about summary refs log tree commit diff
path: root/examples/publication.scm
diff options
context:
space:
mode:
authorMunyoki Kilyungi2023-08-21 14:54:21 +0300
committerMunyoki Kilyungi2023-08-21 14:56:57 +0300
commit51b3c0548c98e0bc05e11a89cbf6b75d31b9f8d5 (patch)
treeab3d7c6f589ed8480f0a9d451566681bcfd8eaaf /examples/publication.scm
parent849874fdfe11003f05abe5f82efde974a8c8a388 (diff)
downloadgn-transform-databases-51b3c0548c98e0bc05e11a89cbf6b75d31b9f8d5.tar.gz
Remove "dump-" prefix
Signed-off-by: Munyoki Kilyungi <me@bonfacemunyoki.com>
Diffstat (limited to 'examples/publication.scm')
-rwxr-xr-xexamples/publication.scm81
1 files changed, 81 insertions, 0 deletions
diff --git a/examples/publication.scm b/examples/publication.scm
new file mode 100755
index 0000000..313ee96
--- /dev/null
+++ b/examples/publication.scm
@@ -0,0 +1,81 @@
+#! /usr/bin/env guile
+!#
+
+(use-modules (srfi srfi-1)
+             (srfi srfi-26)
+             (ice-9 match)
+             (ice-9 regex)
+             (dump strings)
+             (dump sql)
+             (dump triples)
+             (dump special-forms))
+
+
+
+(define %connection-settings
+    (call-with-input-file (list-ref (command-line) 1)
+			  read))
+
+
+
+(define-transformer publication
+  (tables (Publication))
+  (triples
+      (let ((pmid (field
+                   ("IF(Publication.PubMed_ID IS NULL, '', CONVERT(Publication.PubMed_Id, INT))"
+                    pmid)))
+            (publication-id (field Publication Id)))
+        (if (string-null? pmid)
+            (string->identifier "unpublished"
+                                (number->string publication-id))
+            (ontology 'pubmed: pmid)))
+    (set rdf:type 'fabio:ResearchPaper)
+    (set fabio:hasPubMedId
+         (ontology 'pubmed: (field ("IFNULL(PubMed_ID, '')" pubmedId))))
+    (set dct:title (delete-substrings (field Publication Title)
+                                      "Unknown"))
+    (set fabio:Journal (delete-substrings (field Publication Journal)
+                                          "Unknown"))
+    (set prism:volume (delete-substrings (field Publication Volume)
+                                                "Unknown"))
+    (set fabio:page (delete-substrings (field Publication Pages)
+                                       "Unknown"))
+    (set prism:publicationDate (annotate-field
+                                (delete-substrings (field Publication Month)
+                                                   "Unknown")
+                                '^^xsd:gMonth))
+    (set fabio:hasPublicationYear
+         (annotate-field
+          (field
+           ("IF(Publication.Year = 0, NULL, Publication.Year)" Year))
+          '^^xsd:gYear))
+    (multiset dct:creator
+              ;; The authors field is a comma
+              ;; separated list. Split it.
+              (map string-trim (string-split (sanitize-rdf-string (field Publication Authors)) #\,)))
+    (set dct:abstract
+         (sanitize-rdf-string
+          (field Publication Abstract)))))
+
+
+
+(with-documentation
+ (name "Publications Metadata")
+ (connection %connection-settings)
+ (table-metadata? #f)
+ (prefixes
+  '(("gnt:" "<http://genenetwork.org/terms/>")
+    ("fabio:" "<http://purl.org/spar/fabio/>")
+    ("dct:" "<http://purl.org/dc/terms/>")
+    ("prism:" "<http://prismstandard.org/namespaces/basic/2.0/>")
+    ("gn:" "<http://genenetwork.org/id/>")
+    ("gnc:" "<http://genenetwork.org/category/>")
+    ("pubmed:" "<http://rdf.ncbi.nlm.nih.gov/pubmed/>")
+    ("rdfs:" "<http://www.w3.org/2000/01/rdf-schema#>")
+    ("xsd:" "<http://www.w3.org/2001/XMLSchema#>")
+    ("rdf:" "<http://www.w3.org/1999/02/22-rdf-syntax-ns#>")))
+ (inputs
+  (list publication))
+ (outputs
+  '(#:documentation "./docs/publication.md"
+    #:rdf "./verified-data/publication.ttl")))