Add tests and update code

datahub/metadata/migrations/0090_postcodedata.py

@@ -1,4 +1,4 @@
-# Generated by Django 4.2.17 on 2025-02-18 10:35
+# Generated by Django 4.2.17 on 2025-02-19 14:04
 
 from django.db import migrations, models
 import django.db.models.deletion
@@ -18,10 +18,9 @@ class Migration(migrations.Migration):
                 ('disabled_on', models.DateTimeField(blank=True, null=True)),
                 ('id', models.UUIDField(default=uuid.uuid4, primary_key=True, serialize=False)),
                 ('name', models.TextField(blank=True)),
-                ('hashed_uuid', models.CharField(blank=True, default='', max_length=255)),
                 ('postcode', models.CharField(max_length=255)),
                 ('modified_on', models.DateTimeField(auto_now=True, null=True)),
-                ('publication_date', models.DateTimeField()),
+                ('publication_date', models.DateTimeField(blank=True, null=True)),
                 ('postcode_region', models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.SET_NULL, related_name='+', to='metadata.ukregion')),
             ],
             options={

datahub/metadata/models.py

@@ -410,7 +410,6 @@ class ExportBarrierType(BaseOrderedConstantModel):
 class PostcodeData(BaseConstantModel):
     """Postcode data (for the manual addition of a company)."""
 
-    hashed_uuid = models.CharField(max_length=MAX_LENGTH, default='', blank=True)
     postcode = models.CharField(max_length=MAX_LENGTH)
     modified_on = models.DateTimeField(auto_now=True, null=True)
     postcode_region = models.ForeignKey(
@@ -420,4 +419,4 @@ class PostcodeData(BaseConstantModel):
         on_delete=models.SET_NULL,
         related_name='+',
     )
-    publication_date = models.DateTimeField()
+    publication_date = models.DateTimeField(null=True, blank=True)

datahub/metadata/serializers.py

@@ -7,7 +7,7 @@
 from datahub.export_win.models import TeamType
 from datahub.interaction.models import ServiceAnswerOption, ServiceQuestion
 from datahub.metadata.models import (
-    Country, ExchangeRate, OverseasRegion, Service, TeamRole, UKRegion,
+    Country, ExchangeRate, OverseasRegion, PostcodeData, Service, TeamRole, UKRegion,
 )
 
 
@@ -132,8 +132,12 @@ class HVCSerializer(ConstantModelSerializer):
 class PostcodeDataSerializer(ConstantModelSerializer):
     """Postcode data serializer"""
 
-    hashed_uuid = serializers.CharField()
+    id = serializers.UUIDField()
     postcode = serializers.CharField()
     modified_on = serializers.DateTimeField()
     postcode_region = NestedRelatedField(UKRegion, read_only=True)
     publication_date = serializers.DateTimeField()
+
+    class Meta:
+        model = PostcodeData
+        fields = '__all__'

datahub/metadata/tasks.py

@@ -52,7 +52,7 @@ def _should_process_record(self, record: dict) -> bool:
         """Checks whether the record has already been ingested or not."""
         if not self.existing_ids:
             self.existing_ids = set(PostcodeData.objects.values_list(
-                'postcode_data_id', flat=True))
+                'id', flat=True))
 
         postcode_data_id = record.get('id')
         if postcode_data_id in self.existing_ids:
@@ -61,10 +61,6 @@ def _should_process_record(self, record: dict) -> bool:
 
         return True
 
-    def _get_hashed_uuid(self, record: dict) -> str:
-        """Gets the hashed uuid from the incoming record."""
-        return record['hashedUuid']
-
     def _process_record(self, record: dict) -> None:
         """Processes a single record.
 
@@ -73,10 +69,10 @@ def _process_record(self, record: dict) -> None:
         """
         serializer = self.serializer_class(data=record)
         if serializer.is_valid():
-            hashed_uuid = self._get_hashed_uuid(record)
-            queryset = PostcodeData.objects.filter(hashed_uuid=hashed_uuid)
+            primary_key = serializer.validated_data.pop('id')
+            queryset = PostcodeData.objects.filter(pk=primary_key)
             instance, created = queryset.update_or_create(
-                hashed_uuid=hashed_uuid,
+                pk=primary_key,
                 defaults=serializer.validated_data,
             )
             if created:

datahub/metadata/test/factories.py

@@ -1,6 +1,4 @@
-import hashlib
 import uuid
-
 from datetime import timezone
 
 from random import randrange, sample
@@ -129,12 +127,6 @@ class Meta:
         model = 'metadata.AdministrativeArea'
 
 
-def generate_hashed_uuid():
-    new_uuid = uuid.uuid4()
-    hashed_uuid = hashlib.sha256(new_uuid.bytes).hexdigest()
-    return hashed_uuid
-
-
 class PostcodeDataFactory(factory.django.DjangoModelFactory):
     """Postcode data factory"""
 
@@ -149,7 +141,7 @@ class Meta:
 
 def postcode_data_record_faker(overrides: dict | None = None) -> dict:
     data = {
-        'hashedUuid': generate_hashed_uuid(),
+        'id': str(uuid.uuid4()),
         'postcode': fake.postcode(),
         'modified_on': fake.date_time_between(
             start_date='-1y', tzinfo=timezone.utc,

datahub/metadata/test/test_ingest_postcode_data.py

@@ -1,8 +1,11 @@
 import logging
+import uuid
 
 from unittest import mock
+from uuid import uuid4
 
 import pytest
+from faker import Faker
 from moto import mock_aws
 
 from datahub.ingest.boto3 import S3ObjectProcessor
@@ -15,12 +18,13 @@
     PostcodeDataIngestionTask,
 )
 from datahub.metadata.test.factories import (
-    generate_hashed_uuid,
     postcode_data_record_faker,
 )
 
 pytestmark = pytest.mark.django_db
 
+fake = Faker(locale='en_GB')
+
 
 @pytest.fixture
 def postcode_object_key():
@@ -64,16 +68,27 @@ def ingestion_task(self, postcode_object_key):
             serializer_class=PostcodeDataSerializer,
         )
 
-    def test_get_hashed_uuid(self, ingestion_task):
-        record = postcode_data_record_faker()
-        assert ingestion_task._get_hashed_uuid(record) == record['hashedUuid']
+    @pytest.mark.django_db
+    def test_should_process_new_record(self, ingestion_task):
+        new_id = uuid4()
+        record = {'id': new_id}
+
+        assert ingestion_task._should_process_record(record) is True
+
+    @pytest.mark.django_db
+    def test_should_process_existing_record(self, ingestion_task):
+        existing_id = uuid4()
+        PostcodeData.objects.create(id=existing_id)
+        record = {'id': existing_id}
+
+        assert ingestion_task._should_process_record(record) is False
 
     def test_process_record_creates_postcode_data_instance(self, ingestion_task):
-        hashed_uuid = generate_hashed_uuid()
-        record = postcode_data_record_faker({'hashed_uuid': hashed_uuid})
+        primary_key = str(uuid.uuid4())
+        record = postcode_data_record_faker({'id': primary_key})
         ingestion_task._process_record(record)
 
         assert len(ingestion_task.created_ids) == 1
         assert len(ingestion_task.updated_ids) == 0
         assert len(ingestion_task.errors) == 0
-        assert PostcodeData.objects.filter(hashed_uuid=hashed_uuid).exists()
+        assert PostcodeData.objects.filter(pk=primary_key).exists()