HATEOAS feature

NAMESPACE

@@ -11,6 +11,7 @@ export(get_endpoints)
 export(get_fields)
 export(get_ok_pk)
 export(qry_funs)
+export(retrieve_linked_data)
 export(search_pv)
 export(unnest_pv_data)
 export(with_qfuns)

R/search-pv.R

@@ -299,3 +299,41 @@ search_pv <- function(query,
 
   res
 }
+
+#' Get Linked Data
+#'
+#' Some of the endpoints now returns HATEOAS style links to get more data.
+#' ex "inventor": "https://search.patentsview.org/api/v1/inventor/252373/"
+#' 
+#' @param url The link that was returned by the API on a previous call
+#'
+#' @param api_key API key. See \href{https://patentsview.org/apis/keyrequest}{
+#'  Here} for info on creating a key.
+#'
+#' @return A character vector with field names, same as search_pv()
+#'
+#' @examples
+#' \dontrun{
+#'
+#' retrieve_linked_data(
+#'   "https://search.patentsview.org/api/v1/cpc_subgroup/G01S7:4811/"
+#'  )
+#' }
+#'
+#' @export
+retrieve_linked_data <- function(url,
+                                 api_key = Sys.getenv("PATENTSVIEW_API_KEY")
+                                ) {
+
+  # Don't sent the API key to any domain other than patentsview.org
+  if (!grepl("^https://[^/]*\\.*patentsview.org/", url)) {
+    stop2("retrieve_linked_data is only for patentsview.org urls - sends API key")
+  }
+
+  # Go through one_request, it would resend on 429 too many requests 
+  # The API doesn't seeem to mind ?q=&f=&o=&s= appended to the url
+  res = one_request("GET", "", url, list(), api_key)
+  res <- process_resp(res)
+  res
+}
+

tests/testthat/test-search-pv.R

@@ -114,3 +114,61 @@ test_that("Throttled requests are automatically retried", {
 
   expect_identical(built_singly, result_all)
 })
+
+test_that("We won't expose the user's patentsview API key to random websites", {
+  skip_on_cran()
+  skip_on_ci()
+
+  # We will try to call the api that tells us who is currently in space
+  in_space_now_url <- "http://api.open-notify.org/astros.json"
+
+  expect_error(
+    retrieve_linked_data(in_space_now_url)
+  )
+})
+
+
+test_that("We can call all the legitimate HATEOAS endpoints", {
+  skip_on_cran()
+  skip_on_ci()
+
+  # Call the new, Get only endpoints that don't accept q: s:, o:, f:, parameters
+  # The links are returned fully qualified, like below, from some of the endpoints
+  # These queries retrieve one specific row
+  single_item_queries <- c(
+    "https://search.patentsview.org/api/v1/assignee/10/",
+    "https://search.patentsview.org/api/v1/cpc_group/A01B/",
+    "https://search.patentsview.org/api/v1/cpc_subgroup/G01S7:4811/",
+    "https://search.patentsview.org/api/v1/cpc_subsection/A01/",
+    "https://search.patentsview.org/api/v1/inventor/10/",
+    "https://search.patentsview.org/api/v1/nber_category/1/",
+    "https://search.patentsview.org/api/v1/nber_subcategory/11/",
+    "https://search.patentsview.org/api/v1/patent/10757852/",
+    "https://search.patentsview.org/api/v1/uspc_mainclass/30/",
+    "https://search.patentsview.org/api/v1/uspc_subclass/30:100/"
+  )
+
+  # These queries can return more than a single row
+  multi_item_queries <- c(
+    "https://search.patentsview.org/api/v1/application_citation/10966293/",
+    "https://search.patentsview.org/api/v1/patent_citation/10966293/"
+  )
+
+  x <- lapply(single_item_queries, function(q) {
+    Sys.sleep(2)
+    print(q)
+    j <- retrieve_linked_data(q)
+
+    # here all the total hits should be 1
+    expect_equal(j$query_results$total_hits, 1)
+  })
+
+  x <- lapply(multi_item_queries, function(q) {
+    Sys.sleep(2)
+    print(q)
+    j <- retrieve_linked_data(q)
+
+    # here all the total hits should be 1 or more rows
+    expect_true(j$query_results$total_hits >= 1)
+  })
+})