check a page or blog post for valid urls

check-urls.R

library(dplyr)
library(purrr)
library(stringr)
library(rvest)
library(crul)

base_url <- "https://deploy-preview-NN--USERNAME.netlify.app/"
page_url <- file.path(base_url, "blog/post")
page <- read_html(page_url)

res <-
  page %>%
  xml_nodes("a, script, img, figure") %>%
  xml_attrs() %>%
  map_dfr(as.list) %>%
  select(src, href) %>%
  tidyr::pivot_longer(
    everything(),
    names_to = "drop",
    values_to = "href",
    values_drop_na = TRUE
  ) %>%
  count(href) %>%
  filter(!str_detect(href, "^mailto:")) %>%
  mutate(
    url = case_when(
      str_detect(href, "^[/]") ~ file.path(base_url, str_remove(href, "/")),
      !str_detect(href, "^http") ~ file.path(page_url, href),
      TRUE ~ href
    ),
    ok = map_lgl(url, crul::ok)
  )

res %>% filter(!ok)