dv-pipecat-ai 0.0.85.dev840__py3-none-any.whl → 0.0.85.dev842__py3-none-any.whl

{dv_pipecat_ai-0.0.85.dev840.dist-info → dv_pipecat_ai-0.0.85.dev842.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dv-pipecat-ai
-Version: 0.0.85.dev840
+Version: 0.0.85.dev842
 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.dev840.dist-info → dv_pipecat_ai-0.0.85.dev842.dist-info}/RECORD

@@ -1,4 +1,4 @@
-dv_pipecat_ai-0.0.85.dev840.dist-info/licenses/LICENSE,sha256=DWY2QGf2eMCFhuu2ChairtT6CB7BEFffNVhXWc4Od08,1301
+dv_pipecat_ai-0.0.85.dev842.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
@@ -210,7 +210,7 @@ pipecat/services/cartesia/tts.py,sha256=I_OZCINywkDXmYzFL35MjSN8cAuNEaJs7nj0YB_o
 pipecat/services/cerebras/__init__.py,sha256=5zBmqq9Zfcl-HC7ylekVS5qrRedbl1mAeEwUT-T-c_o,259
 pipecat/services/cerebras/llm.py,sha256=-yzSe_6YDGigwzES-LZS4vNXMPugmvsIYEpTySyr5nA,3047
 pipecat/services/deepgram/__init__.py,sha256=IjRtMI7WytRDdmYVpk2qDWClXUiNgdl7ZkvEAWg1eYE,304
-pipecat/services/deepgram/stt.py,sha256=IvdKvo23PxhKoWTJDxuK4Uoo0wCtkFGAE_QrMUoGdYM,13732
+pipecat/services/deepgram/stt.py,sha256=fzKirTjTopwXNQEEPuUOIgk4AMvTJQcrh6H11w13q2c,16185
 pipecat/services/deepgram/tts.py,sha256=H_2WCJEx3_L4ytrHHRNkA-6GKTd1coou_vvTfiEodpQ,3745
 pipecat/services/deepgram/flux/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 pipecat/services/deepgram/flux/stt.py,sha256=yCZodrHAOShgYy_GbdviX8iAuh36dBgDL41gHMXVxEM,25887
@@ -329,7 +329,7 @@ pipecat/services/sarvam/tts.py,sha256=lrwfdC53kZ7f2QPgNRxzryISNkrJCvNtlZ-19-iXg9
 pipecat/services/simli/__init__.py,sha256=cbDcqOaGsEgKbGYKpJ1Vv7LN4ZjOWA04sE84WW5vgQI,257
 pipecat/services/simli/video.py,sha256=Zu2XLvl2Y6VHaWzT9wEdzW9d0EYoZyzYLxjQFyV8vho,8320
 pipecat/services/soniox/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-pipecat/services/soniox/stt.py,sha256=AhJF2YOzmqgB80x22jocgzr3neYCBMyxzP_WjkYR9Gc,15441
+pipecat/services/soniox/stt.py,sha256=Ndml6QvPQ1WZBvdGT3LSg-LLWwrZ8KlqW8wBBFsQrrM,16509
 pipecat/services/speechmatics/__init__.py,sha256=Jgq1fqrZVkpWC21D79L1cn5Ub8PnYgnnCaqC5pOlbIc,89
 pipecat/services/speechmatics/stt.py,sha256=GLGJzlMSeZ1WzTOMjhKXDl5JYkqGhnFTbP3o0ez0hSw,44696
 pipecat/services/tavus/__init__.py,sha256=SNyyi2Xq6tXIihDG2Bwvmg6Srbd-uWd1RwG-NKWcPuI,257
@@ -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.dev840.dist-info/METADATA,sha256=ALfdKasSbWLkqVlc0XSkl9lo6qzO7Wpior0WxVFzWZk,32955
-dv_pipecat_ai-0.0.85.dev840.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-dv_pipecat_ai-0.0.85.dev840.dist-info/top_level.txt,sha256=kQzG20CxGf-nSsHmtXHx3hY2-8zHA3jYg8jk0TajqXc,8
-dv_pipecat_ai-0.0.85.dev840.dist-info/RECORD,,
+dv_pipecat_ai-0.0.85.dev842.dist-info/METADATA,sha256=8uxfODboEJQwB04vWViiyMIo4KyT-pGag_ChfLt8STo,32955
+dv_pipecat_ai-0.0.85.dev842.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+dv_pipecat_ai-0.0.85.dev842.dist-info/top_level.txt,sha256=kQzG20CxGf-nSsHmtXHx3hY2-8zHA3jYg8jk0TajqXc,8
+dv_pipecat_ai-0.0.85.dev842.dist-info/RECORD,,

pipecat/services/deepgram/stt.py

@@ -62,6 +62,8 @@ class DeepgramSTTService(STTService):
         sample_rate: Optional[int] = None,
         live_options: Optional[LiveOptions] = None,
         addons: Optional[Dict] = None,
+        max_connect_retries: int = 3,
+        connect_timeout_s: float = 2.5,
         **kwargs,
     ):
         """Initialize the Deepgram STT service.
@@ -77,6 +79,9 @@ class DeepgramSTTService(STTService):
             sample_rate: Audio sample rate. If None, uses default or live_options value.
             live_options: Deepgram LiveOptions for detailed configuration.
             addons: Additional Deepgram features to enable.
+            max_connect_retries: Maximum number of connection attempts before giving up.
+            connect_timeout_s: Maximum time in seconds to wait for a connection attempt.
+                Connection retries wait 100ms between attempts.
             **kwargs: Additional arguments passed to the parent STTService.
         """
         sample_rate = sample_rate or (live_options.sample_rate if live_options else None)
@@ -121,9 +126,9 @@ class DeepgramSTTService(STTService):
         self._settings = merged_options
         self._addons = addons
 
-        # Connection retry settings
-        self._max_connect_retries = 3
-        self._connect_retry_delay_s = 0.1
+        # Connection retry settings (100ms delay between retries)
+        self._max_connect_retries = max_connect_retries
+        self._connect_timeout_s = connect_timeout_s
 
         self._client = DeepgramClient(
             api_key,
@@ -131,8 +136,8 @@ class DeepgramSTTService(STTService):
                 url=base_url,
                 options={
                     "keepalive": "true",
-                    "open_timeout": 3,  # Max wait for only 3 seconds for the connection to establish #
-                    # "termination_exception_connect": True,  # Enable exception propagation
+                    # Note: Connection timeout is enforced by asyncio.wait_for() in _connect()
+                    # with the connect_timeout_s parameter (default 2.0s)
                 },
                 verbose=logging.ERROR,  # Enable error level and above logging
             ),
@@ -227,6 +232,11 @@ class DeepgramSTTService(STTService):
 
         for attempt in range(self._max_connect_retries):
             try:
+                # Clean up any previous connection attempt in background (non-blocking)
+                if hasattr(self, "_connection") and self._connection is not None:
+                    old_conn = self._connection
+                    asyncio.create_task(self._cleanup_abandoned_connection(old_conn))
+
                 # Create a new connection object for a clean attempt
                 self._connection: AsyncListenWebSocketClient = self._client.listen.asyncwebsocket.v(
                     "1"
@@ -250,10 +260,25 @@ class DeepgramSTTService(STTService):
                         self._on_utterance_end,
                     )
 
-                # Attempt to start the connection (timeout handled by open_timeout config)
-                if await self._connection.start(options=self._settings, addons=self._addons):
-                    self.logger.info("Successfully connected to Deepgram.")
-                    return  # Exit the method on success
+                try:
+                    start_result = await asyncio.wait_for(
+                        self._connection.start(options=self._settings, addons=self._addons),
+                        timeout=self._connect_timeout_s,
+                    )
+                except asyncio.TimeoutError:
+                    self.logger.warning(
+                        f"Deepgram connection attempt {attempt + 1}/{self._max_connect_retries} timed out after {self._connect_timeout_s} second(s)."
+                    )
+                    start_result = False
+                except Exception as start_error:
+                    self.logger.warning(
+                        f"Deepgram connection attempt {attempt + 1}/{self._max_connect_retries} failed with an exception: {start_error}"
+                    )
+                    start_result = False
+                else:
+                    if start_result:
+                        self.logger.info("Successfully connected to Deepgram.")
+                        return  # Exit the method on success
 
                 self.logger.warning(
                     f"Deepgram connection attempt {attempt + 1}/{self._max_connect_retries} failed."
@@ -264,17 +289,21 @@ class DeepgramSTTService(STTService):
                     f"Deepgram connection attempt {attempt + 1}/{self._max_connect_retries} failed with an exception: {e}"
                 )
 
-            # If this is not the last attempt, wait briefly before retrying
+            # If this is not the last attempt, wait 100ms before retrying
             if attempt < self._max_connect_retries - 1:
-                self.logger.info(f"Retrying in {self._connect_retry_delay_s} second(s)...")
-                await asyncio.sleep(self._connect_retry_delay_s)
+                self.logger.info("Retrying in 0.1 second(s)...")
+                await asyncio.sleep(0.1)
 
-        self.logger.error(
+        error_msg = (
             f"{self}: unable to connect to Deepgram after {self._max_connect_retries} attempts."
         )
+        self.logger.error(error_msg)
+        await self.push_error(ErrorFrame(error_msg, fatal=True))
 
     async def _disconnect(self):
-        if self._connection.is_connected:
+        # Guard against missing connection instance and ensure proper async check
+        connection: AsyncListenWebSocketClient = getattr(self, "_connection", None)
+        if connection and await connection.is_connected():
             self.logger.debug("Disconnecting from Deepgram")
             # Deepgram swallows asyncio.CancelledError internally which prevents
             # proper cancellation propagation. This issue was found with
@@ -284,7 +313,25 @@ class DeepgramSTTService(STTService):
             # Deepgram disconnection was still finishing and therefore
             # preventing the task cancellation that occurs during `cleanup()`.
             # GH issue: https://github.com/deepgram/deepgram-python-sdk/issues/570
-            await self._connection.finish()
+            await connection.finish()
+
+    async def _cleanup_abandoned_connection(self, conn: AsyncListenWebSocketClient):
+        """Clean up abandoned connection attempt in background (non-blocking).
+
+        This prevents zombie connections from triggering spurious error events
+        when they eventually timeout and call _on_error().
+
+        Args:
+            conn: The abandoned connection object to clean up.
+        """
+        try:
+            # Try to finish with short timeout
+            await asyncio.wait_for(conn.finish(), timeout=5)
+            self.logger.debug("Successfully cleaned up abandoned connection")
+        except Exception as e:
+            # Ignore all cleanup errors - connection might not be fully started
+            # This is expected and fine - we just want best-effort cleanup
+            self.logger.debug(f"Abandoned connection cleanup failed: {e}")
 
     async def start_metrics(self):
         """Start TTFB and processing metrics collection."""

pipecat/services/soniox/stt.py

@@ -49,6 +49,33 @@ END_TOKEN = "<end>"
 FINALIZED_TOKEN = "<fin>"
 
 
+class SonioxContextGeneralItem(BaseModel):
+    """Represents a key-value pair for structured general context information."""
+
+    key: str
+    value: str
+
+
+class SonioxContextTranslationTerm(BaseModel):
+    """Represents a custom translation mapping for ambiguous or domain-specific terms."""
+
+    source: str
+    target: str
+
+
+class SonioxContextObject(BaseModel):
+    """Context object for models with context_version 2, for Soniox stt-rt-v3-preview and higher.
+
+    Learn more about context in the documentation:
+    https://soniox.com/docs/stt/concepts/context
+    """
+
+    general: Optional[List[SonioxContextGeneralItem]] = None
+    text: Optional[str] = None
+    terms: Optional[List[str]] = None
+    translation_terms: Optional[List[SonioxContextTranslationTerm]] = None
+
+
 class SonioxInputParams(BaseModel):
     """Real-time transcription settings.
 
@@ -60,9 +87,9 @@ class SonioxInputParams(BaseModel):
         audio_format: Audio format to use for transcription.
         num_channels: Number of channels to use for transcription.
         language_hints: List of language hints to use for transcription.
-        context: Customization for transcription.
-        enable_non_final_tokens: Whether to enable non-final tokens. If false, only final tokens will be returned.
-        max_non_final_tokens_duration_ms: Maximum duration of non-final tokens.
+        context: Customization for transcription. String for models with context_version 1 and ContextObject for models with context_version 2.
+        enable_speaker_diarization: Whether to enable speaker diarization. Tokens are annotated with speaker IDs.
+        enable_language_identification: Whether to enable language identification. Tokens are annotated with language IDs.
         client_reference_id: Client reference ID to use for transcription.
     """
 
@@ -72,10 +99,10 @@ class SonioxInputParams(BaseModel):
     num_channels: Optional[int] = 1
 
     language_hints: Optional[List[Language]] = None
-    context: Optional[str] = None
+    context: Optional[SonioxContextObject | str] = None
 
-    enable_non_final_tokens: Optional[bool] = True
-    max_non_final_tokens_duration_ms: Optional[int] = None
+    enable_speaker_diarization: Optional[bool] = False
+    enable_language_identification: Optional[bool] = False
 
     client_reference_id: Optional[str] = None
 
@@ -173,6 +200,10 @@ class SonioxSTTService(STTService):
         # Either one or the other is required.
         enable_endpoint_detection = not self._vad_force_turn_endpoint
 
+        context = self._params.context
+        if isinstance(context, SonioxContextObject):
+            context = context.model_dump()
+
         # Send the initial configuration message.
         config = {
             "api_key": self._api_key,
@@ -182,9 +213,9 @@ class SonioxSTTService(STTService):
             "enable_endpoint_detection": enable_endpoint_detection,
             "sample_rate": self.sample_rate,
             "language_hints": _prepare_language_hints(self._params.language_hints),
-            "context": self._params.context,
-            "enable_non_final_tokens": self._params.enable_non_final_tokens,
-            "max_non_final_tokens_duration_ms": self._params.max_non_final_tokens_duration_ms,
+            "context": context,
+            "enable_speaker_diarization": self._params.enable_speaker_diarization,
+            "enable_language_identification": self._params.enable_language_identification,
             "client_reference_id": self._params.client_reference_id,
         }