handle stt exceptions

livekit-plugins/livekit-plugins-clova/livekit/plugins/clova/stt.py

@@ -11,15 +11,27 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+
+from __future__ import annotations
+
 import io
 import json
+import asyncio
 import os
 import time
 import wave
 from typing import Optional, Union
 
 import aiohttp
-from livekit.agents import stt, utils
+
+from livekit.agents import (
+    APIConnectionError,
+    APIStatusError,
+    APITimeoutError,
+    stt,
+    utils,
+)
+
 from livekit.agents.stt import SpeechEventType, STTCapabilities
 from livekit.agents.utils import AudioBuffer, merge_frames
 from livekit.plugins.clova.constants import CLOVA_INPUT_SAMPLE_RATE
@@ -115,16 +127,21 @@ async def _recognize_impl(
                     )
                 logger.info(f"final event: {response_data}")
                 return self._transcription_to_speech_event(text=text)
-        except Exception as ex:
-            logger.error(f"{ex}")
-            return self._transcription_to_speech_event(
-                event_type=stt.SpeechEventType.FINAL_TRANSCRIPT, text=""
-            )
+
+        except asyncio.TimeoutError as e:
+            raise APITimeoutError() from e
+        except aiohttp.ClientResponseError as e:
+            raise APIStatusError(
+                message=e.message,
+                status_code=e.status,
+                request_id=None,
+                body=None,
+            ) from e
 
     def _transcription_to_speech_event(
         self,
         event_type: SpeechEventType = stt.SpeechEventType.INTERIM_TRANSCRIPT,
-        text: str = None,
+        text: str | None = None,
     ) -> stt.SpeechEvent:
         return stt.SpeechEvent(
             type=event_type,

livekit-plugins/livekit-plugins-google/livekit/plugins/google/stt.py

@@ -20,7 +20,14 @@
 from typing import AsyncIterable, List, Union
 
 from livekit import agents, rtc
-from livekit.agents import stt, utils
+
+from livekit.agents import (
+    APIConnectionError,
+    APIStatusError,
+    APITimeoutError,
+    stt,
+    utils,
+)
 
 from google.auth import default as gauth_default
 from google.auth.exceptions import DefaultCredentialsError
@@ -165,12 +172,27 @@ async def _recognize_impl(
             language_codes=config.languages,
         )
 
-        raw = await self._ensure_client().recognize(
-            cloud_speech.RecognizeRequest(
-                recognizer=self._recognizer, config=config, content=frame.data.tobytes()
+        try:
+            raw = await self._ensure_client().recognize(
+                cloud_speech.RecognizeRequest(
+                    recognizer=self._recognizer,
+                    config=config,
+                    content=frame.data.tobytes(),
+                )
+            )
+
+            return _recognize_response_to_speech_event(raw)
+        except DeadlineExceeded:
+            raise APITimeoutError()
+        except GoogleAPICallError as e:
+            raise APIStatusError(
+                e.message,
+                status_code=e.code or -1,
+                request_id=None,
+                body=None,
             )
-        )
-        return _recognize_response_to_speech_event(raw)
+        except Exception as e:
+            raise APIConnectionError() from e
 
     def stream(
         self, *, language: SpeechLanguages | str | None = None