feat: voice activity support

PiperOrigin-RevId: 856356548

Author: yinghsienwu

google/genai/_live_converters.py

@@ -1151,6 +1151,59 @@ def _LiveSendRealtimeInputParameters_to_vertex(
   return to_object
 
 
+def _LiveServerMessage_from_mldev(
+    from_object: Union[dict[str, Any], object],
+    parent_object: Optional[dict[str, Any]] = None,
+) -> dict[str, Any]:
+  to_object: dict[str, Any] = {}
+  if getv(from_object, ['setupComplete']) is not None:
+    setv(to_object, ['setup_complete'], getv(from_object, ['setupComplete']))
+
+  if getv(from_object, ['serverContent']) is not None:
+    setv(to_object, ['server_content'], getv(from_object, ['serverContent']))
+
+  if getv(from_object, ['toolCall']) is not None:
+    setv(to_object, ['tool_call'], getv(from_object, ['toolCall']))
+
+  if getv(from_object, ['toolCallCancellation']) is not None:
+    setv(
+        to_object,
+        ['tool_call_cancellation'],
+        getv(from_object, ['toolCallCancellation']),
+    )
+
+  if getv(from_object, ['usageMetadata']) is not None:
+    setv(to_object, ['usage_metadata'], getv(from_object, ['usageMetadata']))
+
+  if getv(from_object, ['goAway']) is not None:
+    setv(to_object, ['go_away'], getv(from_object, ['goAway']))
+
+  if getv(from_object, ['sessionResumptionUpdate']) is not None:
+    setv(
+        to_object,
+        ['session_resumption_update'],
+        getv(from_object, ['sessionResumptionUpdate']),
+    )
+
+  if getv(from_object, ['voiceActivityDetectionSignal']) is not None:
+    setv(
+        to_object,
+        ['voice_activity_detection_signal'],
+        getv(from_object, ['voiceActivityDetectionSignal']),
+    )
+
+  if getv(from_object, ['voiceActivity']) is not None:
+    setv(
+        to_object,
+        ['voice_activity'],
+        _VoiceActivity_from_mldev(
+            getv(from_object, ['voiceActivity']), to_object
+        ),
+    )
+
+  return to_object
+
+
 def _LiveServerMessage_from_vertex(
     from_object: Union[dict[str, Any], object],
     parent_object: Optional[dict[str, Any]] = None,
@@ -1198,6 +1251,15 @@ def _LiveServerMessage_from_vertex(
         getv(from_object, ['voiceActivityDetectionSignal']),
     )
 
+  if getv(from_object, ['voiceActivity']) is not None:
+    setv(
+        to_object,
+        ['voice_activity'],
+        _VoiceActivity_from_vertex(
+            getv(from_object, ['voiceActivity']), to_object
+        ),
+    )
+
   return to_object
 
 
@@ -1468,3 +1530,25 @@ def _UsageMetadata_from_vertex(
     setv(to_object, ['traffic_type'], getv(from_object, ['trafficType']))
 
   return to_object
+
+
+def _VoiceActivity_from_mldev(
+    from_object: Union[dict[str, Any], object],
+    parent_object: Optional[dict[str, Any]] = None,
+) -> dict[str, Any]:
+  to_object: dict[str, Any] = {}
+  if getv(from_object, ['type']) is not None:
+    setv(to_object, ['voice_activity_type'], getv(from_object, ['type']))
+
+  return to_object
+
+
+def _VoiceActivity_from_vertex(
+    from_object: Union[dict[str, Any], object],
+    parent_object: Optional[dict[str, Any]] = None,
+) -> dict[str, Any]:
+  to_object: dict[str, Any] = {}
+  if getv(from_object, ['type']) is not None:
+    setv(to_object, ['voice_activity_type'], getv(from_object, ['type']))
+
+  return to_object

google/genai/types.py

@@ -869,6 +869,17 @@ class VadSignalType(_common.CaseInSensitiveEnum):
   """End of sentence signal."""
 
 
+class VoiceActivityType(_common.CaseInSensitiveEnum):
+  """The type of the voice activity signal."""
+
+  TYPE_UNSPECIFIED = 'TYPE_UNSPECIFIED'
+  """The default is VOICE_ACTIVITY_TYPE_UNSPECIFIED."""
+  ACTIVITY_START = 'ACTIVITY_START'
+  """Start of sentence signal."""
+  ACTIVITY_END = 'ACTIVITY_END'
+  """End of sentence signal."""
+
+
 class StartSensitivity(_common.CaseInSensitiveEnum):
   """Start of speech sensitivity."""
 
@@ -16347,6 +16358,24 @@ class VoiceActivityDetectionSignalDict(TypedDict, total=False):
 ]
 
 
+class VoiceActivity(_common.BaseModel):
+  """Voice activity signal."""
+
+  voice_activity_type: Optional[VoiceActivityType] = Field(
+      default=None, description="""The type of the voice activity signal."""
+  )
+
+
+class VoiceActivityDict(TypedDict, total=False):
+  """Voice activity signal."""
+
+  voice_activity_type: Optional[VoiceActivityType]
+  """The type of the voice activity signal."""
+
+
+VoiceActivityOrDict = Union[VoiceActivity, VoiceActivityDict]
+
+
 class LiveServerMessage(_common.BaseModel):
   """Response message for API call."""
 
@@ -16379,7 +16408,13 @@ class LiveServerMessage(_common.BaseModel):
       )
   )
   voice_activity_detection_signal: Optional[VoiceActivityDetectionSignal] = (
-      Field(default=None, description="""Voice activity detection signal.""")
+      Field(
+          default=None,
+          description="""Voice activity detection signal. Allowlisted only.""",
+      )
+  )
+  voice_activity: Optional[VoiceActivity] = Field(
+      default=None, description="""Voice activity signal."""
   )
 
   @property
@@ -16478,7 +16513,10 @@ class LiveServerMessageDict(TypedDict, total=False):
   """Update of the session resumption state."""
 
   voice_activity_detection_signal: Optional[VoiceActivityDetectionSignalDict]
-  """Voice activity detection signal."""
+  """Voice activity detection signal. Allowlisted only."""
+
+  voice_activity: Optional[VoiceActivityDict]
+  """Voice activity signal."""
 
 
 LiveServerMessageOrDict = Union[LiveServerMessage, LiveServerMessageDict]