OBO Compliance suite (replay of 683)

src/oaklib/converters/obo_graph_to_rdf_owl_converter.py

@@ -24,6 +24,7 @@
     HAS_RELATED_SYNONYM,
 )
 from oaklib.types import CURIE
+from oaklib.utilities.format_utilities import RDFLIB_SYNTAX_ALIAS_MAP
 
 TRIPLE = Tuple[rdflib.URIRef, rdflib.URIRef, Any]
 
@@ -45,7 +46,7 @@
 class OboGraphToRdfOwlConverter(DataModelConverter):
     """Converts from OboGraph to OWL layered on RDF."""
 
-    def dump(self, source: GraphDocument, target: str = None, **kwargs) -> None:
+    def dump(self, source: GraphDocument, target: str = None, format="turtle", **kwargs) -> None:
         """
         Dump an OBO Graph Document to a FHIR CodeSystem
 
@@ -54,10 +55,13 @@ def dump(self, source: GraphDocument, target: str = None, **kwargs) -> None:
         :return:
         """
         g = self.convert(source)
+        # TODO: simplify this, see https://github.com/INCATools/ontology-access-kit/issues/687
+        if format in RDFLIB_SYNTAX_ALIAS_MAP:
+            format = RDFLIB_SYNTAX_ALIAS_MAP[format]
         if target is None:
-            print(g.serialize(format="turtle"))
+            print(g.serialize(format=format))
         else:
-            g.serialize(format="turtle", destination=target)
+            g.serialize(format=format, destination=target)
 
     def convert(
         self, source: Union[Graph, GraphDocument], target: rdflib.Graph = None, **kwargs
@@ -157,6 +161,8 @@ def _add_reified(
         target.add((pred, RDF.type, OWL.AnnotationProperty))
 
     def _uri_ref(self, curie: CURIE) -> rdflib.URIRef:
+        if ":" not in curie:
+            curie = f"obo:{curie}"
         if self.curie_converter is None:
             self.curie_converter = BasicOntologyInterface().converter
         uri = self.curie_converter.expand(curie, passthrough=True)

src/oaklib/implementations/simpleobo/simple_obo_implementation.py

@@ -643,7 +643,7 @@ def dump(self, path: Union[str, TextIO] = None, syntax: str = "obo", **kwargs):
             else:
                 self.obo_document.dump(path)
         else:
-            super().dump(path, syntax)
+            super().dump(path, syntax=syntax)
 
     def save(
         self,
@@ -706,6 +706,8 @@ def node(self, curie: CURIE, strict=False, include_metadata=False) -> obograph.N
                     meta.definition = obograph.DefinitionPropertyValue(val=defn)
                 for _, syn in self.synonym_property_values([curie]):
                     meta.synonyms.append(syn)
+                for _, subset in self.terms_subsets([curie]):
+                    meta.subsets.append(subset)
             return obograph.Node(id=curie, lbl=self.label(curie), type=typ, meta=meta)
 
     def as_obograph(self, expand_curies=False) -> Graph:

src/oaklib/implementations/simpleobo/simple_obo_parser.py

@@ -372,7 +372,7 @@ def intersection_of_tuples(self) -> List[Tuple[CURIE, Optional[CURIE]]]:
         pairs = []
         for v in self._values(TAG_INTERSECTION_OF):
             toks = [x for x in v.split(" ") if x]
-            if toks[1].startswith("!"):
+            if len(toks) == 1 or toks[1].startswith("!"):
                 pairs.append((toks[0], None))
             else:
                 pairs.append((toks[0], toks[1]))
@@ -382,7 +382,7 @@ def property_value_tuples(self) -> List[Tuple[CURIE, Optional[CURIE]]]:
         pairs = []
         for v in self._values(TAG_PROPERTY_VALUE):
             toks = [x for x in v.split(" ") if x]
-            if toks[1].startswith("!"):
+            if len(toks) == 1 or toks[1].startswith("!"):
                 pairs.append((toks[0], None))
             else:
                 pairs.append((toks[0], toks[1]))

src/oaklib/implementations/sqldb/sql_implementation.py

@@ -154,6 +154,10 @@
 from oaklib.utilities.axioms.logical_definition_utilities import (
     logical_definition_matches,
 )
+from oaklib.utilities.format_utilities import (
+    OBOGRAPHS_SYNTAX_ALIAS_MAP,
+    RDFLIB_SYNTAX_ALIAS_MAP,
+)
 from oaklib.utilities.graph.relationship_walker import walk_down, walk_up
 from oaklib.utilities.identifier_utils import (
     string_as_base64_curie,
@@ -1180,11 +1184,14 @@ def dump(self, path: str = None, syntax: str = None, **kwargs):
         """
         if syntax is None:
             syntax = "ttl"
-        if syntax == "ttl":
+        if syntax in ["ttl", "rdfxml", "owl"]:
+            if syntax in RDFLIB_SYNTAX_ALIAS_MAP:
+                syntax = RDFLIB_SYNTAX_ALIAS_MAP[syntax]
             g = self.as_rdflib_graph()
             logging.info(f"Dumping to {path}")
             g.serialize(path, format=syntax)
-        elif syntax == "json":
+        elif syntax in OBOGRAPHS_SYNTAX_ALIAS_MAP.keys():
+            syntax = OBOGRAPHS_SYNTAX_ALIAS_MAP[syntax]
             g = self.as_obograph(expand_curies=True)
             gd = obograph.GraphDocument(graphs=[g])
             json_dumper.dump(gd, path)

src/oaklib/interfaces/dumper_interface.py

@@ -25,6 +25,7 @@
     "rdfxml": (OboGraphToRdfOwlConverter, {"format": "xml"}),
     "cx": OboGraphToCXConverter,
     "obojson": None,
+    "json": None,
 }
 
 FORMAT_SYNONYMS = {
@@ -61,15 +62,15 @@ def dump(
             raise ValueError(f"Cannot handle interface: {self}")
         og = self.as_obograph()
         ogdoc = GraphDocument(graphs=[og])
-        if syntax == "obojson":
+        converter_class = OBOGRAPH_CONVERTERS.get(syntax, None)
+        if converter_class is None:
             json_str = json_dumper.dumps(ogdoc, inject_type=False)
             if path:
                 with open(path, "w", encoding="utf-8") as f:
                     f.write(json_str)
             else:
                 print(json_str)
         else:
-            converter_class = OBOGRAPH_CONVERTERS[syntax]
             if isinstance(converter_class, tuple):
                 converter_class, converter_kwargs = converter_class
                 kwargs.update(converter_kwargs)
@@ -78,4 +79,6 @@ def dump(
             logging.info(f"Using {converter}, kwargs={kwargs}")
             converter.curie_converter = self.converter
             kwargs = {k: v for k, v in kwargs.items() if v is not None}
+            if "format" not in kwargs:
+                kwargs["format"] = syntax
             converter.dump(ogdoc, target=path, **kwargs)

src/oaklib/utilities/format_utilities.py

@@ -0,0 +1,68 @@
+"""
+Utilities for working with formats
+
+See https://github.com/INCATools/ontology-access-kit/issues/687
+"""
+
+from enum import Enum
+
+
+class OntologyModel(str, Enum):
+    """
+    Enumerated types for ontology models.
+
+    A model is an abstract representation of an ontology that is potentially
+    serializable in multiple syntaxes.
+    """
+
+    OWL = "owl"
+    """W3C OWL2 Ontology Model"""
+
+    SKOS = "skos"
+    """W3C SKOS Ontology Model"""
+
+    SDO_RDFS = "sdo_rdfs"
+    """Schema.org RDFS Ontology Model"""
+
+    OBOFORMAT = "oboformat"
+    """OBO Format Ontology Model. Always serialized as obo format"""
+
+    OBOGRAPHS = "obographs"
+    """OBO Graphs Ontology Model"""
+
+
+MODEL_ALLOWED_SYNTAXES = {
+    OntologyModel.OWL: [
+        "ttl",
+        "rdfxml",
+        "n3",
+        "nt",
+        "pretty-xml",
+        "trix",
+        "trig",
+        "ofn",
+        "owx",
+        "omn",
+        "obo",
+    ],
+    OntologyModel.SKOS: ["ttl", "rdfxml", "n3", "nt", "pretty-xml", "trix", "trig"],
+    OntologyModel.OBOFORMAT: ["obo"],
+    OntologyModel.OBOGRAPHS: ["json", "yaml"],
+}
+
+
+OBOGRAPHS_SYNTAX_ALIAS_MAP = {
+    "json": "json",
+    "obojson": "json",
+    "yaml": "yaml",
+    "obograph": "json",
+    "obographjson": "json",
+    "obographyaml": "yaml",
+}
+
+
+RDFLIB_SYNTAX_ALIAS_MAP = {
+    "owl": "turtle",
+    "ttl": "turtle",
+    "rdfxml": "xml",
+}