In [1]:

    

using Plots

# this is a custom type that acts just like a Function
immutable F
   f::Function
end


    

In [2]:

    

# We add this type as a Plots "type recipe".  This hook is called after initial Plots
# preprocessing (replacing aliases and other simple steps) but before the core
# data processing.  The return value is a tuple of the new arguments that will
# be passed along.
@recipe p(::Type{F}, f::F) = f.f

# In this example, a call to `plot(f, a, b, xlim=(-5,5))` is equivalent
# to the call `plot(f.f, a, b, xlim(-5,5))`
plot(F(sin), 0, 4π)


    

    




[Plots.jl] Initializing backend: pyplot






    
Out[2]:

In [3]:

    

# we can add another definition for vectors of our custom type
@recipe p{A<:AbstractVector{F}}(::Type{A}, fs::A) = [f.f for f in fs]

v = [F(sin), F(cos)]
plot(v, 0, 4π)


    

    
Out[3]:

In [4]:

    

# 2D parametric
plot(F(sin), F(cos), 0, 4π)


    

    
Out[4]:

In [5]:

    

# 3D parametric
scatter3d(F(_->_+randn()), F(sin), F(cos), 0, 4π)


    

    
Out[5]:

In [6]:

    

# contours/surfaces
x = linspace(-5, 5, 20)
y = x
f = F((x,y) -> x^2 + y^2)
contour(x, y, f)


    

    
Out[6]:

In [7]:

    

surface(x, y, f)


    

    
Out[7]:

In [ ]: