Add schema version in library indexes, improve reindexing command

openedx/core/djangoapps/content_libraries/libraries_index.py

@@ -39,6 +39,8 @@ class ContentLibraryIndexer:
     INDEX_NAME = "content_library_index"
     LIBRARY_DOCUMENT_TYPE = "content_library"
 
+    SCHEMA_VERSION = 0
+
     @classmethod
     def index_libraries(cls, library_keys):
         """
@@ -60,7 +62,10 @@ class ContentLibraryIndexer:
 
             bundle_metadata = get_bundle(ref.bundle_uuid)
 
+            # NOTE: Increment ContentLibraryIndexer.SCHEMA_VERSION if the following schema is updated to avoid dealing
+            # with outdated indexes which might cause errors due to missing/invalid attributes.
             library_dict = {
+                "schema_version": ContentLibraryIndexer.SCHEMA_VERSION,
                 "id": str(library_key),
                 "uuid": str(bundle_metadata.uuid),
                 "title": bundle_metadata.title,
@@ -84,7 +89,10 @@ class ContentLibraryIndexer:
         library_keys_str = [str(key) for key in library_keys]
         response = searcher.search(
             doc_type=cls.LIBRARY_DOCUMENT_TYPE,
-            field_dictionary={"id": library_keys_str},
+            field_dictionary={
+                "id": library_keys_str,
+                "schema_version": ContentLibraryIndexer.SCHEMA_VERSION
+            },
             size=MAX_SIZE,
         )
 

openedx/core/djangoapps/content_libraries/management/commands/reindex_content_library.py

@@ -1,6 +1,8 @@
 """ Management command to update content libraries' search index """
 
 
+import logging
+
 from textwrap import dedent
 
 from django.core.management import BaseCommand
@@ -43,20 +45,29 @@ class Command(BaseCommand):
             dest='all',
             help='Reindex all libraries'
         )
+        parser.add_argument(
+            '--force',
+            action='store_true',
+            dest='force',
+            help='Run command without user prompt for confirmation'
+        )
         parser.add_argument('library_ids', nargs='*')
 
     def handle(self, *args, **options):
         if options['clear-all']:
-            if query_yes_no(self.CONFIRMATION_PROMPT_CLEAR, default="no"):
+            if options['force'] or query_yes_no(self.CONFIRMATION_PROMPT_CLEAR, default="no"):
+                logging.info("Removing all libraries from the index")
                 ContentLibraryIndexer.remove_all_libraries()
             return
 
         if options['all']:
-            if query_yes_no(self.CONFIRMATION_PROMPT_ALL, default="no"):
+            if options['force'] or query_yes_no(self.CONFIRMATION_PROMPT_ALL, default="no"):
+                logging.info("Indexing all libraries")
                 library_keys = [library.library_key for library in ContentLibrary.objects.all()]
             else:
                 return
         else:
+            logging.info("Indexing libraries: {}".format(options['library_ids']))
             library_keys = list(map(LibraryLocatorV2.from_string, options['library_ids']))
 
         ContentLibraryIndexer.index_libraries(library_keys)

openedx/core/djangoapps/content_libraries/tests/test_libraries_index.py

@@ -3,9 +3,11 @@ Testing indexing of blockstore based content libraries
 """
 
 from django.conf import settings
+from django.core.management import call_command
 from django.test.utils import override_settings
-from search.search_engine_base import SearchEngine
+from mock import patch
 from opaque_keys.edx.locator import LibraryLocatorV2
+from search.search_engine_base import SearchEngine
 
 from openedx.core.djangoapps.content_libraries.libraries_index import ContentLibraryIndexer, LibraryNotIndexedException
 from openedx.core.djangoapps.content_libraries.tests.base import ContentLibrariesRestApiTest
@@ -47,6 +49,24 @@ class ContentLibraryIndexerIndexer(ContentLibrariesRestApiTest):
             self.assertEqual(response['has_unpublished_changes'], False)
             self.assertEqual(response['has_unpublished_deletes'], False)
 
+    def test_schema_updates(self):
+        """
+        Test that outdated indexes aren't retrieved
+        """
+        result = self._create_library(slug="test-lib-schemaupdates-1", title="Title 1", description="Description")
+        library_key = LibraryLocatorV2.from_string(result['id'])
+
+        ContentLibraryIndexer.get_libraries([library_key])
+
+        with patch("openedx.core.djangoapps.content_libraries.libraries_index.ContentLibraryIndexer.SCHEMA_VERSION",
+                   new=1):
+            with self.assertRaises(LibraryNotIndexedException):
+                ContentLibraryIndexer.get_libraries([library_key])
+
+            call_command("reindex_content_library", all=True, quiet=True)
+
+            ContentLibraryIndexer.get_libraries([library_key])
+
     def test_remove_all_libraries(self):
         """
         Test if remove_all_libraries() deletes all libraries