dv-pipecat-ai 0.0.85.dev833__tar.gz → 0.0.85.dev837__tar.gz

{dv_pipecat_ai-0.0.85.dev833 → dv_pipecat_ai-0.0.85.dev837}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dv-pipecat-ai
-Version: 0.0.85.dev833
+Version: 0.0.85.dev837
 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.dev833 → dv_pipecat_ai-0.0.85.dev837}/src/dv_pipecat_ai.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dv-pipecat-ai
-Version: 0.0.85.dev833
+Version: 0.0.85.dev837
 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.dev833 → dv_pipecat_ai-0.0.85.dev837}/src/dv_pipecat_ai.egg-info/SOURCES.txt

@@ -247,6 +247,7 @@ src/pipecat/serializers/plivo.py
 src/pipecat/serializers/protobuf.py
 src/pipecat/serializers/telnyx.py
 src/pipecat/serializers/twilio.py
+src/pipecat/serializers/vi.py
 src/pipecat/services/__init__.py
 src/pipecat/services/ai_service.py
 src/pipecat/services/ai_services.py

{dv_pipecat_ai-0.0.85.dev833 → dv_pipecat_ai-0.0.85.dev837}/src/pipecat/processors/aggregators/llm_response.py

@@ -49,7 +49,6 @@ from pipecat.frames.frames import (
     OpenAILLMContextAssistantTimestampFrame,
     SpeechControlParamsFrame,
     StartFrame,
-    StartInterruptionFrame,
     TextFrame,
     TranscriptDropFrame,
     TranscriptionFrame,
@@ -473,8 +472,8 @@ class LLMUserContextAggregator(LLMContextResponseAggregator):
             frame: The frame to process.
             direction: The direction of frame flow in the pipeline.
         """
-        if isinstance(frame, StartInterruptionFrame):
-            self.logger.debug("Received StartInterruptionFrame")
+        if isinstance(frame, InterruptionFrame):
+            self.logger.debug("Received InterruptionFrame")
         await super().process_frame(frame, direction)
 
         if isinstance(frame, StartFrame):
@@ -560,7 +559,7 @@ class LLMUserContextAggregator(LLMContextResponseAggregator):
                         "Triggering interruption - pushing BotInterruptionFrame and aggregation"
                     )
                     # await self.push_frame(BotInterruptionFrame(), FrameDirection.UPSTREAM)
-                    await self.push_frame(StartInterruptionFrame(), FrameDirection.DOWNSTREAM)
+                    await self.push_frame(InterruptionFrame(), FrameDirection.DOWNSTREAM)
                     self.logger.debug("Pushed BotInterruptionFrame")
                 # No interruption config - normal behavior (always push aggregation)
                 await self._process_aggregation()
@@ -596,12 +595,8 @@ class LLMUserContextAggregator(LLMContextResponseAggregator):
         """Notify upstream processors that pending transcripts should be dropped."""
         if self._pending_transcription_ids:
             drop_frame = TranscriptDropFrame(transcript_ids=list(self._pending_transcription_ids))
-            self.logger.debug(
-                f"Dropping {len(self._pending_transcription_ids)} transcript chunk(s) due to {reason}"
-            )
             await self.push_frame(drop_frame, FrameDirection.UPSTREAM)
         self._pending_transcription_ids.clear()
-        self._aggregation = ""
 
     async def _start(self, frame: StartFrame):
         self._create_aggregation_task()

{dv_pipecat_ai-0.0.85.dev833 → dv_pipecat_ai-0.0.85.dev837}/src/pipecat/processors/dtmf_aggregator.py

@@ -1,3 +1,5 @@
+"""DTMF aggregator processor for collecting and flushing DTMF input digits."""
+
 import asyncio
 
 from pipecat.frames.frames import (
@@ -8,8 +10,8 @@ from pipecat.frames.frames import (
     EndFrame,
     Frame,
     InputDTMFFrame,
+    InterruptionFrame,
     StartDTMFCaptureFrame,
-    StartInterruptionFrame,
     TranscriptionFrame,
     WaitForDTMFFrame,
 )
@@ -19,10 +21,11 @@ from pipecat.utils.time import time_now_iso8601
 
 class DTMFAggregator(FrameProcessor):
     """Aggregates DTMF frames using idle wait logic.
+
     The aggregator accumulates digits from incoming InputDTMFFrame instances.
     It flushes the aggregated digits by emitting a TranscriptionFrame when:
       - No new digit arrives within the specified timeout period,
-      - The termination digit (“#”) is received, or
+      - The termination digit ("#") is received, or
       - The number of digits aggregated equals the configured 'digits' value.
     """
 
@@ -34,7 +37,9 @@ class DTMFAggregator(FrameProcessor):
         digits: int = None,
         **kwargs,
     ):
-        """:param timeout: Idle timeout in seconds before flushing the aggregated digits.
+        """Initialize the DTMF aggregator.
+
+        :param timeout: Idle timeout in seconds before flushing the aggregated digits.
         :param digits: Number of digits to aggregate before flushing.
         """
         super().__init__(**kwargs)
@@ -48,6 +53,7 @@ class DTMFAggregator(FrameProcessor):
         self._dtmf_capture_active = False
 
     async def process_frame(self, frame: Frame, direction: FrameDirection) -> None:
+        """Process incoming frames and handle DTMF input aggregation."""
         # Handle DTMF frames.
         await super().process_frame(frame, direction)
 
@@ -69,8 +75,8 @@ class DTMFAggregator(FrameProcessor):
             self._digit_event.set()  # Trigger the timeout handler
             await self._start_dtmf_capture()
             await self.push_frame(frame, direction)
-        elif isinstance(frame, StartInterruptionFrame):
-            self.logger.debug("Received StartInterruptionFrame")
+        elif isinstance(frame, InterruptionFrame):
+            self.logger.debug("Received InterruptionFrame")
             if self._aggregation:
                 await self.flush_aggregation()
             await self._end_dtmf_capture()
@@ -108,9 +114,7 @@ class DTMFAggregator(FrameProcessor):
         if "digits" in settings:
             new_digits = settings["digits"]
             if new_digits != self._digits:
-                self.logger.debug(
-                    f"Updating DTMF digits from {self._digits} to {new_digits}"
-                )
+                self.logger.debug(f"Updating DTMF digits from {self._digits} to {new_digits}")
                 self._digits = new_digits
                 settings_changed = True
 
@@ -125,9 +129,7 @@ class DTMFAggregator(FrameProcessor):
                 new_end_on = set(end_value)
 
             if new_end_on != self._end_on:
-                self.logger.debug(
-                    f"Updating DTMF end_on from {self._end_on} to {new_end_on}"
-                )
+                self.logger.debug(f"Updating DTMF end_on from {self._end_on} to {new_end_on}")
                 self._end_on = new_end_on
                 settings_changed = True
 
@@ -142,9 +144,7 @@ class DTMFAggregator(FrameProcessor):
                 new_reset_on = set(reset_value)
 
             if new_reset_on != self._reset_on:
-                self.logger.debug(
-                    f"Updating DTMF reset_on from {self._reset_on} to {new_reset_on}"
-                )
+                self.logger.debug(f"Updating DTMF reset_on from {self._reset_on} to {new_reset_on}")
                 self._reset_on = new_reset_on
                 settings_changed = True
 
@@ -183,9 +183,7 @@ class DTMFAggregator(FrameProcessor):
     def _create_aggregation_task(self, raise_timeout: bool = False) -> None:
         """Creates the aggregation task if it hasn't been created yet."""
         if not self._aggregation_task:
-            self._aggregation_task = self.create_task(
-                self._aggregation_task_handler(raise_timeout)
-            )
+            self._aggregation_task = self.create_task(self._aggregation_task_handler(raise_timeout))
 
     async def _stop_aggregation_task(self) -> None:
         """Stops the aggregation task."""
@@ -198,9 +196,7 @@ class DTMFAggregator(FrameProcessor):
         while True:
             try:
                 # Wait for a new digit signal with a timeout.
-                await asyncio.wait_for(
-                    self._digit_event.wait(), timeout=self._idle_timeout
-                )
+                await asyncio.wait_for(self._digit_event.wait(), timeout=self._idle_timeout)
                 self._digit_event.clear()
             except asyncio.TimeoutError:
                 # No new digit arrived within the timeout period; flush if needed
@@ -216,7 +212,7 @@ class DTMFAggregator(FrameProcessor):
             aggregated_frame.metadata["push_aggregation"] = True
 
             # Send interruption frame (as per original design)
-            await self.push_frame(StartInterruptionFrame(), FrameDirection.DOWNSTREAM)
+            await self.push_frame(InterruptionFrame(), FrameDirection.DOWNSTREAM)
 
             # Push the transcription frame
             await self.push_frame(aggregated_frame, FrameDirection.DOWNSTREAM)

{dv_pipecat_ai-0.0.85.dev833 → dv_pipecat_ai-0.0.85.dev837}/src/pipecat/serializers/__init__.py

@@ -5,6 +5,7 @@ from .exotel import ExotelFrameSerializer
 from .plivo import PlivoFrameSerializer
 from .telnyx import TelnyxFrameSerializer
 from .twilio import TwilioFrameSerializer
+from .vi import VIFrameSerializer
 
 __all__ = [
     "FrameSerializer",
@@ -15,6 +16,7 @@ __all__ = [
     "PlivoFrameSerializer",
     "TelnyxFrameSerializer",
     "TwilioFrameSerializer",
+    "VIFrameSerializer",
 ]
 
 # Optional imports

{dv_pipecat_ai-0.0.85.dev833 → dv_pipecat_ai-0.0.85.dev837}/src/pipecat/serializers/asterisk.py

@@ -1,4 +1,6 @@
 # asterisk_ws_serializer.py
+"""Frame serializer for Asterisk WebSocket communication."""
+
 import base64
 import json
 from typing import Literal, Optional
@@ -12,8 +14,8 @@ from pipecat.frames.frames import (
     EndFrame,
     Frame,
     InputAudioRawFrame,
+    InterruptionFrame,
     StartFrame,
-    StartInterruptionFrame,
     TransportMessageFrame,
     TransportMessageUrgentFrame,
 )
@@ -21,6 +23,8 @@ from pipecat.serializers.base_serializer import FrameSerializer, FrameSerializer
 
 
 class AsteriskFrameSerializer(FrameSerializer):
+    """Serializes Pipecat frames to/from Asterisk WebSocket JSON messages."""
+
     class InputParams(BaseModel):
         """Configuration parameters for AsteriskFrameSerializer.
 
@@ -39,6 +43,12 @@ class AsteriskFrameSerializer(FrameSerializer):
         auto_hang_up: bool = False  # no-op here; adapter handles hangup
 
     def __init__(self, stream_id: str, params: Optional[InputParams] = None):
+        """Initialize the Asterisk frame serializer.
+
+        Args:
+            stream_id: Unique identifier for the media stream.
+            params: Configuration parameters for the serializer.
+        """
         self._stream_id = stream_id
         self._params = params or AsteriskFrameSerializer.InputParams()
         self._tel_rate = self._params.telephony_sample_rate
@@ -49,13 +59,16 @@ class AsteriskFrameSerializer(FrameSerializer):
 
     @property
     def type(self) -> FrameSerializerType:
+        """Return the serializer type (TEXT for JSON messages)."""
         return FrameSerializerType.TEXT  # we send/recv JSON strings
 
     async def setup(self, frame: StartFrame):
+        """Setup the serializer with audio parameters from the StartFrame."""
         self._sample_rate = self._params.sample_rate or frame.audio_in_sample_rate
 
     # Pipecat -> Adapter (play to caller)
     async def serialize(self, frame: Frame) -> str | bytes | None:
+        """Serialize Pipecat frames to Asterisk WebSocket JSON messages."""
         # On pipeline end, ask bridge to hang up
         if (
             self._params.auto_hang_up
@@ -64,7 +77,7 @@ class AsteriskFrameSerializer(FrameSerializer):
         ):
             self._hangup_sent = True
             return json.dumps({"event": "hangup"})
-        if isinstance(frame, StartInterruptionFrame):
+        if isinstance(frame, InterruptionFrame):
             return json.dumps({"event": "clear", "streamId": self._stream_id})
         if isinstance(frame, AudioRawFrame):
             pcm = frame.audio
@@ -114,6 +127,7 @@ class AsteriskFrameSerializer(FrameSerializer):
 
     # Adapter -> Pipecat (audio from caller)
     async def deserialize(self, data: str | bytes) -> Frame | None:
+        """Deserialize Asterisk WebSocket JSON messages to Pipecat frames."""
         try:
             msg = json.loads(data)
         except Exception:

{dv_pipecat_ai-0.0.85.dev833 → dv_pipecat_ai-0.0.85.dev837}/src/pipecat/serializers/convox.py

@@ -22,9 +22,9 @@ from pipecat.frames.frames import (
     Frame,
     InputAudioRawFrame,
     InputDTMFFrame,
+    InterruptionFrame,
     KeypadEntry,
     StartFrame,
-    StartInterruptionFrame,
     TransportMessageFrame,
     TransportMessageUrgentFrame,
 )
@@ -117,7 +117,7 @@ class ConVoxFrameSerializer(FrameSerializer):
             self._call_ended = True
             # Return the callEnd event to be sent via the WebSocket
             return await self._send_call_end_event()
-        elif isinstance(frame, StartInterruptionFrame):
+        elif isinstance(frame, InterruptionFrame):
             # Clear/interrupt command for ConVox
             message = {
                 "event": "clear",

{dv_pipecat_ai-0.0.85.dev833 → dv_pipecat_ai-0.0.85.dev837}/src/pipecat/serializers/custom.py

@@ -28,8 +28,8 @@ from pipecat.frames.frames import (
     EndFrame,
     Frame,
     InputAudioRawFrame,
+    InterruptionFrame,
     StartFrame,
-    StartInterruptionFrame,
     TransportMessageFrame,
     TransportMessageUrgentFrame,
 )
@@ -121,7 +121,7 @@ class CustomFrameSerializer(FrameSerializer):
         Returns:
             Serialized data as JSON string, or None if the frame isn't handled.
         """
-        if isinstance(frame, StartInterruptionFrame):
+        if isinstance(frame, InterruptionFrame):
             # Send clear event to instruct client to discard buffered audio
             answer = {"event": "clear", "stream_sid": self._stream_sid}
             return json.dumps(answer)

dv_pipecat_ai-0.0.85.dev837/src/pipecat/serializers/vi.py

@@ -0,0 +1,324 @@
+#
+# Copyright (c) 2024–2025, Daily
+#
+# SPDX-License-Identifier: BSD 2-Clause License
+#
+
+"""Vodafone Idea (VI) WebSocket frame serializer for audio streaming and call management."""
+
+import base64
+import json
+from datetime import datetime, timezone
+from typing import Optional
+
+from loguru import logger
+from pydantic import BaseModel
+
+from pipecat.audio.utils import create_default_resampler
+from pipecat.frames.frames import (
+    AudioRawFrame,
+    CancelFrame,
+    EndFrame,
+    Frame,
+    InputAudioRawFrame,
+    InputDTMFFrame,
+    KeypadEntry,
+    StartFrame,
+    StartInterruptionFrame,
+    TransportMessageFrame,
+    TransportMessageUrgentFrame,
+)
+from pipecat.serializers.base_serializer import FrameSerializer, FrameSerializerType
+
+
+class VIFrameSerializer(FrameSerializer):
+    """Serializer for Vodafone Idea (VI) WebSocket protocol.
+
+    This serializer handles converting between Pipecat frames and VI's WebSocket
+    protocol for bidirectional audio streaming. It supports audio conversion, DTMF events,
+    and real-time communication with VI telephony systems.
+
+    VI WebSocket protocol requirements:
+    - PCM audio format at 8kHz sample rate
+    - 16-bit Linear PCM encoding
+    - Base64 encoded audio payloads
+    - JSON message format for control and media events
+    - Bitrate: 128 Kbps
+
+    Events (VI → Endpoint):
+    - connected: WebSocket connection established
+    - start: Stream session started with call/stream IDs
+    - media: Audio data in Base64-encoded PCM
+    - dtmf: Keypad digit pressed
+    - stop: Stream ended
+    - mark: Audio playback checkpoint confirmation
+
+    Events (Endpoint → VI):
+    - media: Send audio back to VI
+    - mark: Request acknowledgment for audio playback
+    - clear: Clear queued audio (interruption)
+    - exit: Terminate session gracefully
+    """
+
+    class InputParams(BaseModel):
+        """Configuration parameters for VIFrameSerializer.
+
+        Attributes:
+            vi_sample_rate: Sample rate used by VI, defaults to 8000 Hz (telephony standard).
+            sample_rate: Optional override for pipeline input sample rate.
+            auto_hang_up: Whether to automatically terminate call on EndFrame.
+        """
+
+        vi_sample_rate: int = 8000
+        sample_rate: Optional[int] = None
+        auto_hang_up: bool = False
+
+    def __init__(
+        self,
+        stream_id: str,
+        call_id: Optional[str] = None,
+        params: Optional[InputParams] = None,
+    ):
+        """Initialize the VIFrameSerializer.
+
+        Args:
+            stream_id: The VI stream identifier.
+            call_id: The associated VI call identifier.
+            params: Configuration parameters.
+        """
+        self._stream_id = stream_id
+        self._call_id = call_id
+        self._params = params or VIFrameSerializer.InputParams()
+
+        self._vi_sample_rate = self._params.vi_sample_rate
+        self._sample_rate = 0  # Pipeline input rate
+        self._call_ended = False
+
+        self._resampler = create_default_resampler()
+
+    @property
+    def type(self) -> FrameSerializerType:
+        """Gets the serializer type.
+
+        Returns:
+            The serializer type as TEXT for JSON WebSocket messages.
+        """
+        return FrameSerializerType.TEXT
+
+    async def setup(self, frame: StartFrame):
+        """Sets up the serializer with pipeline configuration.
+
+        Args:
+            frame: The StartFrame containing pipeline configuration.
+        """
+        self._sample_rate = self._params.sample_rate or frame.audio_in_sample_rate
+
+    async def serialize(self, frame: Frame) -> str | bytes | None:
+        """Serializes a Pipecat frame to VI WebSocket format.
+
+        Handles conversion of various frame types to VI WebSocket messages.
+        For EndFrames, initiates call termination if auto_hang_up is enabled.
+
+        Args:
+            frame: The Pipecat frame to serialize.
+
+        Returns:
+            Serialized data as JSON string, or None if the frame isn't handled.
+        """
+        if (
+            self._params.auto_hang_up
+            and not self._call_ended
+            and isinstance(frame, (EndFrame, CancelFrame))
+        ):
+            self._call_ended = True
+            # Return the exit event to terminate the VI session
+            return await self._send_exit_event()
+
+        elif isinstance(frame, StartInterruptionFrame):
+            # Clear/interrupt command for VI - clears queued audio
+            message = {
+                "event": "clear",
+                "stream_id": self._stream_id,
+                "call_id": self._call_id,
+            }
+            logger.debug(f"VI: Sending clear event for stream_id: {self._stream_id}")
+            return json.dumps(message)
+
+        elif isinstance(frame, AudioRawFrame):
+            if self._call_ended:
+                logger.debug("VI SERIALIZE: Skipping audio - call has ended")
+                return None
+
+            # Convert PCM audio to VI format
+            data = frame.audio
+
+            # Resample to VI sample rate (8kHz)
+            serialized_data = await self._resampler.resample(
+                data, frame.sample_rate, self._vi_sample_rate
+            )
+
+            # Encode as base64 for transmission
+            payload = base64.b64encode(serialized_data).decode("ascii")
+
+            # VI expects media event format with Base64-encoded PCM audio
+            timestamp = datetime.now(timezone.utc).isoformat().replace("+00:00", "Z")
+
+            message = {
+                "event": "media",
+                "stream_id": self._stream_id,
+                "media": {
+                    "timestamp": timestamp,
+                    "chunk": len(serialized_data),  # Chunk size in bytes
+                    "payload": payload,
+                },
+            }
+
+            return json.dumps(message)
+
+        elif isinstance(frame, (TransportMessageFrame, TransportMessageUrgentFrame)):
+            # Pass through transport messages (for mark events, etc.)
+            return json.dumps(frame.message)
+
+        return None
+
+    async def _send_exit_event(self):
+        """Send an exit event to VI to terminate the session gracefully.
+
+        This method is called when auto_hang_up is enabled and an EndFrame or
+        CancelFrame is received. The exit event allows IVR logic to continue
+        after the WebSocket session ends.
+        """
+        try:
+            exit_event = {
+                "event": "exit",
+                "stream_id": self._stream_id,
+                "call_id": self._call_id,
+                "timestamp": datetime.now(timezone.utc).isoformat().replace("+00:00", "Z"),
+            }
+
+            logger.info(
+                f"VI auto_hang_up: Sending exit event for stream_id: {self._stream_id}, call_id: {self._call_id}"
+            )
+            return json.dumps(exit_event)
+        except Exception as e:
+            logger.error(f"VI auto_hang_up: Failed to create exit event: {e}")
+            return None
+
+    async def deserialize(self, data: str | bytes) -> Frame | None:
+        """Deserializes VI WebSocket data to Pipecat frames.
+
+        Handles conversion of VI media events to appropriate Pipecat frames.
+
+        Args:
+            data: The raw WebSocket data from VI.
+
+        Returns:
+            A Pipecat frame corresponding to the VI event, or None if unhandled.
+        """
+        try:
+            message = json.loads(data)
+        except json.JSONDecodeError:
+            logger.error(f"Invalid JSON received from VI: {data}")
+            return None
+
+        # Log all incoming events for debugging and monitoring
+        event = message.get("event")
+        logger.debug(
+            f"VI INCOMING EVENT: {event} - stream_id: {self._stream_id}, call_id: {self._call_id}"
+        )
+
+        if event == "media":
+            # Handle incoming audio data from VI
+            media = message.get("media", {})
+            payload_base64 = media.get("payload")
+
+            if not payload_base64:
+                logger.warning("VI DESERIALIZE: No payload in VI media message")
+                return None
+
+            try:
+                payload = base64.b64decode(payload_base64)
+                chunk_size = len(payload)
+
+                # Log chunk info (optional)
+                logger.debug(
+                    f"VI DESERIALIZE: Received audio from VI - {chunk_size} bytes at {self._vi_sample_rate}Hz"
+                )
+
+            except Exception as e:
+                logger.error(f"VI DESERIALIZE: Error decoding VI audio payload: {e}")
+                return None
+
+            # Convert from VI sample rate (8kHz) to pipeline sample rate
+            deserialized_data = await self._resampler.resample(
+                payload,
+                self._vi_sample_rate,
+                self._sample_rate,
+            )
+
+            audio_frame = InputAudioRawFrame(
+                audio=deserialized_data,
+                num_channels=1,  # VI uses mono audio
+                sample_rate=self._sample_rate,
+            )
+            return audio_frame
+
+        elif event == "dtmf":
+            # Handle DTMF events
+            dtmf_data = message.get("dtmf", {})
+            digit = dtmf_data.get("digit")
+
+            if digit:
+                try:
+                    logger.info(f"VI: Received DTMF digit: {digit}")
+                    return InputDTMFFrame(KeypadEntry(digit))
+                except ValueError:
+                    logger.warning(f"Invalid DTMF digit from VI: {digit}")
+                    return None
+
+        elif event == "connected":
+            # Handle connection event
+            logger.info(f"VI connection established: {message}")
+            return None
+
+        elif event == "start":
+            # Handle stream start event
+            logger.info(f"VI stream started: {message}")
+            return None
+
+        elif event == "stop":
+            # Handle stream stop event
+            logger.info(f"VI stream stopped: {message}")
+            # Don't end the call here, wait for explicit exit or call end
+            return None
+
+        elif event == "mark":
+            # Handle mark event - checkpoint confirming audio playback completion
+            mark_data = message.get("mark", {})
+            mark_name = mark_data.get("name", "unknown")
+            logger.info(f"VI mark event received: {mark_name}")
+            # Mark events are informational, no frame to return
+            return None
+
+        elif event == "error":
+            # Handle error events
+            error_msg = message.get("error", "Unknown error")
+            logger.error(f"VI error: {error_msg}")
+            return None
+
+        elif event == "exit":
+            # Handle exit event from VI
+            logger.info("VI exit event received - terminating session")
+            self._call_ended = True
+            return CancelFrame()
+
+        elif event == "call_end" or event == "callEnd":
+            # Handle call end event (if VI sends this)
+            logger.info("VI call end event received")
+            self._call_ended = True
+            return CancelFrame()
+
+        else:
+            logger.debug(f"VI UNHANDLED EVENT: {event}")
+
+        return None

{dv_pipecat_ai-0.0.85.dev833 → dv_pipecat_ai-0.0.85.dev837}/src/pipecat/services/sarvam/tts.py

@@ -23,7 +23,6 @@ from pipecat.frames.frames import (
     InterruptionFrame,
     LLMFullResponseEndFrame,
     StartFrame,
-    StartInterruptionFrame,
     TTSAudioRawFrame,
     TTSStartedFrame,
     TTSStoppedFrame,

{dv_pipecat_ai-0.0.85.dev833 → dv_pipecat_ai-0.0.85.dev837}/src/pipecat/services/vistaar/llm.py

@@ -14,15 +14,15 @@ from loguru import logger
 from pydantic import BaseModel, Field
 
 from pipecat.frames.frames import (
-    EndFrame,
     CancelFrame,
+    EndFrame,
     Frame,
+    InterruptionFrame,
     LLMFullResponseEndFrame,
     LLMFullResponseStartFrame,
     LLMMessagesFrame,
     LLMTextFrame,
     LLMUpdateSettingsFrame,
-    StartInterruptionFrame,
 )
 from pipecat.processors.aggregators.llm_response import (
     LLMAssistantAggregatorParams,
@@ -391,7 +391,7 @@ class VistaarLLMService(LLMService):
             )
             await self.push_frame(frame, direction)
             return
-        elif isinstance(frame, StartInterruptionFrame):
+        elif isinstance(frame, InterruptionFrame):
             await self._handle_interruption()
             await self.push_frame(frame, direction)
             return
@@ -467,4 +467,4 @@ class VistaarLLMService(LLMService):
 
     def can_generate_metrics(self) -> bool:
         """Check if this service can generate processing metrics."""
-        return True
+        return True