---
title: "Grammar of Graphics for base plot"
author:
  - name: "Guangchuang Yu"
    affiliation: "School of Basic Medical Sciences, Southern Medical University"
date: today
format:
  html:
    toc: true
execute:
  message: false
  warning: false
vignette: >
  %\VignetteIndexEntry{plotbb}
  %\VignetteEngine{quarto::html}
  %\VignetteEncoding{UTF-8}
---

```{r}
#| label: setup
#| include: false
r_libs <- Sys.getenv("R_LIBS")
if (nzchar(r_libs)) {
  libs <- strsplit(r_libs, .Platform$path.sep, fixed = TRUE)[[1]]
  libs <- libs[nzchar(libs)]
  .libPaths(c(libs, .libPaths()))
}
library(dplyr)
library(plotbb)
```

## Aesthetic mapping

`bb_aes()` for aesthetic mapping, that equivalents to `ggplot2::aes()`.

```{r plotbb-aes}
library(plotbb)

p <- bbplot(mtcars, bb_aes(mpg, disp, col = factor(cyl)))
p + bb_grid(col = "grey50", lty = "dashed") + bb_point(pch = 19)
```

## Geometric layer

```{r plotbb-layer}
#| fig-width: 12
p2 <- p + bb_point() + bb_lm(bb_aes(group = cyl), lwd = 2)
p3 <- p2 + bb_lm(col = "red", lwd = 3, lty = "dotted")
p4 <- p + bb_text(bb_aes(label = cyl), cex = 2)
aplot::plot_list(p2, p3, p4, ncol = 3, tag_levels = "A")
```

### Line, segment and errorbar

```{r plotbb-line-seg-errorbar}
#| fig-width: 12
d2 <- group_by(mtcars, cyl) %>%
  summarize(
    x = cyl[1],
    y = mean(disp),
    sd = sd(disp)
  )

g1 <- bbplot(d2, bb_aes(x, y)) +
  bb_line(col = "grey50") +
  bb_point(pch = 19, cex = 1.5) +
  bb_errorbar(bb_aes(ymin = y - sd, ymax = y + sd), width = 0.8)

seg <- data.frame(x = 15, y = 150, xend = 30, yend = 450)
g2 <- bbplot(seg, bb_aes(x, y)) +
  bb_point(pch = 19) +
  bb_segment(bb_aes(xend = xend, yend = yend), lwd = 3, col = "steelblue")

aplot::plot_list(g1, g2, ncol = 2, tag_levels = "A")
```

### Bar plot

```{r plotbb-bar}
#| fig-width: 8
bbplot(mtcars, bb_aes(cyl, mpg, col = factor(cyl))) +
  bb_bar(stat = "count") +
  bb_legend(position = "topleft")
```

### Heatmap

```{r plotbb-heatmap}
#| fig-width: 10
#| fig-height: 10
df <- data.frame(
  x = rep(1:10, 12),
  y = rep(1:12, each = 10),
  values = rnorm(120, mean = 10, sd = 5),
  type = sample(LETTERS[1:5], 120, replace = TRUE),
  stringsAsFactors = FALSE
)

h1 <- bbplot(df, bb_aes(x, y, col = values)) + bb_tile() +
  bb_title("heatmap for continuous numerical values")
h2 <- bbplot(df, bb_aes(x, y, col = values)) + bb_tile() + bb_scale_col_palette("YlOrRd") +
  bb_title("applying a color palette")
h3 <- bbplot(df, bb_aes(x, y, col = type)) + bb_tile() +
  bb_title("heatmap for discrete categorical values")
h4 <- bbplot(df, bb_aes(x, y, col = values)) + bb_tile() + bb_text(col = "black") +
  bb_title("heatmap with text labels") + bb_theme_expand()
aplot::plot_list(h1, h2, h3, h4, ncol = 2, tag_levels = "A")
```

### TODO

+ [x] bb_grid
+ [x] bb_point
+ [x] bb_lm
+ [x] bb_text
+ [x] bb_tile
+ [x] bb_line
+ [x] bb_segment
+ [x] bb_errorbar
+ [x] bb_bar
+ [x] bb_scale_col_manual
+ [x] bb_scale_col_gradient
+ [x] bb_legend
+ [x] bb_facet_wrap
+ more layers need to be added

## Setting labels

```{r plotbb-labs}
p2 + bb_labs(
  title = "hello",
  sub = "just for demo",
  xlab = "this is xlab",
  ylab = "this is ylab"
) +
  bb_title("hello world")
```

## Theme

```{r plotbb-theme}
#| fig-width: 10
g <- p2 +
  bb_theme(
    col.main = "red",
    cex.main = 2,
    mar = c(4, 4, 3, 1)
  ) +
  bb_title("applying graphics::par")
g2 <- p2 + bb_title("theme has no side effect")
aplot::plot_list(g, g2, ncol = 2, tag_levels = "A")
```

`bb_theme` has no side effect and will only apply to the `bbplot` object that it added to.
This is very important for developing pre-defined themes.

```{r plotbb-theme-expand}
#| fig-width: 10
p4 <- p3 + bb_theme_expand()
aplot::plot_list(p4, p3, ncol = 2, tag_levels = "A")
```

```{r plotbb-theme-grey}
p + bb_grid(col = "grey50", lty = "dashed") +
  bb_point(pch = 19) +
  bb_theme_expand() +
  bb_theme_grey()
```

```{r plotbb-theme-blue}
p + bb_point(pch = 19, cex = 2) +
  bb_theme_expand() +
  bb_theme_deepblue()
```

### TODO

+ [x] `bb_theme_expand`
+ [x] `bb_theme_grey`
+ [x] `bb_theme_deepblue`
+ develop more pre-defined themes

## Scale

+ [x] `bb_scale_col_palette`
+ [x] `bb_scale_col_manual`
+ [x] `bb_scale_col_gradient`

Color scales can be added anywhere in the `+` chain and will affect both layers and legends.

```{r plotbb-scale-manual}
pal <- c(`4` = "#1b9e77", `6` = "#d95f02", `8` = "#7570b3")
bbplot(mtcars, bb_aes(mpg, disp, col = factor(cyl))) +
  bb_point(pch = 19) +
  bb_scale_col_manual(pal) +
  bb_legend()
```

```{r plotbb-scale-gradient}
bbplot(mtcars, bb_aes(mpg, disp, col = hp)) +
  bb_point(pch = 19) +
  bb_scale_col_gradient(low = "grey80", high = "firebrick") +
  bb_legend()
```

## Legend

```{r plotbb-legend}
p + bb_point(pch = 19) + bb_legend()
```

## Facet

```{r plotbb-facet}
#| fig-width: 10
p + bb_point(pch = 19) + bb_facet_wrap(~cyl, ncol = 3)
```

```{r plotbb-facet-free}
#| fig-width: 10
p + bb_point(pch = 19) + bb_facet_wrap(~cyl, ncol = 3, scales = "free_y")
```

## Using existing code with plotbb

Suppose we have existing code to plot something:

```r
plot(mtcars$mpg, mtcars$disp)
abline(lm(disp ~ mpg, data = mtcars), col = "red")
```

We can wrap the codes into a function:

```{r}
f <- function() {
  plot(mtcars$mpg, mtcars$disp)
  abline(lm(disp ~ mpg, data = mtcars), col = "red")
}
```

Then we can convert it to a `bbplot` object. The plot produced by the function will be used as the canvas, and we can apply theme and add layers to it:

```{r}
library(dplyr)
d <- group_by(mtcars, cyl) %>%
  summarize(xm = mean(mpg), ym = mean(disp))

pp <- as.bbplot(f) +
  bb_theme_expand() +
  bb_theme_grey() +
  bb_lm(bb_aes(mpg, disp, group = cyl, col = factor(cyl)), data = mtcars, lwd = 2, lty = "dashed") +
  bb_point(bb_aes(xm, ym, col = factor(cyl)), data = d, pch = 19, cex = 2) +
  bb_title("hello plotbb") +
  bb_grid(col = "grey30", lty = "dashed")
```

As there are many features currently not available, `plotbb` supports adding layers using base graphics commands that you are already familiar with. Any base graphics commands in formula, expression or function can be served as layer to be added to a `bbplot` object.

```{r plotbb-base}
pp + (~points(30, 400, pch = 19, col = "red", cex = 3)) +
  ~text(30, 420, label = "hae fun :)", col = "blue", cex = 1.2)
```

Here is another example to use `plotbb` with the `ape` package to visualize phylogenetic tree.

```{r plotbb-ape}
#| fig-width: 8
#| fig-height: 7
require(ape)
set.seed(2020 - 09 - 10)
x = rtree(10)

p = as.bbplot(
  function()
    plot(
      x,
      cex = 2,
      edge.width = 2,
      edge.color = "white",
      tip.color = "purple"
    )
)

p + bb_theme_expand() + bb_theme_deepblue() +
  (~nodelabels(node = 14, pie = matrix(rep(1, 4), 1), cex = 2)) +
  (~nodelabels(node = 15, pie = matrix(rep(1, 3), 1), cex = 2)) +
  (~nodelabels(node = 18, pie = matrix(rep(1, 5), 1), cex = 2)) +
  bb_labs(title = "phylogenetic tree plotted by base graphics") +
  bb_theme(col.main = "firebrick", cex.main = 2)
```