Clean-up phenotype transform.

Signed-off-by: Munyoki Kilyungi <me@bonfacemunyoki.com>

1 files changed, 48 insertions, 6 deletions

diff --git a/examples/phenotype-datasets.scm b/examples/phenotype-datasets.scm
index 410178a..4819627 100755
--- a/examples/phenotype-datasets.scm
+++ b/examples/phenotype-datasets.scm
@@ -15,13 +15,17 @@
 
 
 (define-transformer gn:set->gn:dataset
-  (tables (Datasets
-           (inner-join InfoFiles "ON InfoFiles.DatasetId = Datasets.DatasetId")
-           (inner-join InbredSet "ON InbredSet.Id = InfoFiles.InbredSetId")
+  (tables (Species
+           (inner-join InbredSet "ON InbredSet.SpeciesId = Species.Id")
            (inner-join PublishFreeze "ON PublishFreeze.InbredSetId = InbredSet.Id"))
-          "WHERE PublishFreeze.public > 0 GROUP BY Datasets.DatasetId")
+          "WHERE PublishFreeze.public > 0 AND Species.Name != 'monkey' GROUP BY Species.Name, PublishFreeze.ShortName")
   (triples (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_")
-    (set gnt:has_phenotype_data (string->identifier "dataset" (field PublishFreeze Name) #:separator "_"))))
+    (multiset gnt:has_phenotype_data
+              (map (cut string->identifier "dataset" <> #:separator "_")
+                   (string-split
+                    (field ("GROUP_CONCAT(PublishFreeze.Name SEPARATOR ',')"
+                            dataset_name))
+                    #\,)))))
 
 (define-transformer gn:dataset->gn:set
   (tables (Datasets
@@ -32,6 +36,40 @@
   (triples (string->identifier "dataset" (field PublishFreeze Name) #:separator "_")
     (set gnt:has_strain (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_"))))
 
+(define-transformer gn:dataset->metadata
+  (tables (PublishXRef
+           (inner-join InbredSet "ON InbredSet.InbredSetId = PublishXRef.InbredSetId")
+           (inner-join Species "ON InbredSet.SpeciesId = Species.Id")
+           (inner-join PublishFreeze "ON PublishFreeze.InbredSetId = InbredSet.Id")
+           (inner-join Publication "ON Publication.Id = PublishXRef.PublicationId")
+           (inner-join Phenotype "ON Phenotype.Id = PublishXRef.PhenotypeId"))
+          "WHERE InbredSet.public > 0 GROUP BY Species.Name, PublishFreeze.Name")
+  (triples (string->identifier "dataset" (field PublishFreeze Name) #:separator "_")
+    (set dct:created (annotate-field (field PublishFreeze CreateTime) '^^xsd:datetime))
+    (set gnt:has_strain (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_"))))
+
+(define-transformer gn:dataset->gn:trait
+  (tables (PublishXRef
+           (inner-join InbredSet "ON InbredSet.InbredSetId = PublishXRef.InbredSetId")
+           (inner-join Species "ON InbredSet.SpeciesId = Species.Id")
+           (inner-join PublishFreeze "ON PublishFreeze.InbredSetId = InbredSet.Id")
+           (inner-join Publication "ON Publication.Id = PublishXRef.PublicationId")
+           (inner-join Phenotype "ON Phenotype.Id = PublishXRef.PhenotypeId"))
+          "WHERE InbredSet.public > 0")
+  (triples (string->identifier "dataset" (field PublishFreeze Name) #:separator "_")
+    (set gnt:has_phenotype_trait
+         (let ((post-abbrev (blank-p (field Phenotype Post_publication_abbreviation)))
+               (pre-abbrev (blank-p (field Phenotype Pre_publication_abbreviation)))
+               (post-desc (blank-p (field Phenotype Post_publication_description)))
+               (pre-desc (blank-p (field Phenotype Post_publication_description))))
+           (string->identifier
+            "trait"
+            (format #f "~a_~a" (field PublishFreeze Name)
+                    (or post-abbrev pre-abbrev post-desc pre-desc))
+            #:separator "_")))
+    (set dct:created (annotate-field (field PublishFreeze CreateTime) '^^xsd:datetime))
+    (set gnt:has_strain (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_"))))
+
 
 (let* ((option-spec
         '((settings (single-char #\s) (value #t))
@@ -61,7 +99,11 @@
       ("xkos:" "<http://rdf-vocabulary.ddialliance.org/xkos#>")
       ("xsd:" "<http://www.w3.org/2001/XMLSchema#>")))
    (inputs
-    (list gn:set->gn:dataset gn:dataset->gn:set))
+    (list
+     gn:set->gn:dataset
+     gn:dataset->gn:set
+     gn:dataset->metadata
+     gn:dataset->gn:trait))
    (outputs
     `(#:documentation ,documentation
       #:rdf ,output))))