.as_leading_term(x) + g = (arg1 - f)/ f + res1 = sqrt(S.One + g)._eval_nseries(x, n=n, logx=logx) + res = (res1.removeO()*sqrt(f)).expand() + return ser.removeO().subs(t, res).expand().powsimp() + O(x**n, x) + + if arg0 is S.NegativeOne: + t = Dummy('t', positive=True) + ser = asec(S.NegativeOne - t**2).rewrite(log).nseries(t, 0, 2*n) + arg1 = S.NegativeOne - self.args[0] + f = arg1.as_leading_term(x) + g = (arg1 - f)/ f + res1 = sqrt(S.One + g)._eval_nseries(x, n=n, logx=logx) + res = (res1.removeO()*sqrt(f)).expand() + return ser.removeO().subs(t, res).expand().powsimp() + O(x**n, x) + + res = Function._eval_nseries(self, x, n=n, logx=logx) + if arg0 is S.ComplexInfinity: + return res + # Handling points lying on branch cuts (-1, 1) + if arg0.is_real and (1 - arg0**2).is_positive: + ndir = self.args[0].dir(x, cdir if cdir else 1) + if im(ndir).is_negative: + if arg0.is_positive: + return -res + elif im(ndir).is_positive: + if arg0.is_negative: + return 2*pi - res + else: + return self.rewrite(log)._eval_nseries(x, n, logx=logx, cdir=cdir) + return res + + def _eval_is_extended_real(self): + x = self.args[0] + if x.is_extended_real is False: + return False + return fuzzy_or(((x - 1).is_nonnegative, (-x - 1).is_nonnegative)) + + def _eval_rewrite_as_log(self, arg, **kwargs): + return pi/2 + S.ImaginaryUnit*log(S.ImaginaryUnit/arg + sqrt(1 - 1/arg**2)) + + _eval_rewrite_as_tractable = _eval_rewrite_as_log + + def _eval_rewrite_as_asin(self, arg, **kwargs): + return pi/2 - asin(1/arg) + + def _eval_rewrite_as_acos(self, arg, **kwargs): + return acos(1/arg) + + def _eval_rewrite_as_atan(self, x, **kwargs): + sx2x = sqrt(x**2)/x + return pi/2*(1 - sx2x) + sx2x*atan(sqrt(x**2 - 1)) + + def _eval_rewrite_as_acot(self, x, **kwargs): + sx2x = sqrt(x**2)/x + return pi/2*(1 - sx2x) + sx2x*acot(1/sqrt(x**2 - 1)) + + def _eval_rewrite_as_acsc(self, arg, **kwargs): + return pi/2 - acsc(arg) + + +class acsc(InverseTrigonometricFunction): + r""" + The inverse cosecant function. + + Returns the arc cosecant of x (measured in radians). + + Explanation + =========== + + ``acsc(x)`` will evaluate automatically in the cases + $x \in \{\infty, -\infty, 0, 1, -1\}$` and for some instances when the + result is a rational multiple of $\pi$ (see the ``eval`` class method). + + Examples + ======== + + >>> from sympy import acsc, oo + >>> acsc(1) + pi/2 + >>> acsc(-1) + -pi/2 + >>> acsc(oo) + 0 + >>> acsc(-oo) == acsc(oo) + True + >>> acsc(0) + zoo + + See Also + ======== + + sin, csc, cos, sec, tan, cot + asin, acos, asec, atan, acot, atan2 + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Inverse_trigonometric_functions + .. [2] https://dlmf.nist.gov/4.23 + .. [3] https://functions.wolfram.com/ElementaryFunctions/ArcCsc + + """ + + @classmethod + def eval(cls, arg): + if arg.is_zero: + return S.ComplexInfinity + if arg.is_Number: + if arg is S.NaN: + return S.NaN + elif arg is S.One: + return pi/2 + elif arg is S.NegativeOne: + return -pi/2 + if arg in [S.Infinity, S.NegativeInfinity, S.ComplexInfinity]: + return S.Zero + + if arg.could_extract_minus_sign(): + return -cls(-arg) + + if arg.is_infinite: + return S.Zero + + if arg.is_number: + acsc_table = cls._acsc_table() + if arg in acsc_table: + return acsc_table[arg] + + if isinstance(arg, csc): + ang = arg.args[0] + if ang.is_comparable: + ang %= 2*pi # restrict to [0,2*pi) + if ang > pi: # restrict to (-pi,pi] + ang = pi - ang + + # restrict to [-pi/2,pi/2] + if ang > pi/2: + ang = pi - ang + if ang < -pi/2: + ang = -pi - ang + + return ang + + if isinstance(arg, sec): # asec(x) + acsc(x) = pi/2 + ang = arg.args[0] + if ang.is_comparable: + return pi/2 - asec(arg) + + def fdiff(self, argindex=1): + if argindex == 1: + return -1/(self.args[0]**2*sqrt(1 - 1/self.args[0]**2)) + else: + raise ArgumentIndexError(self, argindex) + + def inverse(self, argindex=1): + """ + Returns the inverse of this function. + """ + return csc + + @staticmethod + @cacheit + def taylor_term(n, x, *previous_terms): + if n == 0: + return pi/2 - S.ImaginaryUnit*log(2) + S.ImaginaryUnit*log(x) + elif n < 0 or n % 2 == 1: + return S.Zero + else: + x = sympify(x) + if len(previous_terms) > 2 and n > 2: + p = previous_terms[-2] + return p * ((n - 1)*(n-2)) * x**2/(4 * (n//2)**2) + else: + k = n // 2 + R = RisingFactorial(S.Half, k) * n + F = factorial(k) * n // 2 * n // 2 + return S.ImaginaryUnit * R / F * x**n / 4 + + def _eval_as_leading_term(self, x, logx=None, cdir=0): # acsc + arg = self.args[0] + x0 = arg.subs(x, 0).cancel() + # Handling branch points + if x0 in (-S.One, S.One, S.Zero): + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir).expand() + if x0 is S.ComplexInfinity: + return (1/arg).as_leading_term(x) + # Handling points lying on branch cuts (-1, 1) + if x0.is_real and (1 - x0**2).is_positive: + ndir = arg.dir(x, cdir if cdir else 1) + if im(ndir).is_negative: + if x0.is_positive: + return pi - self.func(x0) + elif im(ndir).is_positive: + if x0.is_negative: + return -pi - self.func(x0) + else: + return self.rewrite(log)._eval_as_leading_term(x, logx=logx, cdir=cdir).expand() + return self.func(x0) + + def _eval_nseries(self, x, n, logx, cdir=0): # acsc + from sympy.series.order import O + arg0 = self.args[0].subs(x, 0) + # Handling branch points + if arg0 is S.One: + t = Dummy('t', positive=True) + ser = acsc(S.One + t**2).rewrite(log).nseries(t, 0, 2*n) + arg1 = S.NegativeOne + self.args[0] + f = arg1.as_leading_term(x) + g = (arg1 - f)/ f + res1 = sqrt(S.One + g)._eval_nseries(x, n=n, logx=logx) + res = (res1.removeO()*sqrt(f)).expand() + return ser.removeO().subs(t, res).expand().powsimp() + O(x**n, x) + + if arg0 is S.NegativeOne: + t = Dummy('t', positive=True) + ser = acsc(S.NegativeOne - t**2).rewrite(log).nseries(t, 0, 2*n) + arg1 = S.NegativeOne - self.args[0] + f = arg1.as_leading_term(x) + g = (arg1 - f)/ f + res1 = sqrt(S.One + g)._eval_nseries(x, n=n, logx=logx) + res = (res1.removeO()*sqrt(f)).expand() + return ser.removeO().subs(t, res).expand().powsimp() + O(x**n, x) + + res = Function._eval_nseries(self, x, n=n, logx=logx) + if arg0 is S.ComplexInfinity: + return res + # Handling points lying on branch cuts (-1, 1) + if arg0.is_real and (1 - arg0**2).is_positive: + ndir = self.args[0].dir(x, cdir if cdir else 1) + if im(ndir).is_negative: + if arg0.is_positive: + return pi - res + elif im(ndir).is_positive: + if arg0.is_negative: + return -pi - res + else: + return self.rewrite(log)._eval_nseries(x, n, logx=logx, cdir=cdir) + return res + + def _eval_rewrite_as_log(self, arg, **kwargs): + return -S.ImaginaryUnit*log(S.ImaginaryUnit/arg + sqrt(1 - 1/arg**2)) + + _eval_rewrite_as_tractable = _eval_rewrite_as_log + + def _eval_rewrite_as_asin(self, arg, **kwargs): + return asin(1/arg) + + def _eval_rewrite_as_acos(self, arg, **kwargs): + return pi/2 - acos(1/arg) + + def _eval_rewrite_as_atan(self, x, **kwargs): + return sqrt(x**2)/x*(pi/2 - atan(sqrt(x**2 - 1))) + + def _eval_rewrite_as_acot(self, arg, **kwargs): + return sqrt(arg**2)/arg*(pi/2 - acot(1/sqrt(arg**2 - 1))) + + def _eval_rewrite_as_asec(self, arg, **kwargs): + return pi/2 - asec(arg) + + +class atan2(InverseTrigonometricFunction): + r""" + The function ``atan2(y, x)`` computes `\operatorname{atan}(y/x)` taking + two arguments `y` and `x`. Signs of both `y` and `x` are considered to + determine the appropriate quadrant of `\operatorname{atan}(y/x)`. + The range is `(-\pi, \pi]`. The complete definition reads as follows: + + .. math:: + + \operatorname{atan2}(y, x) = + \begin{cases} + \arctan\left(\frac y x\right) & \qquad x > 0 \\ + \arctan\left(\frac y x\right) + \pi& \qquad y \ge 0, x < 0 \\ + \arctan\left(\frac y x\right) - \pi& \qquad y < 0, x < 0 \\ + +\frac{\pi}{2} & \qquad y > 0, x = 0 \\ + -\frac{\pi}{2} & \qquad y < 0, x = 0 \\ + \text{undefined} & \qquad y = 0, x = 0 + \end{cases} + + Attention: Note the role reversal of both arguments. The `y`-coordinate + is the first argument and the `x`-coordinate the second. + + If either `x` or `y` is complex: + + .. math:: + + \operatorname{atan2}(y, x) = + -i\log\left(\frac{x + iy}{\sqrt{x^2 + y^2}}\right) + + Examples + ======== + + Going counter-clock wise around the origin we find the + following angles: + + >>> from sympy import atan2 + >>> atan2(0, 1) + 0 + >>> atan2(1, 1) + pi/4 + >>> atan2(1, 0) + pi/2 + >>> atan2(1, -1) + 3*pi/4 + >>> atan2(0, -1) + pi + >>> atan2(-1, -1) + -3*pi/4 + >>> atan2(-1, 0) + -pi/2 + >>> atan2(-1, 1) + -pi/4 + + which are all correct. Compare this to the results of the ordinary + `\operatorname{atan}` function for the point `(x, y) = (-1, 1)` + + >>> from sympy import atan, S + >>> atan(S(1)/-1) + -pi/4 + >>> atan2(1, -1) + 3*pi/4 + + where only the `\operatorname{atan2}` function reurns what we expect. + We can differentiate the function with respect to both arguments: + + >>> from sympy import diff + >>> from sympy.abc import x, y + >>> diff(atan2(y, x), x) + -y/(x**2 + y**2) + + >>> diff(atan2(y, x), y) + x/(x**2 + y**2) + + We can express the `\operatorname{atan2}` function in terms of + complex logarithms: + + >>> from sympy import log + >>> atan2(y, x).rewrite(log) + -I*log((x + I*y)/sqrt(x**2 + y**2)) + + and in terms of `\operatorname(atan)`: + + >>> from sympy import atan + >>> atan2(y, x).rewrite(atan) + Piecewise((2*atan(y/(x + sqrt(x**2 + y**2))), Ne(y, 0)), (pi, re(x) < 0), (0, Ne(x, 0)), (nan, True)) + + but note that this form is undefined on the negative real axis. + + See Also + ======== + + sin, csc, cos, sec, tan, cot + asin, acsc, acos, asec, atan, acot + + References + ========== + + .. [1] https://en.wikipedia.org/wiki/Inverse_trigonometric_functions + .. [2] https://en.wikipedia.org/wiki/Atan2 + .. [3] https://functions.wolfram.com/ElementaryFunctions/ArcTan2 + + """ + + @classmethod + def eval(cls, y, x): + from sympy.functions.special.delta_functions import Heaviside + if x is S.NegativeInfinity: + if y.is_zero: + # Special case y = 0 because we define Heaviside(0) = 1/2 + return pi + return 2*pi*(Heaviside(re(y))) - pi + elif x is S.Infinity: + return S.Zero + elif x.is_imaginary and y.is_imaginary and x.is_number and y.is_number: + x = im(x) + y = im(y) + + if x.is_extended_real and y.is_extended_real: + if x.is_positive: + return atan(y/x) + elif x.is_negative: + if y.is_negative: + return atan(y/x) - pi + elif y.is_nonnegative: + return atan(y/x) + pi + elif x.is_zero: + if y.is_positive: + return pi/2 + elif y.is_negative: + return -pi/2 + elif y.is_zero: + return S.NaN + if y.is_zero: + if x.is_extended_nonzero: + return pi*(S.One - Heaviside(x)) + if x.is_number: + return Piecewise((pi, re(x) < 0), + (0, Ne(x, 0)), + (S.NaN, True)) + if x.is_number and y.is_number: + return -S.ImaginaryUnit*log( + (x + S.ImaginaryUnit*y)/sqrt(x**2 + y**2)) + + def _eval_rewrite_as_log(self, y, x, **kwargs): + return -S.ImaginaryUnit*log((x + S.ImaginaryUnit*y)/sqrt(x**2 + y**2)) + + def _eval_rewrite_as_atan(self, y, x, **kwargs): + return Piecewise((2*atan(y/(x + sqrt(x**2 + y**2))), Ne(y, 0)), + (pi, re(x) < 0), + (0, Ne(x, 0)), + (S.NaN, True)) + + def _eval_rewrite_as_arg(self, y, x, **kwargs): + if x.is_extended_real and y.is_extended_real: + return arg_f(x + y*S.ImaginaryUnit) + n = x + S.ImaginaryUnit*y + d = x**2 + y**2 + return arg_f(n/sqrt(d)) - S.ImaginaryUnit*log(abs(n)/sqrt(abs(d))) + + def _eval_is_extended_real(self): + return self.args[0].is_extended_real and self.args[1].is_extended_real + + def _eval_conjugate(self): + return self.func(self.args[0].conjugate(), self.args[1].conjugate()) + + def fdiff(self, argindex): + y, x = self.args + if argindex == 1: + # Diff wrt y + return x/(x**2 + y**2) + elif argindex == 2: + # Diff wrt x + return -y/(x**2 + y**2) + else: + raise ArgumentIndexError(self, argindex) + + def _eval_evalf(self, prec): + y, x = self.args + if x.is_extended_real and y.is_extended_real: + return super()._eval_evalf(prec)