Remove example notebooks

NEWS.md

@@ -1,5 +1,25 @@
 # IntervalConstraintProgramming.jl
 
+# v0.4
+- `@function f(x) = 4x` defines a function
+- Functions may be used inside constraints
+- Functions may be iterated
+- Functions may be multi-dimensional
+- Local variables may be introduced
+- Simple plotting solution for the results of `pave` using `Plots.jl` recipes
+(via `ValidatedNumerics.jl`):
+```
+using Plots
+gr()  # preferred (fast) backed for `Plots.jl`
+plot(paving.inner)
+plot!(paving.boundary)
+```
+- Major internals rewrite
+- Unary minus and `power_rev` with odd powers now work correctly
+- Examples updated
+- Basic documentation using `Documenter.jl`
+
+
 # v0.3
 - Renamed `setinverse` to `pave`
 

README.md

@@ -11,96 +11,11 @@ The package is based on interval arithmetic using the author's
 [`ValidatedNumerics.jl`](https://github.com/dpsanders/ValidatedNumerics.jl) package,
 in particular multi-dimensional `IntervalBox`es (i.e. Cartesian products of one-dimensional intervals).
 
-The goal is to impose constraints, given by inequalities, and find the set that
-satisfies the constraints, known as the **feasible set**.
+## Documentation
+Documentation for the package is available [here](http://dpsanders.github.io/IntervalConstraintProgramming.jl/latest/)
 
-The method used to do this is known as *interval constraint programming*, in particular the
-so-called "forward--backward contractor". This is implemented in terms of *separators*; see
-[Jaulin & Desrochers].
+See, in particular, the [example notebooks](https://github.com/dpsanders/IntervalConstraintProgramming.jl/tree/master/examples) for usage examples.
 
-## Constraints
-First we define a constraint using the `@constraint` macro:
-```julia
-S = @constraint x^2 + y^2 <= 1
-```
-and an initial interval in the $x$--$y$ plane, `X`:
-```julia
-x = y = -100..100
-X = IntervalBox(x, y)
-```
-
-The `@constraint` macro defines an object `S`, of type `Separator`.
-This is a function which,
-when applied to the box $X = x \times y$
-in the x--y plane, applies two *contractors*, an inner one and an outer one.
-
-The inner contractor tries to shrink down, or *contract*, the box, to the smallest subbox
-of $X$ that contains the part of $X$ that satisfies the constraint; the
-outer contractor tries to contract $X$ to the smallest subbox that contains the
-region where the constraint is not satisfied.
-
-When `S` is applied to the box `X`, it returns the result of the inner and outer contractors:
-```julia
-julia> inner, outer = S(X);
-
-julia> inner
-([-1, 1],[-1, 1])
-
-julia> outer
-([-100, 100],[-100, 100])
-```
-
-## Set inversion: finding the feasible set
-
-To make progress, we must recursively bisect and apply the contractors, keeping
-track of the region proved to be inside the feasible set, and the region that is
-on the boundary ("both inside and outside"). This is done by the `pave` function,
-that takes a separator, a domain to search inside, and an optional tolerance:
-
-```julia
-julia> S = @constraint 1 <= x^2 + y^2 <= 3
-julia> paving = pave(S, X, 0.125);
-```
-
-`pave` returns an object of type `Paving`. This contains: the separator itself;
-an `inner` approximation, of type `SubPaving`, which is an alias for a `Vector` of `IntervalBox`es;
-a `SubPaving` representing the boxes on the boundary that could not be assigned either to the inside or outside of the set;
-and the tolerance.
-
-We may draw the result using the code in the `draw_boxes` file in the examples directory,
-which uses `PyPlot.jl`:
-```julia
-julia> filename = joinpath(Pkg.dir("IntervalConstraintProgramming"), "examples", "draw_boxes.jl");
-julia> include(filename);
-julia> draw(paving)
-```
-
-We can get more control with
-```
-julia> draw(paving.inner, "green", 0.5, 1)
-julia> draw(paving.boundary, "grey", 0.2)
-```
-The second argument is the color; the third (optional) is the alpha value (transparency);
-and the fourth is the linewidth (default is 0).
-
-The output should look like this:
-
-![Ring](examples/ring.png)
-
-
-The green boxes have been **rigorously** proved to be contained within the feasible set,
-and the white boxes to be outside the set. The grey boxes show those that lie on the boundary, whose status is unknown.
-
-### 3D
-
-The package works in any number of dimensions, although it suffers from the usual exponential slowdown ("combinatorial explosion") in higher dimensions. In 3D, it is still relatively fast. There are sample 3D calculations in the `examples` directory, in particular in the [solid torus notebook](examples/Solid torus.ipynb), which uses [`GLVisualize.gl`](https://github.com/JuliaGL/GLVisualize.jl) to provide a 3D visualization which may be rotated and zoomed. The output for the solid torus looks like this:
-
-![Coloured solid torus](examples/coloured_solid_torus.png)
-
-
-## Set operations
-Separators may be combined using the operators `!` (complement), `∩` and `∪` to make
-more complicated sets; see the [notebook](examples/Set inversion.ipynb) for several examples.
 
 ## Author
 
@@ -110,8 +25,7 @@ Departamento de Física, Facultad de Ciencias, Universidad Nacional Autónoma de
 
 ## References:
 - *Applied Interval Analysis*, Luc Jaulin, Michel Kieffer, Olivier Didrit, Eric Walter (2001)
-- Introduction to the Algebra of Separators with Application to Path Planning, Luc Jaulin and Benoît Desrochers,
-*Engineering Applications of Artificial Intelligence* **33**, 141–147 (2014)
+- Introduction to the Algebra of Separators with Application to Path Planning, Luc Jaulin and Benoît Desrochers, *Engineering Applications of Artificial Intelligence* **33**, 141–147 (2014)
 
 ## Acknowledements
 Financial support is acknowledged from DGAPA-UNAM PAPIME grants PE-105911 and PE-107114, and DGAPA-UNAM PAPIIT grant IN-117214, and from a CONACYT-Mexico sabbatical fellowship. The author thanks Alan Edelman and the Julia group for hospitality during his sabbatical visit. He also thanks Luc Jaulin and Jordan Ninin for the [IAMOOC](http://iamooc.ensta-bretagne.fr/) online course, which introduced him to this subject.

REQUIRE

@@ -2,3 +2,4 @@ julia 0.4
 ValidatedNumerics 0.5.1  # plot recipes from 0.5.1
 MacroTools 0.3
 Compat 0.7.14
+FixedSizeArrays 0.1.2