Update post-transform methods to handle None values (#147)

Why these changes are being introduced:
* When a field method returns "None", the resulting "fields" dictionary will
set a key-value pair, where key = <field_name> and value = None. This will result
in the .setdefault() method being ignored during the post-transform methods
as it will only set the value in the "fields" dictionary to an empty list if
the key is not present.

How this addresses that need:
* Check if field methods that are updated by post-transform methods are
already set to None. If None, set to an empty list.

Side effects of this change:
* None

Relevant ticket(s):
* https://mitlibraries.atlassian.net/browse/GDT-217
* https://mitlibraries.atlassian.net/browse/GDT-210

Author: jonavellecuerdo

tests/sources/test_transformer.py

@@ -46,7 +46,24 @@ def test_create_dates_and_locations_from_publishers_success():
     }
 
 
-def test_create_locations_from_spatial_subjects():
+def test_create_dates_and_locations_from_publishers_when_fields_are_none_success():
+    fields = {
+        "publishers": [
+            timdex.Publisher(name="Publisher", date="Date", location="Location")
+        ],
+        "dates": None,
+        "locations": None,
+    }
+    assert Transformer.create_dates_and_locations_from_publishers(fields) == {
+        "publishers": [
+            timdex.Publisher(name="Publisher", date="Date", location="Location")
+        ],
+        "dates": [timdex.Date(kind="Publication date", value="Date")],
+        "locations": [timdex.Location(value="Location", kind="Place of Publication")],
+    }
+
+
+def test_create_locations_from_spatial_subjects_success():
     fields = {
         "subjects": [
             timdex.Subject(
@@ -70,6 +87,31 @@ def test_create_locations_from_spatial_subjects():
     }
 
 
+def test_create_locations_from_spatial_subjects_when_field_is_none_success():
+    fields = {
+        "subjects": [
+            timdex.Subject(
+                value=["Some city, Some country"], kind="Dublin Core; Spatial"
+            ),
+            timdex.Subject(value=["City 1", "City 2"], kind="Dublin Core; Spatial"),
+        ],
+        "locations": None,
+    }
+    assert Transformer.create_locations_from_spatial_subjects(fields) == {
+        "subjects": [
+            timdex.Subject(
+                value=["Some city, Some country"], kind="Dublin Core; Spatial"
+            ),
+            timdex.Subject(value=["City 1", "City 2"], kind="Dublin Core; Spatial"),
+        ],
+        "locations": [
+            timdex.Location(value="Some city, Some country", kind="Place Name"),
+            timdex.Location(value="City 1", kind="Place Name"),
+            timdex.Location(value="City 2", kind="Place Name"),
+        ],
+    }
+
+
 def test_xmltransformer_initializes_with_expected_attributes(oai_pmh_records):
     transformer = XMLTransformer("cool-repo", oai_pmh_records)
     assert transformer.source == "cool-repo"

transmogrifier/sources/transformer.py

@@ -384,13 +384,17 @@ def create_dates_and_locations_from_publishers(fields: dict) -> dict:
                 publisher_date = timdex.Date(
                     kind="Publication date", value=publisher.date
                 )
-                if publisher_date not in fields.setdefault("dates", []):
+                if fields.get("dates") is None:
+                    fields["dates"] = []
+                if publisher_date not in fields["dates"]:
                     fields["dates"].append(publisher_date)
             if publisher.location:
                 publisher_location = timdex.Location(
                     kind="Place of Publication", value=publisher.location
                 )
-                if publisher_location not in fields.setdefault("locations", []):
+                if fields.get("locations") is None:
+                    fields["locations"] = []
+                if publisher_location not in fields["locations"]:
                     fields["locations"].append(publisher_location)
         return fields
 
@@ -414,6 +418,8 @@ def create_locations_from_spatial_subjects(fields: dict) -> dict:
         for subject in spatial_subjects:
             for place_name in subject.value:
                 subject_location = timdex.Location(value=place_name, kind="Place Name")
-                if subject_location not in fields.setdefault("locations", []):
+                if fields.get("locations") is None:
+                    fields["locations"] = []
+                if subject_location not in fields["locations"]:
                     fields["locations"].append(subject_location)
         return fields