Merge branch 'develop' for new release

Author: dwinter

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: rentrez
-Version: 1.2.1
-Date: 2018-02-12
+Version: 1.2.2
+Date: 2019-05-02
 Title: 'Entrez' in R
 Authors@R: c(
  person("David", "Winter", role=c("aut", "cre"), 

NEWS

@@ -1,3 +1,14 @@
+Version 1.2.2
+------------------
+Maintenance release containing a number of bug fixes. 
+
+* A fix #127 forces curl to use http/1.1, as NCBI produces errors when sing http/2
+  (thanks @hammer for bug report and fix)
+* fix #131 stops pubmed records with reference sections production > 1
+  value for the 'pmid' field. Thanks @agbarnett for the report.
+* Added clarification around the expectd input for `extract_from_esummary` to
+  the docs.
+
 Version 1.2.1
 ------------------
 

R/base.r

@@ -49,6 +49,23 @@ make_entrez_query <- function(util, config, interface=".fcgi?", by_id=FALSE, deb
         return( list(uri = uri, args=args ) )
     }
 
+    #Seemingly, NCBI moved to https but not http v2.0?
+    # (thatnks Jeff Hammerbacher for report/solution)
+    #
+    # if no httr::config was passed we will add one
+    if(is.null(config)){
+         config = httr::config(http_version = 2)
+    # otherwise add http version, checkign we aren't overwriting something
+    # passed in (seems unlikely, but worth checking?)
+    # 
+    } else {
+        if ("http_version" %in% names(config$options)) {
+            warning("Over-writing httr config options for 'http_version', as NCBI servers require v1.1")
+        }
+        config$options$http_version <- 2        
+    }
+
+    
     if(length(args$id) > 200){ 
         response <- httr::POST(uri, body=args, config= config)
     } else {

R/entrez_summary.r

@@ -81,9 +81,11 @@ entrez_summary <- function(db, id=NULL, web_history=NULL,
 
 #' Extract elements from a list of esummary records
 #'@export
-#'@param esummaries A list of esummary objects
+#'@param esummaries Either an esummary or an esummary_list (as returned by
+#' entrez_summary).
 #'@param elements the names of the element to extract
 #'@param simplify logical, if possible return a vector
+#'@seealso \code{\link{entrez_summary}} for examples of this function in action.
 #'@return List or vector containing requested elements 
 extract_from_esummary <- function(esummaries, elements, simplify=TRUE){
     UseMethod("extract_from_esummary", esummaries)

R/parse_pubmed_xml.r

@@ -58,8 +58,8 @@ parse_one_pubmed <- function(paper){
     res$issue <- get_value(".//JournalIssue/Issue")
     res$pages <- get_value(".//MedlinePgn")
     res$key_words <- get_value(".//DescriptorName")
-    res$doi <- get_value(".//ArticleId[@IdType='doi']")
-    res$pmid <- get_value(".//ArticleId[@IdType='pubmed']")
+    res$doi <- get_value(".//PubmedData/ArticleIdList/ArticleId[@IdType='doi']")
+    res$pmid <- get_value(".//PubmedData/ArticleIdList/ArticleId[@IdType='pubmed']")
     res$abstract <- get_value(".//AbstractText")
 
     structure(res, class="pubmed_record", empty=FALSE) 

man/extract_from_esummary.Rd

@@ -1,4 +1,5 @@
 context("Network")
 test_that("The NCBI is contactable from this comptuter /",{
-    expect_true(!httr::http_error("http://eutils.ncbi.nlm.nih.gov/entrez/eutils/"))
+    response <- httr::GET("https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi")
+    expect_lt(response$status_code, 400)
 })

tests/testthat/test_net.r

@@ -10,6 +10,10 @@ parsed_raw <- parse_pubmed_xml(raw_rec)
 parsed_rec <- parse_pubmed_xml(xml_rec)
 parsed_multi <- parse_pubmed_xml(multi_rec)
 
+multi_pmid_eg <- parse_pubmed_xml(
+       entrez_fetch(db="pubmed", id='29743284', rettype="xml")
+)
+
 test_that("pubmed file parsers work",{
     expect_that(raw_rec, is_a("character"))
 
@@ -27,6 +31,12 @@ test_that("pubmed file parsers work",{
     # re-introduced there will be 25 authors in each record and this will fail
     expect_that(length(parsed_multi[[1]]$authors), equals(1))
 
+    # Bug #131 related to extracting all PMIDs from a pubmed xml that contained
+    # references, not just the one true pm od teh artcle. Testing we don't
+    # regress 
+
+    expect_that(length(multi_pmid_eg$pmid), equals(1))
+
 })
 
 test_that("we can print pubmed records", {