Merge pull request #278 from cthoyt/pass-converters

Enable passing a converter to RDFLib dumper

linkml_runtime/dumpers/rdflib_dumper.py

@@ -4,6 +4,7 @@
 from typing import Optional, Any, Dict, Union
 from pydantic import BaseModel
 
+from curies import Converter
 from rdflib import Graph, URIRef, XSD
 from rdflib.term import Node, BNode, Literal
 from rdflib.namespace import RDF
@@ -24,7 +25,12 @@ class RDFLibDumper(Dumper):
     This requires a SchemaView object
 
     """
-    def as_rdf_graph(self, element: Union[BaseModel, YAMLRoot], schemaview: SchemaView, prefix_map: Dict[str, str] = None) -> Graph:
+    def as_rdf_graph(
+        self,
+        element: Union[BaseModel, YAMLRoot],
+        schemaview: SchemaView,
+        prefix_map: Union[Dict[str, str], Converter, None] = None,
+    ) -> Graph:
         """
         Dumps from element to an rdflib Graph,
         following a schema
@@ -35,6 +41,9 @@ def as_rdf_graph(self, element: Union[BaseModel, YAMLRoot], schemaview: SchemaVi
         :return:
         """
         g = Graph()
+        if isinstance(prefix_map, Converter):
+            # TODO replace with `prefix_map = prefix_map.bimap` after making minimum requirement on python 3.8
+            prefix_map = {record.prefix: record.uri_prefix for record in prefix_map.records}
         logging.debug(f'PREFIXMAP={prefix_map}')
         if prefix_map:
             for k, v in prefix_map.items():
@@ -137,10 +146,15 @@ def inject_triples(self, element: Any, schemaview: SchemaView, graph: Graph, tar
             graph.add((element_uri, RDF.type, URIRef(schemaview.get_uri(cn, expand=True))))
         return element_uri
 
-    def dump(self, element: Union[BaseModel, YAMLRoot],
-             to_file: str,
-             schemaview: SchemaView = None,
-             fmt: str = 'turtle', prefix_map: Dict[str, str] = None, **args) -> None:
+    def dump(
+        self,
+        element: Union[BaseModel, YAMLRoot],
+        to_file: str,
+        schemaview: SchemaView = None,
+        fmt: str = 'turtle',
+        prefix_map: Union[Dict[str, str], Converter, None] = None,
+        **args,
+    ) -> None:
         """
         Write element as rdf to to_file
 
@@ -153,8 +167,13 @@ def dump(self, element: Union[BaseModel, YAMLRoot],
         """
         super().dump(element, to_file, schemaview=schemaview, fmt=fmt, prefix_map=prefix_map)
 
-    def dumps(self, element: Union[BaseModel, YAMLRoot], schemaview: SchemaView = None,
-              fmt: Optional[str] = 'turtle', prefix_map: Dict[str, str] = None) -> str:
+    def dumps(
+        self,
+        element: Union[BaseModel, YAMLRoot],
+        schemaview: SchemaView = None,
+        fmt: Optional[str] = 'turtle',
+        prefix_map: Union[Dict[str, str], Converter, None] = None,
+    ) -> str:
         """
         Convert element into an RDF graph guided by the schema
 

linkml_runtime/loaders/rdflib_loader.py

@@ -4,6 +4,7 @@
 from dataclasses import dataclass
 from typing import Optional, Any, Dict, Type, Union, TextIO, List, Tuple, Set
 
+from curies import Converter
 from hbreader import FileInfo
 from rdflib import Graph, URIRef
 from rdflib.term import Node, BNode, Literal
@@ -30,11 +31,15 @@ class RDFLibLoader(Loader):
 
     Note: this is a more complete replacement for rdf_loader
     """
-    def from_rdf_graph(self, graph: Graph, schemaview: SchemaView, target_class: Type[Union[BaseModel, YAMLRoot]],
-                       prefix_map: Dict[str, str] = None,
-                       cast_literals: bool = True,
-                       allow_unprocessed_triples: bool = True,
-                       ignore_unmapped_predicates: bool = False) -> List[Union[BaseModel, YAMLRoot]]:
+    def from_rdf_graph(
+        self, graph: Graph,
+        schemaview: SchemaView,
+        target_class: Type[Union[BaseModel, YAMLRoot]],
+        prefix_map: Union[Dict[str, str], Converter, None] = None,
+        cast_literals: bool = True,
+        allow_unprocessed_triples: bool = True,
+        ignore_unmapped_predicates: bool = False,
+    ) -> List[Union[BaseModel, YAMLRoot]]:
         """
         Loads objects from graph into lists of the python target_class structure,
         recursively walking RDF graph from instances of target_class.
@@ -58,6 +63,9 @@ def from_rdf_graph(self, graph: Graph, schemaview: SchemaView, target_class: Typ
                     logging.error(f'Inconsistent URI to class map: {uri} -> {c2.name}, {c.name}')
             uri_to_class_map[uri] = c
         # data prefix map: supplements or overrides existing schema prefix map
+        if isinstance(prefix_map, Converter):
+            # TODO replace with `prefix_map = prefix_map.bimap` after making minimum requirement on python 3.8
+            prefix_map = {record.prefix: record.uri_prefix for record in prefix_map.records}
         if prefix_map:
             for k, v in prefix_map.items():
                 namespaces[k] = v
@@ -224,12 +232,16 @@ def _uri_to_id(self, node: VALID_SUBJECT, id_slot: SlotDefinition, schemaview: S
             return schemaview.namespaces().curie_for(node)
 
 
-    def load(self, source: Union[str, TextIO, Graph], target_class: Type[Union[BaseModel, YAMLRoot]], *,
-             schemaview: SchemaView = None,
-             prefix_map: Dict[str, str] = None,
-             fmt: Optional[str] = 'turtle',
-             metadata: Optional[FileInfo] = None,
-             **kwargs) -> Union[BaseModel, YAMLRoot]:
+    def load(
+        self,
+        source: Union[str, TextIO, Graph],
+        target_class: Type[Union[BaseModel, YAMLRoot]], *,
+        schemaview: SchemaView = None,
+        prefix_map: Union[Dict[str, str], Converter, None] = None,
+        fmt: Optional[str] = 'turtle',
+        metadata: Optional[FileInfo] = None,
+        **kwargs,
+    ) -> Union[BaseModel, YAMLRoot]:
         """
         Load the RDF in source into the python target_class structure
 

tests/test_loaders_dumpers/test_rdflib_dumper.py

@@ -4,6 +4,7 @@
 import logging
 from pathlib import Path
 
+from curies import Converter
 from rdflib import Graph, Literal, URIRef
 from rdflib.namespace import RDF, SKOS, XSD
 from rdflib import Namespace
@@ -26,7 +27,7 @@
 DATA_ROUNDTRIP = os.path.join(OUTPUT_DIR, 'example_personinfo_data.roundtrip-rdf.yaml')
 UNMAPPED_ROUNDTRIP = os.path.join(OUTPUT_DIR, 'example_personinfo_data.unmapped-preds.yaml')
 
-prefix_map = {
+PREFIX_MAP = {
     'CODE': 'http://example.org/code/',
     'ROR': 'http://example.org/ror/',
     'P': 'http://example.org/P/',
@@ -105,12 +106,14 @@
 WD = Namespace('http://www.wikidata.org/entity/')
 
 class RdfLibDumperTestCase(unittest.TestCase):
+    def setUp(self) -> None:
+        self.prefix_map = PREFIX_MAP
 
     def test_rdflib_dumper(self):
         view = SchemaView(SCHEMA)
         container = yaml_loader.load(DATA, target_class=Container)
         self._check_objs(view, container)
-        rdflib_dumper.dump(container, schemaview=view, to_file=OUT, prefix_map=prefix_map)
+        rdflib_dumper.dump(container, schemaview=view, to_file=OUT, prefix_map=self.prefix_map)
         g = Graph()
         g.parse(OUT, format='ttl')
 
@@ -134,7 +137,7 @@ def test_rdflib_dumper(self):
         self.assertIn((container, INFO.organizations, ROR['2']), g)
         self.assertIn((container, INFO.persons, P['001']), g)
         self.assertIn((container, INFO.persons, P['002']), g)
-        container: Container = rdflib_loader.load(OUT, target_class=Container, schemaview=view, prefix_map=prefix_map)
+        container: Container = rdflib_loader.load(OUT, target_class=Container, schemaview=view, prefix_map=self.prefix_map)
         self._check_objs(view, container)
         #print(yaml_dumper.dumps(container))
         #person = next(p for p in container.persons if p.id == 'P:002')
@@ -148,7 +151,7 @@ def test_enums(self):
         assert org1type2.meaning
         org1 = Organization('ROR:1', categories=[org1type1, org1type2])
         print(org1.categories)
-        g = rdflib_dumper.as_rdf_graph(org1, schemaview=view, prefix_map=prefix_map)
+        g = rdflib_dumper.as_rdf_graph(org1, schemaview=view, prefix_map=self.prefix_map)
         print(g)
         cats = list(g.objects(ROR['1'], INFO['categories']))
         print(cats)
@@ -173,7 +176,7 @@ def test_rdflib_loader(self):
         tests loading from an RDF graph
         """
         view = SchemaView(SCHEMA)
-        container: Container = rdflib_loader.load(DATA_TTL, target_class=Container, schemaview=view, prefix_map=prefix_map)
+        container: Container = rdflib_loader.load(DATA_TTL, target_class=Container, schemaview=view, prefix_map=self.prefix_map)
         self._check_objs(view, container)
         yaml_dumper.dump(container, to_file=DATA_ROUNDTRIP)
 
@@ -186,10 +189,10 @@ def test_unmapped_predicates(self):
         # default behavior is to raise error on unmapped predicates
         with self.assertRaises(MappingError) as context:
             rdflib_loader.loads(unmapped_predicates_test_ttl, target_class=Person,
-                                schemaview=view, prefix_map=prefix_map)
+                                schemaview=view, prefix_map=self.prefix_map)
         # called can explicitly allow unmapped predicates to be dropped
         person: Person = rdflib_loader.loads(unmapped_predicates_test_ttl, target_class=Person,
-                                                   schemaview=view, prefix_map=prefix_map,
+                                                   schemaview=view, prefix_map=self.prefix_map,
                                                    ignore_unmapped_predicates=True)
         self.assertEqual(person.id, 'P:001')
         self.assertEqual(person.age_in_years, 33)
@@ -204,7 +207,7 @@ def test_any_of_enum(self):
         view = SchemaView(SCHEMA)
         # default behavior is to raise error on unmapped predicates
         person = rdflib_loader.loads(enum_union_type_test_ttl, target_class=Person,
-                                schemaview=view, prefix_map=prefix_map)
+                                schemaview=view, prefix_map=self.prefix_map)
         self.assertEqual(person.id, 'P:001')
         self.assertEqual(person.age_in_years, 33)
         yaml_dumper.dump(person, to_file=UNMAPPED_ROUNDTRIP)
@@ -225,11 +228,11 @@ def test_unmapped_type(self):
         # default behavior is to raise error on unmapped predicates
         with self.assertRaises(DataNotFoundError) as context:
             rdflib_loader.loads(unmapped_type_test_ttl, target_class=Person,
-                                schemaview=view, prefix_map=prefix_map)
+                                schemaview=view, prefix_map=self.prefix_map)
         graph = Graph()
         graph.parse(data=unmapped_type_test_ttl, format='ttl')
         objs = rdflib_loader.from_rdf_graph(graph, target_class=Person,
-                                     schemaview=view, prefix_map=prefix_map)
+                                     schemaview=view, prefix_map=self.prefix_map)
         self.assertEqual(len(objs), 0)
 
     def test_blank_node(self):
@@ -238,7 +241,7 @@ def test_blank_node(self):
         """
         view = SchemaView(SCHEMA)
         address: Address = rdflib_loader.loads(blank_node_test_ttl, target_class=Address,
-                                             schemaview=view, prefix_map=prefix_map,
+                                             schemaview=view, prefix_map=self.prefix_map,
                                              ignore_unmapped_predicates=True)
         self.assertEqual(address.city, 'foo city')
         ttl = rdflib_dumper.dumps(address, schemaview=view)
@@ -368,7 +371,7 @@ def test_phenopackets(self):
             pkt = Phenopacket(id='id with spaces',
                               metaData=MetaData(resources=[Resource(id='id with spaces')]),
                               phenotypicFeatures=[pf])
-            ttl = rdflib_dumper.dumps(pkt, view, prefix_map=prefix_map)
+            ttl = rdflib_dumper.dumps(pkt, view, prefix_map=self.prefix_map)
             g = Graph()
             g.parse(data=ttl, format='ttl')
             self.assertIn(Literal(test_label), list(g.objects(URIRef(expected_uri))),
@@ -380,6 +383,12 @@ def test_phenopackets(self):
 
 
 
+class RDFLibConverterDumperTestCase(RdfLibDumperTestCase):
+    """A test case that uses a :class:`curies.Converter` for testing loading and dumping with RDFLib."""
+
+    def setUp(self) -> None:
+        """Set up the test case using a :class:`curies.Converter` instead of a simple prefix map."""
+        self.prefix_map = Converter.from_prefix_map(PREFIX_MAP)