Fix

Author: nick863

infra/core/ai/hub-dependencies.bicep

@@ -142,6 +142,7 @@ module searchService '../search/search-services.bicep' =
       location: location
       tags: tags
       name: searchServiceName
+      semanticSearch: 'free'
       authOptions: { aadOrApiKey: { aadAuthFailureMode: 'http401WithBearerChallenge'}}
     }
   }

src/api/search_index_manager.py

@@ -167,6 +167,7 @@ async def semantic_search(self, message: str) -> str:
         response = await self._get_client().search(
             search_text=message,
             query_type="semantic",
+            search_fields=['token', 'document_reference'],
             semantic_configuration_name=SearchIndexManager._SEMANTIC_CONFIG,
         )
         return await self._format_search_results(response)
@@ -277,8 +278,10 @@ async def _index_create(self, vector_index_dimensions: int) -> SearchIndex:
                     SemanticConfiguration(
                         name=SearchIndexManager._SEMANTIC_CONFIG,
                         prioritized_fields=SemanticPrioritizedFields(
-                            title_field=SemanticField(field_name="embedId"),
-                            content_fields=[SemanticField(field_name="token")]
+                            title_field=SemanticField(field_name="document_reference"),
+                            content_fields=[
+                                SemanticField(field_name="token"),
+                            ]
                         )
                     )
                 ] 

src/gunicorn.conf.py

@@ -124,8 +124,6 @@ async def get_available_toolset(
     :param creds: The credentials, used for the index.
     :return: The tool set, available based on the environment.
     """
-    # File name -> {"id": file_id, "path": file_path}
-    files: Dict[str, Dict[str, str]] = {}
     # First try to get an index search.
     conn_id = ""
     if os.environ.get('AZURE_AI_SEARCH_INDEX_NAME'):
@@ -145,20 +143,11 @@ async def get_available_toolset(
 
         toolset.add(ai_search)
         logger.info("agent: initialized index")
-        # Populate file links.
-        embeddings_path = os.path.join(
-            os.path.dirname(__file__), 'data', 'embeddings.csv')
-        with open(embeddings_path, newline='') as fp:
-            reader = csv.DictReader(fp)
-            for row in reader:
-                if row['document_reference'] in FILES_NAMES:
-                    files[row['document_reference']] = {
-                        "id": row['document_reference'],
-                        "path": _get_file_path(row['document_reference'])
-                    }
     else:
         logger.info(
             "agent: index was not initialized, falling back to file search.")
+        # File name -> {"id": file_id, "path": file_path}
+        files: Dict[str, Dict[str, str]] = {}
         # Upload files for file search
         for file_name in FILES_NAMES:
             file_path = _get_file_path(file_name)
@@ -167,6 +156,9 @@ async def get_available_toolset(
             # Store both file id and the file path using the file name as key.
             files[file_name] = {"id": file.id, "path": file_path}
 
+        # Serialize and store files information in the environment variable (so
+        # workers see it)
+        os.environ["UPLOADED_FILE_MAP"] = json.dumps(files)
         logger.info(
             f"Set env UPLOADED_FILE_MAP = {os.environ['UPLOADED_FILE_MAP']}")
 
@@ -179,9 +171,7 @@ async def get_available_toolset(
 
         file_search_tool = FileSearchTool(vector_store_ids=[vector_store.id])
         toolset.add(file_search_tool)
-    # Serialize and store files information in the environment variable (so
-    # workers see it)
-    os.environ["UPLOADED_FILE_MAP"] = json.dumps(files)
+
     return toolset
 
 

tests/test_search_index_manager.py

@@ -121,7 +121,7 @@ async def test_exception_different_dimmensions(self):
                 "from dimensions provided to constructor."):
             await rag.create_index(vector_index_dimensions=42)
 
-    @unittest.skip("Only for live tests.")
+    #@unittest.skip("Only for live tests.")
     async def test_e2e(self):
         """Run search end to end."""
         async with DefaultAzureCredential() as creds:
@@ -153,9 +153,13 @@ async def test_e2e(self):
                 result = await rag.search(
                     "What is the temperature rating "
                     "of the cozynights sleeping bag?")
+                result_semantic = await rag.semantic_search(
+                    "What is the temperature rating "
+                    "of the cozynights sleeping bag?")
                 await rag.delete_index()
                 await rag.close()
-                self.assertTrue(bool(result))
+                self.assertTrue(bool(result), "The regular search is empty.")
+                self.assertTrue(bool(result_semantic), "The semantic search is empty.")
 
     async def test_life_cycle_mock(self):
         """Test create, upload, search and delete"""