about summary refs log tree commit diff
path: root/examples/genotype-datasets.scm
diff options
context:
space:
mode:
Diffstat (limited to 'examples/genotype-datasets.scm')
-rwxr-xr-xexamples/genotype-datasets.scm53
1 files changed, 33 insertions, 20 deletions
diff --git a/examples/genotype-datasets.scm b/examples/genotype-datasets.scm
index f140600..38d524b 100755
--- a/examples/genotype-datasets.scm
+++ b/examples/genotype-datasets.scm
@@ -15,29 +15,39 @@
 
 
 (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")
-  (schema-triples
-   (gnt:has_genotype_data rdf:type owl:ObjectProperty)
-   (gnt:has_genotype_data rdfs:label "this resources has genotype data.")
-   (gnt:has_genotype_data rdfs:comment "Associates a resource with its genotype data.")
-   (gnt:has_genotype_data rdfs:domain gnc:set)
-   (gnt:has_genotype_data rdfs:range dcat:Dataset)
-   (gnt:has_genotype_data rdfs:subPropertyOf dct:relation))
+          "WHERE GenoFreeze.public > 0 AND GenoFreeze.confidentiality < 1 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 GenoFreeze.confidentiality < 1 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 dct:created (annotate-field (field GenoFreeze CreateTime) '^^xsd:datetime))))
+
+(define-transformer gn:dataset->marker/snp-count
+  (tables (GenoFreeze
+           (inner-join InbredSet "ON InbredSet.Id = GenoFreeze.InbredSetId")
+           (inner-join Species "ON InbredSet.SpeciesId = Species.Id")
+           (inner-join Geno "ON Geno.SpeciesId = Species.Id"))
+          "WHERE GenoFreeze.public > 0 AND GenoFreeze.confidentiality < 1 AND Species.Name != 'monkey' GROUP BY GenoFreeze.Name")
   (triples (string->identifier "dataset" (field GenoFreeze Name) #:separator "_")
-    (set gnt:has_strain (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_"))))
+    (set gnt:has_marker_count
+         (string->symbol
+          (format #f "'~s'^^xsd:integer"
+                  (field
+                   ("COUNT(DISTINCT Geno.Marker_Name)" MarkerCount)))))))
 
 
 (let* ((option-spec
@@ -68,7 +78,10 @@
       ("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
+     gn:dataset->marker/snp-count))
    (outputs
     `(#:documentation ,documentation
       #:rdf ,output))))