From 0ecdcdeb7b515ce27e0a3dbc2905e5d72a058d3c Mon Sep 17 00:00:00 2001
From: Diego H <diego.hernangomezherrero@gmail.com>
Date: Wed, 6 Apr 2022 13:28:35 +0200
Subject: [PATCH] Improving package description auto-linking (#1315)

Fixes #1265. Fixes #1164.
---
 NEWS.md                              |  3 ++
 R/object-defaults.R                  |  1 +
 R/object-package.R                   | 29 ++++++++++++++++++
 tests/testthat/test-object-package.R | 44 ++++++++++++++++++++++++++++
 4 files changed, 77 insertions(+)

diff --git a/NEWS.md b/NEWS.md
index 799e2186a..043e5992d 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -1,5 +1,8 @@
 # roxygen2 (development version)
 
+* DOIs, arXiv links, and urls in the `Description` field of the `DESCRIPTION`
+  are now converted to the appropriate Rd markup (@dieghernan, #1265, #1164).
+
 * Arguments containing syntactically significant whitespace (e.g anonymous
   functions) are now wrapped correctly (#1281).
 
diff --git a/R/object-defaults.R b/R/object-defaults.R
index 59736e82a..a48ce69dc 100644
--- a/R/object-defaults.R
+++ b/R/object-defaults.R
@@ -41,6 +41,7 @@ object_defaults.package <- function(x) {
   desc <- x$value$desc
 
   description <- as.character(desc$Description)
+  description <- package_url_parse(description)
   logo_path <- file.path(x$value$path, "man", "figures", "logo.png")
   if (file.exists(logo_path)) {
     fig <- "\\if{html}{\\figure{logo.png}{options: align='right' alt='logo' width='120'}}"
diff --git a/R/object-package.R b/R/object-package.R
index c91cf278a..26f6e1397 100644
--- a/R/object-package.R
+++ b/R/object-package.R
@@ -374,3 +374,32 @@ itemize <- function(header, x) {
     "}\n"
   )
 }
+
+package_url_parse <- function(x) {
+  # <doi:XX.XXX> -> \doi{XX.XXX} to avoid CRAN Notes, etc.
+  x <- str_replace_all(x, "<(doi|DOI):(.*?)>", function(match) {
+    match <- str_remove_all(match, "^<(doi|DOI):|>$")
+    paste0("\\doi{", escape(match), "}")
+  })
+
+  # <http:XX.XXX> -> \url{http:XX.XXX}
+  x <- str_replace_all(x, "<(http|https):\\/\\/(.*?)>", function(match) {
+    match <- str_remove_all(match, "^<|>$")
+    paste0("\\url{", escape(match), "}")
+  })
+
+  # <arxiv:XXX> -> \href{https://arxiv.org/abs/XXX}{arXiv:XXX}
+  # https://github.com/wch/r-source/blob/trunk/src/library/tools/R/Rd2pdf.R#L149-L151
+  patt_arxiv <- "<(arXiv:|arxiv:)([[:alnum:]/.-]+)([[:space:]]*\\[[^]]+\\])?>"
+  x <- str_replace_all(x, patt_arxiv, function(match) {
+    match <- str_remove_all(match, "^<(arXiv:|arxiv:)|>$")
+    # Special cases has <arxiv:id [code]>.
+    # See https://CRAN.R-project.org/package=ciccr
+    # Extract arxiv id, split by space
+    arxiv_id <- str_split_fixed(match, " ", n = 2)[, 1]
+
+    paste0("\\href{https://arxiv.org/abs/", escape(arxiv_id), "}{arXiv:", match, "}")
+  })
+
+  x
+}
diff --git a/tests/testthat/test-object-package.R b/tests/testthat/test-object-package.R
index 76bccd604..6b6b44666 100644
--- a/tests/testthat/test-object-package.R
+++ b/tests/testthat/test-object-package.R
@@ -25,3 +25,47 @@ test_that("can convert DOIs in url", {
     "\\doi{10.5281/zenodo.1485309}"
   )
 })
+
+test_that("can autolink urls on package Description", {
+  expect_equal(
+    package_url_parse("x <https://x.com> y"),
+    "x \\url{https://x.com} y"
+  )
+  expect_equal(
+    package_url_parse("x <https://x.com/%3C-%3E> y"),
+    "x \\url{https://x.com/\\%3C-\\%3E} y"
+  )
+})
+
+test_that("can autolink DOIs", {
+  expect_equal(package_url_parse("x <doi:abcdef> y"), "x \\doi{abcdef} y")
+  expect_equal(package_url_parse("x <DOI:abcdef> y"), "x \\doi{abcdef} y")
+  expect_equal(package_url_parse("x <DOI:%3C-%3E> y"), "x \\doi{\\%3C-\\%3E} y")
+})
+
+test_that("can autolink arxiv", {
+  expect_equal(
+    package_url_parse("x <arXiv:abc> y"),
+    "x \\href{https://arxiv.org/abs/abc}{arXiv:abc} y"
+  )
+  expect_equal(
+    package_url_parse("x <arxiv:abc> y"),
+    "x \\href{https://arxiv.org/abs/abc}{arXiv:abc} y"
+  )
+  expect_equal(
+    package_url_parse("x <arxiv:abc [def]> y"),
+    "x \\href{https://arxiv.org/abs/abc}{arXiv:abc [def]} y"
+  )
+})
+
+test_that("autolink several matching patterns", {
+  text <- "url <http://a.com> doi <doi:xx> arxiv <arXiv:xx>"
+  expect_equal(
+    package_url_parse(text),
+    paste(
+      "url \\url{http://a.com}",
+      "doi \\doi{xx}",
+      "arxiv \\href{https://arxiv.org/abs/xx}{arXiv:xx}"
+    )
+  )
+})