Dataset object populated with properties from h5ad uns (#70)

* Updated version to 0.1.13

* Added dataset object to rdf representation

pandasaurus_cxg/graph_generator/graph_generator.py

@@ -28,6 +28,8 @@
 from pandasaurus_cxg.graph_generator.graph_predicates import (
     CLUSTER,
     CONSIST_OF,
+    DATASET,
+    HAS_SOURCE,
     SUBCLUSTER_OF,
 )
 from pandasaurus_cxg.utils.exceptions import (
@@ -119,21 +121,32 @@ def generate_rdf_graph(self):
             if temp_dict not in grouped_dict_uuid.values():
                 grouped_dict_uuid[str(uuid.uuid4())] = temp_dict
 
+        # generate dataset entity and has_source property
+        dataset_class = URIRef(self.ns[str(uuid.uuid4())])
+        self.graph.add((dataset_class, RDF.type, URIRef(DATASET.get("iri"))))
+        self.graph.add((dataset_class, RDFS.label, Literal(DATASET.get("label"))))
+        uns = self.ea.enricher_manager.anndata.uns
+        for key, value in uns.items():
+            if not isinstance(value, str):
+                continue
+            self.graph.add((dataset_class, URIRef(self.ns[key]), Literal(value)))
+        has_source = URIRef(HAS_SOURCE["iri"])
+        self.graph.add((has_source, RDFS.label, Literal(HAS_SOURCE["label"])))
+
         # generate a resource for each free-text cell_type annotation and cell_type_ontology_term annotation
-        # cell_set_class = self.ns["CellSet"]
         cell_set_class = URIRef(CLUSTER.get("iri"))
         self.graph.add((cell_set_class, RDF.type, OWL.Class))
         self.graph.add((cell_set_class, RDFS.label, Literal(CLUSTER.get("label"))))
         for _uuid, inner_dict in grouped_dict_uuid.items():
             resource = self.ns[_uuid]
             self.graph.add((resource, RDF.type, cell_set_class))
+            self.graph.add((resource, has_source, dataset_class))
             for k, v in inner_dict.items():
                 if k == "subcluster_of":
                     continue
                 self.graph.add((resource, self.ns[remove_special_characters(k)], Literal(v)))
 
         # add relationship between each resource based on their predicate in the co_annotation_report
-        # subcluster = self.ns["subcluster_of"]
         subcluster = URIRef(SUBCLUSTER_OF.get("iri"))
         self.graph.add((subcluster, RDFS.label, Literal(SUBCLUSTER_OF.get("label"))))
         self.graph.add((subcluster, RDF.type, OWL.ObjectProperty))

pandasaurus_cxg/graph_generator/graph_predicates.py

@@ -3,3 +3,7 @@
 SUBCLUSTER_OF = {"iri": "http://purl.obolibrary.org/obo/RO_0015003", "label": "subcluster of"}
 
 CLUSTER = {"iri": "http://purl.obolibrary.org/obo/PCL_0010001", "label": "cell cluster"}
+
+DATASET = {"iri": "https://schema.org/Dataset", "label": "dataset"}
+
+HAS_SOURCE = {"iri": "http://purl.org/dc/terms/source", "label": "has_source"}

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "pandasaurus-cxg"
-version = "0.1.12"
+version = "0.1.13"
 description = "Ontology enrichment tool for CxG standard AnnData files."
 authors = ["Ugur Bayindir <ugur@ebi.ac.uk>"]
 license = "http://www.apache.org/licenses/LICENSE-2.0"

test/graph_generator/test_graph_generator.py

@@ -90,13 +90,13 @@ def test_graph_generator_init_with_valid_input(enrichment_analyzer_instance_for_
 def test_generate_rdf_graph(graph_generator_instance_for_kidney):
     graph_generator = graph_generator_instance_for_kidney
     graph_generator.generate_rdf_graph()
-    assert len(graph_generator.graph) == 530
+    assert len(graph_generator.graph) == 629
     assert (
-        len([[s, p, o] for s, p, o in graph_generator.graph.triples((None, RDF.type, None))]) == 143
+        len([[s, p, o] for s, p, o in graph_generator.graph.triples((None, RDF.type, None))]) == 144
     )
     assert (
         len([[s, p, o] for s, p, o in graph_generator.graph.triples((None, RDFS.label, None))])
-        == 19
+        == 21
     )
     assert (
         len(
@@ -152,11 +152,11 @@ def test_enrich_rdf_graph(graph_generator_instance_for_kidney):
     graph_generator = graph_generator_instance_for_kidney
     graph_generator.generate_rdf_graph()
 
-    assert len(graph_generator.graph) == 530
+    assert len(graph_generator.graph) == 629
 
     graph_generator.enrich_rdf_graph()
 
-    assert len(graph_generator.graph) == 1025
+    assert len(graph_generator.graph) == 1124
     assert (
         URIRef(CONSIST_OF.get("iri")),
         RDFS.label,