Merge pull request #27 from SimonDanisch/master

Make Contour.jl compatible with GLVisualize

Author: darwindarak

.travis.yml

@@ -1,9 +1,15 @@
 language: julia
 sudo: false
+os:
+  - linux
+  - osx
 julia:
   - nightly
-  - release
   - 0.4
-  - 0.3
 notifications:
   - email: false
+script:
+    - if [[ -a .git/shallow ]]; then git fetch --unshallow; fi
+    - julia --inline=no -e 'Pkg.clone(pwd()); Pkg.test("Contour"; coverage=true)'
+after_success:
+  - julia -e 'cd(Pkg.dir("Contour")); Pkg.add("Coverage"); using Coverage; Coveralls.submit(process_folder()); Codecov.submit(process_folder())'

README.md

@@ -21,10 +21,10 @@ and `z` is a matrix arranged such that `z[xi,yi]` correspond to the location
 
 
 ```julia
-x = [-3:0.01:3]
-y = [-4:0.02:5]
+x = -3:0.01:3
+y = -4:0.02:5
 
-z = [(xi^2 + yi^2)::Float64 for xi in x, yi in y]
+z = [Float64((xi^2 + yi^2)) for xi in x, yi in y]
 ```
 
 Let's find the contour line corresponding to `z = 4.0`:
@@ -50,13 +50,13 @@ julia> c.lines
  [0.05,1.99937],[0.04,1.9996],[0.03,1.99977],[0.02,1.9999],[0.01,1.99997],
  [0.0,2.0]])
  ```
- 
+
 The format of the output data is intented to give as extensive information as possible about the contour line, in a format that can be generalized in the future, if/when something like a [`Geometry.jl` package](https://groups.google.com/forum/#!topic/julia-dev/vZpZ8NBX_z8) is created. Each contour level is represented by an instance of
 
 ```julia
-type ContourLevel
-    level::Float64
-    lines::Vector{Curve2}
+type ContourLevel{T}
+    level::T
+    lines::Vector{Curve2{T}}
 end
 ```
 
@@ -73,8 +73,8 @@ julia> plot(xs, ys) # using your favorite plotting tool
 
 `Contour.jl` makes sure that the coordinates are ordered correctly, and contours that close on themselves are given cyclically, so that e.g. `xs[1]==xs[end]` - in other words, plotting the contour does not require you to add the first point at the end manually to close the curve.
 
-We can also find the contours at multiple levels using `contours`, 
-which returns an array of `ContourLevel` types. 
+We can also find the contours at multiple levels using `contours`,
+which returns an array of `ContourLevel` types.
 
 ```julia
 julia> h = [4.0, 5.0, 6.0];
@@ -92,7 +92,7 @@ specify the number of levels we want.
 julia> N = 3;
 julia> c = contours(x, y, z, N)
 3-element Array{ContourLevel,1}:
- ContourLevel(8.5,[Curve2{Float64}([[0.62,2.84877],…,[0.62,2.84877]])]) 
+ ContourLevel(8.5,[Curve2{Float64}([[0.62,2.84877],…,[0.62,2.84877]])])
  ContourLevel(17.0,[Curve2{Float64}([[3.0,2.82841],…,[3.0,-2.82841]])])
  ContourLevel(25.5,[Curve2{Float64}([[3.0,4.06201],…,[-3.0,4.06201]])])
 ```

REQUIRE

@@ -1,3 +1,3 @@
-julia 0.3
+julia 0.4
 Compat
-ImmutableArrays
+FixedSizeArrays

src/Contour.jl

@@ -1,20 +1,20 @@
 module Contour
 
-using Compat, ImmutableArrays
+using Compat, FixedSizeArrays
 
 export ContourLevel, Curve2, contour, contours, coordinates
 
 type Curve2{T}
-    vertices::Vector{Vector2{T}}
+    vertices::Vector{Point{2,T}}
 end
-Curve2{T}(::Type{T}) = Curve2(Vector2{T}[])
+Curve2{T}(::Type{T}) = Curve2(Point{2, T}[])
 
-type ContourLevel
-    level::Float64
-    lines::Vector{Curve2{Float64}}
+type ContourLevel{T}
+    level::T
+    lines::Vector{Curve2{T}}
 end
-ContourLevel(h::Float64) = ContourLevel(h, Curve2{Float64}[])
-ContourLevel(h::Real) = ContourLevel(@compat Float64(h))
+ContourLevel{T<:AbstractFloat}(h::T) = ContourLevel(h, Curve2{T}[])
+ContourLevel{T}(h::T) = ContourLevel(Float64(h))
 
 function contour(x, y, z, level::Number)
     # Todo: size checking on x,y,z
@@ -32,10 +32,10 @@ function contourlevels(z,n)
     range(zmin+dz,dz,n)
 end
 
-function coordinates(c::Curve2)
+function coordinates{T}(c::Curve2{T})
     N = length(c.vertices)
-    xlist = Array(Float64,N)
-    ylist = Array(Float64,N)
+    xlist = Array(T,N)
+    ylist = Array(T,N)
 
     for (i,v) in enumerate(c.vertices)
         xlist[i] = v[1]
@@ -144,9 +144,9 @@ const fwd, rev = (@compat UInt8(0)), (@compat UInt8(1))
 
 function add_vertex!{T}(curve::Curve2{T}, pos::(@compat Tuple{T,T}), dir::UInt8)
     if dir == fwd
-        push!(curve.vertices, Vector2{T}(pos...))
+        push!(curve.vertices, Point{2,T}(pos...))
     else
-        unshift!(curve.vertices, Vector2{T}(pos...))
+        unshift!(curve.vertices, Point{2,T}(pos...))
     end
 end
 
@@ -233,7 +233,7 @@ function trace_contour(x, y, z, h::Number, cells::Dict{(@compat Tuple{Int,Int}),
     # It then tries to trace the contour in the opposite direction.
 
     while length(cells) > 0
-        contour = Curve2(Float64)
+        contour = Curve2(promote_type(map(eltype, (x,y,z))...))
 
         # Pick initial box
         (xi_0, yi_0), cell = first(cells)

test/verify_vertices.jl

@@ -1,5 +1,3 @@
-using ImmutableArrays
-
 # Setup test axes that will be shared among the tests
 
 # Shift the axes so that they do not line up with
@@ -111,7 +109,7 @@ end
 
 # Issue #12
 x = float(collect(1:3));
-y = copy(x); 
+y = copy(x);
 z = eye(3,3);
 contours(x,y,z)
 
@@ -139,7 +137,7 @@ for line in contourlevels.lines
     ys .== [v[2] for v in line.vertices]
 end
 
-# Test that closed contours are identified correctly 
+# Test that closed contours are identified correctly
 # when ambiguous cells are involved
 
 Z = float([1 1 1 1 1 1