Scale.alpha_continuous & Scale.alpha_discrete #1257
Conversation
Codecov Report
@@ Coverage Diff @@
## master #1257 +/- ##
=========================================
- Coverage 85.74% 85.5% -0.25%
=========================================
Files 35 35
Lines 4119 4139 +20
=========================================
+ Hits 3532 3539 +7
- Misses 587 600 +13
Continue to review full report at Codecov.
|
src/scale.jl
Outdated
| @@ -313,7 +325,7 @@ anything in the data that's not respresented in `levels` will be set to | |||
| default order. | |||
|
|
|||
| See also [`group_discrete`](@ref), [`shape_discrete`](@ref), | |||
| [`size_discrete`](@ref), and [`linestyle_discrete`](@ref). | |||
| [`size_discrete`](@ref), and [`linestyle_discrete`](@ref), and [`alpha_discrete`](@ref). | |||
There was a problem hiding this comment.
the "and" before linestyle_discrete could be removed as it was in #1252
| for (k,g) in enumerate(ug) | ||
| i = groups.==[g] | ||
| polys[k] = polygon_points(aes_x[i], aes_y[i], geom.preserve_order) | ||
| colors[k] = first(aes_color[i]) | ||
| line_styles[k] = mod1(first(aes_linestyle[i]), linestyle_palette_length) | ||
| line_styles[k] = mod1(first(aes_linestyle[i]), linestyle_palette_length) | ||
| alphas[k] = first(alpha_discrete ? theme.alphas[aes_alpha[i]] : aes_alpha[i]) |
There was a problem hiding this comment.
what git weirdness is causing the line_style[k] line to be highlight as a change?
src/geom/ribbon.jl
Outdated
| @@ -10,41 +10,42 @@ RibbonGeometry(default_statistic=Gadfly.Stat.identity(); fill=true, tag=empty_ta | |||
| Geom.ribbon | |||
|
|
|||
| Draw a ribbon at the positions in `x` bounded above and below by `ymax` and | |||
| `ymin`, respectively. Optionally draw multiple ribbons by grouping with `color`. | |||
| `ymin`, respectively. Optionally draw multiple ribbons by grouping with `color` and `alpha` or `linestyle`. | |||
There was a problem hiding this comment.
what does "color and alpha or linestyle" mean? the boolean logic is unclear
There was a problem hiding this comment.
I've clarified this. It differs whether the ribbon is filled or not filled:
Lines 75 to 76 in 99f54e4
| Theme(default_color="green", discrete_highlight_color=c->"gray", | ||
| point_size=2mm, alphas=[0.0,0.2,0.4,0.6,0.8,1.0]) | ||
| ) | ||
| hstack(p1,p2) |
There was a problem hiding this comment.
both of these tests are near identical to the docs, the only difference being the hard-coded random numbers. seems redundant. ideally i like to see real data, say from RDatasets, used in the docs that is particularly pretty, and very simple made up cases for the tests with focus on execution speed (we time out sometimes) and code coverage. this is a bit of a nit-pick, but can you at least make the docs and the tests less redundant?
There was a problem hiding this comment.
I like this example because it illustrates e.g. "color aesthetic, fixed alpha" and "alpha aesthetic, fixed color". If I used an example from RDatasets to illustrate the alpha aesthetic, I would use something like the Diamonds dataset (53940 points). In my next PR I'm going to halve the time it takes to plot that: Geom.point currently suffers from using a new context for every point.
NEWS.mdSame as #1252, but now with
alphainfo in the docstrings forGeom.point,Geom.polygonandGeom.ribbon