From 9193597f12631bcf33ce768f2842132bd22a41c3 Mon Sep 17 00:00:00 2001
From: Munyoki Kilyungi
Date: Thu, 15 Jun 2023 15:40:58 +0300
Subject: Fetch genotypes from virtuoso

* gn3/api/metadata.py: Import get_genotype_metadata.
(genotype): New end-point.
* gn3/db/rdf.py (get_phenotype_metadata): New function.

Signed-off-by: Munyoki Kilyungi <me@bonfacemunyoki.com>
---
 gn3/db/rdf.py | 43 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)

(limited to 'gn3/db/rdf.py')

diff --git a/gn3/db/rdf.py b/gn3/db/rdf.py
index 83bc46b..ad2e3be 100644
--- a/gn3/db/rdf.py
+++ b/gn3/db/rdf.py
@@ -26,6 +26,7 @@ PREFIX up: <http://purl.uniprot.org/core/>
 PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
 PREFIX publication: <http://genenetwork.org/publication/>
 PREFIX phenotype: <http://genenetwork.org/phenotype/>
+PREFIX genotype: <http://genenetwork.org/genotype/>
 """
 
 
@@ -243,3 +244,45 @@ CONSTRUCT {
     )[0].items():
         result[key] = value
     return result
+
+
+def get_genotype_metadata(
+        sparql_conn: SPARQLWrapper, name: str
+):
+    """Return info about a phenotype with a given NAME"""
+    __metadata_query = """
+$prefix
+
+CONSTRUCT {
+    ?genotype ?pPredicate ?pValue .
+    ?genotype gn:speciesName ?speciesName .
+    ?genotype gn:inbredSetName ?inbredSetBinomialName .
+    ?genotype gn:datasetName ?datasetFullName .
+} WHERE {
+    ?genotype ?pPredicate ?pValue .
+    OPTIONAL {
+        ?genotype gn:genotypeOfDataset ?dataset .
+        ?dataset gn:fullName ?datasetFullName .
+    }.
+    OPTIONAL {
+        ?genotype gn:genotypeOfDataset ?dataset .
+        ?dataset gn:datasetOfInbredSet ?inbredSet .
+        ?inbredSet gn:binomialName ?inbredSetBinomialName .
+        ?inbredSet gn:inbredSetOfSpecies ?species .
+        ?species gn:displayName ?speciesName .
+    } .
+    FILTER( ?genotype = genotype:$name ) .
+    MINUS {
+         ?genotype rdf:type ?pValue .
+    }
+}
+"""
+    result: MonadicDict = MonadicDict()
+    for key, value in sparql_query(
+            sparql_conn,
+            Template(__metadata_query)
+            .substitute(name=name,
+                        prefix=RDF_PREFIXES)
+    )[0].items():
+        result[key] = value
+    return result
-- 
cgit v1.2.3