Merge pull request #118 from TidierOrg/you-wouldnt-pirate-a-type

Massive refactor. Adds: calculations in @aes macro broadcasting support fill scales (not complete) ggsave for GGPlotGrids Removes: All type piracy A lot of complexity from the calculation system (now handled entirely by DataFrames.jl) closes #71 closes #100 closes #108 closes #97
TidierOrg · Sep 17, 2024 · a8c6589 · a8c6589
2 parents de5808b + c0b754e
commit a8c6589
Show file tree

Hide file tree

Showing 35 changed files with 592 additions and 1,016 deletions.
diff --git a/Project.toml b/Project.toml
@@ -1,19 +1,20 @@
 name = "TidierPlots"
 uuid = "337ecbd1-5042-4e2a-ae6f-ca776f97570a"
-authors = [
-    "Randall Boyes <[email protected]> and contributors",
-]
-version = "0.7.8"
+authors = ["Randall Boyes <[email protected]> and contributors"]
+version = "0.8.0"
 
 [deps]
 CairoMakie = "13f3f980-e62b-5c42-98c6-ff1f3baf88f0"
 CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
 ColorSchemes = "35d6a980-a343-548e-a6ea-1d62b119f2f4"
+ColorTypes = "3da002f7-5984-5a60-b8a6-cbb66c0b333f"
 Colors = "5ae59095-9a9b-59fe-a467-6f913c188581"
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 Dates = "ade2ca70-3891-5945-98fb-dc099432e06a"
+FixedPointNumbers = "53c48c17-4a7d-5ca2-90c5-79b7896eea93"
 Format = "1fa38f19-a742-5d3f-a2b9-30dd87b9d5f8"
 GLM = "38e38edf-8417-5370-95a0-9cbb8c7f171a"
+Infiltrator = "5903a43b-9cc3-4c30-8d17-598619ec4e9b"
 KernelDensity = "5ab0869b-81aa-558d-bb23-cbf5423bbe9b"
 Loess = "4345ca2d-374a-55d4-8d30-97f9976e7612"
 Makie = "ee78f7c6-11fb-53f2-987a-cfe4a2b5a57a"

diff --git a/README.md b/README.md
@@ -14,20 +14,20 @@ TidierPlots.jl is a 100% Julia implementation of the R package [ggplot2](https:/
 1.  **Stick as closely to tidyverse syntax and behaviour as possible:** Whereas other
     meta-packages introduce Julia-centric idioms for working with
     plots, this package’s goal is to reimplement ggplot
-    in Julia. This currently just means that `TidierPlots.jl` gives the option for specifying `aes` with the macro `@es` to allow unquoted column references, but the use of macros may need to expand as more features are added. 
+    in Julia. This currently just means that `TidierPlots.jl` gives the option for specifying `aes` with the macro `@es` to allow unquoted column references, but the use of macros may need to expand as more features are added.
 
 2.  **Stay as compatible as possible with Makie.jl** This package is meant
-    to be a thin wrapper around Makie's SpecApi syntax to help introduce R users to plotting in 
-    Julia. 
+    to be a thin wrapper around Makie's SpecApi syntax to help introduce R users to plotting in
+    Julia.
 
 3. **To Extend ggplot using julia-specific features where appropriate** as long as this does
     not confict with the first two goals. The package aims to behave exactly like ggplot
-    unless told otherwise. Additional options and parameters that are not present in ggplot 
-    may be added, but options that are present in R's ggplot should behave the way they do in R. 
+    unless told otherwise. Additional options and parameters that are not present in ggplot
+    may be added, but options that are present in R's ggplot should behave the way they do in R.
 
 ## Installation
 
-For the "stable" version, access the Pkg interface by pressing `]` at the `julia>` prompt, then type `add TidierPlots`. 
+For the "stable" version, access the Pkg interface by pressing `]` at the `julia>` prompt, then type `add TidierPlots`.
 
 For the development version:
 
@@ -62,7 +62,7 @@ Makie Themes:
 Colour Scales:
 
 - `scale_color_manual()` - set `values = c(c1, c2, c3, ...)`, accepts anything that can be parsed as a color by Colors.jl (named colors, hex values, etc.)
-- `scale_color_[discrete|continuous|binned]()` - set `palette =` a [ColorSchemes.jl palette](https://juliagraphics.github.io/ColorSchemes.jl/stable/catalogue/) as a string or symbol. Also accepts ColorSchemes.jl color scheme objects. 
+- `scale_color_[discrete|continuous|binned]()` - set `palette =` a [ColorSchemes.jl palette](https://juliagraphics.github.io/ColorSchemes.jl/stable/catalogue/) as a string or symbol. Also accepts ColorSchemes.jl color scheme objects.
 
 Additional Elements:
 
@@ -77,7 +77,7 @@ Use the function `TidierPlots_set(option::String, value::Bool)` to control displ
 - "plot_show" (default true). Enables `ggplot`-like behaviour where plots are displayed when created.
 - "plot_log" (default true). Prints a text summary of the properties of the ggplot
 
-You will likely want to disable both of these if you are working in a notebook environment. In [Pluto.jl](https://github.com/fonsp/Pluto.jl), you can get interactive plots (scroll, zoom, labels, etc.) using `WGLMakie` by including `WGLMakie.activate!()` as the first cell after your imports. 
+You will likely want to disable both of these if you are working in a notebook environment. In [Pluto.jl](https://github.com/fonsp/Pluto.jl), you can get interactive plots (scroll, zoom, labels, etc.) using `WGLMakie` by including `WGLMakie.activate!()` as the first cell after your imports.
 
 ## Differences from ggplot2
 
@@ -87,30 +87,33 @@ The goal of this package is to allow you to write code that is as similar to ggp
 - Option 2: `@aes` (or `@es`) macro, aes as in ggplot, e.g. `@aes(x = x, y = y)` or `@aes(x, y)`
 - Option 3 (Deprecated): `aes` function, column names as strings, e.g. `aes(x = "x", y = "y")` or `aes("x", "y")`
 
-If you use Option 1, you get experimental support for calculations inside aes, including `+`, `-`, `*`, `/` and function application. Functions can be applied to columns with the `>>` operator, or wrapped for aes use with the `aesthetics_function()` command. The following geom_point specifications are equivalent: 
+If you use Option 1, functions can be applied to columns with the `=>` operator to form a `Pair{Symbol, Function}`, similar to how `DataFrames.jl` functions work.
 
 ```julia
 my_fn(x) = x ./ 10
-my_aes_fn = aesthetics_function(my_fn)
 
-geom_point(aes(x = :x/10))
-geom_point(aes(x = :x >> my_fn))
-geom_point(aes(x = my_aes_fn(:x)))
+geom_point(aes(x = :x => my_fn))
 ```
 
-Functions can take multiple columns as input (up to two, currently). The following `geom_point` specifications are equivalent, and result in `x / y` (where `x` and `y` are the names of columns in a DataFrame) being plotted as the x axis of the graph: 
+Functions can take multiple columns as input. The following `geom_point` specification results in `x / y` (where `x` and `y` are the names of columns in a DataFrame) being plotted as the x axis of the graph:
 
 ```julia
 my_new_fn(x, y) = x ./ y
-my_new_aes_fn = aesthetics_function(my_new_fn)
 
-geom_point(aes(x = :x/:y))
-geom_point(aes(x = my_new_aes_fn(:x, :y)))
+geom_point(aes(x = [:x, :y] => my_new_fn))
 ```
 
-## Why would I use this instead of ggplot2? 
+With Option 2, functions will be interpreted using `TidierData.jl`:
 
-**Right now, you probably wouldn't.** This package is still early in development, and is not ready for production use. However, there are a couple of advantages already and the list will hopefully get longer over time. 
+```julia
+# Macro aes equivalents to the above examples
+geom_point(@aes(x = x / 10))
+geom_point(@aes(x = x / y))
+```
+
+## Why would I use this instead of ggplot2?
+
+**Right now, you probably wouldn't.** This package is still early in development, and is not ready for production use. However, there are a couple of advantages already and the list will hopefully get longer over time.
 
 ### Easier Factor Handling
 
@@ -147,21 +150,21 @@ beautiful_makie_theme = Attributes(
     fonts=(;regular="CMU Serif"),
 )
 
-ggplot(df) + 
+ggplot(df) +
     geom_point(aes(x = :x, y = :y, size = :size, color = :x), alpha = 0.8) +
-    scale_x_log10() + 
-    scale_y_log10() + 
+    scale_x_log10() +
+    scale_y_log10() +
     labs(x = "x", y = "y") +
     lims(y = c(.1, 100)) +
     scale_color_continuous(palette = "Hiroshige", name = "") +
     theme(
-        xminorticksvisible=true, 
-        xminorgridvisible=true, 
-        yminorticksvisible=true, 
-        yminorgridvisible=true, 
-        xminorticks=IntervalsBetween(9), 
+        xminorticksvisible=true,
+        xminorgridvisible=true,
+        yminorticksvisible=true,
+        yminorgridvisible=true,
+        xminorticks=IntervalsBetween(9),
         yminorticks=IntervalsBetween(9),
-        backgroundcolor = :transparent, 
+        backgroundcolor = :transparent,
         xgridstyle=:dash,
         ygridstyle=:dash
     ) + beautiful_makie_theme
@@ -170,37 +173,37 @@ ggplot(df) +
 
 ### Built-in Support for Plot Layouts
 
-Combine plots with a `{patchwork}`-inspired syntax to create complex layouts (adapted from [beautiful.makie.org](https://beautiful.makie.org/examples/2d/histogram/hists_on_the_sides)): 
+Combine plots with a `{patchwork}`-inspired syntax to create complex layouts (adapted from [beautiful.makie.org](https://beautiful.makie.org/examples/2d/histogram/hists_on_the_sides)):
 
 ```julia
 Random.seed!(123)
 n = 200
 df = DataFrame(x = randn(n) / 2, y = randn(n))
 
-top = ggplot(df) + 
-    geom_histogram(aes(x = :x), color = (:orangered, 0.5), strokewidth = 0.5) + 
-    lims(x = c(-4, 4)) + 
-    theme(xticklabelsvisible = false, xgridvisible = false) + 
+top = ggplot(df) +
+    geom_histogram(aes(x = :x), color = (:orangered, 0.5), strokewidth = 0.5) +
+    lims(x = c(-4, 4)) +
+    theme(xticklabelsvisible = false, xgridvisible = false) +
     beautiful_makie_theme
 
-right = ggplot(df) + 
-    geom_histogram(aes(:y), color = (:dodgerblue, 0.5), 
-        direction = :x, strokewidth = 0.5) + 
-    lims(y = c(-3, 3)) + 
+right = ggplot(df) +
+    geom_histogram(aes(:y), color = (:dodgerblue, 0.5),
+        direction = :x, strokewidth = 0.5) +
+    lims(y = c(-3, 3)) +
     theme(yticklabelsvisible = false, ygridvisible = false) +
     beautiful_makie_theme
 
-middle = ggplot(df) + geom_point(aes(:x, :y), size = 10) + 
-    lims(x = c(-4, 4), y = c(-3, 3)) + labs(x = "x", y = "y") + 
+middle = ggplot(df) + geom_point(aes(:x, :y), size = 10) +
+    lims(x = c(-4, 4), y = c(-3, 3)) + labs(x = "x", y = "y") +
     beautiful_makie_theme
 
-blank = ggplot() + 
+blank = ggplot() +
     theme(xticklabelsvisible = false, xgridvisible = false, yticklabelsvisible = false,
-        ygridvisible = false, xtickcolor = :transparent, ytickcolor = :transparent, 
-        bottomspinevisible = false, topspinevisible = false, rightspinevisible = false, 
+        ygridvisible = false, xtickcolor = :transparent, ytickcolor = :transparent,
+        bottomspinevisible = false, topspinevisible = false, rightspinevisible = false,
         leftspinevisible = false) + beautiful_makie_theme
 
-top + blank + middle + right + 
+top + blank + middle + right +
     plot_layout(ncol = 2, nrow = 2, widths = c(3, 1), heights = c(1, 2))
 ```
 
@@ -213,21 +216,20 @@ Add basic support for any Makie plot using `geom_template(name, required_aes, ma
 ```julia
 geom_raincloud = geom_template("geom_raincloud", ["x", "y"], :RainClouds)
 
-ggplot(penguins) + 
+ggplot(penguins) +
     geom_raincloud(aes(x = :species, y = :bill_depth_mm/10, color = :species), size = 4) +
-    scale_y_continuous(labels = "{:.1f} cm") + 
+    scale_y_continuous(labels = "{:.1f} cm") +
     labs(title = "Bill Depth by Species", x = "Species", y = "Bill Depth") +
     theme_minimal()
 ```
 ![](assets/raincloud.png)
 
-See the [documentation](https://tidierorg.github.io/TidierPlots.jl/latest) for more information and examples. 
+See the [documentation](https://tidierorg.github.io/TidierPlots.jl/latest) for more information and examples.
 
 # What's New
 
 See [NEWS.md](https://github.com/TidierOrg/TidierPlots.jl/blob/main/NEWS.md) for the latest updates.
 
 # What's Missing
 
-Lots! Please feel free to file an issue and/or submit a pull request to add additional ggplot-based features. If it is in ggplot, we want to add it. 
-
+Lots! Please feel free to file an issue and/or submit a pull request to add additional ggplot-based features. If it is in ggplot, we want to add it.
diff --git a/src/TidierPlots.jl b/src/TidierPlots.jl
@@ -12,6 +12,7 @@ using Reexport
 # Needed for color scales to work correctly
 using Colors
 using ColorSchemes
+using FixedPointNumbers
 
 # Needed for label functions
 using Format
@@ -33,32 +34,29 @@ using KernelDensity
 include("structs.jl")
 
 include("addplots.jl")
-include("aes_ops.jl")
 include("aes.jl")
 include("attributes.jl")
+include("broadcasting.jl")
 include("draw.jl")
 include("extract_aes.jl")
 include("facets.jl")
 include("geom.jl")
 include("ggplot.jl")
 include("ggsave.jl")
-include("grouping.jl")
 include("labs.jl")
 include("label_functions.jl")
 include("legend.jl")
 include("patchwork.jl")
 include("scales_colour.jl")
 include("scales_numeric.jl")
 include("themes.jl")
-include("transforms.jl")
 include("show.jl")
 include("util.jl")
 
 include("geoms/geom_template.jl")
 
 include("geoms/geom_bar.jl")
 include("geoms/geom_boxplot.jl")
-include("geoms/geom_contour.jl")
 include("geoms/geom_density.jl")
 include("geoms/geom_errorbar.jl")
 include("geoms/geom_hvline.jl")
@@ -120,12 +118,7 @@ export scale_x_pseudolog10, scale_y_pseudolog10, scale_x_symlog10, scale_y_symlo
 export scale_x_reverse, scale_y_reverse, scale_x_sqrt, scale_y_sqrt
 export scale_colour_continuous, scale_colour_discrete, scale_colour_manual, scale_colour_binned
 export scale_color_continuous, scale_color_discrete, scale_color_manual, scale_color_binned
-
-# transforms
-
-export cat_inseq, cat_inorder, number_on_axis, as_is, discard, verbatim, kernel_density_2d
-export as_color
-export aesthetics_function
+export scale_fill_continuous, scale_fill_discrete, scale_fill_manual, scale_fill_binned
 
 # default options
 

diff --git a/src/addplots.jl b/src/addplots.jl
@@ -8,13 +8,21 @@ function get_options(geom_list)
 end
 
 function Base.:+(x::GGPlot, y::Union{Geom, Vector{Geom}, Aesthetics, AxisOptions, FacetOptions, Attributes}...)::GGPlot
-    
+
     themes = [i for i in y if i isa Attributes]
     theme = length(themes) == 0 ? Makie.theme_ggplot2() : themes[end]
 
     facet = [i for i in y if i isa FacetOptions]
     facet_options = length(facet) == 0 ? nothing : facet[end]
 
+    color = [i.color_palette for i in y if i isa AxisOptions &&
+        !isnothing(i.color_palette)]
+    color_palette = length(color) == 0 ? nothing : color[end]
+
+    fill = [i.fill_palette for i in y if i isa AxisOptions &&
+        !isnothing(i.fill_palette)]
+    fill_palette = length(fill) == 0 ? nothing : fill[end]
+
     result = GGPlot(
         vcat(x.geoms, # if there are geoms or lists of geoms, append them to the ggplot's geoms
             [i              for i in y if i isa Geom],
@@ -27,11 +35,11 @@ function Base.:+(x::GGPlot, y::Union{Geom, Vector{Geom}, Aesthetics, AxisOptions
             [get_options(i) for i in y if i isa Vector{Geom}]...,
             [i.opt          for i in y if i isa AxisOptions]...),
         theme,
-        merge(x.column_transformations,
-            [i.column_transformations for i in y if i isa AxisOptions]...),
         merge(x.legend_options,
             [i.legend_options for i in y if i isa AxisOptions]...),
-        facet_options
+        facet_options,
+        color_palette,
+        fill_palette
     )
 
     return result