Fixed path issues for donttest

R/grab_crosswalk.R

@@ -18,54 +18,68 @@
 #' # Download and load current geographic crosswalk for Alaska
 #' alaska_xwalk <- grab_crosswalk('AK')
 #' 
-#' # Download and load current geographic crosswalk for New England
-#' 
+#' # Download and load current geographic crosswalk for New England#' 
 #' new_england_xwalk <- grab_crosswalk(c('CT', 'RI', 'MA', 'VT', 'NH', 'ME'))
 #' 
 #' }
 
 grab_crosswalk <- function(state, 
-                           download_dir = file.path(tools::R_user_dir("lehdr", which="cache"))){
+                           download_dir = file.path(tools::R_user_dir("lehdr",
+                                                                      which="cache"))){
 
   states <- tolower(state)
 
   urls <- glue::glue("https://lehd.ces.census.gov/data/lodes/LODES8/{states}/{states}_xwalk.csv.gz")
 
   for (url in urls){
     httr::stop_for_status(httr::HEAD(url),
-                          paste0("download crosswalk. Data for this state was not found on LODES.\n  URL: ", url))
+                          paste0("Data for this state was not found on LODES.\nURL: ", url))
   }
 
-
   vdownload_xwalk(url = urls, download_dir = download_dir) %>%
     vread_xwalk() %>%
     dplyr::bind_rows()
 
 }
 
 download_xwalk <- function(url, download_dir){
+  # Set download directory, check for cache
+  download_dir <- path.expand(download_dir)
+  if (!dir.exists(download_dir))
+    dir.create(download_dir, recursive=TRUE)
+  fil <- normalizePath(file.path(download_dir, basename(url)), mustWork = FALSE)
 
-  fil <- file.path(path.expand(download_dir), basename(url))
-
+  # Read from FTP site
   if (file.exists(fil)) {
-
     message(glue::glue("Cached version of file found in {fil}\n"))
-
   } else {
-
-    message(glue::glue("Downloading {url} to {fil} now..."))
+    message(glue::glue("Downloading {url} to {fil}"))
     res <- httr::GET(url, httr::write_disk(fil))
+    message(glue::glue("Download complete."))
   }
 
   return(fil)
 
 }
 
 read_xwalk <- function(filepath){
-
   res <- suppressMessages(readr::read_csv(filepath, col_types = readr::cols(.default = 'c')))
+  download_dir <- dirname(filepath)
+
+  # Remove cached files now that they're read in
+  # Unlink returns 0 for success, 1 for failure 
+  if(unlink(filepath)) {
+    message(glue::glue("Could not clear crosswalk cache."))
+  } else {
+    message(glue::glue("Crosswalk cache cleared."))
+    # Now check to see if the cache directory is empty, remove it if it is
+    if(length(list.files(download_dir)) == 0) {
+      unlink(download_dir, recursive = TRUE)
+    }
+
+  }
 
-  res
+  return(res)
 }
 
 vdownload_xwalk <- Vectorize(download_xwalk, vectorize.args = 'url')

R/lehdr.R

@@ -79,8 +79,9 @@ grab_lodes <- function(state, year,
                                    "SE03", "SI01", "SI02", "SI03"),  
                        agg_geo = c("block", "bg", "tract", "county", "state"),
                        state_part = c("","main","aux"), 
-                       download_dir = file.path(tools::R_user_dir("lehdr", which="cache")),
-                       use_cache = FALSE) { # Thanks Kyle Walker for this
+                       download_dir = normalizePath(file.path(tools::R_user_dir("lehdr", which="cache")), 
+                                                    mustWork = FALSE),
+                       use_cache = FALSE) { # Thanks to Kyle Walker for this
 
   if (length(state) > 1 | length(year) > 1) {
     ## Handle multiple states x years
@@ -114,7 +115,7 @@ grab_lodes <- function(state, year,
   # If someone uses od, but doesn't set state_part
   if(lodes_type == "od" && !(state_part %in% c("main","aux"))) {
     state_part <- "main"
-    warning("state_part is required when setting lodes_type =\"od\", defaulting to state_part=\"main\"")
+    warning("'state_part' is required when setting lodes_type =\"od\", defaulting to state_part=\"main\"")
   }
 
   # Setup col_types and url variables, used in the if statement