scales dict

[Mattriks]

• I've added an entry to NEWS.md
• I've run the regression tests
• I've built the docs and confirmed these changes don't cause new errors

This PR

In Gadfly, the scales dict contains aesthetic=>Scale pairs, but what aesthetics are added to scales is poorly defined.
This PR provides more sensible membership. The scales dict should contain:

1. Aesthetics that are mapped by the user e.g. color=:Species
2. Aesthetics which are not mapped by user, but are mapped by some Stat e.g. Stat.contour generates a color scale.
3. Aesthetics in Geom.subplot_grid layers (as for 1 and 2).

As a result, this PR:

• Provides a versatile alternative to Scale.color_none (plot p1 & p2 below).
• fixes subplot_grid not working with Scale #1218 (plot p3 below).

Future PR

Scale.color_none could be deprecated (see #527). Any objections?

Examples

x1, x2 = 1:10., -2:0.2:2
zs = Float64[x^2*log(y) for x in x1, y in x1]
D = DataFrame(x=repeat(x2, outer=length(x2)), y=repeat(x2, inner=length(x2)))
D.z = D.x.*exp.(-(D.x.^2 + D.y.^2))
D2 = DataFrame(x=0.8*randn(10), y=0.8*randn(10), r=rand(10))

p1 = plot(x=x1, y=x1, z=zs, Geom.contour, color=[colorant"steelblue3"])
p2 = plot(D, x=:x, y=:y,  layer(D2, color=:r, Geom.point),
    layer(z=:z, Geom.contour(levels=[-0.1:-0.1:-0.5;]), linestyle=[:dash], color=[colorant"black"]),
    layer(z=:z, Geom.contour(levels=[0.1:0.1:0.5;]), color=[colorant"black"]),
)
hstack(p1, p2)

N = 5000
df1 = DataFrame(
  x1 = rand(Normal(1,3), N),   x3 = rand(Normal(2, 6), N),
  x2 = [sample(["High", "Medium", "Low"], pweights([0.25,0.45,0.30])) for i=1:N],
)
p3 = plot(df1, xgroup =:x2,  x=:x1,  y=:x3,
    Geom.subplot_grid(Geom.hexbin),
    Scale.color_continuous(colormap=Scale.lab_gradient("blue", "gold3", "orange")),
)

[codecov-commenter]

Codecov Report

Merging #1468 into master will decrease coverage by 0.90%.
The diff coverage is 97.29%.

@@            Coverage Diff             @@
##           master    #1468      +/-   ##
==========================================
- Coverage   90.23%   89.32%   -0.91%     
==========================================
  Files          39       39              
  Lines        4197     4395     +198     
==========================================
+ Hits         3787     3926     +139     
- Misses        410      469      +59     

Impacted Files	Coverage Δ
src/Gadfly.jl	75.48% <96.66%> (-2.33%)	⬇️
src/geom/subplot.jl	94.70% <100.00%> (-0.67%)	⬇️
src/statistics.jl	96.45% <100.00%> (+0.02%)	⬆️
src/poetry.jl	90.74% <0.00%> (-5.26%)	⬇️
src/theme.jl	65.07% <0.00%> (-4.42%)	⬇️
src/geom/label.jl	96.26% <0.00%> (-2.74%)	⬇️
src/geom/beeswarm.jl	90.00% <0.00%> (-2.65%)	⬇️
src/aesthetics.jl	77.77% <0.00%> (-2.06%)	⬇️
src/misc.jl	65.06% <0.00%> (-2.06%)	⬇️
... and 20 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 f759d9c...fa8e587. Read the comment docs.