Merge branch 'corpus-worker-version-list' into 'master'

Add ListCorpusWorkerVersions endpoint

Closes #537

See merge request !1095

arkindex/dataimport/api.py

@@ -4,7 +4,7 @@ from uuid import UUID
 from django.conf import settings
 from django.core.exceptions import PermissionDenied
 from django.db import transaction
-from django.db.models import F, Max, Q
+from django.db.models import Count, F, Max, Q
 from django.http.response import Http404
 from django.shortcuts import get_object_or_404
 from rest_framework import status
@@ -743,6 +743,27 @@ class WorkerVersionList(ListCreateAPIView):
         return Response(WorkerVersionSerializer(version).data, status=reponse_status)
 
 
+class CorpusWorkerVersionList(CorpusACLMixin, ListAPIView):
+    """
+    List worker versions used by elements of a given corpus.
+    """
+    permission_classes = (IsVerified, )
+    pagination_class = None
+    serializer_class = WorkerVersionSerializer
+    openapi_overrides = {
+        'operationId': 'ListCorpusWorkerVersions',
+        'tags': ['ml'],
+    }
+
+    def get_queryset(self):
+        return WorkerVersion.objects \
+            .filter(elements__corpus_id=self.kwargs['pk']) \
+            .select_related('revision__repo', 'worker__repository') \
+            .prefetch_related('revision__refs', 'revision__versions') \
+            .order_by('-revision__created') \
+            .annotate(element_count=Count('id'))
+
+
 class WorkerVersionRetrieve(CorpusACLMixin, RetrieveUpdateAPIView):
     """
     Retrieve a specific worker version

arkindex/dataimport/serializers/workers.py

@@ -34,6 +34,7 @@ class WorkerVersionSerializer(serializers.ModelSerializer):
     # ModelSerializer does not yet support Django 3.1's JSONField
     # https://github.com/encode/django-rest-framework/pull/7467
     configuration = serializers.JSONField()
+    element_count = serializers.IntegerField(read_only=True)
 
     class Meta:
         model = WorkerVersion
@@ -46,6 +47,7 @@ class WorkerVersionSerializer(serializers.ModelSerializer):
             'docker_image_name',
             'state',
             'worker',
+            'element_count',
         )
         read_only_fields = ('docker_image_name',)
         # Avoid loading all revisions and all Ponos artifacts when opening this endpoint in a browser

arkindex/dataimport/tests/test_workers.py

+from unittest import expectedFailure
+
 from django.urls import reverse
 from rest_framework import status
 
@@ -489,3 +491,58 @@ class TestWorkersWorkerVersions(FixtureAPITestCase):
         self.version_1.configuration = {'docker': {'command': 'mysupercommand'}}
         self.assertEqual(self.version_1.docker_command, 'mysupercommand')
         self.version_1.configuration = {"test": "test1"}
+
+    def test_corpus_worker_version_requires_login(self):
+        with self.assertNumQueries(0):
+            response = self.client.get(reverse('api:corpus-versions', kwargs={'pk': self.corpus.id}))
+            self.assertEqual(response.status_code, status.HTTP_403_FORBIDDEN)
+
+    @expectedFailure
+    def test_corpus_worker_version_requires_verified(self):
+        """
+        This test fails due to a bug in the IsVerified permission class.
+        https://gitlab.com/arkindex/backend/-/issues/554
+        """
+        self.user.verified_email = False
+        self.user.save()
+        self.client.force_login(self.user)
+        with self.assertNumQueries(2):
+            response = self.client.get(reverse('api:corpus-versions', kwargs={'pk': self.corpus.id}))
+            self.assertEqual(response.status_code, status.HTTP_403_FORBIDDEN)
+
+    def test_corpus_worker_version_list(self):
+        self.client.force_login(self.user)
+
+        self.corpus.elements.filter(type__slug='word').update(worker_version=self.version_1)
+
+        with self.assertNumQueries(5):
+            response = self.client.get(reverse('api:corpus-versions', kwargs={'pk': self.corpus.id}))
+            self.assertEqual(response.status_code, status.HTTP_200_OK)
+
+        self.assertListEqual(response.json(), [
+            {
+                'id': str(self.version_1.id),
+                'configuration': {'test': 42},
+                'revision': {
+                    'id': str(self.version_1.revision_id),
+                    'state': 'available',
+                    'hash': '1337',
+                    'author': 'Test user',
+                    'message': 'My w0rk3r',
+                    'created': '2020-02-02T01:23:45.678000Z',
+                    'commit_url': 'http://my_repo.fake/workers/worker/commit/1337',
+                    'refs': []
+                },
+                'docker_image': str(self.version_1.docker_image_id),
+                'docker_image_iid': None,
+                'docker_image_name': self.version_1.docker_image_name,
+                'state': 'available',
+                'worker': {
+                    'id': str(self.worker_1.id),
+                    'name': 'Recognizer',
+                    'slug': 'reco',
+                    'type': 'recognizer'
+                },
+                'element_count': 9
+            }
+        ])

arkindex/project/api_v1.py

@@ -6,6 +6,7 @@ from rest_framework.schemas import get_schema_view
 
 from arkindex.dataimport.api import (
     AvailableRepositoriesList,
+    CorpusWorkerVersionList,
     CorpusWorkflow,
     DataFileCreate,
     DataFileList,
@@ -139,6 +140,7 @@ api = [
     path('corpus/<uuid:pk>/roles/', CorpusRoles.as_view(), name='corpus-roles'),
     path('corpus/<uuid:pk>/ml-stats/', CorpusMLStats.as_view(), name='corpus-ml-stats'),
     path('corpus/<uuid:pk>/allowed-metadata/', CorpusAllowedMetaData.as_view(), name='corpus-allowed-metadata'),
+    path('corpus/<uuid:pk>/versions/', CorpusWorkerVersionList.as_view(), name='corpus-versions'),
 
     # Moderation
     path('ml-classes/', MLClassList.as_view(), name='mlclass-list'),