Skip children fetching parameter for thumbnails generation only

arkindex/dataimport/models.py

@@ -122,6 +122,11 @@ class DataImport(IndexableModel):
             if elements:
                 command = ' '.join([command, '--dataimport-id {}'.format(str(self.id))])
 
+            # Do not retrieve elements children in case there of thumbnails generation only
+            thumbnails = self.payload.get('thumbnails')
+            if thumbnails and len(self.ml_tools) == 0:
+                command = ' '.join([command, '--skip-children'.format(str(self.id))])
+
             tasks = {
                 import_task_name: {
                     'image': settings.ARKINDEX_TASKS_IMAGE,

arkindex/dataimport/tests/test_workflows_api.py

@@ -499,3 +499,31 @@ class TestWorkflows(FixtureAPITestCase):
             'python -m arkindex_tasks.init_elements --corpus-id {} --chunks-number 1 --dataimport-id {}'
             .format(str(self.corpus.id), str(dataimport.id))
         )
+
+    def test_thumbnails_generation_only(self, ml_get_mock):
+        """
+        Generating thumbnails without any workflow must generate an import task
+        which do not to retrieve filtered elements children
+        """
+        self.client.force_login(self.user)
+        corpus_type = self.corpus.types.first()
+        response = self.client.post(
+            reverse('api:corpus-workflow'),
+            {
+                'ml_tools': [],
+                'corpus': str(self.corpus.id),
+                'chunks': 3,
+                'type': corpus_type.slug,
+                'thumbnails': True
+            },
+            format='json'
+        )
+        self.assertEqual(response.status_code, status.HTTP_201_CREATED)
+        data = response.json()
+        dataimport = DataImport.objects.get(id=data['id'])
+        workflow = dataimport.workflow
+        self.assertEqual(
+            workflow.recipes['initialisation'].command,
+            'python -m arkindex_tasks.init_elements --corpus-id {} --chunks-number 3 --type {} --skip-children'
+            .format(str(self.corpus.id), corpus_type.slug)
+        )