From 2684d58401c694e30551f424eb2404f3a5674769 Mon Sep 17 00:00:00 2001
From: Munyoki Kilyungi
Date: Thu, 17 Aug 2023 16:54:18 +0300
Subject: Update how phenotypes are transformed

Signed-off-by: Munyoki Kilyungi <me@bonfacemunyoki.com>
---
 examples/dump-phenotype.scm | 51 +++++++++++++--------------------------------
 1 file changed, 15 insertions(+), 36 deletions(-)

(limited to 'examples')

diff --git a/examples/dump-phenotype.scm b/examples/dump-phenotype.scm
index ed10b24..8d04a80 100755
--- a/examples/dump-phenotype.scm
+++ b/examples/dump-phenotype.scm
@@ -48,16 +48,11 @@
             #:proc string-capitalize-first))))
 
 (define-dump dump-phenotypes
-  (tables (Phenotype
-           (left-join PublishXRef "ON Phenotype.Id = PublishXRef.PhenotypeId")
-           (left-join Publication "ON Publication.Id = PublishXRef.PublicationId")
-           ;; We need this join so as to construct the trait's skos:altLabel
+  (tables (PublishXRef
            (left-join InbredSet "ON InbredSet.InbredSetId = PublishXRef.InbredSetId")
-           (left-join PublishFreeze "ON PublishFreeze.InbredSetId = PublishXRef.InbredSetId")
-           (left-join InfoFiles "ON InfoFiles.InfoPageName = PublishFreeze.Name"))
-          ;; Only dump public traits; Ignore "hanging" traits
-          ;; I.e. traits that have no associated vectors
-          "WHERE PublishFreeze.public > 0 AND PublishFreeze.confidentiality < 1 AND PublishFreeze.Id IS NOT NULL")
+           (left-join Publication "ON Publication.Id = PublishXRef.PublicationId")
+           (left-join Phenotype "ON Phenotype.Id = PublishXRef.PhenotypeId"))
+          "WHERE PublishXRef.InbredSetId IN (SELECT PublishFreeze.InbredSetId FROM PublishFreeze)")
   (schema-triples
    (gnc:phenotype a skos:Concept)
    (gnc:phenotype skos:description "This is a set of controlled terms that are used to describe a given phenotype")
@@ -80,22 +75,13 @@
    (gnt:sequence rdfs:domain gnc:phenotype)
    (gnt:sequence rdfs:range xsd:integer))
   (triples (string->identifier
-            ""
-            (regexp-substitute/global #f "[^A-Za-z0-9:]"
-                                      (field ("CONCAT(IF(PublishFreeze.Name IS NULL, '', CONCAT(PublishFreeze.Name, '_')), IF(Phenotype.Post_publication_abbreviation IS NULL, IF(Phenotype.Pre_publication_abbreviation IS NULL, Phenotype.Id, Pre_publication_abbreviation), Phenotype.Post_publication_abbreviation))" abbrev))
-                                      'pre "_" 'post)
-            #:separator ""
-            #:proc string-capitalize-first)
+            "trait"
+            (field ("CONCAT(IFNULL(InbredSet.Name, PublishXRef.InbredSetId), '_', PublishXRef.Id)"
+                    Phenotype)))
     (set rdf:type 'gnc:phenotype)
-    (set skos:prefLabel (sanitize-rdf-string
-                         (field
-                          ("IF(Phenotype.Post_publication_abbreviation IS NULL, IF(Phenotype.Pre_publication_abbreviation IS NULL, Phenotype.Id, Phenotype.Pre_publication_abbreviation), Phenotype.Post_publication_abbreviation)"
-                           PhenotypeName))))
-    ;; Add an alternative name for this resources.  This is how GN
-    ;; currently labels phenotypes
-    (set skos:altLabel (field
-                        ("CONCAT(InbredSet.Name, '_', PublishXRef.Id)"
-                         phenotypeAltName)))
+    (set rdfs:label
+         (field ("CONCAT(IFNULL(InbredSet.Name, PublishXRef.InbredSetId), '_', PublishXRef.Id)"
+                 Phenotype)))
     ;; All phenotypes have a post-publication description
     (set dct:description
          (sanitize-rdf-string
@@ -105,10 +91,10 @@
     (set gnt:labCode (field Phenotype Lab_code))
     (set gnt:submitter
          (sanitize-rdf-string (field Phenotype Submitter)))
-    (set dct:contributor (sanitize-rdf-string (field Phenotype Owner)))
-    (multiset dct:contributor (string-split
-                               (sanitize-rdf-string (field Phenotype Owner))
-                               #\,))
+    (multiset dct:contributor
+              (string-split
+               (sanitize-rdf-string (field Phenotype Owner))
+               #\,))
     (set gnt:mean (annotate-field (field ("IFNULL(PublishXRef.mean, '')" mean))
                                   '^^xsd:double))
     (set gnt:locus (field PublishXRef Locus))
@@ -119,14 +105,6 @@
          (annotate-field (field ("IFNULL(PublishXRef.additive, '')" additive))
                          '^^xsd:double))
     (set gnt:sequence (annotate-field (field PublishXRef Sequence) '^^xsd:integer))
-    (set gnt:belongsToDataset
-         (string->identifier
-          ""
-          (regexp-substitute/global #f "[^A-Za-z0-9:]"
-                                    (field InfoFiles InfoPageName)
-                                    'pre "_" 'post)
-          #:separator ""
-          #:proc string-capitalize-first))
     (set dct:isReferencedBy
          (let ((pmid (field
                       ("IF(Publication.PubMed_ID IS NULL, '', CONVERT(Publication.PubMed_Id, INT))"
@@ -136,6 +114,7 @@
                (string->identifier "unpublished"
                                    (number->string publication-id))
                (ontology 'pubmed: pmid))))))
+
 
 
 (dump-with-documentation
-- 
cgit v1.2.3