dv-pipecat-ai 0.0.85.dev852__py3-none-any.whl → 0.0.85.dev854__py3-none-any.whl

{dv_pipecat_ai-0.0.85.dev852.dist-info → dv_pipecat_ai-0.0.85.dev854.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dv-pipecat-ai
-Version: 0.0.85.dev852
+Version: 0.0.85.dev854
 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-0.0.85.dev852.dist-info → dv_pipecat_ai-0.0.85.dev854.dist-info}/RECORD

@@ -1,4 +1,4 @@
-dv_pipecat_ai-0.0.85.dev852.dist-info/licenses/LICENSE,sha256=DWY2QGf2eMCFhuu2ChairtT6CB7BEFffNVhXWc4Od08,1301
+dv_pipecat_ai-0.0.85.dev854.dist-info/licenses/LICENSE,sha256=DWY2QGf2eMCFhuu2ChairtT6CB7BEFffNVhXWc4Od08,1301
 pipecat/__init__.py,sha256=j0Xm6adxHhd7D06dIyyPV_GlBYLlBnTAERVvD_jAARQ,861
 pipecat/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 pipecat/adapters/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -164,7 +164,7 @@ pipecat/serializers/plivo.py,sha256=ie6VUhZDTJ7KlAuJyHNeIeMtJ3ScDq_2js1SZtz7jLI,
 pipecat/serializers/protobuf.py,sha256=L0jSqvgTdkfxsu6JWjYK8QSTVji9nhzmgRsEEbGU7xY,5223
 pipecat/serializers/telnyx.py,sha256=eFkC7dExDFildYLR8DPvgfHbgXlCwdSPd1vc11yxyok,10847
 pipecat/serializers/twilio.py,sha256=0emSzXVw8DU_N5RPruMekbBKku9Q429-0z1PMuYejSk,10823
-pipecat/serializers/vi.py,sha256=Q7kMXvKM493RIuOUc99LKZWgVmvd8_owAzIK_oEktfw,11150
+pipecat/serializers/vi.py,sha256=XdnyNwhpxo8CvgS4v1_9DmetF0BkRmx8KoSVdjCdPPQ,11247
 pipecat/services/__init__.py,sha256=8e3Ta-8_BOPozhDB3l0GJkNXs5PWhib6yqZQUof2Kvw,1209
 pipecat/services/ai_service.py,sha256=yE386fm2Id-yD4fCNfkmEMtg0lTA7PB17n2x_A_jwTg,5896
 pipecat/services/ai_services.py,sha256=_RrDWfM8adV17atzY9RxK0nXRVM5kbUkKrvN90GAWYM,795
@@ -280,7 +280,7 @@ pipecat/services/nim/llm.py,sha256=o4WPGI6kOmSiMV7WwOZ0cNEAoq9hW4Aqs2R8X7c9i94,4
 pipecat/services/ollama/__init__.py,sha256=aw-25zYsR8LR74OFFlMKMTnJjaKwOzdPWVsClueNRkI,255
 pipecat/services/ollama/llm.py,sha256=rfpG92LRHGJlpENKhF6ld8CLVS9DxlKW-WRVNldOIGs,1605
 pipecat/services/openai/__init__.py,sha256=V0ZVa8PzEm3hmcStYICbAsYwfgk4ytZ6kiQoq9UZPmI,354
-pipecat/services/openai/base_llm.py,sha256=jOiWacimREywCMZZwAwH8RAHCbwnnXvbqAjWQUYA0yM,20727
+pipecat/services/openai/base_llm.py,sha256=mrHRwYL-0lYr0BhJMg_ByTfqY2no6T8DOrhqMDztBlY,21518
 pipecat/services/openai/image.py,sha256=3e3h-dVQ6DQuQE7fp8akXwRMd-oYOdGuZg7RCOjHu9A,2994
 pipecat/services/openai/llm.py,sha256=_aKtz1VebSFUUenT3tH6mBW9pSCm65_u45cDu_dkTzs,7396
 pipecat/services/openai/stt.py,sha256=Idf0k73kxFyDgNRBt62MFpoKKNsBV9bwvJteJ6MGWzQ,2419
@@ -416,7 +416,7 @@ pipecat/utils/tracing/service_decorators.py,sha256=fwzxFpi8DJl6BJbK74G0UEB4ccMJg
 pipecat/utils/tracing/setup.py,sha256=7TEgPNpq6M8lww8OQvf0P9FzYc5A30xICGklVA-fua0,2892
 pipecat/utils/tracing/turn_context_provider.py,sha256=ikon3plFOx0XbMrH6DdeHttNpb-U0gzMZIm3bWLc9eI,2485
 pipecat/utils/tracing/turn_trace_observer.py,sha256=dma16SBJpYSOE58YDWy89QzHyQFc_9gQZszKeWixuwc,9725
-dv_pipecat_ai-0.0.85.dev852.dist-info/METADATA,sha256=L_uFM2KLucwhFvtLcQ9dWL_DQicbrpPyiHOlW81e9LM,32955
-dv_pipecat_ai-0.0.85.dev852.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-dv_pipecat_ai-0.0.85.dev852.dist-info/top_level.txt,sha256=kQzG20CxGf-nSsHmtXHx3hY2-8zHA3jYg8jk0TajqXc,8
-dv_pipecat_ai-0.0.85.dev852.dist-info/RECORD,,
+dv_pipecat_ai-0.0.85.dev854.dist-info/METADATA,sha256=bO3NZWI6qyHGO6asSfaEk4uIxnYPnOAp3BQDzr-sf1w,32955
+dv_pipecat_ai-0.0.85.dev854.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+dv_pipecat_ai-0.0.85.dev854.dist-info/top_level.txt,sha256=kQzG20CxGf-nSsHmtXHx3hY2-8zHA3jYg8jk0TajqXc,8
+dv_pipecat_ai-0.0.85.dev854.dist-info/RECORD,,

pipecat/serializers/vi.py

@@ -173,6 +173,8 @@ class VIFrameSerializer(FrameSerializer):
                 },
             }
 
+            logger.debug(f"VI: Sending media event {message} for stream_id: {self._stream_id}")
+
             return json.dumps(message)
 
         elif isinstance(frame, (TransportMessageFrame, TransportMessageUrgentFrame)):

pipecat/services/openai/base_llm.py

@@ -18,6 +18,7 @@ from openai import (
     APITimeoutError,
     AsyncOpenAI,
     AsyncStream,
+    BadRequestError,
     DefaultAsyncHttpxClient,
 )
 from openai.types.chat import ChatCompletionChunk, ChatCompletionMessageParam
@@ -100,6 +101,7 @@ class BaseOpenAILLMService(LLMService):
         params: Optional[InputParams] = None,
         retry_timeout_secs: Optional[float] = 5.0,
         retry_on_timeout: Optional[bool] = False,
+        enable_warmup: bool = False,
         **kwargs,
     ):
         """Initialize the BaseOpenAILLMService.
@@ -114,6 +116,7 @@ class BaseOpenAILLMService(LLMService):
             params: Input parameters for model configuration and behavior.
             retry_timeout_secs: Request timeout in seconds. Defaults to 5.0 seconds.
             retry_on_timeout: Whether to retry the request once if it times out.
+            enable_warmup: Whether to enable LLM cache warmup. Defaults to False.
             **kwargs: Additional arguments passed to the parent LLMService.
         """
         super().__init__(**kwargs)
@@ -133,6 +136,7 @@ class BaseOpenAILLMService(LLMService):
         }
         self._retry_timeout_secs = retry_timeout_secs
         self._retry_on_timeout = retry_on_timeout
+        self._enable_warmup = enable_warmup
         self.set_model_name(model)
         self._client = self.create_client(
             api_key=api_key,
@@ -477,6 +481,11 @@ class BaseOpenAILLMService(LLMService):
                 await self.stop_processing_metrics()
                 await self.push_frame(LLMFullResponseEndFrame())
 
+    def _is_gpt5_model(self) -> bool:
+        """Check if the current model is a GPT-5 series model that requires max_completion_tokens."""
+        model = (self.model_name or "").lower()
+        return model.startswith("gpt-5")
+
     async def _handle_warmup_frame(self, frame: WarmupLLMFrame):
         """Handle WarmupLLMFrame to prime the LLM cache without emitting responses.
 
@@ -486,18 +495,30 @@ class BaseOpenAILLMService(LLMService):
         Args:
             frame: WarmupLLMFrame containing the messages to cache.
         """
+        # Skip warmup if disabled
+        if not self._enable_warmup:
+            self.logger.debug("LLM warmup is disabled, skipping")
+            return
+
         try:
             # Use the provided messages for warmup
             messages: List[ChatCompletionMessageParam] = frame.messages  # type: ignore
 
             # Make a non-streaming call to warm the cache
-            # We use a minimal max_tokens to reduce latency and cost
-            await self._client.chat.completions.create(
-                model=self.model_name,  # Use the property, not self._model
-                messages=messages,
-                max_tokens=10,  # Minimal response
-                stream=False,
-            )
+            # We use a minimal token limit to reduce latency and cost
+            # GPT-5 series models require max_completion_tokens instead of max_tokens
+            warmup_params = {
+                "model": self.model_name,
+                "messages": messages,
+                "stream": False,
+            }
+
+            if self._is_gpt5_model():
+                warmup_params["max_completion_tokens"] = 10
+            else:
+                warmup_params["max_tokens"] = 10
+
+            await self._client.chat.completions.create(**warmup_params)
 
             self.logger.info("LLM cache warmed successfully")
             # Intentionally don't emit any frames - this is a silent warmup