Update

src/sssom/io.py

@@ -99,8 +99,8 @@ def parse_file(
     # else:
     #     doc = parse_func(
     #         input_path,
-    #         prefix_map=metadata.prefix_map,
-    #         meta=metadata.metadata,
+    #         prefix_map=converter,
+    #         meta=meta,
     #     )
     if clean_prefixes:
         # We do this because we got a lot of prefixes from the default SSSOM prefixes!

src/sssom/parsers.py

@@ -322,7 +322,11 @@ def _address_multivalued_slot(k: str, v: Any) -> Union[str, List[str]]:
 
 def _init_mapping_set(meta: Optional[MetadataType] = None) -> MappingSet:
     _metadata = dict(ChainMap(meta or {}, get_default_metadata()))
-    return MappingSet(mapping_set_id=_metadata["mapping_set_id"], license=_metadata["license"])
+    mapping_set = MappingSet(
+        mapping_set_id=_metadata["mapping_set_id"], license=_metadata["license"]
+    )
+    _set_metadata_in_mapping_set(mapping_set=mapping_set, metadata=meta)
+    return mapping_set
 
 
 def _get_mapping_dict(row: pd.Series, bad_attrs: Counter) -> Dict[str, Any]:
@@ -451,7 +455,6 @@ def from_sssom_rdf(
         _add_valid_mapping_to_list(mdict, mlist, flip_superclass_assertions=True)
 
     ms.mappings = mlist  # type: ignore
-    _set_metadata_in_mapping_set(mapping_set=ms, metadata=meta)
     mdoc = MappingSetDocument(mapping_set=ms, converter=converter)
     return to_mapping_set_dataframe(mdoc)
 
@@ -527,7 +530,6 @@ def from_alignment_minidom(
                     ms[OBJECT_SOURCE] = e.firstChild.nodeValue
 
     ms.mappings = mlist  # type: ignore
-    _set_metadata_in_mapping_set(mapping_set=ms, metadata=meta)
     mapping_set_document = MappingSetDocument(mapping_set=ms, converter=converter)
     return to_mapping_set_dataframe(mapping_set_document)
 
@@ -662,7 +664,6 @@ def from_obographs(
         raise Exception("No graphs element in obographs file, wrong format?")
 
     ms.mappings = mlist  # type: ignore
-    _set_metadata_in_mapping_set(mapping_set=ms, metadata=meta)
     mdoc = MappingSetDocument(mapping_set=ms, converter=converter)
     return to_mapping_set_dataframe(mdoc)
 
@@ -799,7 +800,6 @@ def _get_mapping_set_from_df(df: pd.DataFrame, meta: Optional[MetadataType] = No
         _add_valid_mapping_to_list(mapping_dict, mapping_set.mappings)
     for k, v in bad_attrs.items():
         logging.warning(f"No attr for {k} [{v} instances]")
-    _set_metadata_in_mapping_set(mapping_set=mapping_set, metadata=meta)
     return mapping_set
 
 

src/sssom/util.py

@@ -1009,7 +1009,7 @@ def get_file_extension(file: Union[str, Path, TextIO]) -> str:
     return "tsv"
 
 
-def _extract_global_metadata(msdoc: MappingSetDocument):
+def _extract_global_metadata(msdoc: MappingSetDocument) -> MetadataType:
     """Extract metadata.
 
     :param msdoc: MappingSetDocument object

tests/test_parsers.py

@@ -13,6 +13,7 @@
 import numpy as np
 import pandas as pd
 import yaml
+from curies import Converter
 from rdflib import Graph
 
 from sssom.constants import CURIE_MAP, DEFAULT_LICENSE, SSSOM_URI_PREFIX, get_default_metadata
@@ -64,7 +65,7 @@ def setUp(self) -> None:
 
         with open(f"{test_data_dir}/basic-meta-external.yml") as file:
             df_meta = yaml.safe_load(file)
-            self.df_prefix_map = df_meta.pop(CURIE_MAP)
+            self.df_converter = Converter.from_prefix_map(df_meta.pop(CURIE_MAP))
             self.df_meta = df_meta
 
         self.alignmentxml_file = f"{test_data_dir}/oaei-ordo-hp.rdf"
@@ -117,21 +118,23 @@ def test_parse_obographs(self):
         """Test parsing OBO Graph JSON."""
         msdf = from_obographs(
             jsondoc=self.obographs,
-            prefix_map=self.converter.bimap,
+            prefix_map=self.converter,
             meta=self.metadata,
         )
         path = os.path.join(test_out_dir, "test_parse_obographs.tsv")
         with open(path, "w") as file:
             write_table(msdf, file)
         self.assertEqual(
+            # this number went up from 8099 when the curies.Converter was introduced
+            # since it was able to handle CURIE prefix and URI prefix synonyms
+            8488,
             len(msdf.df),
-            8099,
             f"{self.obographs_file} has the wrong number of mappings.",
         )
 
     def test_parse_tsv(self):
         """Test parsing TSV."""
-        msdf = from_sssom_dataframe(df=self.df, prefix_map=self.df_prefix_map, meta=self.df_meta)
+        msdf = from_sssom_dataframe(df=self.df, prefix_map=self.df_converter, meta=self.df_meta)
         path = os.path.join(test_out_dir, "test_parse_tsv.tsv")
         with open(path, "w") as file:
             write_table(msdf, file)
@@ -145,7 +148,7 @@ def test_parse_alignment_minidom(self):
         """Test parsing an alignment XML."""
         msdf = from_alignment_minidom(
             dom=self.alignmentxml,
-            prefix_map=self.converter.bimap,
+            prefix_map=self.converter,
             meta=self.metadata,
         )
         path = os.path.join(test_out_dir, "test_parse_alignment_minidom.tsv")
@@ -244,7 +247,7 @@ def test_parse_alignment_xml(self):
 
     def test_parse_sssom_rdf(self):
         """Test parsing RDF."""
-        msdf = from_sssom_rdf(g=self.rdf_graph, prefix_map=self.df_prefix_map, meta=self.metadata)
+        msdf = from_sssom_rdf(g=self.rdf_graph, prefix_map=self.df_converter, meta=self.metadata)
         path = os.path.join(test_out_dir, "test_parse_sssom_rdf.tsv")
         with open(path, "w") as file:
             write_table(msdf, file)
@@ -258,7 +261,7 @@ def test_parse_sssom_json(self):
         """Test parsing JSON."""
         msdf = from_sssom_json(
             jsondoc=self.json,
-            prefix_map=self.df_prefix_map,
+            prefix_map=self.df_converter,
             meta=self.metadata,
         )
         path = os.path.join(test_out_dir, "test_parse_sssom_json.tsv")