Add optional confidence

README.md

@@ -55,7 +55,7 @@ http://ark.localhost:8000/api/v1/oauth/providers/gitlab/callback/
 https://ark.localhost/api/v1/oauth/providers/gitlab/callback/
 ```
 
-Once the application is created, GitLab will provide you with an application ID and a secret. Use the `arkindex/config.yaml` file to set them:
+Once the application is created, GitLab will provide you with an application ID and a secret. Use the `arkindex/config.yml` file to set them:
 
 ```yaml
 gitlab:

arkindex/documents/api/entities.py

@@ -226,7 +226,7 @@ class EntityLinkCreate(CreateAPIView):
 ))
 class TranscriptionEntityCreate(CreateAPIView):
     """
-    Link an existing Entity to a given transcription with its position
+    Link an existing Entity to a given transcription with its position.
     """
     permission_classes = (IsVerified, )
     serializer_class = TranscriptionEntitySerializer

arkindex/documents/migrations/0042_transcription_entity_confidence.py

+# Generated by Django 3.2.6 on 2021-08-09 11:34
+
+import django.core.validators
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('dataimport', '0034_worker_run_config'),
+        ('documents', '0041_rotation'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='transcriptionentity',
+            name='confidence',
+            field=models.FloatField(blank=True, null=True, validators=[django.core.validators.MinValueValidator(0), django.core.validators.MaxValueValidator(1)]),
+        ),
+    ]

arkindex/documents/models.py

@@ -531,6 +531,8 @@ class TranscriptionEntity(models.Model):
         blank=True,
     )
 
+    confidence = models.FloatField(validators=[MinValueValidator(0), MaxValueValidator(1)], null=True, blank=True)
+
     class Meta:
         unique_together = (
             ('transcription', 'entity', 'offset', 'length', 'worker_version'),

arkindex/documents/serializers/entities.py

@@ -212,10 +212,11 @@ class TranscriptionEntitySerializer(serializers.ModelSerializer):
         default=None,
         style={'base_template': 'input.html'},
     )
+    confidence = serializers.FloatField(min_value=0, max_value=1, default=None)
 
     class Meta:
         model = TranscriptionEntity
-        fields = ('entity', 'offset', 'length', 'worker_version_id')
+        fields = ('entity', 'offset', 'length', 'worker_version_id', 'confidence')
 
     def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)

arkindex/documents/tests/test_entities_api.py

@@ -82,6 +82,12 @@ class TestEntitiesAPI(FixtureAPITestCase):
             'offset': 4,
             'length': len(self.entity.name)
         }
+        self.tr_entities_confidence_sample = {
+            'entity': str(self.entity.id),
+            'offset': 4,
+            'length': len(self.entity.name),
+            'confidence': 1.0
+        }
         self.tr_entities_version_sample = {
             'entity': str(self.entity.id),
             'offset': 4,
@@ -421,6 +427,34 @@ class TestEntitiesAPI(FixtureAPITestCase):
                 'offset': transcription_entity.offset,
                 'length': transcription_entity.length,
                 'worker_version_id': None,
+                'confidence': None
+            }
+        )
+
+    def test_create_transcription_entity_with_confidence(self):
+        self.client.force_login(self.user)
+        response = self.client.post(
+            reverse('api:transcription-entity-create', kwargs={'pk': str(self.transcription.id)}),
+            data=self.tr_entities_confidence_sample,
+            format='json'
+        )
+        self.assertEqual(response.status_code, status.HTTP_201_CREATED)
+        transcription_entity = TranscriptionEntity.objects.get(
+            transcription=self.transcription,
+            entity=self.entity,
+            offset=self.tr_entities_confidence_sample['offset'],
+            length=self.tr_entities_confidence_sample['length'],
+            confidence=self.tr_entities_confidence_sample['confidence']
+        )
+        self.assertIsNotNone(transcription_entity)
+        self.assertDictEqual(
+            response.json(),
+            {
+                'entity': str(transcription_entity.entity.id),
+                'offset': transcription_entity.offset,
+                'length': transcription_entity.length,
+                'worker_version_id': None,
+                'confidence': transcription_entity.confidence
             }
         )
 
@@ -492,6 +526,42 @@ class TestEntitiesAPI(FixtureAPITestCase):
             {'length': ['Ensure this value is greater than or equal to 1.']}
         )
 
+    def test_create_transcription_entity_wrong_high_confidence(self):
+        self.client.force_login(self.user)
+        response = self.client.post(
+            reverse('api:transcription-entity-create', kwargs={'pk': str(self.transcription.id)}),
+            data={
+                'entity': str(self.entity.id),
+                'offset': 4,
+                'length': 1,
+                'confidence': 2.0
+            },
+            format='json'
+        )
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+        self.assertDictEqual(
+            response.json(),
+            {'confidence': ['Ensure this value is less than or equal to 1.']}
+        )
+
+    def test_create_transcription_entity_wrong_low_confidence(self):
+        self.client.force_login(self.user)
+        response = self.client.post(
+            reverse('api:transcription-entity-create', kwargs={'pk': str(self.transcription.id)}),
+            data={
+                'entity': str(self.entity.id),
+                'offset': 4,
+                'length': 1,
+                'confidence': -0.5
+            },
+            format='json'
+        )
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+        self.assertDictEqual(
+            response.json(),
+            {'confidence': ['Ensure this value is greater than or equal to 0.']}
+        )
+
     def test_create_transcription_entity_different_corpus(self):
         self.client.force_login(self.user)
         ent = Entity.objects.create(
@@ -578,6 +648,7 @@ class TestEntitiesAPI(FixtureAPITestCase):
                 'length': self.transcriptionentity.length,
                 'offset': self.transcriptionentity.offset,
                 'worker_version_id': None,
+                'confidence': None
             }]
         )
 
@@ -628,6 +699,7 @@ class TestEntitiesAPI(FixtureAPITestCase):
                 'length': 8,
                 'offset': 8,
                 'worker_version_id': None,
+                'confidence': None
             }]
         )
 
@@ -748,6 +820,7 @@ class TestEntitiesAPI(FixtureAPITestCase):
                     'offset': t.offset,
                     'length': t.length,
                     'worker_version_id': None,
+                    'confidence': None
                 } for t in TranscriptionEntity.objects.filter(transcription=self.transcription).order_by('offset')],
                 'metadata': [{
                     'entity': {