#16 more work on cff_reader - add tests and checks for required eleme…

…nts of references

R/a_types.R

@@ -0,0 +1,54 @@
+#' cff references types
+#' @references http://bit.ly/2PRK1Vt
+#' @details cff citation format types for references
+#' @name cff_reference_types
+#' @docType data
+cff_reference_types <- c(
+  "art",
+  "article",
+  "audiovisual",
+  "bill",
+  "blog",
+  "book",
+  "catalogue",
+  "conference",
+  "conference-paper",
+  "data",
+  "database",
+  "dictionary",
+  "edited-work",
+  "encyclopedia",
+  "film-broadcast",
+  "generic",
+  "government-document",
+  "grant",
+  "hearing",
+  "historical-work",
+  "legal-case",
+  "legal-rule",
+  "magazine-article",
+  "manual",
+  "map",
+  "multimedia",
+  "music",
+  "newspaper-article",
+  "pamphlet",
+  "patent",
+  "personal-communication",
+  "proceedings",
+  "report",
+  "serial",
+  "slides",
+  "software",
+  "software-code",
+  "software-container",
+  "software-executable",
+  "software-virtual-machine",
+  "sound-recording",
+  "standard",
+  "statute",
+  "thesis",
+  "unpublished",
+  "video",
+  "website"
+)

R/cff_reader.R

@@ -7,9 +7,36 @@
 #' @family cff
 #' @references CFF format:
 #' https://github.com/citation-file-format/citation-file-format/blob/master/README.md
+#' @details CFF only supports one citation, so `many` will always be
+#' `FALSE`. You can though have many references in your CFF file
+#' associated with the citation.
+#' 
+#' `references` is an optional component in cff files. If included, we check
+#' the following:
+#' - each reference must have the 3 required fields: type, authors, title
+#' - type must be in the allowed set, see [cff_reference_types]
+#' - the elements within authors must each be an entity or person object
+#' https://github.com/citation-file-format/citation-file-format/blob/master/README.md#entity-objects
+#' https://github.com/citation-file-format/citation-file-format/blob/master/README.md#person-objects
+#' - title must be a string
 #' @examples
 #' (z <- system.file('extdata/citation.cff', package = "handlr"))
-#' cff_reader(x = z)
+#' res <- cff_reader(x = z)
+#' res
+#' res$cff_version
+#' res$b_version
+#' res$message
+#' res$id
+#' res$doi
+#' res$title
+#' res$author
+#' res$references
+#' 
+#' # no references
+#' (z <- system.file('extdata/citation-norefs.cff', package = "handlr"))
+#' out <- cff_reader(x = z)
+#' out
+#' out$references
 cff_reader <- function(x) {
   assert(x, "character")
   txt <- if (is_file(x)) yaml::yaml.load_file(x) else yaml::yaml.load(x)
@@ -26,40 +53,91 @@ cff_read_one <- function(x) {
       type = "Person",
       name = pcsp(pcsp(z$`given-names`), pcsp(z$`family-names`)),
       givenName = pcsp(z$`given-names`),
-      familyName = pcsp(z$`family-names`)
+      familyName = pcsp(z$`family-names`),
+      orcid = pcsp(z$orcid)
     )
   })
   state <- if (!is.null(doi)) "findable" else "not_found"
-
+  type <- "SoftwareSourceCode"
   list(
     "cff_version" = x$`cff-version`,
     "message" = x$message,
     # "key" = attr(x, "key"),
     "id" = normalize_doi(doi),
-    # "type" = type,
-    # "bibtex_type" = type,
-    # "citeproc_type" = CFF_TO_CP_TRANSLATIONS[[type]] %||% "misc",
-    # "ris_type" = CFF_TO_RIS_TRANSLATIONS[[type]] %||% "GEN",
-    # "resource_type_general" = SO_TO_DC_TRANSLATIONS[[type]],
+    "type" = type,
     # "additional_type" = CFF_TO_CR_TRANSLATIONS[[type]] %||% type,
+    "citeproc_type" = SO_TO_CP_TRANSLATIONS[[type]] %||% "article-journal",
+    "bibtex_type" = SO_TO_BIB_TRANSLATIONS[[type]] %||% "misc",
+    "ris_type" = SO_TO_RIS_TRANSLATIONS[[type]] %||% "GEN",
+    "resource_type_general" = SO_TO_DC_TRANSLATIONS[[type]] %||% "Other",
+    "identifier" = doi,
     "doi" = doi,
     "b_url" = x$url %||% NULL,
     "title" = x$title %||% NULL,
     "author" = author,
-    "publisher" = x$publisher %||% NULL,
-    "is_part_of" = NULL,
     "date_published" = x$`date-released` %||% NULL,
     "b_version" = x$version %||% NULL,
-    "volume" = x$volume %||% NULL,
-    "first_page" = NULL,
-    "last_page" = NULL,
-    # "description" = list(text = x$abstract %||% NULL && sanitize(x$abstract)),
     "description" = list(text = x$abstract %||% NULL),
-    "license" = list(id = x$copyright %||% NULL),
-    "state" = state
+    "license" = list(id = x$license %||% NULL),
+    "state" = state,
+    "references" = process_refs(x$references)
+    # "description" = list(text = x$abstract %||% NULL && sanitize(x$abstract)),
+    # "first_page" = NULL,
+    # "last_page" = NULL,
+    # "publisher" = x$publisher %||% NULL,
+    # "is_part_of" = NULL,
+    # "volume" = x$volume %||% NULL,
   )
 }
 
+process_refs <- function(w) {
+  if (is.null(w)) return(NULL)
+
+  # check that required fields are given
+  cff_required_nms <- c('type', 'authors', 'title')
+  cff_required_nms_c <- paste0(cff_required_nms, collapse = ", ")
+  for (i in seq_along(w)) {
+    mtch <- all(names(w[[i]]) %in% cff_required_nms)
+    if (!mtch) {
+      stop("reference ", i, " malformed; must have required fields: ",
+        cff_required_nms_c)
+    }
+  }
+
+  # check that title field is a string
+  for (i in w) if (!is.character(i$title)) stop("'title' must be a string")
+
+  # check that type values are within allowed set
+  types <- vapply(w, "[[", "", "type")
+  mtch_type <- types %in% cff_reference_types
+  if (!all(mtch_type)) {
+    stop("these reference types not in allowed set: ",
+      paste0(types[!mtch_type], collapse = ", "),
+      " (see ?cff_reader)", call. = FALSE)
+  }
+
+  # check that authors is a list of type entity's or person's
+  auths <- unlist(lapply(w, "[[", "authors"), FALSE)
+  for (i in auths) {
+    if (!is_cff_entity(i) && !is_cff_person(i))
+      stop("each element in 'authors' must be of type entity or person\n",
+        "  see ?cff_reader Details", call. = FALSE)
+  }
+
+  return(w)
+}
+
+# check that ALL list elements are named
+is_named <- function(x) all(nzchar(names(x)))
+is_cff_entity <- function(x) {
+  is.list(x) && is_named(x) && "name" %in% names(x)
+}
+is_cff_person <- function(x) {
+  is.list(x) &&
+  is_named(x) &&
+  all(c("family-names", "given-names") %in% names(x))
+}
+
 # CFF_TO_CP_TRANSLATIONS <- list(
 #   article = "article-journal",
 #   phdthesis = "thesis"

R/client.R

@@ -1,4 +1,4 @@
-handlr_readers <- c('citeproc', 'ris', 'bibtex', 'codemeta')
+handlr_readers <- c('citeproc', 'ris', 'bibtex', 'codemeta', 'cff')
 handlr_writers <- c('citeproc', 'ris', 'bibtex', 'schema_org',
   'rdfxml', 'codemeta')
 
@@ -102,6 +102,15 @@ handlr_writers <- c('citeproc', 'ris', 'bibtex', 'schema_org',
 #' x$read("codemeta")
 #' x$parsed
 #' x$write("codemeta")
+#' 
+#' # cff
+#' (z <- system.file('extdata/citation.cff', package = "handlr"))
+#' (x <- HandlrClient$new(x = z))
+#' x$path
+#' x$format_guessed
+#' x$read("cff")
+#' x$parsed
+#' x$write("codemeta")
 #'
 #' # > 1
 #' z <- system.file('extdata/citeproc-many.json', package = "handlr")
@@ -201,6 +210,7 @@ HandlrClient <- R6::R6Class(
         ris = ris_reader(self$path %||% self$string, ...),
         bibtex = bibtex_reader(self$path %||% self$string, ...),
         codemeta = codemeta_reader(self$path %||% self$string, ...),
+        cff = cff_reader(self$path %||% self$string, ...),
         stop("format must be one of ",
           paste(handlr_readers, collapse = ", "))
       )
@@ -284,6 +294,10 @@ HandlrClient <- R6::R6Class(
               # && self$ext == "bib"
             ) {
               return("bibtex")
+            } else {
+              # decide between ris and cff
+              fmt <- if (!is.null(cff_reader(x)$cff_version)) "cff" else "ris"
+              return(fmt)
             }
           }
         }

inst/extdata/citation-norefs.cff

@@ -0,0 +1,10 @@
+cff-version: 1.1.0
+message: If you use this software, please cite it as below.
+authors:
+  - family-names: Druskat
+    given-names: Stephan
+    orcid: https://orcid.org/0000-0003-4925-7248
+title: "My Research Software"
+version: 2.0.4
+doi: 10.5281/zenodo.1234
+date-released: 2017-12-18

inst/extdata/citation.cff

@@ -8,3 +8,16 @@ title: "My Research Software"
 version: 2.0.4
 doi: 10.5281/zenodo.1234
 date-released: 2017-12-18
+references:
+  - type: book
+    authors:
+      - family-names: Doe
+        given-names: Jane
+      - name: "Foo Bar Working Group"
+        website: https://foo-bar.com
+    title: The science of citation
+  - type: software
+    authors:
+      - family-names: Doe
+        given-names: John
+    title: Software Citation Tool