MoH_media_releases

# works through 
# https://www.health.govt.nz/news-media/media-releases?page=1 and subsequent pages

library(rvest)
library(dplyr)
library(stringr)

table_fetch <- function(x){
  url_loc <- paste0("https://www.health.govt.nz",x)
  page_obj <- url_loc %>% read_html() 
  table_objs <- page_obj %>% html_nodes("table")
  update_date <- page_obj %>% 
    html_nodes("span.date-display-single") %>% 
    html_attr("content")
  # recent tableformat reports have the cases table and a footer table
  # so assumng 2 tables with cases being table 1
  if(length(table_objs) < 2) {
    return(data.frame(page_url = x, 
                      update_datetime = update_date,
                      stringsAsFactors = FALSE))
  }
  tbl <- (table_objs %>% html_table(fill = TRUE))[[1]] 
  tbl$page_url <- x
  tbl$update_datetime = update_date
  return(tbl)
}


getlinks <- function(x){
  webpage <- paste0("https://www.health.govt.nz/news-media/media-releases?page=",x)
  links <- unique(webpage %>% read_html() %>% html_nodes("a") %>% html_attr("href"))
  links <- links[str_starts(links, fixed("/news-media/media-releases/")) & !is.na(links)]
  return(links)
}

pages_back_to_go <- 15
all_links <- lapply(1:pages_back_to_go, getlinks) %>% unlist() %>% unique()
all_table_1s <- lapply(all_links, table_fetch)
aggregate_tables <- bind_rows(all_table_1s)
write.csv(aggregate_tables, file="MoHcheck_by_hand.csv", row.names=FALSE)