dv-pipecat-ai 0.0.74.dev577__tar.gz → 0.0.74.dev589__tar.gz

{dv_pipecat_ai-0.0.74.dev577 → dv_pipecat_ai-0.0.74.dev589}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dv-pipecat-ai
-Version: 0.0.74.dev577
+Version: 0.0.74.dev589
 Summary: An open source framework for voice (and multimodal) assistants
 License-Expression: BSD-2-Clause
 Project-URL: Source, https://github.com/pipecat-ai/pipecat

{dv_pipecat_ai-0.0.74.dev577 → dv_pipecat_ai-0.0.74.dev589}/src/dv_pipecat_ai.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dv-pipecat-ai
-Version: 0.0.74.dev577
+Version: 0.0.74.dev589
 Summary: An open source framework for voice (and multimodal) assistants
 License-Expression: BSD-2-Clause
 Project-URL: Source, https://github.com/pipecat-ai/pipecat

{dv_pipecat_ai-0.0.74.dev577 → dv_pipecat_ai-0.0.74.dev589}/src/pipecat/serializers/twilio.py

@@ -139,6 +139,9 @@ class TwilioFrameSerializer(FrameSerializer):
             serialized_data = await pcm_to_ulaw(
                 data, frame.sample_rate, self._twilio_sample_rate, self._resampler
             )
+            if serialized_data is None or len(serialized_data) == 0:
+                # Ignoring in case we don't have audio
+                return None
             payload = base64.b64encode(serialized_data).decode("utf-8")
             answer = {
                 "event": "media",

{dv_pipecat_ai-0.0.74.dev577 → dv_pipecat_ai-0.0.74.dev589}/src/pipecat/services/google/llm_openai.py

@@ -16,6 +16,8 @@ from pipecat.utils.asyncio.watchdog_async_iterator import WatchdogAsyncIterator
 # Suppress gRPC fork warnings
 os.environ["GRPC_ENABLE_FORK_SUPPORT"] = "false"
 
+import uuid
+
 from loguru import logger
 
 from pipecat.frames.frames import LLMTextFrame
@@ -84,7 +86,11 @@ class GoogleLLMOpenAIBetaService(OpenAILLMService):
                 # yield a frame containing the function name and the arguments.
                 logger.debug(f"Tool call: {chunk.choices[0].delta.tool_calls}")
                 tool_call = chunk.choices[0].delta.tool_calls[0]
-                if tool_call.index != func_idx:
+                if tool_call.id:
+                    tool_call_id = tool_call.id
+                elif not tool_call_id:
+                    tool_call_id = f"call_{uuid.uuid4()}"
+                if tool_call.index is not None and tool_call.index != func_idx:
                     functions_list.append(function_name)
                     arguments_list.append(arguments)
                     tool_id_list.append(tool_call_id)
@@ -94,7 +100,7 @@ class GoogleLLMOpenAIBetaService(OpenAILLMService):
                     func_idx += 1
                 if tool_call.function and tool_call.function.name:
                     function_name += tool_call.function.name
-                    tool_call_id = tool_call.id
+                    # tool_call_id = tool_call.id Note: We have removed this
                 if tool_call.function and tool_call.function.arguments:
                     # Keep iterating through the response to collect all the argument fragments
                     arguments += tool_call.function.arguments

{dv_pipecat_ai-0.0.74.dev577 → dv_pipecat_ai-0.0.74.dev589}/src/pipecat/services/tts_service.py

@@ -832,7 +832,7 @@ class AudioContextWordTTSService(WebsocketWordTTSService):
                 del self._contexts[context_id]
 
                 # Append some silence between sentences.
-                silence = b"\x00" * self.sample_rate
+                silence = b"\x00" * (self.sample_rate // 2)
                 frame = TTSAudioRawFrame(
                     audio=silence, sample_rate=self.sample_rate, num_channels=1
                 )

{dv_pipecat_ai-0.0.74.dev577 → dv_pipecat_ai-0.0.74.dev589}/src/pipecat/transports/base_output.py

@@ -42,7 +42,8 @@ from pipecat.transports.base_transport import TransportParams
 from pipecat.utils.asyncio.watchdog_priority_queue import WatchdogPriorityQueue
 from pipecat.utils.time import nanoseconds_to_seconds
 
-BOT_VAD_STOP_SECS = 0.35
+# TODO: When we use GeminiMultimodalLiveLLMService, we need to change this to 0.35 but that creates issue for faster TTS.
+BOT_VAD_STOP_SECS = 0.30
 
 
 class BaseOutputTransport(FrameProcessor):

{dv_pipecat_ai-0.0.74.dev577 → dv_pipecat_ai-0.0.74.dev589}/src/pipecat/transports/network/fastapi_websocket.py

@@ -107,7 +107,6 @@ class FastAPIWebsocketClient:
         self._leave_counter -= 1
         if self._leave_counter > 0:
             return
-
         if self.is_connected and not self.is_closing:
             self._closing = True
             await self._websocket.close()