[pull] dev from home-assistant:dev

homeassistant/components/google_generative_ai_conversation/__init__.py

@@ -124,7 +124,11 @@ async def generate_content(call: ServiceCall) -> ServiceResponse:
                 f"Error generating content due to content violations, reason: {response.prompt_feedback.block_reason_message}"
             )
 
-        if not response.candidates[0].content.parts:
+        if (
+            not response.candidates
+            or not response.candidates[0].content
+            or not response.candidates[0].content.parts
+        ):
             raise HomeAssistantError("Unknown error generating content")
 
         return {"text": response.text}

homeassistant/components/google_generative_ai_conversation/config_flow.py

@@ -377,7 +377,7 @@ async def google_generative_ai_config_option_schema(
             value=api_model.name,
         )
         for api_model in sorted(
-            api_models, key=lambda x: x.name.lstrip("models/") or ""
+            api_models, key=lambda x: (x.name or "").lstrip("models/")
         )
         if (
             api_model.name

homeassistant/components/google_generative_ai_conversation/entity.py

@@ -4,7 +4,7 @@
 
 import asyncio
 import codecs
-from collections.abc import AsyncGenerator, Callable
+from collections.abc import AsyncGenerator, AsyncIterator, Callable
 from dataclasses import replace
 import mimetypes
 from pathlib import Path
@@ -15,6 +15,7 @@
 from google.genai.types import (
     AutomaticFunctionCallingConfig,
     Content,
+    ContentDict,
     File,
     FileState,
     FunctionDeclaration,
@@ -23,9 +24,11 @@
     GoogleSearch,
     HarmCategory,
     Part,
+    PartUnionDict,
     SafetySetting,
     Schema,
     Tool,
+    ToolListUnion,
 )
 import voluptuous as vol
 from voluptuous_openapi import convert
@@ -237,7 +240,7 @@ def _convert_content(
 
 
 async def _transform_stream(
-    result: AsyncGenerator[GenerateContentResponse],
+    result: AsyncIterator[GenerateContentResponse],
 ) -> AsyncGenerator[conversation.AssistantContentDeltaDict]:
     new_message = True
     try:
@@ -342,7 +345,7 @@ async def _async_handle_chat_log(
         """Generate an answer for the chat log."""
         options = self.subentry.data
 
-        tools: list[Tool | Callable[..., Any]] | None = None
+        tools: ToolListUnion | None = None
         if chat_log.llm_api:
             tools = [
                 _format_tool(tool, chat_log.llm_api.custom_serializer)
@@ -373,7 +376,7 @@ async def _async_handle_chat_log(
         else:
             raise HomeAssistantError("Invalid prompt content")
 
-        messages: list[Content] = []
+        messages: list[Content | ContentDict] = []
 
         # Google groups tool results, we do not. Group them before sending.
         tool_results: list[conversation.ToolResultContent] = []
@@ -400,7 +403,10 @@ async def _async_handle_chat_log(
         # The SDK requires the first message to be a user message
         # This is not the case if user used `start_conversation`
         # Workaround from https://github.com/googleapis/python-genai/issues/529#issuecomment-2740964537
-        if messages and messages[0].role != "user":
+        if messages and (
+            (isinstance(messages[0], Content) and messages[0].role != "user")
+            or (isinstance(messages[0], dict) and messages[0]["role"] != "user")
+        ):
             messages.insert(
                 0,
                 Content(role="user", parts=[Part.from_text(text=" ")]),
@@ -440,14 +446,14 @@ async def _async_handle_chat_log(
         )
         user_message = chat_log.content[-1]
         assert isinstance(user_message, conversation.UserContent)
-        chat_request: str | list[Part] = user_message.content
+        chat_request: list[PartUnionDict] = [user_message.content]
         if user_message.attachments:
             files = await async_prepare_files_for_prompt(
                 self.hass,
                 self._genai_client,
                 [a.path for a in user_message.attachments],
             )
-            chat_request = [chat_request, *files]
+            chat_request = [*chat_request, *files]
 
         # To prevent infinite loops, we limit the number of iterations
         for _iteration in range(MAX_TOOL_ITERATIONS):
@@ -464,15 +470,17 @@ async def _async_handle_chat_log(
                 error = ERROR_GETTING_RESPONSE
                 raise HomeAssistantError(error) from err
 
-            chat_request = _create_google_tool_response_parts(
-                [
-                    content
-                    async for content in chat_log.async_add_delta_content_stream(
-                        self.entity_id,
-                        _transform_stream(chat_response_generator),
-                    )
-                    if isinstance(content, conversation.ToolResultContent)
-                ]
+            chat_request = list(
+                _create_google_tool_response_parts(
+                    [
+                        content
+                        async for content in chat_log.async_add_delta_content_stream(
+                            self.entity_id,
+                            _transform_stream(chat_response_generator),
+                        )
+                        if isinstance(content, conversation.ToolResultContent)
+                    ]
+                )
             )
 
             if not chat_log.unresponded_tool_results:
@@ -559,13 +567,13 @@ async def wait_for_file_processing(uploaded_file: File) -> None:
                 await asyncio.sleep(FILE_POLLING_INTERVAL_SECONDS)
 
             uploaded_file = await client.aio.files.get(
-                name=uploaded_file.name,
+                name=uploaded_file.name or "",
                 config={"http_options": {"timeout": TIMEOUT_MILLIS}},
             )
 
         if uploaded_file.state == FileState.FAILED:
             raise HomeAssistantError(
-                f"File `{uploaded_file.name}` processing failed, reason: {uploaded_file.error.message}"
+                f"File `{uploaded_file.name}` processing failed, reason: {uploaded_file.error.message if uploaded_file.error else 'unknown'}"
             )
 
     prompt_parts = await hass.async_add_executor_job(upload_files)

homeassistant/components/google_generative_ai_conversation/manifest.json

@@ -8,5 +8,5 @@
   "documentation": "https://www.home-assistant.io/integrations/google_generative_ai_conversation",
   "integration_type": "service",
   "iot_class": "cloud_polling",
-  "requirements": ["google-genai==1.7.0"]
+  "requirements": ["google-genai==1.29.0"]
 }

homeassistant/components/google_generative_ai_conversation/tts.py

@@ -146,15 +146,41 @@ async def async_get_tts_audio(
                 )
             )
         )
+
+        def _extract_audio_parts(
+            response: types.GenerateContentResponse,
+        ) -> tuple[bytes, str]:
+            if (
+                not response.candidates
+                or not response.candidates[0].content
+                or not response.candidates[0].content.parts
+                or not response.candidates[0].content.parts[0].inline_data
+            ):
+                raise ValueError("No content returned from TTS generation")
+
+            data = response.candidates[0].content.parts[0].inline_data.data
+            mime_type = response.candidates[0].content.parts[0].inline_data.mime_type
+
+            if not isinstance(data, bytes):
+                raise TypeError(
+                    f"Expected bytes for audio data, got {type(data).__name__}"
+                )
+            if not isinstance(mime_type, str):
+                raise TypeError(
+                    f"Expected str for mime_type, got {type(mime_type).__name__}"
+                )
+
+            return data, mime_type
+
         try:
             response = await self._genai_client.aio.models.generate_content(
                 model=self.subentry.data.get(CONF_CHAT_MODEL, RECOMMENDED_TTS_MODEL),
                 contents=message,
                 config=config,
             )
-            data = response.candidates[0].content.parts[0].inline_data.data
-            mime_type = response.candidates[0].content.parts[0].inline_data.mime_type
-        except (APIError, ClientError, ValueError) as exc:
+
+            data, mime_type = _extract_audio_parts(response)
+        except (APIError, ClientError, ValueError, TypeError) as exc:
             LOGGER.error("Error during TTS: %s", exc, exc_info=True)
             raise HomeAssistantError(exc) from exc
         return "wav", convert_to_wav(data, mime_type)

tests/components/google_generative_ai_conversation/__init__.py

@@ -1,43 +1,16 @@
 """Tests for the Google Generative AI Conversation integration."""
 
-from unittest.mock import Mock
-
 from google.genai.errors import APIError, ClientError
-import httpx
 
 API_ERROR_500 = APIError(
     500,
-    Mock(
-        __class__=httpx.Response,
-        json=Mock(
-            return_value={
-                "message": "Internal Server Error",
-                "status": "internal-error",
-            }
-        ),
-    ),
+    {"message": "Internal Server Error", "status": "internal-error"},
 )
 CLIENT_ERROR_BAD_REQUEST = ClientError(
     400,
-    Mock(
-        __class__=httpx.Response,
-        json=Mock(
-            return_value={
-                "message": "Bad Request",
-                "status": "invalid-argument",
-            }
-        ),
-    ),
+    {"message": "Bad Request", "status": "invalid-argument"},
 )
 CLIENT_ERROR_API_KEY_INVALID = ClientError(
     400,
-    Mock(
-        __class__=httpx.Response,
-        json=Mock(
-            return_value={
-                "message": "'reason': API_KEY_INVALID",
-                "status": "unauthorized",
-            }
-        ),
-    ),
+    {"message": "'reason': API_KEY_INVALID", "status": "unauthorized"},
 )

tests/components/google_generative_ai_conversation/snapshots/test_init.ambr

@@ -128,8 +128,14 @@
       dict({
         'contents': list([
           'Describe this image from my doorbell camera',
-          File(name='doorbell_snapshot.jpg', display_name=None, mime_type=None, size_bytes=None, create_time=None, expiration_time=None, update_time=None, sha256_hash=None, uri=None, download_uri=None, state=<FileState.ACTIVE: 'ACTIVE'>, source=None, video_metadata=None, error=None),
-          File(name='context.txt', display_name=None, mime_type=None, size_bytes=None, create_time=None, expiration_time=None, update_time=None, sha256_hash=None, uri=None, download_uri=None, state=<FileState.PROCESSING: 'PROCESSING'>, source=None, video_metadata=None, error=None),
+          File(
+    name='doorbell_snapshot.jpg',
+    state=<FileState.ACTIVE: 'ACTIVE'>
+  ),
+          File(
+    name='context.txt',
+    state=<FileState.PROCESSING: 'PROCESSING'>
+  ),
         ]),
         'model': 'models/gemini-2.5-flash',
       }),
@@ -145,8 +151,14 @@
       dict({
         'contents': list([
           'Describe this image from my doorbell camera',
-          File(name='doorbell_snapshot.jpg', display_name=None, mime_type=None, size_bytes=None, create_time=None, expiration_time=None, update_time=None, sha256_hash=None, uri=None, download_uri=None, state=<FileState.ACTIVE: 'ACTIVE'>, source=None, video_metadata=None, error=None),
-          File(name='context.txt', display_name=None, mime_type=None, size_bytes=None, create_time=None, expiration_time=None, update_time=None, sha256_hash=None, uri=None, download_uri=None, state=<FileState.ACTIVE: 'ACTIVE'>, source=None, video_metadata=None, error=None),
+          File(
+    name='doorbell_snapshot.jpg',
+    state=<FileState.ACTIVE: 'ACTIVE'>
+  ),
+          File(
+    name='context.txt',
+    state=<FileState.ACTIVE: 'ACTIVE'>
+  ),
         ]),
         'model': 'models/gemini-2.5-flash',
       }),

tests/components/google_generative_ai_conversation/test_conversation.py

@@ -195,10 +195,13 @@ async def test_function_call(
             "response": {
                 "result": "Test response",
             },
+            "scheduling": None,
+            "will_continue": None,
         },
         "inline_data": None,
         "text": None,
         "thought": None,
+        "thought_signature": None,
         "video_metadata": None,
     }
 

tests/components/google_generative_ai_conversation/test_tts.py

@@ -37,7 +37,7 @@
 from tests.components.tts.common import retrieve_media
 from tests.typing import ClientSessionGenerator
 
-API_ERROR_500 = APIError("test", response=MagicMock())
+API_ERROR_500 = APIError("test", response_json={})
 TEST_CHAT_MODEL = "models/some-tts-model"