Move guess_category to transform_utils

kg_covid_19/transform_utils/sars_cov_2_gene_annot/sars_cov_2_gene_annot.py

@@ -4,7 +4,7 @@
 import os
 from typing import Generator, TextIO, List, Optional
 
-from kg_covid_19.utils.transform_utils import get_item_by_priority, ItemInDictNotFound
+from kg_covid_19.utils.transform_utils import get_item_by_priority, ItemInDictNotFound, guess_category
 
 from kg_covid_19.transform_utils.transform import Transform
 from kg_covid_19.utils import write_node_edge_item
@@ -66,30 +66,17 @@ def run(self, data_file: str = None):
                     edge_data = self.gpa_to_edge_data(rec)
                     subject_node = edge_data[0]
                     if subject_node not in seen:
-                        subject_node_data = [subject_node, self.guess_category(subject_node)] + [""] * 4 + [self.source_name]
+                        subject_node_data = [subject_node, guess_category(subject_node)] + [""] * 4 + [self.source_name]
                         write_node_edge_item(node, self.node_header, subject_node_data)
                         seen.add(subject_node)
                     object_node = edge_data[2]
                     if object_node not in seen:
-                        object_node_data = [object_node, self.guess_category(object_node)] + [""] * 4 + [self.source_name]
+                        object_node_data = [object_node, guess_category(object_node)] + [""] * 4 + [self.source_name]
                         write_node_edge_item(node, self.node_header, object_node_data)
                         seen.add(object_node)
 
                     write_node_edge_item(edge, self.edge_header, edge_data)
 
-    def guess_category(self, identifier):
-        """Guess category for a given identifier.
-
-        """
-        prefix = identifier.split(':')[0]
-        if prefix in {'UniProtKB', 'ComplexPortal'}:
-            category = 'biolink:Protein'
-        elif prefix in {'GO'}:
-            category = 'biolink:OntologyClass'
-        else:
-            category = 'biolink:NamedThing'
-        return category
-
     def gpa_to_edge_data(self, rec: dict) -> list:
         """given a parsed gpa entry, return an edge with the annotations
 

kg_covid_19/utils/transform_utils.py

@@ -154,3 +154,25 @@ def parse_header(header_string: str, sep: str = '\t') -> List:
 def unzip_to_tempdir(zip_file_name: str, tempdir: str) -> None:
     with zipfile.ZipFile(zip_file_name, 'r') as z:
         z.extractall(tempdir)
+
+
+def guess_category(identifier: str):
+    """Guess category for a given identifier.
+
+    Note: This is a temporary solution and should not be used long term.
+
+    Args:
+        identifier: A CURIE
+
+    Returns:
+        The category for the given CURIE
+
+    """
+    prefix = identifier.split(':')[0]
+    if prefix in {'UniProtKB', 'ComplexPortal'}:
+        category = 'biolink:Protein'
+    elif prefix in {'GO'}:
+        category = 'biolink:OntologyClass'
+    else:
+        category = 'biolink:NamedThing'
+    return category