Merge pull request #441 from biolink/find_dependency

validation: speed up log_error method

kgx/error_detection.py

@@ -119,14 +119,9 @@ def log_error(
 
         if error not in self.errors[level]:
             self.errors[level][error] = dict()
-
-        # don't record duplicate instances of error type and
-        # messages for entity identifiers...
-        if message not in self.errors[level][error]:
-            self.errors[level][error][message] = [entity]
-        else:
-            if entity not in self.errors[level][error][message]:
-                self.errors[level][error][message].append(entity)
+
+        self.errors[level][error][message] = [entity]
+        self.errors[level][error][message].append(entity)
 
     def get_errors(self, level: str = None) -> Dict:
         """

kgx/sink/rdf_sink.py

@@ -237,8 +237,12 @@ def write_edge(self, record: Dict) -> None:
                     value_uri = self._prepare_object(prop, prop_type, value)
                     self._write_triple(URIRef(n), prop_uri, value_uri)
         else:
+            if "type" in record:
+                for type in record["type"]:
+                    if type in associations:
+                        at_least_one_type_in_associations = True
             if (
-                ("type" in record and record["type"] in associations)
+                ("type" in record and at_least_one_type_in_associations)
                 or (
                     "association_type" in record
                     and record["association_type"] in associations

kgx/utils/kgx_utils.py

@@ -21,7 +21,6 @@
 import numpy as np
 from prefixcommons.curie_util import contract_uri
 from prefixcommons.curie_util import expand_uri
-
 from kgx.config import get_logger, get_jsonld_context, get_biolink_model_schema
 from kgx.graph.base_graph import BaseGraph
 
@@ -34,6 +33,8 @@
 CORE_EDGE_PROPERTIES = {"id", "subject", "predicate", "object", "type"}
 XSD_STRING = "xsd:string"
 
+tk = Toolkit()
+
 
 class GraphEntityType(Enum):
     GRAPH = "graph"
@@ -53,6 +54,7 @@ class GraphEntityType(Enum):
     "provided_by": list,
 }
 
+
 column_types = {
     "publications": list,
     "qualifiers": list,
@@ -833,6 +835,7 @@ def _sanitize_import_property(key: str, value: Any, list_delimiter: str) -> Any:
         Sanitized value
 
     """
+
     if key in column_types:
         if column_types[key] == list:
             if isinstance(value, (list, set, tuple)):
@@ -843,7 +846,7 @@ def _sanitize_import_property(key: str, value: Any, list_delimiter: str) -> Any:
                 new_value = list(value)
             elif isinstance(value, str):
                 value = value.replace("\n", " ").replace("\t", " ")
-                new_value = [x for x in value.split(list_delimiter) if x] if list_delimiter else value
+                new_value = [x for x in value.split(list_delimiter) if x] if list_delimiter else [value]
             else:
                 new_value = [str(value).replace("\n", " ").replace("\t", " ")]
             # remove duplication in the list
@@ -871,9 +874,12 @@ def _sanitize_import_property(key: str, value: Any, list_delimiter: str) -> Any:
             ]
             new_value = list(value)
         elif isinstance(value, str):
+            multivalued_slots = [sentencecase_to_snakecase(x) for x in tk.get_all_multivalued_slots()]
             if list_delimiter and list_delimiter in value:
                 value = value.replace("\n", " ").replace("\t", " ")
                 new_value = [x for x in value.split(list_delimiter) if x]
+            elif key in multivalued_slots:
+                new_value = [value]
             else:
                 new_value = value.replace("\n", " ").replace("\t", " ")
         elif isinstance(value, bool):