A function to get around the youtube api. Scrape number of videos for each account ( up to 20000) without any API access

get_accessible_videos.R

# video_counts <- c("UCtahKSp0CdvDv8CMS7RXsiQ", "UC3XTzVzaHQEd30rQbuvCtTQ", "UC-i2qb4sL10OdR6PyMEU_5Q", 
#        "UCaXkIU1QidjPwiAYu6GcHjg", "UCn8zNIfYAQNdrFRrr8oibKw") %>%
#   purrr::map_dfr(get_accessible_videos)


get_accessible_videos <- function(channel_id){
  data <- glue::glue("https://www.youtube.com/channel/{channel_id}/videos") %>%
    xml2::read_html() %>%
    rvest::html_nodes("script") %>%
    .[28] %>%
    rvest::html_text()
  
  n <- stringr::str_count(data, "gridVideoRenderer")
  
  if(n > 20){
    playlist <- stringr::str_extract(data, '/playlist\\?.*?\"')
    
    play <- glue::glue("https://www.youtube.com{playlist}") %>%
      stringr::str_replace_all("\\\\u0026", "&") %>%
      stringr::str_extract("(?<=list=).*?(?=&)")
    
    n <- glue::glue("https://www.youtube.com/watch?v=rBu0BRTx2x8&list={play}&index=1") %>%
      xml2::read_html() %>%
      rvest::html_nodes("script") %>%
      as.character %>%
      stringr::str_subset("videoCountText") %>%
      stringr::str_extract("videoCountText.{100}") %>%
      stringr::str_extract('(?<=\\"\\}\\,\\{"text"\\:")\\d+') %>%
      as.numeric
    
  }
    
  return(tibble::tibble(channel_id = channel_id, n_videos = n))
}

# 
# parse_youtube_data <- function(yt_data){
#   yt_data %>%
#     str_remove('(.|\\s)*?(?=\\{"gridVideoRenderer")') %>%
#     str_remove('(?<=\\]\\}\\})(\\]\\,"(trackingParams|continuations)")(.|\\s)+$') %>%
#     str_split('\\{"gridVideoRenderer":') %>%
#     .[[1]] %>% 
#     .[-1] %>%
#     str_remove("\\}\\,?$") %>% #bashR::simule_map(30)
#     imap_dfr(~{
#       .x %>%
#         jsonlite::fromJSON() %>%
#         rlist::list.flatten() %>%
#         imap_dfc(~{
#           if(length(.x) > 1){
#             a = list(.x)
#           } else {
#             a = .x
#           }
#           
#           tibble(a) %>%
#             set_names(.y)
#         })
#     })
#   
# }