Merge branch 'main' of https://github.com/Azure-Samples/chat-with-you…

…r-data-solution-accelerator into PSL-BUG-5536

code/backend/batch/utilities/search/search_handler_base.py

@@ -12,10 +12,15 @@ def __init__(self, env_helper: EnvHelper):
         self.env_helper = env_helper
         self.search_client = self.create_search_client()
 
-    def search_with_facets(self, query: str, facets: list[str]):
+    def search_with_facets(self, query: str, facet: str, facet_count: int):
         if self.search_client is None:
             return None
-        return self.search_client.search(query, facets=facets)
+
+        # Construct facet parameter based on facet_count
+        facet_param = f"{facet},count:{facet_count}"
+
+        # Perform search with facets and facet_param
+        return self.search_client.search(query, facets=[facet_param])
 
     def get_unique_files(self, results, facet_key: str):
         if results:

code/backend/pages/02_Explore_Data.py

@@ -40,7 +40,7 @@ def load_css(file_path):
 try:
     search_handler = Search.get_search_handler(env_helper)
 
-    results = search_handler.search_with_facets("*", ["title"])
+    results = search_handler.search_with_facets("*", "title", facet_count=0)
     unique_files = search_handler.get_unique_files(results, "title")
     filename = st.selectbox("Select your file:", unique_files)
     st.write("Showing chunks for:", filename)

code/backend/pages/04_Configuration.py

@@ -208,7 +208,7 @@ def validate_documents():
     )
     example_user_question_help = "The example user question."
     example_answer_help = "The expected answer."
-    with st.expander("", expanded=True):
+    with st.expander("Assistant type configuration", expanded=True):
         cols = st.columns([2, 4])
         with cols[0]:
             st.selectbox(

code/create_app.py

@@ -9,7 +9,7 @@
 from os import path
 import sys
 import requests
-from openai import AzureOpenAI, Stream, RateLimitError
+from openai import AzureOpenAI, Stream, RateLimitError, APIStatusError
 from openai.types.chat import ChatCompletionChunk
 from flask import Flask, Response, request, Request, jsonify
 from dotenv import load_dotenv
@@ -24,7 +24,6 @@
 ERROR_GENERIC_MESSAGE = "An error occurred. Please try again. If the problem persists, please contact the site administrator."
 logger = logging.getLogger(__name__)
 
-
 def stream_with_data(response: Stream[ChatCompletionChunk]):
     """This function streams the response from Azure OpenAI with data."""
     response_obj = {
@@ -345,10 +344,15 @@ def conversation_azure_byod():
                 return conversation_with_data(request, env_helper)
             else:
                 return conversation_without_data(request, env_helper)
-        except RateLimitError as e:
+        except APIStatusError as e:
             error_message = str(e)
             logger.exception("Exception in /api/conversation | %s", error_message)
-            return jsonify({"error": ERROR_429_MESSAGE}), 429
+            response_json = e.response.json()
+            response_message = response_json.get("error", {}).get("message", "")
+            response_code = response_json.get("error", {}).get("code", "")
+            if response_code == "429" or "429" in response_message:
+                return jsonify({"error": ERROR_429_MESSAGE}), 429
+            return jsonify({"error": ERROR_GENERIC_MESSAGE}), 500
         except Exception as e:
             error_message = str(e)
             logger.exception("Exception in /api/conversation | %s", error_message)
@@ -384,10 +388,15 @@ async def conversation_custom():
 
             return jsonify(response_obj), 200
 
-        except RateLimitError as e:
+        except APIStatusError as e:
             error_message = str(e)
             logger.exception("Exception in /api/conversation | %s", error_message)
-            return jsonify({"error": ERROR_429_MESSAGE}), 429
+            response_json = e.response.json()
+            response_message = response_json.get("error", {}).get("message", "")
+            response_code = response_json.get("error", {}).get("code", "")
+            if response_code == "429" or "429" in response_message:
+                return jsonify({"error": ERROR_429_MESSAGE}), 429
+            return jsonify({"error": ERROR_GENERIC_MESSAGE}), 500
         except Exception as e:
             error_message = str(e)
             logger.exception("Exception in /api/conversation | %s", error_message)

code/tests/search_utilities/test_azure_search_handler.py

@@ -364,6 +364,49 @@ def test_semantic_search_with_advanced_image_processing(
         top=handler.env_helper.AZURE_SEARCH_TOP_K,
     )
 
+@pytest.fixture
+def search_handler():
+    env_helper=Mock()
+    return AzureSearchHandler(env_helper)
+
+def test_search_with_facets_no_search_client(search_handler):
+
+    search_handler.search_client = None
+    result = search_handler.search_with_facets("query", "facet1", 10)
+    assert result is None
+
+
+def test_search_with_facets_valid(search_handler):
+    mock_search_client = MagicMock()
+    search_handler.search_client = mock_search_client
+    mock_search_client.search.return_value = "search_results"
+    result = search_handler.search_with_facets("query", "facet1", 10)
+    mock_search_client.search.assert_called_once_with(
+        "query", facets=["facet1,count:10"]
+    )
+    assert result == "search_results"
+
+
+
+def test_get_unique_files_no_results(search_handler):
+    results = None
+    facet_key = "facet_key"
+    result = search_handler.get_unique_files(results, facet_key)
+    assert(result, [])
+
+
+def test_get_unique_files_with_results(search_handler):
+    mock_results = MagicMock()
+    mock_results.get_facets.return_value = {
+        "facet_key": [
+            {"value": "file1"},
+            {"value": "file2"},
+        ]
+    }
+    facet_key = "facet_key"
+    result = search_handler.get_unique_files(mock_results, facet_key)
+    assert(result, ["file1", "file2"])
+
 
 def test_delete_from_index(handler, mock_search_client):
     # given

code/tests/test_app.py

@@ -3,7 +3,9 @@
 """
 
 from unittest.mock import AsyncMock, MagicMock, Mock, patch, ANY
-from openai import RateLimitError
+
+from httpx import Response
+from openai import RateLimitError, BadRequestError, InternalServerError
 import pytest
 from flask.testing import FlaskClient
 from backend.batch.utilities.helpers.config.conversation_flow import ConversationFlow
@@ -332,6 +334,15 @@ def test_conversation_custom_returns_error_response_on_rate_limit_error(
         # given
         response_mock = Mock()
         response_mock.status_code = 429
+        response_mock.json.return_value = {
+            'error': {
+                'code': "429",
+                'message': 'Requests to the Embeddings_Create Operation under Azure OpenAI API version 2024-02-01 '
+                           'have exceeded call rate limit of your current OpenAI S0 pricing tier. Please retry after '
+                           '2 seconds. Please go here: https://aka.ms/oai/quotaincrease if you would like to further '
+                           'increase the default rate limit.'
+            }
+        }
         body_mock = {"error": "Rate limit exceeded"}
 
         rate_limit_error = RateLimitError("Rate limit exceeded", response=response_mock, body=body_mock)
@@ -351,6 +362,30 @@ def test_conversation_custom_returns_error_response_on_rate_limit_error(
                      "Please wait a moment and try again."
         }
 
+    @patch("create_app.get_orchestrator_config")
+    def test_conversation_custom_returns_500_when_internalservererror_occurs(
+            self, get_orchestrator_config_mock, env_helper_mock, client
+    ):
+        """Test that an error response is returned when an exception occurs."""
+        # given
+        response_mock = MagicMock()
+        response_mock.status_code = 500
+        get_orchestrator_config_mock.side_effect = InternalServerError("Test exception", response=response_mock, body="")
+
+        # when
+        response = client.post(
+            "/api/conversation",
+            headers={"content-type": "application/json"},
+            json=self.body,
+        )
+
+        # then
+        assert response.status_code == 500
+        assert response.json == {
+            "error": "An error occurred. Please try again. If the problem persists, please contact the site "
+                     "administrator."
+        }
+
     @patch("create_app.get_message_orchestrator")
     @patch("create_app.get_orchestrator_config")
     def test_conversation_custom_allows_multiple_messages_from_user(
@@ -719,18 +754,51 @@ def test_conversation_azure_byod_returns_500_when_exception_occurs(
             "error": "An error occurred. Please try again. If the problem persists, please contact the site administrator."
         }
 
+    @patch("create_app.conversation_with_data")
+    def test_conversation_azure_byod_returns_500_when_internalservererror_occurs(
+            self, conversation_with_data_mock, env_helper_mock, client
+    ):
+        """Test that an error response is returned when an exception occurs."""
+        # given
+        response_mock = MagicMock()
+        response_mock.status_code = 500
+        conversation_with_data_mock.side_effect = InternalServerError("Test exception", response=response_mock, body="")
+        env_helper_mock.CONVERSATION_FLOW = ConversationFlow.BYOD.value
+
+        # when
+        response = client.post(
+            "/api/conversation",
+            headers={"content-type": "application/json"},
+            json=self.body,
+        )
+
+        # then
+        assert response.status_code == 500
+        assert response.json == {
+            "error": "An error occurred. Please try again. If the problem persists, please contact the site "
+                     "administrator."
+        }
+
     @patch("create_app.conversation_with_data")
     def test_conversation_azure_byod_returns_429_on_rate_limit_error(
-        self, conversation_with_data_mock, env_helper_mock, client
+            self, conversation_with_data_mock, env_helper_mock, client
     ):
         """Test that a 429 response is returned on RateLimitError for BYOD conversation."""
         # given
-        response_mock = Mock()
-        response_mock.status_code = 429
-        body_mock = {"error": "Rate limit exceeded"}
+        response_mock = MagicMock()
+        response_mock.status_code = 400
+        response_mock.json.return_value = {
+            'error': {
+                'requestid': 'f30740e1-c6e1-48ab-ab1e-35469ed41ba4',
+                'code': "400",
+                'message': 'An error occurred when calling Azure OpenAI: Rate limit reached for AOAI embedding '
+                           'resource: Server responded with status 429. Error message: {"error":{"code":"429",'
+                           '"message": "Rate limit is exceeded. Try again in 44 seconds."}}'
+            }
+        }
 
-        rate_limit_error = RateLimitError("Rate limit exceeded", response=response_mock, body=body_mock)
-        conversation_with_data_mock.side_effect = rate_limit_error
+        conversation_with_data_mock.side_effect = BadRequestError(message="Error code: 400", response=response_mock,
+                                                                  body="")
         env_helper_mock.CONVERSATION_FLOW = ConversationFlow.BYOD.value
 
         # when

code/tests/utilities/helpers/test_azure_computer_vision_client.py

@@ -176,22 +176,11 @@ def test_returns_text_vectors(
     assert actual_vectors == expected_vectors
 
 
+@mock.patch("backend.batch.utilities.helpers.azure_computer_vision_client.requests")
 def test_vectorize_image_calls_computer_vision_timeout(
-    httpserver: HTTPServer, azure_computer_vision_client: AzureComputerVisionClient
+        mock_requests: MagicMock, azure_computer_vision_client: AzureComputerVisionClient
 ):
-    # given
-    def handler(_) -> werkzeug.Response:
-        time.sleep(0.3)
-        return werkzeug.Response(
-            json.dumps({"modelVersion": "2022-04-11", "vector": [1.0, 2.0, 3.0]}),
-            status=200,
-        )
-
-    httpserver.expect_request(
-        COMPUTER_VISION_VECTORIZE_IMAGE_PATH,
-        COMPUTER_VISION_VECTORIZE_IMAGE_REQUEST_METHOD,
-    ).respond_with_handler(handler)
-
+    mock_requests.post.side_effect = ReadTimeout("An error occurred")
     # when
     with pytest.raises(Exception) as exec_info:
         azure_computer_vision_client.vectorize_image(IMAGE_URL)

pyproject.toml

@@ -36,7 +36,7 @@ azure-monitor-opentelemetry = "^1.6.0"
 opentelemetry-instrumentation-httpx = "^0.46b0"
 pillow = "10.4.0"
 azure-mgmt-cognitiveservices = "^13.5.0"
-jsonschema = "^4.22.0"
+jsonschema = "^4.23.0"
 semantic-kernel = {version = "1.1.2", python = "<3.13"}
 azure-ai-ml = "^1.17.1"