Rework structure for 8bit -> 24bit mode

[t-bltg]

• use UInt32 instead of Union{Nothing,UInt8} for canvas color (allows packing r, g, b channels for truecolor or 256 colors), blending rgb, ...
• use more Crayons: reduces test files by about 25%
• switch between 24bit and 8bit at runtime on COLORTERM env variable (generic) or UP_COLORMODE
• these changes will allow a following PR for saving a UnicodePlots as a png image (Save Plot as png #239)
• get rid of printstyled in Base and delegate colors to Crayons (more flexible for style change (e.g. bold, italic, ...))
• this PR allows a finer (increased resolution) color range in 24bit mode for heatmaps or when using colorbars:
  
• in order to ease transition, USE_LUT is set to false by default, hence using 256 colors when using named colors or integers in 24bit mode.

perf regression

using BenchmarkTools, UnicodePlots

main() = begin
  @show ENV["COLORTERM"]
  io = IOContext(PipeBuffer(), :color=>true)
  @btime show($io, surfaceplot(0:0.5:(2π), 0:0.5:(2π), (x, y) -> sin(x) + cos(y), color = :yellow))
  @btime show($io, lineplot([cos, sin, tan], -π / 2, 2π))
  @btime show($io, barplot(["I need a\nbreak", "I dont but could", :Hello, :Again], [30, 40, 20, 10]))
  @btime show($io, surfaceplot(-2:2, -2:2, (x, y) -> 15sinc(√(x^2 + y^2) / π), zscale=z -> 0, lines=true, colormap=:jet))
  @btime show($io, heatmap(collect(0:30) * transpose(collect(0:30)), xfact=.1, yfact=.1, xoffset=-1.5, colormap=:inferno))
end

main()

pr

ENV["COLORTERM"] = "truecolor"
  300.874 μs (2411 allocations: 127.09 KiB)
  368.014 μs (2886 allocations: 127.16 KiB)
  83.622 μs (584 allocations: 25.94 KiB)
  592.752 μs (5182 allocations: 221.33 KiB)
  1.348 ms (17366 allocations: 792.30 KiB)
ENV["COLORTERM"] = "nothing"
  242.127 μs (1635 allocations: 90.72 KiB)
  323.428 μs (2034 allocations: 87.22 KiB)
  73.201 μs (496 allocations: 21.81 KiB)
  561.642 μs (5662 allocations: 214.48 KiB)
  1.127 ms (14799 allocations: 668.06 KiB)

master

  706.757 μs (9052 allocations: 502.95 KiB)
  795.607 μs (9363 allocations: 496.50 KiB)
  73.877 μs (601 allocations: 35.06 KiB)
  1.271 ms (13516 allocations: 915.23 KiB)
  1.268 ms (14826 allocations: 1021.61 KiB)

visual inspection

root="~/.julia/dev/UnicodePlots.jl"  # fork

for f in $(find $root/test/references_8 -name '*.txt'); do
  tmp=$(mktemp)
  fm=${f/references_8/references}
  url="https://raw.githubusercontent.com/JuliaPlots/UnicodePlots.jl/master/${fm#$root/}"
  wget -q -O $tmp $url
  if ! diff $f $tmp 1>/dev/null; then
    echo; echo
    echo "== remote($url) =="
    cat $tmp; echo
    echo "== local8($f) =="
    cat $f; echo
    f24=${f/references_8/references_24}
    echo "== local24($f24) =="
    cat $f24; echo
  else
    echo "== same($f) =="
  fi
  rm $tmp
done

[codecov-commenter]

Codecov Report

Merging #238 (10eacb8) into master (406f5cc) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master     #238   +/-   ##
=======================================
  Coverage   99.93%   99.93%           
=======================================
  Files          28       28           
  Lines        1501     1572   +71     
=======================================
+ Hits         1500     1571   +71     
  Misses          1        1           

Impacted Files	Coverage Δ
src/canvas.jl	100.00% <100.00%> (ø)
src/canvas/asciicanvas.jl	100.00% <100.00%> (ø)
src/canvas/blockcanvas.jl	100.00% <100.00%> (ø)
src/canvas/braillecanvas.jl	98.11% <100.00%> (ø)
src/canvas/densitycanvas.jl	100.00% <100.00%> (ø)
src/canvas/dotcanvas.jl	100.00% <100.00%> (ø)
src/canvas/heatmapcanvas.jl	100.00% <100.00%> (ø)
src/canvas/lookupcanvas.jl	100.00% <100.00%> (ø)
src/colormaps.jl	100.00% <100.00%> (ø)
src/common.jl	100.00% <100.00%> (ø)
... and 4 more

---

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 406f5cc...10eacb8. Read the comment docs.

[t-bltg]

@johnnychen94, this PR is massive (due to the changes and addition of reference files), and I don't know if you have the time or the will to review it. If not, no worries, tell me so and I'll merge it later this week.

[johnnychen94]

I don't know if you have the time or the will to review it.

Not really... Sorry.. I'm starting to put more time into my own work here to make my phd career not that a failure :sad:

For the long-term goal that might never be achieved, I'm expecting to work closely with Term.jl, and eventually build a terminal-based IDE for Julia. This is why I was writing a decoder as well in JuliaImages/ImageInTerminal.jl#62. But I have to confess that I haven't looked into Term.jl at all.