R/wander.R

In nrennie/aRt: Generative Art with R

#' Wander
#'
#' This function generates a generative art ggplot
#' object using random walks
#'
#' @param n_lines Number of lines. Default 100.
#' @param n_points Number of points. Default 350.
#' @param r_outer Radius of outer circle. Default 8.
#' @param r_inner Radius of inner circle. Default 3.
#' @param line_var Variance of random walk noise. Default 0.01.
#' @param deg_jitter Degree of jitter for multiple lines. Default 0.1.
#' @param linewidth Width of lines. Default 0.1.
#' @param bg_col Background colour. Default "#462255".
#' @param col_palette Vector of colours. Default `c("#FF8811", "#9DD9D2", "#046E8F", "#D44D5C")`
#' @param n_cols Number of colours to create. Default 20.
#' @param s Random seed. Default 123.
#' @return A ggplot object.
#' @examples
#' wander()
#' @export

wander <- function(n_lines = 100,
                   n_points = 350,
                   r_outer = 8,
                   r_inner = 3,
                   line_var = 0.01,
                   deg_jitter = 0.1,
                   linewidth = 0.1,
                   bg_col = "#462255",
                   col_palette = c("#FF8811", "#9DD9D2", "#046E8F", "#D44D5C"),
                   n_cols = 20,
                   s = 123) {
  p <- withr::with_seed(
    seed = s,
    code = {
      theta <- seq(0, 2 * pi, length.out = n_lines)
      x_inner <- r_inner * cos(theta)
      y_inner <- r_inner * sin(theta)
      x_outer <- r_outer * cos(theta)
      y_outer <- r_outer * sin(theta)
      all_cols <- grDevices::colorRampPalette(col_palette)(n_cols)
      # generate lines
      line_data <- purrr::map(
        .x = seq_len(length(theta)),
        .f = ~ data.frame(
          x0 = seq(x_inner[.x], x_outer[.x], length.out = n_points),
          y0 = seq(y_inner[.x], y_outer[.x], length.out = n_points),
          grp = .x,
          col = sample(all_cols, size = 1)
        )
      ) |>
        dplyr::bind_rows()
      # generate noise
      line_data$noise_x <- replicate(
        length(theta),
        cumsum(c(0, stats::rnorm(n = n_points - 1, mean = 0, sd = sqrt(line_var))))
      ) |>
        as.vector()
      line_data$noise_y <- replicate(
        length(theta),
        cumsum(c(0, stats::rnorm(n = n_points - 1, mean = 0, sd = sqrt(line_var))))
      ) |>
        as.vector()

      # add noise
      plot_data <- line_data |>
        tibble::as_tibble() |>
        dplyr::mutate(
          x = .data$x0 + .data$noise_x,
          y = .data$y0 + .data$noise_y
        )
      lims <- ceiling(max(abs(c(plot_data$x, plot_data$y))))

      # plot
      p <- ggplot2::ggplot() +
        ggfx::with_blur(
          ggplot2::geom_path(
            data = plot_data,
            mapping = ggplot2::aes(
              x = .data$x,
              y = .data$y,
              group = .data$grp,
              colour = .data$col
            ),
            position = ggplot2::position_jitter(
              width = deg_jitter * 2,
              height = deg_jitter * 2
            ),
            linewidth = linewidth * 1.5,
            alpha = 0.3
          ),
          sigma = 0.5
        ) +
        ggplot2::geom_path(
          data = plot_data,
          mapping = ggplot2::aes(
            x = .data$x,
            y = .data$y,
            group = .data$grp,
            colour = .data$col
          ),
          position = ggplot2::position_jitter(
            width = deg_jitter,
            height = deg_jitter
          ),
          alpha = 0.5,
          linewidth = linewidth
        ) +
        ggplot2::geom_path(
          data = plot_data,
          mapping = ggplot2::aes(
            x = .data$x,
            y = .data$y,
            group = .data$grp,
            colour = .data$col
          ),
          alpha = 0.7,
          linewidth = linewidth
        ) +
        ggplot2::scale_x_continuous(limits = c(-lims, lims)) +
        ggplot2::scale_y_continuous(limits = c(-lims, lims)) +
        ggplot2::scale_colour_identity() +
        ggplot2::coord_fixed(expand = FALSE) +
        theme_aRt(bg_col, 0.5)
      p
    }
  )
  return(p)
}