From fa2ce410d4730b136f24555c049cef3d6dac1102 Mon Sep 17 00:00:00 2001
From: Munyoki Kilyungi
Date: Thu, 8 Jun 2023 14:44:38 +0300
Subject: Fetch phenotypes from virtuoso

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

Signed-off-by: Munyoki Kilyungi <me@bonfacemunyoki.com>
---
 gn3/api/metadata.py | 16 ++++++++++++++++
 gn3/db/rdf.py       | 37 +++++++++++++++++++++++++++++++++++++
 2 files changed, 53 insertions(+)

diff --git a/gn3/api/metadata.py b/gn3/api/metadata.py
index 49c6798..a2eff05 100644
--- a/gn3/api/metadata.py
+++ b/gn3/api/metadata.py
@@ -10,6 +10,7 @@ from SPARQLWrapper import SPARQLWrapper
 
 from gn3.db.rdf import get_dataset_metadata
 from gn3.db.rdf import get_publication_metadata
+from gn3.db.rdf import get_phenotype_metadata
 from gn3.db.rdf import sparql_query
 from gn3.db.rdf import RDF_PREFIXES
 
@@ -51,6 +52,21 @@ def publication(name):
         return jsonify({})
 
 
+@metadata.route("/phenotype/<name>", methods=["GET"])
+def phenotype(name):
+    """Fetch a phenotype's metadata given it's name"""
+    try:
+        return jsonify(
+            get_phenotype_metadata(
+                SPARQLWrapper(current_app.config.get("SPARQL_ENDPOINT")),
+                name,
+            ).data
+        )
+    # The virtuoso server is misconfigured or it isn't running at all
+    except (RemoteDisconnected, URLError):
+        return jsonify({})
+
+
 @metadata.route("/genewiki/<symbol>", methods=["GET"])
 def get_genewiki_entries(symbol):
     """Fetch the GN and NCBI GeneRIF entries"""
diff --git a/gn3/db/rdf.py b/gn3/db/rdf.py
index e92ccdb..0e472ce 100644
--- a/gn3/db/rdf.py
+++ b/gn3/db/rdf.py
@@ -26,6 +26,7 @@ PREFIX ncbiTaxon: <https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mode
 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/>
 """
 
 
@@ -198,12 +199,48 @@ CONSTRUCT {
     return response
 
 
+def get_phenotype_metadata(
+        sparql_conn: SPARQLWrapper, name: str
 ):
+    """Return info about a phenotype with a given NAME"""
     __metadata_query = """
+$prefix
 
+CONSTRUCT {
+    ?phenotype ?pPredicate ?pValue .
+    ?phenotype ?publicationTerm ?publicationValue .
+    ?phenotype gn:speciesName ?speciesName .
+    ?phenotype gn:inbredSetName ?inbredSetBinomialName .
+    ?phenotype gn:datasetName ?datasetFullName .
+} WHERE {
+    ?phenotype ?pPredicate ?pValue .
+    OPTIONAL {
+        ?phenotype gn:phenotypeOfPublication ?publication .
+        ?publication ?publicationTerm ?publicationValue .
+    } .
+    OPTIONAL {
+        ?phenotype gn:phenotypeOfDataset ?dataset .
+        ?dataset gn:name ?datasetFullName .
+        ?dataset gn:datasetOfInbredSet ?inbredSet .
+        ?inbredSet gn:binomialName ?inbredSetBinomialName .
+        ?inbredSet gn:inbredSetOfSpecies ?species .
+        ?species gn:displayName ?speciesName .
+    } .
+    FILTER( ?phenotype = phenotype:$name ) .
+    MINUS {
+         ?phenotype rdf:type ?pValue .
+    }
+    MINUS {
+         ?publication rdf:type ?publicationValue .
+    }
 }
 """
     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