refactor Balanced using the functor API

[johnnychen94]

This is a PR-commit for PR #29 that refactors the whole codebase to use the new functor API, which
is discussed in more details in #26 and explained in my GSoC 2019 blog

Changes:

• support Color3 input
• move implementations into functor
• enhance the test codes with wider test coverage

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (api@2951a80). Click here to learn what that means.
The diff coverage is 100%.

@@          Coverage Diff           @@
##             api      #32   +/-   ##
======================================
  Coverage       ?   41.27%           
======================================
  Files          ?       20           
  Lines          ?      172           
  Branches       ?        0           
======================================
  Hits           ?       71           
  Misses         ?      101           
  Partials       ?        0

Impacted Files	Coverage Δ
src/balanced.jl	100% <100%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2951a80...36c8f13. Read the comment docs.

[johnnychen94]

If remove this, the test would fail on RGB images. Looks like build_histogram doesn't support Color3 images yet? @zygmuntszpak

diff --git a/src/balanced.jl b/src/balanced.jl
index a72291b..52a6ba9 100644
--- a/src/balanced.jl
+++ b/src/balanced.jl
@@ -80,7 +80,8 @@ img_binary = binarize(Balanced(), img)
 """
 struct Balanced <: AbstractImageBinarizationAlgorithm end

-function (f::Balanced)(out::GenericGrayImage, img::GenericGrayImage)
+function (f::Balanced)(out::GenericGrayImage,
+                       img::AbstractArray{T, 2}) where T<:Union{Number, Colorant}
     edges, counts = build_histogram(img,  256)
     t = find_threshold(HistogramThresholding.Balanced(), counts[1:end], edges)
     for i in CartesianIndices(img)
@@ -88,6 +89,3 @@ function (f::Balanced)(out::GenericGrayImage, img::GenericGrayImage)
     end
     out
 end
-
-(f::Balanced)(out::GenericGrayImage, img::AbstractArray{<:Color3}) =
-    f(out, of_eltype(Gray, img))

The error message

Types: Error During Test at /home/jc/Documents/Julia/ImageBinarization.jl/test/balanced.jl:35
  Got exception outside of a @test
  TypeError: in keyword argument minval, expected Union{Real, Color{T,1} where T}, got RGB{Float32}
  Stacktrace:
   [1] (::getfield(ImageContrastAdjustment, Symbol("#kw##build_histogram")))(::NamedTuple{(:minval, :maxval),Tuple{RGB{Float32},RGB{Float32}}}, ::typeof(ImageContrastAdjustment.build_histogram), ::Array{RGB{Float32},2}, ::Int64) at ./none:0
   [2] build_histogram(::Array{RGB{Float32},2}, ::Int64) at /home/jc/.julia/packages/ImageContrastAdjustment/RY6r2/src/contrastadjustment.jl:150
   [3] (::Balanced)(::Array{Gray{Float32},2}, ::Array{RGB{Float32},2}) at /home/jc/Documents/Julia/ImageBinarization.jl/src/balanced.jl:85
   [4] #binarize!#1(::Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{(),Tuple{}}}, ::Function, ::Array{Gray{Float32},2}, ::Array{RGB{Float32},2}, ::Balanced) at /home/jc/Documents/Julia/ImageBinarization.jl/src/BinarizationAPI/binarize.jl:24
   [5] binarize!(::Array{Gray{Float32},2}, ::Array{RGB{Float32},2}, ::Balanced) at /home/jc/Documents/Julia/ImageBinarization.jl/src/BinarizationAPI/binarize.jl:24
   [6] #binarize#3(::Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{(),Tuple{}}}, ::Function, ::Type{Gray{Float32}}, ::Array{RGB{Float32},2}, ::Balanced) at /home/jc/Documents/Julia/ImageBinarization.jl/src/BinarizationAPI/binarize.jl:41
   [7] binarize(::Type{Gray{Float32}}, ::Array{RGB{Float32},2}, ::Balanced) at /home/jc/Documents/Julia/ImageBinarization.jl/src/BinarizationAPI/binarize.jl:40
   [8] #binarize#4(::Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{(),Tuple{}}}, ::Function, ::Array{RGB{Float32},2}, ::Balanced) at /home/jc/Documents/Julia/ImageBinarization.jl/src/BinarizationAPI/binarize.jl:47
   [9] binarize(::Array{RGB{Float32},2}, ::Balanced) at /home/jc/Documents/Julia/ImageBinarization.jl/src/BinarizationAPI/binarize.jl:47
   [10] top-level scope at /home/jc/Documents/Julia/ImageBinarization.jl/test/balanced.jl:53
   [11] top-level scope at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Test/src/Test.jl:1083
   [12] top-level scope at /home/jc/Documents/Julia/ImageBinarization.jl/test/balanced.jl:37
   [13] top-level scope at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Test/src/Test.jl:1083
   [14] top-level scope at /home/jc/Documents/Julia/ImageBinarization.jl/test/balanced.jl:2
   [15] include at ./boot.jl:326 [inlined]
   [16] include_relative(::Module, ::String) at ./loading.jl:1038
   [17] include(::Module, ::String) at ./sysimg.jl:29
   [18] include(::String) at ./client.jl:403
   [19] top-level scope at /home/jc/Documents/Julia/ImageBinarization.jl/test/runtests.jl:13
   [20] top-level scope at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Test/src/Test.jl:1083
   [21] top-level scope at /home/jc/Documents/Julia/ImageBinarization.jl/test/runtests.jl:10
   [22] include at ./boot.jl:326 [inlined]
   [23] include_relative(::Module, ::String) at ./loading.jl:1038
   [24] include(::Module, ::String) at ./sysimg.jl:29
   [25] include(::String) at ./client.jl:403
   [26] top-level scope at none:0
   [27] eval at ./boot.jl:328 [inlined]
   [28] repleval(::Module, ::Expr) at /home/jc/.julia/packages/Atom/BXRAC/src/repl.jl:135
   [29] (::getfield(Atom, Symbol("##168#170")){Module})() at /home/jc/.julia/packages/Atom/BXRAC/src/repl.jl:157
   [30] with_logstate(::getfield(Atom, Symbol("##168#170")){Module}, ::Base.CoreLogging.LogState) at ./logging.jl:395
   [31] with_logger at ./logging.jl:491 [inlined]
   [32] evalrepl(::Module, ::String) at /home/jc/.julia/packages/Atom/BXRAC/src/repl.jl:148
   [33] top-level scope at /home/jc/.julia/packages/Atom/BXRAC/src/repl.jl:190
   [34] eval(::Module, ::Any) at ./boot.jl:328
   [35] eval_user_input(::Any, ::REPL.REPLBackend) at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/REPL/src/REPL.jl:85
   [36] run_backend(::REPL.REPLBackend) at /home/jc/.julia/packages/Revise/agmgx/src/Revise.jl:949
   [37] (::getfield(Revise, Symbol("##75#77")){REPL.REPLBackend})() at ./task.jl:259

[zygmuntszpak]

Currently the build_histogram method converts an image to a single channel (Gray) before it computes the histogram. If one wants to compute the histogram for an RGB image one would need to call build_histogram for each channel separately.

It looks like I forgot to make sure that the image is converted to Grayscale when the user specifies the minval option. For example, I currently have

function build_histogram(img::AbstractArray, nbins::Integer = 256)
    build_histogram(img, nbins, minval =  minfinite(img), maxval = maxfinite(img))
end

function build_histogram(img::AbstractArray{T}, edges::AbstractRange) where {T<:Color3}
    build_histogram(Gray.(img),edges)
end

I forgot to add something like

function build_histogram(img::AbstractArray, nbins::Integer = 256) where {T<:Color3}
    img_gray = Gray.(img)
    build_histogram(img_gray, nbins, minval =  minfinite(img_gray), maxval = maxfinite(img_gray))
end

[johnnychen94]

@zygmuntszpak the test passed so we can safely merge them.

All the following PRs are created in this pattern so that you can review them easily. For non-trivial cases that worth attention, I'll add a comment to them like in this PR.

There can be some further changes on the codebase, but at least for now, I'd like to keep changes relatively small in these sequences of PRs which simply wrap the implementations in the functor.