Add model option to speak action for ElevenLabs (#133902)

homeassistant/components/elevenlabs/const.py

@@ -1,5 +1,7 @@
 """Constants for the ElevenLabs text-to-speech integration."""
 
+ATTR_MODEL = "model"
+
 CONF_VOICE = "voice"
 CONF_MODEL = "model"
 CONF_CONFIGURE_VOICE = "configure_voice"

homeassistant/components/elevenlabs/tts.py

@@ -24,6 +24,7 @@ from homeassistant.helpers.entity_platform import AddEntitiesCallback
 
 from . import ElevenLabsConfigEntry
 from .const import (
+    ATTR_MODEL,
     CONF_OPTIMIZE_LATENCY,
     CONF_SIMILARITY,
     CONF_STABILITY,
@@ -85,7 +86,7 @@ async def async_setup_entry(
 class ElevenLabsTTSEntity(TextToSpeechEntity):
     """The ElevenLabs API entity."""
 
-    _attr_supported_options = [ATTR_VOICE]
+    _attr_supported_options = [ATTR_VOICE, ATTR_MODEL]
     _attr_entity_category = EntityCategory.CONFIG
 
     def __init__(
@@ -141,13 +142,14 @@ class ElevenLabsTTSEntity(TextToSpeechEntity):
         _LOGGER.debug("Getting TTS audio for %s", message)
         _LOGGER.debug("Options: %s", options)
         voice_id = options.get(ATTR_VOICE, self._default_voice_id)
+        model = options.get(ATTR_MODEL, self._model.model_id)
         try:
             audio = await self._client.generate(
                 text=message,
                 voice=voice_id,
                 optimize_streaming_latency=self._latency,
                 voice_settings=self._voice_settings,
-                model=self._model.model_id,
+                model=model,
             )
             bytes_combined = b"".join([byte_seg async for byte_seg in audio])
         except ApiError as exc:

tests/components/elevenlabs/test_tts.py

@@ -13,6 +13,7 @@ import pytest
 
 from homeassistant.components import tts
 from homeassistant.components.elevenlabs.const import (
+    ATTR_MODEL,
     CONF_MODEL,
     CONF_OPTIMIZE_LATENCY,
     CONF_SIMILARITY,
@@ -163,6 +164,16 @@ async def mock_config_entry_setup(
 @pytest.mark.parametrize(
     ("setup", "tts_service", "service_data"),
     [
+        (
+            "mock_config_entry_setup",
+            "speak",
+            {
+                ATTR_ENTITY_ID: "tts.mock_title",
+                tts.ATTR_MEDIA_PLAYER_ENTITY_ID: "media_player.something",
+                tts.ATTR_MESSAGE: "There is a person at the front door.",
+                tts.ATTR_OPTIONS: {},
+            },
+        ),
         (
             "mock_config_entry_setup",
             "speak",
@@ -173,6 +184,26 @@ async def mock_config_entry_setup(
                 tts.ATTR_OPTIONS: {tts.ATTR_VOICE: "voice2"},
             },
         ),
+        (
+            "mock_config_entry_setup",
+            "speak",
+            {
+                ATTR_ENTITY_ID: "tts.mock_title",
+                tts.ATTR_MEDIA_PLAYER_ENTITY_ID: "media_player.something",
+                tts.ATTR_MESSAGE: "There is a person at the front door.",
+                tts.ATTR_OPTIONS: {ATTR_MODEL: "model2"},
+            },
+        ),
+        (
+            "mock_config_entry_setup",
+            "speak",
+            {
+                ATTR_ENTITY_ID: "tts.mock_title",
+                tts.ATTR_MEDIA_PLAYER_ENTITY_ID: "media_player.something",
+                tts.ATTR_MESSAGE: "There is a person at the front door.",
+                tts.ATTR_OPTIONS: {tts.ATTR_VOICE: "voice2", ATTR_MODEL: "model2"},
+            },
+        ),
     ],
     indirect=["setup"],
 )
@@ -206,11 +237,13 @@ async def test_tts_service_speak(
         await retrieve_media(hass, hass_client, calls[0].data[ATTR_MEDIA_CONTENT_ID])
         == HTTPStatus.OK
     )
+    voice_id = service_data[tts.ATTR_OPTIONS].get(tts.ATTR_VOICE, "voice1")
+    model_id = service_data[tts.ATTR_OPTIONS].get(ATTR_MODEL, "model1")
 
     tts_entity._client.generate.assert_called_once_with(
         text="There is a person at the front door.",
-        voice="voice2",
-        model="model1",
+        voice=voice_id,
+        model=model_id,
         voice_settings=tts_entity._voice_settings,
         optimize_streaming_latency=tts_entity._latency,
     )