fix(yandex_station): adapt intent listener to handle mapping in event…

… data

This commit updates the `listener` function within the Yandex Station integration to support a change introduced in version 2024.4.0, where event data can be passed directly as a `Mapping[str, Any]` instead of an `Event` object. The update involves checking the type of the `event_data` parameter and extracting the data from it if it's an instance of `Event`. This change ensures compatibility with Home Assistant's evolving event model while maintaining performance by using a direct type check rather than `isinstance`, which is slower.

The adjustment is critical for ensuring that the Yandex Station integration continues to function correctly without breaking due to the structural changes in how events are passed to listeners in Home Assistant versions 2024.4.0 and onwards.

custom_components/yandex_station/intent.py

@@ -2,8 +2,10 @@
 import logging
 from typing import Optional
 
+from collections.abc import Mapping
 from homeassistant.core import Event, HomeAssistant
 from homeassistant.helpers.intent import Intent, IntentHandler, IntentResponse
+from typing import Any
 
 from .core.const import DATA_SPEAKERS, DOMAIN
 
@@ -28,11 +30,19 @@ async def async_setup_intents(hass: HomeAssistant) -> None:
     if not handlers:
         return
 
-    async def listener(event: Event):
-        request_id = event.data["request_id"]
+    async def listener(event_data: Mapping[str, Any] | Event):
+
+        # Breaking change in 2024.4.0, check for Event for versions prior to this
+        if (
+            type(event_data) is Event
+        ):  # Intentionally avoid `isinstance` because it's slow and we trust `Event` is not subclassed
+            event_data = event_data.data
+
+        request_id = event_data["request_id"]
+
         for handler in handlers:
             if handler.request_id == request_id:
-                handler.response_text = event.data["text"]
+                handler.response_text = event_data["text"]
                 handler.response_waiter.set()
                 return