The integrals module in SymPy implements methods to calculate definite and indefinite integrals of expressions.
Principal method in this module is integrate()
 integrate(f, x) returns the indefinite integral \(\int f\,dx\)
 integrate(f, (x, a, b)) returns the definite integral \(\int_{a}^{b} f\,dx\)
SymPy can integrate a vast array of functions. It can integrate polynomial functions:
>>> from sympy import *
>>> init_printing(use_unicode=False, wrap_line=False, no_global=True)
>>> x = Symbol('x')
>>> integrate(x**2 + x + 1, x)
3 2
x x
 +  + x
3 2
Rational functions:
>>> integrate(x/(x**2+2*x+1), x)
1
log(x + 1) + 
x + 1
Exponentialpolynomial functions. These multiplicative combinations of polynomials and the functions exp, cos and sin can be integrated by hand using repeated integration by parts, which is an extremely tedious process. Happily, SymPy will deal with these integrals.
>>> integrate(x**2 * exp(x) * cos(x), x)
2 x 2 x x x
x *e *sin(x) x *e *cos(x) x e *sin(x) e *cos(x)
 +   x*e *sin(x) +   
2 2 2 2
even a few nonelementary integrals (in particular, some integrals involving the error function) can be evaluated:
>>> integrate(exp(x**2)*erf(x), x)
____ 2
\/ pi *erf (x)

4
SymPy has special support for definite integrals, and integral transforms.
Compute the Mellin transform \(F(s)\) of \(f(x)\),
The Mellin transform is related via change of variables to the Fourier transform, and also to the (bilateral) Laplace transform.
This function returns (F, (a, b), cond) where F is the Mellin transform of f, (a, b) is the fundamental strip (as above), and cond are auxiliary convergence conditions.
If the integral cannot be computed in closed form, this function returns an unevaluated MellinTransform object.
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit(). If noconds=False, then only \(F\) will be returned (i.e. not cond, and also not the strip (a, b)).
>>> from sympy.integrals.transforms import mellin_transform
>>> from sympy import exp
>>> from sympy.abc import x, s
>>> mellin_transform(exp(x), x, s)
(gamma(s), (0, oo), True)
Compute the inverse Mellin transform of \(F(s)\) over the fundamental strip given by strip=(a, b).
This can be defined as
for any \(c\) in the fundamental strip. Under certain regularity conditions on \(F\) and/or \(f\), this recovers \(f\) from its Mellin transform \(F\) (and vice versa), for positive real \(x\).
One of \(a\) or \(b\) may be passed as None; a suitable \(c\) will be inferred.
If the integral cannot be computed in closed form, this function returns an unevaluated InverseMellinTransform object.
Note that this function will assume x to be positive and real, regardless of the sympy assumptions!
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit().
>>> from sympy.integrals.transforms import inverse_mellin_transform
>>> from sympy import oo, gamma
>>> from sympy.abc import x, s
>>> inverse_mellin_transform(gamma(s), s, x, (0, oo))
exp(x)
The fundamental strip matters:
>>> f = 1/(s**2  1)
>>> inverse_mellin_transform(f, s, x, (oo, 1))
(x/2  1/(2*x))*Heaviside(x  1)
>>> inverse_mellin_transform(f, s, x, (1, 1))
x*Heaviside(x + 1)/2  Heaviside(x  1)/(2*x)
>>> inverse_mellin_transform(f, s, x, (1, oo))
(x/2 + 1/(2*x))*Heaviside(x + 1)
Compute the Laplace Transform \(F(s)\) of \(f(t)\),
For all “sensible” functions, this converges absolutely in a half plane \(a < \operatorname{Re}(s)\).
This function returns (F, a, cond) where F is the Laplace transform of f, \(\operatorname{Re}(s) > a\) is the halfplane of convergence, and cond are auxiliary convergence conditions.
If the integral cannot be computed in closed form, this function returns an unevaluated LaplaceTransform object.
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit(). If noconds=True, only \(F\) will be returned (i.e. not cond, and also not the plane a).
>>> from sympy.integrals import laplace_transform
>>> from sympy.abc import t, s, a
>>> laplace_transform(t**a, t, s)
(s**(a)*gamma(a + 1)/s, 0, re(a) < 1)
Compute the inverse Laplace transform of \(F(s)\), defined as
for \(c\) so large that \(F(s)\) has no singularites in the halfplane \(\operatorname{Re}(s) > c\epsilon\).
The plane can be specified by argument plane, but will be inferred if passed as None.
Under certain regularity conditions, this recovers \(f(t)\) from its Laplace Transform \(F(s)\), for nonnegative \(t\), and vice versa.
If the integral cannot be computed in closed form, this function returns an unevaluated InverseLaplaceTransform object.
Note that this function will always assume \(t\) to be real, regardless of the sympy assumption on \(t\).
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit().
>>> from sympy.integrals.transforms import inverse_laplace_transform
>>> from sympy import exp, Symbol
>>> from sympy.abc import s, t
>>> a = Symbol('a', positive=True)
>>> inverse_laplace_transform(exp(a*s)/s, s, t)
Heaviside(a + t)
Compute the unitary, ordinaryfrequency Fourier transform of \(f\), defined as
If the transform cannot be computed in closed form, this function returns an unevaluated FourierTransform object.
For other Fourier transform conventions, see the function sympy.integrals.transforms._fourier_transform().
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import fourier_transform, exp
>>> from sympy.abc import x, k
>>> fourier_transform(exp(x**2), x, k)
sqrt(pi)*exp(pi**2*k**2)
>>> fourier_transform(exp(x**2), x, k, noconds=False)
(sqrt(pi)*exp(pi**2*k**2), True)
Compute the unitary, ordinaryfrequency inverse Fourier transform of \(F\), defined as
If the transform cannot be computed in closed form, this function returns an unevaluated InverseFourierTransform object.
For other Fourier transform conventions, see the function sympy.integrals.transforms._fourier_transform().
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import inverse_fourier_transform, exp, sqrt, pi
>>> from sympy.abc import x, k
>>> inverse_fourier_transform(sqrt(pi)*exp((pi*k)**2), k, x)
exp(x**2)
>>> inverse_fourier_transform(sqrt(pi)*exp((pi*k)**2), k, x, noconds=False)
(exp(x**2), True)
Compute the unitary, ordinaryfrequency sine transform of \(f\), defined as
If the transform cannot be computed in closed form, this function returns an unevaluated SineTransform object.
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import sine_transform, exp
>>> from sympy.abc import x, k, a
>>> sine_transform(x*exp(a*x**2), x, k)
sqrt(2)*k*exp(k**2/(4*a))/(4*a**(3/2))
>>> sine_transform(x**(a), x, k)
2**(a + 1/2)*k**(a  1)*gamma(a/2 + 1)/gamma(a/2 + 1/2)
Compute the unitary, ordinaryfrequency inverse sine transform of \(F\), defined as
If the transform cannot be computed in closed form, this function returns an unevaluated InverseSineTransform object.
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import inverse_sine_transform, exp, sqrt, gamma, pi
>>> from sympy.abc import x, k, a
>>> inverse_sine_transform(2**((12*a)/2)*k**(a  1)*
... gamma(a/2 + 1)/gamma((a+1)/2), k, x)
x**(a)
>>> inverse_sine_transform(sqrt(2)*k*exp(k**2/(4*a))/(4*sqrt(a)**3), k, x)
x*exp(a*x**2)
Compute the unitary, ordinaryfrequency cosine transform of \(f\), defined as
If the transform cannot be computed in closed form, this function returns an unevaluated CosineTransform object.
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import cosine_transform, exp, sqrt, cos
>>> from sympy.abc import x, k, a
>>> cosine_transform(exp(a*x), x, k)
sqrt(2)*a/(sqrt(pi)*(a**2 + k**2))
>>> cosine_transform(exp(a*sqrt(x))*cos(a*sqrt(x)), x, k)
a*exp(a**2/(2*k))/(2*k**(3/2))
Compute the unitary, ordinaryfrequency inverse cosine transform of \(F\), defined as
If the transform cannot be computed in closed form, this function returns an unevaluated InverseCosineTransform object.
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import inverse_cosine_transform, exp, sqrt, pi
>>> from sympy.abc import x, k, a
>>> inverse_cosine_transform(sqrt(2)*a/(sqrt(pi)*(a**2 + k**2)), k, x)
exp(a*x)
>>> inverse_cosine_transform(1/sqrt(k), k, x)
1/sqrt(x)
Compute the Hankel transform of \(f\), defined as
If the transform cannot be computed in closed form, this function returns an unevaluated HankelTransform object.
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import hankel_transform, inverse_hankel_transform
>>> from sympy import gamma, exp, sinh, cosh
>>> from sympy.abc import r, k, m, nu, a
>>> ht = hankel_transform(1/r**m, r, k, nu)
>>> ht
2*2**(m)*k**(m  2)*gamma(m/2 + nu/2 + 1)/gamma(m/2 + nu/2)
>>> inverse_hankel_transform(ht, k, r, nu)
r**(m)
>>> ht = hankel_transform(exp(a*r), r, k, 0)
>>> ht
a/(k**3*(a**2/k**2 + 1)**(3/2))
>>> inverse_hankel_transform(ht, k, r, 0)
exp(a*r)
Compute the inverse Hankel transform of \(F\) defined as
If the transform cannot be computed in closed form, this function returns an unevaluated InverseHankelTransform object.
For a description of possible hints, refer to the docstring of sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import hankel_transform, inverse_hankel_transform, gamma
>>> from sympy import gamma, exp, sinh, cosh
>>> from sympy.abc import r, k, m, nu, a
>>> ht = hankel_transform(1/r**m, r, k, nu)
>>> ht
2*2**(m)*k**(m  2)*gamma(m/2 + nu/2 + 1)/gamma(m/2 + nu/2)
>>> inverse_hankel_transform(ht, k, r, nu)
r**(m)
>>> ht = hankel_transform(exp(a*r), r, k, 0)
>>> ht
a/(k**3*(a**2/k**2 + 1)**(3/2))
>>> inverse_hankel_transform(ht, k, r, 0)
exp(a*r)
There is a general method for calculating antiderivatives of elementary functions, called the Risch algorithm. The Risch algorithm is a decision procedure that can determine whether an elementary solution exists, and in that case calculate it. It can be extended to handle many nonelementary functions in addition to the elementary ones.
SymPy currently uses a simplified version of the Risch algorithm, called the RischNorman algorithm. This algorithm is much faster, but may fail to find an antiderivative, although it is still very powerful. SymPy also uses pattern matching and heuristics to speed up evaluation of some types of integrals, e.g. polynomials.
For nonelementary definite integrals, SymPy uses socalled Meijer Gfunctions. Details are described here:
Compute definite or indefinite integral of one or more variables using RischNorman algorithm and table lookup. This procedure is able to handle elementary algebraic and transcendental functions and also a huge class of special functions, including Airy, Bessel, Whittaker and Lambert.
var can be:
a symbol – indefinite integration
given with \(a\) replacing \(symbol\)
a tuple (symbol, a, b) – definite integration
Several variables can be specified, in which case the result is multiple integration. (If var is omitted and the integrand is univariate, the indefinite integral in that variable will be performed.)
Indefinite integrals are returned without terms that are independent of the integration variables. (see examples)
Definite improper integrals often entail delicate convergence conditions. Pass conds=’piecewise’, ‘separate’ or ‘none’ to have these returned, respectively, as a Piecewise function, as a separate result (i.e. result will be a tuple), or not at all (default is ‘piecewise’).
Strategy
SymPy uses various approaches to definite integration. One method is to find an antiderivative for the integrand, and then use the fundamental theorem of calculus. Various functions are implemented to integrate polynomial, rational and trigonometric functions, and integrands containing DiracDelta terms.
SymPy also implements the part of the Risch algorithm, which is a decision procedure for integrating elementary functions, i.e., the algorithm can either find an elementary antiderivative, or prove that one does not exist. There is also a (very successful, albeit somewhat slow) general implementation of the heuristic Risch algorithm. This algorithm will eventually be phased out as more of the full Risch algorithm is implemented. See the docstring of Integral._eval_integral() for more details on computing the antiderivative using algebraic methods.
The option risch=True can be used to use only the (full) Risch algorithm. This is useful if you want to know if an elementary function has an elementary antiderivative. If the indefinite Integral returned by this function is an instance of NonElementaryIntegral, that means that the Risch algorithm has proven that integral to be nonelementary. Note that by default, additional methods (such as the Meijer G method outlined below) are tried on these integrals, as they may be expressible in terms of special functions, so if you only care about elementary answers, use risch=True. Also note that an unevaluated Integral returned by this function is not necessarily a NonElementaryIntegral, even with risch=True, as it may just be an indication that the particular part of the Risch algorithm needed to integrate that function is not yet implemented.
Another family of strategies comes from rewriting the integrand in terms of socalled Meijer Gfunctions. Indefinite integrals of a single Gfunction can always be computed, and the definite integral of a product of two Gfunctions can be computed from zero to infinity. Various strategies are implemented to rewrite integrands as Gfunctions, and use this information to compute integrals (see the meijerint module).
The option manual=True can be used to use only an algorithm that tries to mimic integration by hand. This algorithm does not handle as many integrands as the other algorithms implemented but may return results in a more familiar form. The manualintegrate module has functions that return the steps used (see the module docstring for more information).
In general, the algebraic methods work best for computing antiderivatives of (possibly complicated) combinations of elementary functions. The Gfunction methods work best for computing definite integrals from zero to infinity of moderately complicated combinations of special functions, or indefinite integrals of very simple combinations of special functions.
The strategy employed by the integration code is as follows:
The option meijerg=True, False, None can be used to, respectively: always use Gfunction methods and no others, never use Gfunction methods, or use all available methods (in order as described above). It defaults to None.
See also
Integral, Integral.doit
Examples
>>> from sympy import integrate, log, exp, oo
>>> from sympy.abc import a, x, y
>>> integrate(x*y, x)
x**2*y/2
>>> integrate(log(x), x)
x*log(x)  x
>>> integrate(log(x), (x, 1, a))
a*log(a)  a + 1
>>> integrate(x)
x**2/2
Terms that are independent of x are dropped by indefinite integration:
>>> from sympy import sqrt
>>> integrate(sqrt(1 + x), (x, 0, x))
2*(x + 1)**(3/2)/3  2/3
>>> integrate(sqrt(1 + x), x)
2*(x + 1)**(3/2)/3
>>> integrate(x*y)
Traceback (most recent call last):
...
ValueError: specify integration variables to integrate x*y
Traceback (most recent call last):
...
ValueError: specify integration variables to integrate x*y
Note that integrate(x) syntax is meant only for convenience in interactive sessions and should be avoided in library code.
>>> integrate(x**a*exp(x), (x, 0, oo)) # same as conds='piecewise'
Piecewise((gamma(a + 1), re(a) < 1),
(Integral(x**a*exp(x), (x, 0, oo)), True))
>>> integrate(x**a*exp(x), (x, 0, oo), conds='none')
gamma(a + 1)
>>> integrate(x**a*exp(x), (x, 0, oo), conds='separate')
(gamma(a + 1), re(a) < 1)
Compute the line integral.
See also
integrate, Integral
Examples
>>> from sympy import Curve, line_integrate, E, ln
>>> from sympy.abc import x, y, t
>>> C = Curve([E**t + 1, E**t  1], (t, 0, ln(2)))
>>> line_integrate(x + y, C, [x, y])
3*sqrt(2)
The idea for integration is the following:
If we are dealing with a DiracDelta expression, i.e. DiracDelta(g(x)), we try to simplify it.
If we could simplify it, then we integrate the resulting expression. We already know we can integrate a simplified expression, because only simple DiracDelta expressions are involved.
If we couldn’t simplify it, there are two cases:
If the node is a multiplication node having a DiracDelta term:
First we expand it.
If the expansion did work, the we try to integrate the expansion.
If not, we try to extract a simple DiracDelta term, then we have two cases:
See also
sympy.functions.special.delta_functions.DiracDelta, sympy.integrals.integrals.Integral
Examples
>>> from sympy.abc import x, y, z
>>> from sympy.integrals.deltafunctions import deltaintegrate
>>> from sympy import sin, cos, DiracDelta, Heaviside
>>> deltaintegrate(x*sin(x)*cos(x)*DiracDelta(x  1), x)
sin(1)*cos(1)*Heaviside(x  1)
>>> deltaintegrate(y**2*DiracDelta(x  z)*DiracDelta(y  z), y)
z**2*DiracDelta(x  z)*Heaviside(y  z)
Performs indefinite integration of rational functions.
Given a field \(K\) and a rational function \(f = p/q\), where \(p\) and \(q\) are polynomials in \(K[x]\), returns a function \(g\) such that \(f = g'\).
>>> from sympy.integrals.rationaltools import ratint
>>> from sympy.abc import x
>>> ratint(36/(x**5  2*x**4  2*x**3 + 4*x**2 + x  2), x)
(12*x + 6)/(x**2  1) + 4*log(x  2)  4*log(x + 1)
See also
sympy.integrals.integrals.Integral.doit, ratint_logpart, ratint_ratpart
References
[Bro05]  M. Bronstein, Symbolic Integration I: Transcendental Functions, Second Edition, SpringerVerlag, 2005, pp. 3570 
Compute indefinite integral using heuristic Risch algorithm.
This is a heuristic approach to indefinite integration in finite terms using the extended heuristic (parallel) Risch algorithm, based on Manuel Bronstein’s “Poor Man’s Integrator”.
The algorithm supports various classes of functions including transcendental elementary or special functions like Airy, Bessel, Whittaker and Lambert.
Note that this algorithm is not a decision procedure. If it isn’t able to compute the antiderivative for a given function, then this is not a proof that such a functions does not exist. One should use recursive Risch algorithm in such case. It’s an open question if this algorithm can be made a full decision procedure.
This is an internal integrator procedure. You should use toplevel ‘integrate’ function in most cases, as this procedure needs some preprocessing steps and otherwise may fail.
See also
sympy.integrals.integrals.Integral.doit, sympy.integrals.integrals.Integral, components
Examples
>>> from sympy import tan
>>> from sympy.integrals.heurisch import heurisch
>>> from sympy.abc import x, y
>>> heurisch(y*tan(x), x)
y*log(tan(x)**2 + 1)/2
See Manuel Bronstein’s “Poor Man’s Integrator”:
[1] http://wwwsop.inria.fr/cafe/Manuel.Bronstein/pmint/index.html
For more information on the implemented algorithm refer to:
Specification
heurisch(f, x, rewrite=False, hints=None)
 where
f : expression x : symbol
rewrite > force rewrite ‘f’ in terms of ‘tan’ and ‘tanh’ hints > a list of functions that may appear in antiderivate
 hints = None –> no suggestions at all
 hints = [ ] –> try to figure out
 hints = [f1, ..., fn] –> we know better
A wrapper around the heurisch integration algorithm.
This method takes the result from heurisch and checks for poles in the denominator. For each of these poles, the integral is reevaluated, and the final integration result is given in terms of a Piecewise.
See also
heurisch
Examples
>>> from sympy.core import symbols
>>> from sympy.functions import cos
>>> from sympy.integrals.heurisch import heurisch, heurisch_wrapper
>>> n, x = symbols('n x')
>>> heurisch(cos(n*x), x)
sin(n*x)/n
>>> heurisch_wrapper(cos(n*x), x)
Piecewise((x, n == 0), (sin(n*x)/n, True))
Integrate f = Mul(trig) over x
>>> from sympy import Symbol, sin, cos, tan, sec, csc, cot >>> from sympy.integrals.trigonometry import trigintegrate >>> from sympy.abc import x>>> trigintegrate(sin(x)*cos(x), x) sin(x)**2/2>>> trigintegrate(sin(x)**2, x) x/2  sin(x)*cos(x)/2>>> trigintegrate(tan(x)*sec(x), x) 1/cos(x)>>> trigintegrate(sin(x)*tan(x), x) log(sin(x)  1)/2 + log(sin(x) + 1)/2  sin(x)http://en.wikibooks.org/wiki/Calculus/Integration_techniques
See also
sympy.integrals.integrals.Integral.doit, sympy.integrals.integrals.Integral
Compute indefinite integral of a single variable using an algorithm that resembles what a student would do by hand.
Unlike integrate, var can only be a single symbol.
See also
sympy.integrals.integrals.integrate, sympy.integrals.integrals.Integral.doit, sympy.integrals.integrals.Integral
Examples
>>> from sympy import sin, cos, tan, exp, log, integrate
>>> from sympy.integrals.manualintegrate import manualintegrate
>>> from sympy.abc import x
>>> manualintegrate(1 / x, x)
log(x)
>>> integrate(1/x)
log(x)
>>> manualintegrate(log(x), x)
x*log(x)  x
>>> integrate(log(x))
x*log(x)  x
>>> manualintegrate(exp(x) / (1 + exp(2 * x)), x)
atan(exp(x))
>>> integrate(exp(x) / (1 + exp(2 * x)))
RootSum(4*_z**2 + 1, Lambda(_i, _i*log(2*_i + exp(x))))
>>> manualintegrate(cos(x)**4 * sin(x), x)
cos(x)**5/5
>>> integrate(cos(x)**4 * sin(x), x)
cos(x)**5/5
>>> manualintegrate(cos(x)**4 * sin(x)**3, x)
cos(x)**7/7  cos(x)**5/5
>>> integrate(cos(x)**4 * sin(x)**3, x)
cos(x)**7/7  cos(x)**5/5
>>> manualintegrate(tan(x), x)
log(cos(x))
>>> integrate(tan(x), x)
log(sin(x)**2  1)/2
Returns the steps needed to compute an integral.
This function attempts to mirror what a student would do by hand as closely as possible.
SymPy Gamma uses this to provide a stepbystep explanation of an integral. The code it uses to format the results of this function can be found at https://github.com/sympy/sympy_gamma/blob/master/app/logic/intsteps.py.
Returns :  rule : namedtuple


Examples
>>> from sympy import exp, sin, cos
>>> from sympy.integrals.manualintegrate import integral_steps
>>> from sympy.abc import x
>>> print(repr(integral_steps(exp(x) / (1 + exp(2 * x)), x)))
URule(u_var=_u, u_func=exp(x), constant=1,
substep=ArctanRule(context=1/(_u**2 + 1), symbol=_u),
context=exp(x)/(exp(2*x) + 1), symbol=x)
>>> print(repr(integral_steps(sin(x), x)))
TrigRule(func='sin', arg=x, context=sin(x), symbol=x)
>>> print(repr(integral_steps((x**2 + 3)**2 , x)))
RewriteRule(rewritten=x**4 + 6*x**2 + 9,
substep=AddRule(substeps=[PowerRule(base=x, exp=4, context=x**4, symbol=x),
ConstantTimesRule(constant=6, other=x**2,
substep=PowerRule(base=x, exp=2, context=x**2, symbol=x),
context=6*x**2, symbol=x),
ConstantRule(constant=9, context=9, symbol=x)],
context=x**4 + 6*x**2 + 9, symbol=x), context=(x**2 + 3)**2, symbol=x)
The class \(Integral\) represents an unevaluated integral and has some methods that help in the integration of an expression.
Represents unevaluated integral.
Returns whether all the free symbols in the integral are commutative.
Approximates the definite integral by a sum.
method ... one of: left, right, midpoint, trapezoid
These are all basically the rectangle method [1], the only difference is where the function value is taken in each interval to define the rectangle.
[1] http://en.wikipedia.org/wiki/Rectangle_method
See also
Examples
>>> from sympy import sin, sqrt
>>> from sympy.abc import x
>>> from sympy.integrals import Integral
>>> e = Integral(sin(x), (x, 3, 7))
>>> e
Integral(sin(x), (x, 3, 7))
For demonstration purposes, this interval will only be split into 2 regions, bounded by [3, 5] and [5, 7].
The lefthand rule uses function evaluations at the left of each interval:
>>> e.as_sum(2, 'left')
2*sin(5) + 2*sin(3)
The midpoint rule uses evaluations at the center of each interval:
>>> e.as_sum(2, 'midpoint')
2*sin(4) + 2*sin(6)
The righthand rule uses function evaluations at the right of each interval:
>>> e.as_sum(2, 'right')
2*sin(5) + 2*sin(7)
The trapezoid rule uses function evaluations on both sides of the intervals. This is equivalent to taking the average of the left and right hand rule results:
>>> e.as_sum(2, 'trapezoid')
2*sin(5) + sin(3) + sin(7)
>>> (e.as_sum(2, 'left') + e.as_sum(2, 'right'))/2 == _
True
All but the trapexoid method may be used when dealing with a function with a discontinuity. Here, the discontinuity at x = 0 can be avoided by using the midpoint or righthand method:
>>> e = Integral(1/sqrt(x), (x, 0, 1))
>>> e.as_sum(5).n(4)
1.730
>>> e.as_sum(10).n(4)
1.809
>>> e.doit().n(4) # the actual value is 2
2.000
The left or trapezoid method will encounter the discontinuity and return oo:
>>> e.as_sum(5, 'left')
oo
>>> e.as_sum(5, 'trapezoid')
oo
Perform the integration using any hints given.
See also
sympy.integrals.trigonometry.trigintegrate, sympy.integrals.risch.heurisch, sympy.integrals.rationaltools.ratint
Examples
>>> from sympy import Integral
>>> from sympy.abc import x, i
>>> Integral(x**i, (i, 1, 3)).doit()
Piecewise((2, log(x) == 0), (x**3/log(x)  x/log(x), True))
This method returns the symbols that will exist when the integral is evaluated. This is useful if one is trying to determine whether an integral depends on a certain symbol or not.
See also
function, limits, variables
Examples
>>> from sympy import Integral
>>> from sympy.abc import x, y
>>> Integral(x, (x, y, 1)).free_symbols
set([y])
Performs a change of variables from \(x\) to \(u\) using the relationship given by \(x\) and \(u\) which will define the transformations \(f\) and \(F\) (which are inverses of each other) as follows:
Once f and F have been identified, the transformation is made as follows:
where \(F(x)\) is the inverse of \(f(x)\) and the limits and integrand have been corrected so as to retain the same value after integration.
See also
Notes
The mappings, F(x) or f(u), must lead to a unique integral. Linear or rational linear expression, \(2*x\), \(1/x\) and \(sqrt(x)\), will always work; quadratic expressions like \(x**2  1\) are acceptable as long as the resulting integrand does not depend on the sign of the solutions (see examples).
The integral will be returned unchanged if \(x\) is not a variable of integration.
\(x\) must be (or contain) only one of of the integration variables. If \(u\) has more than one free symbol then it should be sent as a tuple (\(u\), \(uvar\)) where \(uvar\) identifies which variable is replacing the integration variable. XXX can it contain another integration variable?
Examples
>>> from sympy.abc import a, b, c, d, x, u, y
>>> from sympy import Integral, S, cos, sqrt
>>> i = Integral(x*cos(x**2  1), (x, 0, 1))
transform can change the variable of integration
>>> i.transform(x, u)
Integral(u*cos(u**2  1), (u, 0, 1))
transform can perform usubstitution as long as a unique integrand is obtained:
>>> i.transform(x**2  1, u)
Integral(cos(u)/2, (u, 1, 0))
This attempt fails because x = +/sqrt(u + 1) and the sign does not cancel out of the integrand:
>>> Integral(cos(x**2  1), (x, 0, 1)).transform(x**2  1, u)
Traceback (most recent call last):
...
ValueError:
The mapping between F(x) and f(u) did not give a unique integrand.
Traceback (most recent call last):
...
ValueError:
transform can do a substitution. Here, the previous result is transformed back into the original expression using “usubstitution”:
>>> ui = _
>>> _.transform(sqrt(u + 1), x) == i
True
We can accomplish the same with a regular substitution:
>>> ui.transform(u, x**2  1) == i
True
If the \(x\) does not contain a symbol of integration then the integral will be returned unchanged. Integral \(i\) does not have an integration variable \(a\) so no change is made:
>>> i.transform(a, x) == i
True
When \(u\) has more than one free symbol the symbol that is replacing \(x\) must be identified by passing \(u\) as a tuple:
>>> Integral(x, (x, 0, 1)).transform(x, (u + a, u))
Integral(a + u, (u, a, a + 1))
>>> Integral(x, (x, 0, 1)).transform(x, (u + a, a))
Integral(a + u, (a, u, u + 1))
There are still lots of functions that SymPy does not know how to integrate. For bugs related to this module, see https://github.com/sympy/sympy/issues?labels=Integration
SymPy has functions to calculate points and weights for Gaussian quadrature of any order and any precision:
Computes the GaussLegendre quadrature [R291] points and weights.
The GaussLegendre quadrature approximates the integral:
The nodes \(x_i\) of an order \(n\) quadrature rule are the roots of \(P_n\) and the weights \(w_i\) are given by:
Parameters :  n : the order of quadrature n_digits : number of significant digits of the points and weights to return 

Returns :  (x, w) : the x and w are lists of points and weights as Floats.

See also
gauss_laguerre, gauss_gen_laguerre, gauss_hermite, gauss_chebyshev_t, gauss_chebyshev_u, gauss_jacobi
References
[R291]  (1, 2) http://en.wikipedia.org/wiki/Gaussian_quadrature 
[R292]  http://people.sc.fsu.edu/~jburkardt/cpp_src/legendre_rule/legendre_rule.html 
Examples
>>> from sympy.integrals.quadrature import gauss_legendre
>>> x, w = gauss_legendre(3, 5)
>>> x
[0.7746, 0, 0.7746]
>>> w
[0.55556, 0.88889, 0.55556]
>>> x, w = gauss_legendre(4, 5)
>>> x
[0.86114, 0.33998, 0.33998, 0.86114]
>>> w
[0.34786, 0.65215, 0.65215, 0.34786]
Computes the GaussLaguerre quadrature [R293] points and weights.
The GaussLaguerre quadrature approximates the integral:
The nodes \(x_i\) of an order \(n\) quadrature rule are the roots of \(L_n\) and the weights \(w_i\) are given by:
Parameters :  n : the order of quadrature n_digits : number of significant digits of the points and weights to return 

Returns :  (x, w) : the x and w are lists of points and weights as Floats.

See also
gauss_legendre, gauss_gen_laguerre, gauss_hermite, gauss_chebyshev_t, gauss_chebyshev_u, gauss_jacobi
References
[R293]  (1, 2) http://en.wikipedia.org/wiki/Gauss%E2%80%93Laguerre_quadrature 
[R294]  http://people.sc.fsu.edu/~jburkardt/cpp_src/laguerre_rule/laguerre_rule.html 
Examples
>>> from sympy.integrals.quadrature import gauss_laguerre
>>> x, w = gauss_laguerre(3, 5)
>>> x
[0.41577, 2.2943, 6.2899]
>>> w
[0.71109, 0.27852, 0.010389]
>>> x, w = gauss_laguerre(6, 5)
>>> x
[0.22285, 1.1889, 2.9927, 5.7751, 9.8375, 15.983]
>>> w
[0.45896, 0.417, 0.11337, 0.010399, 0.00026102, 8.9855e7]
Computes the GaussHermite quadrature [R295] points and weights.
The GaussHermite quadrature approximates the integral:
The nodes \(x_i\) of an order \(n\) quadrature rule are the roots of \(H_n\) and the weights \(w_i\) are given by:
Parameters :  n : the order of quadrature n_digits : number of significant digits of the points and weights to return 

Returns :  (x, w) : the x and w are lists of points and weights as Floats.

See also
gauss_legendre, gauss_laguerre, gauss_gen_laguerre, gauss_chebyshev_t, gauss_chebyshev_u, gauss_jacobi
References
[R295]  (1, 2) http://en.wikipedia.org/wiki/GaussHermite_Quadrature 
[R296]  http://people.sc.fsu.edu/~jburkardt/cpp_src/hermite_rule/hermite_rule.html 
[R297]  http://people.sc.fsu.edu/~jburkardt/cpp_src/gen_hermite_rule/gen_hermite_rule.html 
Examples
>>> from sympy.integrals.quadrature import gauss_hermite
>>> x, w = gauss_hermite(3, 5)
>>> x
[1.2247, 0, 1.2247]
>>> w
[0.29541, 1.1816, 0.29541]
>>> x, w = gauss_hermite(6, 5)
>>> x
[2.3506, 1.3358, 0.43608, 0.43608, 1.3358, 2.3506]
>>> w
[0.00453, 0.15707, 0.72463, 0.72463, 0.15707, 0.00453]
Computes the generalized GaussLaguerre quadrature [R298] points and weights.
The generalized GaussLaguerre quadrature approximates the integral:
The nodes \(x_i\) of an order \(n\) quadrature rule are the roots of \(L^{\alpha}_n\) and the weights \(w_i\) are given by:
Parameters :  n : the order of quadrature alpha : the exponent of the singularity, \(\alpha > 1\) n_digits : number of significant digits of the points and weights to return 

Returns :  (x, w) : the x and w are lists of points and weights as Floats.

See also
gauss_legendre, gauss_laguerre, gauss_hermite, gauss_chebyshev_t, gauss_chebyshev_u, gauss_jacobi
References
[R298]  (1, 2) http://en.wikipedia.org/wiki/Gauss%E2%80%93Laguerre_quadrature 
[R299]  http://people.sc.fsu.edu/~jburkardt/cpp_src/gen_laguerre_rule/gen_laguerre_rule.html 
Examples
>>> from sympy import S
>>> from sympy.integrals.quadrature import gauss_gen_laguerre
>>> x, w = gauss_gen_laguerre(3, S.Half, 5)
>>> x
[0.19016, 1.7845, 5.5253]
>>> w
[1.4493, 0.31413, 0.00906]
>>> x, w = gauss_gen_laguerre(4, 3*S.Half, 5)
>>> x
[0.97851, 2.9904, 6.3193, 11.712]
>>> w
[0.53087, 0.67721, 0.11895, 0.0023152]
Computes the GaussChebyshev quadrature [R300] points and weights of the first kind.
The GaussChebyshev quadrature of the first kind approximates the integral:
The nodes \(x_i\) of an order \(n\) quadrature rule are the roots of \(T_n\) and the weights \(w_i\) are given by:
Parameters :  n : the order of quadrature n_digits : number of significant digits of the points and weights to return 

Returns :  (x, w) : the x and w are lists of points and weights as Floats.

See also
gauss_legendre, gauss_laguerre, gauss_hermite, gauss_gen_laguerre, gauss_chebyshev_u, gauss_jacobi
References
[R300]  (1, 2) http://en.wikipedia.org/wiki/Chebyshev%E2%80%93Gauss_quadrature 
[R301]  http://people.sc.fsu.edu/~jburkardt/cpp_src/chebyshev1_rule/chebyshev1_rule.html 
Examples
>>> from sympy import S
>>> from sympy.integrals.quadrature import gauss_chebyshev_t
>>> x, w = gauss_chebyshev_t(3, 5)
>>> x
[0.86602, 0, 0.86602]
>>> w
[1.0472, 1.0472, 1.0472]
>>> x, w = gauss_chebyshev_t(6, 5)
>>> x
[0.96593, 0.70711, 0.25882, 0.25882, 0.70711, 0.96593]
>>> w
[0.5236, 0.5236, 0.5236, 0.5236, 0.5236, 0.5236]
Computes the GaussChebyshev quadrature [R302] points and weights of the second kind.
The GaussChebyshev quadrature of the second kind approximates the integral:
The nodes \(x_i\) of an order \(n\) quadrature rule are the roots of \(U_n\) and the weights \(w_i\) are given by:
Parameters :  n : the order of quadrature n_digits : number of significant digits of the points and weights to return 

Returns :  (x, w) : the x and w are lists of points and weights as Floats.

See also
gauss_legendre, gauss_laguerre, gauss_hermite, gauss_gen_laguerre, gauss_chebyshev_t, gauss_jacobi
References
[R302]  (1, 2) http://en.wikipedia.org/wiki/Chebyshev%E2%80%93Gauss_quadrature 
[R303]  http://people.sc.fsu.edu/~jburkardt/cpp_src/chebyshev2_rule/chebyshev2_rule.html 
Examples
>>> from sympy import S
>>> from sympy.integrals.quadrature import gauss_chebyshev_u
>>> x, w = gauss_chebyshev_u(3, 5)
>>> x
[0.70711, 0, 0.70711]
>>> w
[0.3927, 0.7854, 0.3927]
>>> x, w = gauss_chebyshev_u(6, 5)
>>> x
[0.90097, 0.62349, 0.22252, 0.22252, 0.62349, 0.90097]
>>> w
[0.084489, 0.27433, 0.42658, 0.42658, 0.27433, 0.084489]
Computes the GaussJacobi quadrature [R304] points and weights.
The GaussJacobi quadrature of the first kind approximates the integral:
The nodes \(x_i\) of an order \(n\) quadrature rule are the roots of \(P^{(\alpha,\beta)}_n\) and the weights \(w_i\) are given by:
Parameters :  n : the order of quadrature alpha : the first parameter of the Jacobi Polynomial, \(\alpha > 1\) beta : the second parameter of the Jacobi Polynomial, \(\beta > 1\) n_digits : number of significant digits of the points and weights to return 

Returns :  (x, w) : the x and w are lists of points and weights as Floats.

See also
gauss_legendre, gauss_laguerre, gauss_hermite, gauss_gen_laguerre, gauss_chebyshev_t, gauss_chebyshev_u
References
[R304]  (1, 2) http://en.wikipedia.org/wiki/Gauss%E2%80%93Jacobi_quadrature 
[R305]  http://people.sc.fsu.edu/~jburkardt/cpp_src/jacobi_rule/jacobi_rule.html 
[R306]  http://people.sc.fsu.edu/~jburkardt/cpp_src/gegenbauer_rule/gegenbauer_rule.html 
Examples
>>> from sympy import S
>>> from sympy.integrals.quadrature import gauss_jacobi
>>> x, w = gauss_jacobi(3, S.Half, S.Half, 5)
>>> x
[0.90097, 0.22252, 0.62349]
>>> w
[1.7063, 1.0973, 0.33795]
>>> x, w = gauss_jacobi(6, 1, 1, 5)
>>> x
[0.87174, 0.5917, 0.2093, 0.2093, 0.5917, 0.87174]
>>> w
[0.050584, 0.22169, 0.39439, 0.39439, 0.22169, 0.050584]