Added format_utilities.py -- see #687

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]
 
@@ -54,8 +55,9 @@ def dump(self, source: GraphDocument, target: str = None, format="turtle", **kwa
         :return:
         """
         g = self.convert(source)
-        if format == "rdfxml":
-            format = "xml"
+        # 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=format))
         else:

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,
@@ -1181,10 +1185,13 @@ def dump(self, path: str = None, syntax: str = None, **kwargs):
         if syntax is None:
             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 in ["json", "obojson"]:
+        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/utilities/format_utilities.py

@@ -0,0 +1,67 @@
+"""
+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",
+}

tests/test_cli.py

@@ -534,6 +534,9 @@ def test_dump(self):
             if conf_path:
                 cmd.extend(["-c", conf_path])
             result = self.runner.invoke(main, cmd)
+            if result.exit_code != 0:
+                print("STDOUT", result.stdout)
+                print("STDERR", result.stderr)
             self.assertEqual(0, result.exit_code, f"input={input}, output_format={output_format}")
             if output_format == "obojson":
                 obj: obograph.GraphDocument