coffee_functions.R

coffee_colors <- tibble::tribble(
  ~coffee_type,  ~color_code,
  "Coffee A",  "#E41A1C",
  "Coffee B",  "#377EB8",
  "Coffee C",  "#4DAF4A",
  "Coffee D",  "#984EA3"
)

coffee_pal <- coffee_colors$color_code
names(coffee_pal) <- coffee_colors$coffee_type
# tmaptools::get_brewer_pal("Set1", n = 10) |> list()

match_colors = tibble::tribble(~ids, ~color2,
                               "A  D", "#FFB716",
                               "A   ", "#E41A1C", # A
                               "A CD", "#9ecae1",
                               "ABC ", "#FF7F00", # 
                               " B  ", "#377EB8", # B
                               "  C ", "#4DAF4A", # C
                               "   D", "#984EA3", # D
                               "ABCD" ,"#808080",
                               " BC ", "#A65628",
                               NA, "#e5e5e5")

flavor_colors = tibble::tribble(~ids, ~color,
                                "Fruity",  "#ee3087",
                                "Sour/Acid", "#e6550d",         #
                                "Alcohol/Fermented",  "#9ecae1",
                                "Green/Vegetative", "#3fa746", #
                                "Stale/Papery", "#fee6ce",  #
                                "Earthy", "#ff3d19",   ##
                                "Chemical",  "#9ecae1",  ###
                                "Roasted" , "#747e7d",
                                "Cereal", "#eedaac",
                                "Spices", "#ff6a6d",
                                "Nutty", "#d3a778",
                                "Cocoa" , "#3b281e", # chocolate?
                                "Sweet" , "#f2f4f4",
                                "Floral",  "#abb9dc")

gg_bars <-
  function(df,
           x_axis,
           y_axis,
           fill = NULL, # what to fill by
           y_label_type = scales::percent,
           bar_label_type = scales::percent_format(accuracy = 1),
           single_color = FALSE, # If bars are one color or follow a palette
           bar_color = NULL, # single_color must be TRUE
           base_size = 12) {
    max_lim = max(df[[{{  y_axis  }}]], na.rm = TRUE) * 1.1
    
    if (single_color == FALSE) {
     p <- ggplot(df, aes(!!sym(x_axis), !!sym(y_axis), fill = !!sym(fill))) +
       geom_col() +
      geom_text(
        aes(label = bar_label_type(!!sym(y_axis)), color = !!sym(fill)),
        fontface = "bold",
        vjust = -0.4,
        size = base_size * (4/14)
      ) 
    }
    else{
      p <- ggplot(df, aes(!!sym(x_axis), !!sym(y_axis))) +
        geom_col(fill = bar_color) +
        geom_text(
          aes(label = bar_label_type(!!sym(y_axis))),
          color = bar_color,
          fontface = "bold",
          vjust = -0.4,
          size = base_size * (4/14)
        ) 
    }
    return(
      p + theme_coffee(base_size) +
        scale_y_continuous(
          expand = c(0, 0),
          limits = c(0, max_lim),
          labels = y_label_type
        )
    )
  }

# Used to save wordcloud
save_widget <- function(wd) {
  htmlwidgets::saveWidget(
    widget = wd,
    file = paste0("assets/widgets/",deparse(substitute(wd)), ".html"),
    selfcontained = TRUE
  )
}

gt_add_coffee_color <- function(gt, coff_col){
  gt |> gt::text_transform(
    locations = cells_body(
      columns = !!sym(coff_col)
    ),
    fn = function(x){
      color = filter(coffee_colors, coffee_type == x)$color_code
      name <- x
      # team <- word(x, -1)
      glue::glue(
        "<div><span style='font-weight:bold;color:{color}'>{name}</div>"
      )
    })
}

find_two_words <- function(str, two_word){
  c = stringi::stri_count_regex(str, pattern = two_word)
  rep(two_word, c)
}

count_cupping_notes <-
  function(cup_list,
           count_min = 20) {
    
    remove_words = c("nice", "and", "aftertaste", "better", "bit", "but", "drink", "favorite", "feel", "felt",
                     "finish", "first", "for", "good", "got", "too", "out", "that", "than", "the", "thin", "this",
                     "very", "one", "had", "when", "more", "high", "little", "was", "would", "some", "not", "would",
                     "with", "really", "flavor", "mouth", "notes", "taste", "less", "most", "well", "diner", "after",
                     "much", "tasted", "like", "almost", "tastes", "cup", "flavors", "slightly", "slight", "liked",
                     "just", "tasting", "note", "notes", "also", "what", "maybe", "from", "similar", "its", "didnt",
                     "could", "all", "coffee", "coffees", "which", "probably", "something", "way", "dont", "did",
                     "lot", "over", "have", "process", "time", "down", "brown", "green", "mouthfeel", "body", "cool",
                     "cools", "cooled")
    
    # The first part of each regex will be used to count later on (i.e. black tea)
    two_words = c(
      "brown sugar",
      "black tea|tea like|tea-like|tealike",
      "olive oil|olives",
      "dark chocolate|dark choc|dark chocolatey",
      "green apple|green apples",
      "full body|high body|heavy body|heavier body|fuller body|full bodied|full-bodied",
      "medium body|medium bodied",
      "light body|lighter body|little body|low body|light bodied"
    )
    
    # find_two_words(word, "brown sugar")
    # two_words_counts = purrr::map(two_words, find_two_words, str = word) |> unlist()
    # two_words_counts = sub("\\|.*", "", two_words_counts)
    
    # Put all words in a single long string
    cup_words = cup_list |> stringr::str_to_lower() |>
      paste0(collapse = " ")
    
    two_words_counts = purrr::map(two_words, find_two_words, str = cup_words) |> unlist()
    two_words_counts = sub("\\|.*", "", two_words_counts)
    
    # Also remove dark chocolate bc we will count chocolate
    cup_words = stringi::stri_replace_all(str = cup_words, regex = "dark chocolate", replacement = "")
    
    cup_words = cup_words |> stringr::str_split(pattern = " |,|\n|/") |> unlist()
    
    cup_words = stringi::stri_replace_all_regex(
      cup_words,
      pattern = c(
        "[[:punct:]]|[0-9]+",
        "raspberries",
        "strawberries",
        "blueberries",
        "^berries$",
        "^cherries",
        "^fruit$|^fruit$|fruitful|fruitiest|fruitiness|fruitier|fruityness|fruits$|fruite$|fruty$",
        "chocolatey",
        "acidic|acidity",
        "bitterness",
        "citrusy",
        "^smokey$|^smoke$",
        "^wood$",
        "^balance$",
        "^roast$|roasty",
        "^toast$|toasty",
        "^nuts$|^nut$",
        "^darker$",
        "sweetness",
        # "^cool$|^cools$",
        "fermenty",
        "^funk$",
        "^love$",
        "juice$",
        "tealike",
        "^apples$",
        "^stone",
        "^wine$",
        "^cloves$",
        "paper",
        "^lemony$|lemonade",
        "herby|herbed|herb|herbs", # replace dark chocolate with nothing
        "peanuts",
        "hazelnuts",
        "grainy|grains",
        "malty",
        "vegetable|vegetables|vegetal|vegetabley",
        "raisins",
        "bananas"
      ),
      replacement = c(
        "",
        "raspberry",
        "strawberry",
        "blueberry",
        "berry",
        "cherry",
        "fruity",
        "chocolate",
        "acid",
        "bitter",
        "citrus",
        "smoky",
        "woody",
        "balanced",
        "roasted",
        "toasted",
        "nutty",
        "dark",
        "sweet",
        # "cooled",
        "fermented",
        "funky",
        "loved",
        "juicy",
        "tea",
        "apple",
        "stone fruit",
        "winey",
        "clove",
        "papery",
        "lemon",
        "herb-like",
        "peanut",
        "hazelnut",
        "grain",
        "malt",
        "vegetative",
        "raisin",
        "banana"
      ),
      vectorize_all = FALSE
    )
    
    cup_words = cup_words[!cup_words %in% remove_words]
    cup_words = cup_words[nchar(cup_words) > 2] # Above removes punctuation
    
    # Add the two words counts
    cup_words = c(cup_words, two_words_counts)
    
    cup_words_table = cup_words |> table()
    
    more_than_x = cup_words_table[cup_words_table > count_min]
    
    return(more_than_x[order(more_than_x, decreasing = TRUE)])
  }

prepare_words <- function(coffee_words, color = "#000") {
  coffee_df = data.frame(coffee_words) |> mutate(cup_words = stringr::str_to_title(cup_words),
                                                 Freq = color)
  coffee_df = coffee_df |> mutate(
    cup_words = recode(
      cup_words,
      "Acid" = "Sour/Acid",
      "Caramel" = "Caramelized",
      "Citrus" = "Citrus Fruit",
      Vegetal = "Vegetative",
      "Papery" = "Stale/Papery",
      "Apple" = "Green Apple"
    )
  )
  return(coffee_df)
}

coffee_sunburst <- function(coffee_cup_df, mcolor = ~color, hover = ~new_label) {
  # coffee_cup_df <-
  #   coffee_cup_df |> mutate(new_label = paste0(
  #     "<b>",
  #     end_name,
  #     ": ",
  #     "</b>",
  #     ifelse(is.na(labels), "", stringr::str_wrap(labels, width = 30))
  #   ))

  coffee_cup_df |> plot_ly() |> add_trace(
    marker = list(colors = mcolor),
    type = 'sunburst',
    ids = ~ ids,
    labels = ~ end_name,
    parents = ~ parents,
    text = hover,
    textinfo = 'label',
    # What shows on the chart
    hovertemplate = '%{text}<extra></extra>',
    domain = list(column = 1),
    maxdepth = 3,
    insidetextorientation = 'radial'
  ) |>
    layout(margin = list(
      l = 0,
      r = 0,
      b = 0,
      t = 0
    )) |>
    as_widget()
}

coffee_words_join <-
  function(coffee_words,
           coffee_cup_df,
           color = "#fee6ce") {
    coffee_join = left_join(coffee_cup_df, coffee_words, by = c("end_name" = "cup_words"))
    
    coffee_join = coffee_join |> mutate(color = ifelse(is.na(Freq), color, Freq))
    return(coffee_join)
  }