From 9c43d6dea557d11ed936950cd190172883a34595 Mon Sep 17 00:00:00 2001
From: Munyoki Kilyungi
Date: Fri, 6 Feb 2026 05:54:42 +0300
Subject: Clean up genotype transform.

Signed-off-by: Munyoki Kilyungi <me@bonfacemunyoki.com>
---
 examples/genotype-datasets.scm | 32 +++++++++++++++++++-------------
 1 file changed, 19 insertions(+), 13 deletions(-)

(limited to 'examples')

diff --git a/examples/genotype-datasets.scm b/examples/genotype-datasets.scm
index 6f3a1cb..ba31088 100755
--- a/examples/genotype-datasets.scm
+++ b/examples/genotype-datasets.scm
@@ -15,22 +15,26 @@
 
 
 (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 GenoFreeze "ON GenoFreeze.InbredSetId = InbredSet.Id"))
-          "WHERE GenoFreeze.public > 0 GROUP BY Datasets.DatasetId")
+          "WHERE GenoFreeze.public > 0 AND Species.Name != 'monkey' GROUP BY Species.Name, GenoFreeze.ShortName")
   (triples (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_")
-    (set gnt:has_genotype_data (string->identifier "dataset" (field GenoFreeze Name) #:separator "_"))))
+    (multiset gnt:has_genotype_data
+              (map (cut string->identifier "dataset" <> #:separator "_")
+                   (string-split
+                    (field ("GROUP_CONCAT(GenoFreeze.Name SEPARATOR ',')"
+                            dataset_name))
+                    #\,)))))
 
-(define-transformer gn:dataset->gn:set
-  (tables (Datasets
-           (inner-join InfoFiles "ON InfoFiles.DatasetId = Datasets.DatasetId")
-           (inner-join InbredSet "ON InbredSet.Id = InfoFiles.InbredSetId")
-           (inner-join GenoFreeze "ON GenoFreeze.InbredSetId = InbredSet.Id"))
-          "WHERE GenoFreeze.public > 0 GROUP BY Datasets.DatasetId")
+(define-transformer gn:dataset->metadata
+  (tables (GenoFreeze
+           (inner-join InbredSet "ON InbredSet.Id = GenoFreeze.InbredSetId")
+           (inner-join Species "ON InbredSet.SpeciesId = Species.Id"))
+          "WHERE GenoFreeze.public > 0 AND Species.Name != 'monkey'")
   (triples (string->identifier "dataset" (field GenoFreeze Name) #:separator "_")
-    (set gnt:has_strain (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_"))))
+    (set gnt:has_strain (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_"))
+    (set dct:created (annotate-field (field GenoFreeze CreateTime) '^^xsd:datetime))))
 
 
 (let* ((option-spec
@@ -61,7 +65,9 @@
       ("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->metadata))
    (outputs
     `(#:documentation ,documentation
       #:rdf ,output))))
-- 
cgit 1.4.1