library(tidyverse)

# Create tree of categories and subcategories
categories <-
  list(
    "category_1" = c(
      "subcategory_1",
      "subcategory_2",
      "subcategory_3",
      "subcategory_4",
      "subcategory_5"
    ),
    "category_2" = c(
      "subcategory_6",
      "subcategory_7",
      "subcategory_8",
      "subcategory_9",
      "subcategory_10"
    ),
    "category_3" = c(
      "subcategory_11",
      "subcategory_12",
      "subcategory_13",
      "subcategory_14",
      "subcategory_15"
    )
  )


# Create dataframe of mixed categories and cateogires `subcategory`) along with
# expected category (parent)
df <-
  tibble(
    id = seq(1:5),
    subcategory = c(
      "category_2",
      "subcategory_13",
      "subcategory_4",
      "category_3",
      "category_1"
    ),
    expected_category = c(
      "category_2",
      "category_3",
      "category_1",
      "category_3",
      "category_1"
    )
  )


#' @title find_cateogry
#' @description return parent categories of subcategories
#' @details Given a fector of catgories will return a vector of aprent
#'   categories.
find_category <- function(x) {
  y <- x %in% names(categories)
  y[y == TRUE] <- x[x %in% names(categories)]
  y[y == "FALSE"] <- map_int(x[y == FALSE], ~grep(.x, categories)) %>%
    names(categories)[.]
  return(y)
}

# debug(find_category)

# Test it returns vector of length 5
find_category(df$subcategory)

df %>%
  mutate(category = find_category(subcategory),
         match = if_else(category == expected_category, TRUE, FALSE))