test: add fallback for unsupported languages in elastic search indexing

huridocs · Jan 7, 2025 · a63cb39 · a63cb39
1 parent f8bdc5d
commit a63cb39
Show file tree

Hide file tree

Showing 2 changed files with 45 additions and 0 deletions.
diff --git a/app/api/search/specs/entitiesIndex.spec.ts b/app/api/search/specs/entitiesIndex.spec.ts
@@ -4,6 +4,8 @@ import { UserInContextMockFactory } from 'api/utils/testingUserInContext';
 import db from 'api/utils/testing_db';
 import { AccessLevels, PermissionType } from 'shared/types/permissionSchema';
 import { UserRole } from 'shared/types/userSchema';
+import { EntitySchema } from 'shared/types/entityType';
+import { FileType } from 'shared/types/fileType';
 import { elastic } from '../elastic';
 import { reindexAll, updateMapping } from '../entitiesIndex';
 import { search } from '../search';
@@ -159,4 +161,30 @@ describe('entitiesIndex', () => {
       ).toBeUndefined();
     });
   });
+
+  it('should fallback to "other" language if the language is not fully supported by elastic search', async () => {
+    const sharedId = db.id().toString();
+    const entities: EntitySchema[] = [{ sharedId, title: 'Entity 1', language: 'en' }];
+    const files: FileType[] = [
+      {
+        entity: sharedId,
+        originalname: 'file1',
+        filename: 'file1',
+        type: 'document',
+        mimetype: 'application/pdf',
+        language: 'ukr', // Ukrainian is not fully supported by elastic search
+        fullText: {},
+        totalPages: 0,
+      },
+    ];
+
+    await db.setupFixturesAndContext({ entities, files });
+
+    await elasticTesting.reindex();
+
+    const [indexedFile] = await elasticTesting.getIndexedFullTextFromFiles();
+
+    expect(indexedFile.fullText_other).toBeDefined();
+    expect(indexedFile.fullText_undefined).not.toBeDefined();
+  });
 });
diff --git a/app/api/utils/elastic_testing.ts b/app/api/utils/elastic_testing.ts
@@ -27,6 +27,23 @@ const elasticTesting = {
   async getIndexedEntities(sort = 'title.sort') {
     return (await elastic.search({ sort: [sort] })).body.hits.hits.map(i => i._source);
   },
+
+  async getIndexedFullTextFromFiles() {
+    const result = await elastic.search({
+      body: {
+        query: {
+          has_parent: {
+            parent_type: 'entity',
+            query: {
+              match_all: {},
+            },
+          },
+        },
+      },
+    });
+
+    return result.body.hits.hits.map(i => i._source);
+  },
 };
 
 export { elasticTesting };