vignette updates

Author: mustberuss

vignettes/citation-networks.Rmd.orig

@@ -47,8 +47,10 @@ pat_lst <- unnest_pv_data(pv_out$data, pk = "patent_id")
 citing_query <- qry_funs$eq(patent_id = patent_ids)
 cited_query <- qry_funs$eq(citation_patent_id = patent_ids)
 
-# Create a list of fields to pull from the API.  Note that, with a recent API chanage,
-# these aren't the only fields we receive back from the API
+# Create a list of fields to pull from the API.  Note that, with a recent chanage
+# to the R package, these aren't the only fields we'll receive back from the API.
+# This is due to the way that the API's paging works now.  There is more about this
+# in the [Result Set Paging](result-set-paging.html) vignette.
 fields <- c(
   "patent_id",
   "citation_patent_id"
@@ -64,7 +66,7 @@ res2 <- search_pv(cited_query,
   endpoint = "patent/us_patent_citation", method = "POST", size = 1000
 )
 
-# Unnest the data found in the two lists of columns
+# Unnest the data found in the three lists of columns
 # (citation_sequence being a field we didn't request)
 res_lst <- unnest_pv_data(res$data, pk = "patent_id")
 res_lst
@@ -88,7 +90,7 @@ pat_title <- function(title, number) {
 edges <-
   res_lst$us_patent_citations %>%
   semi_join(x = ., y = ., by = c("citation_patent_id" = "patent_id")) %>%
-  select(patent_id, citation_patent_id) %>%  # discard citation_sequence we didn't ask for
+  select(-citation_sequence) %>%  # discard citation_sequence that we don't need here
   set_colnames(c("from", "to"))
 
 nodes <-

vignettes/getting-started.Rmd.orig

@@ -26,9 +26,11 @@ Let's start with a basic example of how to use the package's primary function, `
 ```{r}
 library(patentsview)
 
+fields <- c("patent_id", "patent_date", "patent_title")
+
 search_pv(
   query = '{"_gte":{"patent_date":"2007-01-01"}}',
-  endpoint = "patent"
+  endpoint = "patent", fields = fields
 )
 ```
 
@@ -102,10 +104,12 @@ By default, `search_pv()` returns 1,000 records per page and only gives you the
 You can download all pages of output in one call by setting `all_pages = TRUE`. This will set `size` equal to 1,000 and loop over all pages of output:
 
 ```{r}
+fields <- c("patent_id", "inventors.inventor_name_last", "inventors.inventor_name_first")
 search_pv(
   query = qry_funs$eq(inventors.inventor_name_last = "Chambers"),
-  all_pages = TRUE, size = 1000
+  all_pages = TRUE, size = 1000, fields = fields
 )
+
 ```
 See the [result set paging vignette](result-set-paging.html) for information on custom paging.
 

vignettes/state-of-the-api.Rmd

@@ -504,6 +504,54 @@ but the build will silently fail on r-universe.  Don't ask how I know that.
 there isn't the 100,000 row limitation now.  The API's ```after``` is now exposed 
 in search_pv() so users could do their own paging. See the new [paging vignette](result-set-paging.html)
 
+## Odd Demo
+   Here we demonstrate a weird interaction between the API and R package when no sort parmeter
+is specified.  
+
+```{r}
+fields <- c("patent_id", "patent_date", "patent_title")
+
+search_pv(
+  query = '{"_gte":{"patent_date":"2007-01-01"}}',
+  endpoint = "patent", fields = fields
+)
+```
+
+Note that the first patent_date returned above is "2020-04-21" and that what gets 
+sent to the API is `&s=[]&o=`
+
+```{r}
+url <- httr2::last_request()$url
+cat(utils::URLdecode(url))
+```
+
+
+Now we'll send in our own GET with `&s=&o` to see that the API return is different,
+seemingly applying a default sort?
+
+```{r}
+library(httr2)
+
+  url <- sub("&s=\\[\\]", "&s=", url)
+  cat(utils::URLdecode(url))
+  api_key <- Sys.getenv("PATENTSVIEW_API_KEY")
+
+  req <- httr2::request(url) |>
+    httr2::req_method("GET")
+
+  resp <- req |>
+    httr2::req_user_agent("https://github.com/ropensci/patentsview") |>
+    httr2::req_retry(max_tries = 20) |> # automatic 429 Retry-After
+    httr2::req_headers("X-Api-Key" = api_key, .redact = "X-Api-Key") |>
+    httr2::req_perform()
+
+  json <- httr2::last_response() |> httr2::resp_body_json()
+  json$patents[[1]]
+
+```
+For a while POSTs and GETs behaved differently when a sort wasn't specified.  Currently
+an API bug prevents seeing what the behavior is on POSTS. 
+
 [^1]: Observation sent to the API team.
 [^2]: Observation sent to the API team.
 [^3]: There isn't a data dictionary for the API so I suggested they create one.