dv-pipecat-ai 0.0.85.dev7__py3-none-any.whl → 0.0.85.dev698__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Potentially problematic release.
This version of dv-pipecat-ai might be problematic. Click here for more details.
- {dv_pipecat_ai-0.0.85.dev7.dist-info → dv_pipecat_ai-0.0.85.dev698.dist-info}/METADATA +78 -117
- {dv_pipecat_ai-0.0.85.dev7.dist-info → dv_pipecat_ai-0.0.85.dev698.dist-info}/RECORD +156 -122
- pipecat/adapters/base_llm_adapter.py +38 -1
- pipecat/adapters/services/anthropic_adapter.py +9 -14
- pipecat/adapters/services/aws_nova_sonic_adapter.py +5 -0
- pipecat/adapters/services/bedrock_adapter.py +236 -13
- pipecat/adapters/services/gemini_adapter.py +12 -8
- pipecat/adapters/services/open_ai_adapter.py +19 -7
- pipecat/adapters/services/open_ai_realtime_adapter.py +5 -0
- pipecat/audio/filters/krisp_viva_filter.py +193 -0
- pipecat/audio/filters/noisereduce_filter.py +15 -0
- pipecat/audio/turn/base_turn_analyzer.py +9 -1
- pipecat/audio/turn/smart_turn/base_smart_turn.py +14 -8
- pipecat/audio/turn/smart_turn/data/__init__.py +0 -0
- pipecat/audio/turn/smart_turn/data/smart-turn-v3.0.onnx +0 -0
- pipecat/audio/turn/smart_turn/http_smart_turn.py +6 -2
- pipecat/audio/turn/smart_turn/local_smart_turn.py +1 -1
- pipecat/audio/turn/smart_turn/local_smart_turn_v2.py +1 -1
- pipecat/audio/turn/smart_turn/local_smart_turn_v3.py +124 -0
- pipecat/audio/vad/data/README.md +10 -0
- pipecat/audio/vad/vad_analyzer.py +13 -1
- pipecat/extensions/voicemail/voicemail_detector.py +5 -5
- pipecat/frames/frames.py +120 -87
- pipecat/observers/loggers/debug_log_observer.py +3 -3
- pipecat/observers/loggers/llm_log_observer.py +7 -3
- pipecat/observers/loggers/user_bot_latency_log_observer.py +22 -10
- pipecat/pipeline/runner.py +12 -4
- pipecat/pipeline/service_switcher.py +64 -36
- pipecat/pipeline/task.py +85 -24
- pipecat/processors/aggregators/dtmf_aggregator.py +28 -22
- pipecat/processors/aggregators/{gated_openai_llm_context.py → gated_llm_context.py} +9 -9
- pipecat/processors/aggregators/gated_open_ai_llm_context.py +12 -0
- pipecat/processors/aggregators/llm_response.py +6 -7
- pipecat/processors/aggregators/llm_response_universal.py +19 -15
- pipecat/processors/aggregators/user_response.py +6 -6
- pipecat/processors/aggregators/vision_image_frame.py +24 -2
- pipecat/processors/audio/audio_buffer_processor.py +43 -8
- pipecat/processors/filters/stt_mute_filter.py +2 -0
- pipecat/processors/frame_processor.py +103 -17
- pipecat/processors/frameworks/langchain.py +8 -2
- pipecat/processors/frameworks/rtvi.py +209 -68
- pipecat/processors/frameworks/strands_agents.py +170 -0
- pipecat/processors/logger.py +2 -2
- pipecat/processors/transcript_processor.py +4 -4
- pipecat/processors/user_idle_processor.py +3 -6
- pipecat/runner/run.py +270 -50
- pipecat/runner/types.py +2 -0
- pipecat/runner/utils.py +51 -10
- pipecat/serializers/exotel.py +5 -5
- pipecat/serializers/livekit.py +20 -0
- pipecat/serializers/plivo.py +6 -9
- pipecat/serializers/protobuf.py +6 -5
- pipecat/serializers/telnyx.py +2 -2
- pipecat/serializers/twilio.py +43 -23
- pipecat/services/ai_service.py +2 -6
- pipecat/services/anthropic/llm.py +2 -25
- pipecat/services/asyncai/tts.py +2 -3
- pipecat/services/aws/__init__.py +1 -0
- pipecat/services/aws/llm.py +122 -97
- pipecat/services/aws/nova_sonic/__init__.py +0 -0
- pipecat/services/aws/nova_sonic/context.py +367 -0
- pipecat/services/aws/nova_sonic/frames.py +25 -0
- pipecat/services/aws/nova_sonic/llm.py +1155 -0
- pipecat/services/aws/stt.py +1 -3
- pipecat/services/aws_nova_sonic/__init__.py +19 -1
- pipecat/services/aws_nova_sonic/aws.py +11 -1151
- pipecat/services/aws_nova_sonic/context.py +13 -355
- pipecat/services/aws_nova_sonic/frames.py +13 -17
- pipecat/services/azure/realtime/__init__.py +0 -0
- pipecat/services/azure/realtime/llm.py +65 -0
- pipecat/services/azure/stt.py +15 -0
- pipecat/services/cartesia/tts.py +2 -2
- pipecat/services/deepgram/__init__.py +1 -0
- pipecat/services/deepgram/flux/__init__.py +0 -0
- pipecat/services/deepgram/flux/stt.py +636 -0
- pipecat/services/elevenlabs/__init__.py +2 -1
- pipecat/services/elevenlabs/stt.py +254 -276
- pipecat/services/elevenlabs/tts.py +5 -5
- pipecat/services/fish/tts.py +2 -2
- pipecat/services/gemini_multimodal_live/events.py +38 -524
- pipecat/services/gemini_multimodal_live/file_api.py +23 -173
- pipecat/services/gemini_multimodal_live/gemini.py +41 -1403
- pipecat/services/gladia/stt.py +56 -72
- pipecat/services/google/__init__.py +1 -0
- pipecat/services/google/gemini_live/__init__.py +3 -0
- pipecat/services/google/gemini_live/file_api.py +189 -0
- pipecat/services/google/gemini_live/llm.py +1582 -0
- pipecat/services/google/gemini_live/llm_vertex.py +184 -0
- pipecat/services/google/llm.py +15 -11
- pipecat/services/google/llm_openai.py +3 -3
- pipecat/services/google/llm_vertex.py +86 -16
- pipecat/services/google/tts.py +7 -3
- pipecat/services/heygen/api.py +2 -0
- pipecat/services/heygen/client.py +8 -4
- pipecat/services/heygen/video.py +2 -0
- pipecat/services/hume/__init__.py +5 -0
- pipecat/services/hume/tts.py +220 -0
- pipecat/services/inworld/tts.py +6 -6
- pipecat/services/llm_service.py +15 -5
- pipecat/services/lmnt/tts.py +2 -2
- pipecat/services/mcp_service.py +4 -2
- pipecat/services/mem0/memory.py +6 -5
- pipecat/services/mistral/llm.py +29 -8
- pipecat/services/moondream/vision.py +42 -16
- pipecat/services/neuphonic/tts.py +2 -2
- pipecat/services/openai/__init__.py +1 -0
- pipecat/services/openai/base_llm.py +27 -20
- pipecat/services/openai/realtime/__init__.py +0 -0
- pipecat/services/openai/realtime/context.py +272 -0
- pipecat/services/openai/realtime/events.py +1106 -0
- pipecat/services/openai/realtime/frames.py +37 -0
- pipecat/services/openai/realtime/llm.py +829 -0
- pipecat/services/openai/tts.py +16 -8
- pipecat/services/openai_realtime/__init__.py +27 -0
- pipecat/services/openai_realtime/azure.py +21 -0
- pipecat/services/openai_realtime/context.py +21 -0
- pipecat/services/openai_realtime/events.py +21 -0
- pipecat/services/openai_realtime/frames.py +21 -0
- pipecat/services/openai_realtime_beta/azure.py +16 -0
- pipecat/services/openai_realtime_beta/openai.py +17 -5
- pipecat/services/playht/tts.py +31 -4
- pipecat/services/rime/tts.py +3 -4
- pipecat/services/sarvam/tts.py +2 -6
- pipecat/services/simli/video.py +2 -2
- pipecat/services/speechmatics/stt.py +1 -7
- pipecat/services/stt_service.py +34 -0
- pipecat/services/tavus/video.py +2 -2
- pipecat/services/tts_service.py +9 -9
- pipecat/services/vision_service.py +7 -6
- pipecat/tests/utils.py +4 -4
- pipecat/transcriptions/language.py +41 -1
- pipecat/transports/base_input.py +17 -42
- pipecat/transports/base_output.py +42 -26
- pipecat/transports/daily/transport.py +199 -26
- pipecat/transports/heygen/__init__.py +0 -0
- pipecat/transports/heygen/transport.py +381 -0
- pipecat/transports/livekit/transport.py +228 -63
- pipecat/transports/local/audio.py +6 -1
- pipecat/transports/local/tk.py +11 -2
- pipecat/transports/network/fastapi_websocket.py +1 -1
- pipecat/transports/smallwebrtc/connection.py +98 -19
- pipecat/transports/smallwebrtc/request_handler.py +204 -0
- pipecat/transports/smallwebrtc/transport.py +65 -23
- pipecat/transports/tavus/transport.py +23 -12
- pipecat/transports/websocket/client.py +41 -5
- pipecat/transports/websocket/fastapi.py +21 -11
- pipecat/transports/websocket/server.py +14 -7
- pipecat/transports/whatsapp/api.py +8 -0
- pipecat/transports/whatsapp/client.py +47 -0
- pipecat/utils/base_object.py +54 -22
- pipecat/utils/string.py +12 -1
- pipecat/utils/tracing/service_decorators.py +21 -21
- {dv_pipecat_ai-0.0.85.dev7.dist-info → dv_pipecat_ai-0.0.85.dev698.dist-info}/WHEEL +0 -0
- {dv_pipecat_ai-0.0.85.dev7.dist-info → dv_pipecat_ai-0.0.85.dev698.dist-info}/licenses/LICENSE +0 -0
- {dv_pipecat_ai-0.0.85.dev7.dist-info → dv_pipecat_ai-0.0.85.dev698.dist-info}/top_level.txt +0 -0
- /pipecat/services/{aws_nova_sonic → aws/nova_sonic}/ready.wav +0 -0
pipecat/serializers/plivo.py
CHANGED
|
@@ -22,10 +22,10 @@ from pipecat.frames.frames import (
|
|
|
22
22
|
Frame,
|
|
23
23
|
InputAudioRawFrame,
|
|
24
24
|
InputDTMFFrame,
|
|
25
|
+
InterruptionFrame,
|
|
26
|
+
OutputTransportMessageFrame,
|
|
27
|
+
OutputTransportMessageUrgentFrame,
|
|
25
28
|
StartFrame,
|
|
26
|
-
StartInterruptionFrame,
|
|
27
|
-
TransportMessageFrame,
|
|
28
|
-
TransportMessageUrgentFrame,
|
|
29
29
|
)
|
|
30
30
|
from pipecat.serializers.base_serializer import FrameSerializer, FrameSerializerType
|
|
31
31
|
|
|
@@ -122,7 +122,7 @@ class PlivoFrameSerializer(FrameSerializer):
|
|
|
122
122
|
self._hangup_attempted = True
|
|
123
123
|
await self._hang_up_call()
|
|
124
124
|
return None
|
|
125
|
-
elif isinstance(frame,
|
|
125
|
+
elif isinstance(frame, InterruptionFrame):
|
|
126
126
|
answer = {"event": "clearAudio", "streamId": self._stream_id}
|
|
127
127
|
return json.dumps(answer)
|
|
128
128
|
elif isinstance(frame, AudioRawFrame):
|
|
@@ -148,7 +148,7 @@ class PlivoFrameSerializer(FrameSerializer):
|
|
|
148
148
|
}
|
|
149
149
|
|
|
150
150
|
return json.dumps(answer)
|
|
151
|
-
elif isinstance(frame, (
|
|
151
|
+
elif isinstance(frame, (OutputTransportMessageFrame, OutputTransportMessageUrgentFrame)):
|
|
152
152
|
return json.dumps(frame.message)
|
|
153
153
|
|
|
154
154
|
# Return None for unhandled frames
|
|
@@ -178,10 +178,7 @@ class PlivoFrameSerializer(FrameSerializer):
|
|
|
178
178
|
return
|
|
179
179
|
|
|
180
180
|
# Plivo API endpoint for hanging up calls
|
|
181
|
-
|
|
182
|
-
endpoint = f"https://api.plivo.com/v1/Account/{auth_id}/Call/{call_id}/Stream/{self._stream_id}/"
|
|
183
|
-
else:
|
|
184
|
-
endpoint = f"https://api.plivo.com/v1/Account/{auth_id}/Call/{call_id}/"
|
|
181
|
+
endpoint = f"https://api.plivo.com/v1/Account/{auth_id}/Call/{call_id}/"
|
|
185
182
|
|
|
186
183
|
# Create basic auth from auth_id and auth_token
|
|
187
184
|
auth = aiohttp.BasicAuth(auth_id, auth_token)
|
pipecat/serializers/protobuf.py
CHANGED
|
@@ -15,11 +15,12 @@ import pipecat.frames.protobufs.frames_pb2 as frame_protos
|
|
|
15
15
|
from pipecat.frames.frames import (
|
|
16
16
|
Frame,
|
|
17
17
|
InputAudioRawFrame,
|
|
18
|
+
InputTransportMessageFrame,
|
|
18
19
|
OutputAudioRawFrame,
|
|
20
|
+
OutputTransportMessageFrame,
|
|
21
|
+
OutputTransportMessageUrgentFrame,
|
|
19
22
|
TextFrame,
|
|
20
23
|
TranscriptionFrame,
|
|
21
|
-
TransportMessageFrame,
|
|
22
|
-
TransportMessageUrgentFrame,
|
|
23
24
|
)
|
|
24
25
|
from pipecat.serializers.base_serializer import FrameSerializer, FrameSerializerType
|
|
25
26
|
|
|
@@ -82,7 +83,7 @@ class ProtobufFrameSerializer(FrameSerializer):
|
|
|
82
83
|
Serialized frame as bytes, or None if frame type is not serializable.
|
|
83
84
|
"""
|
|
84
85
|
# Wrapping this messages as a JSONFrame to send
|
|
85
|
-
if isinstance(frame, (
|
|
86
|
+
if isinstance(frame, (OutputTransportMessageFrame, OutputTransportMessageUrgentFrame)):
|
|
86
87
|
frame = MessageFrame(
|
|
87
88
|
data=json.dumps(frame.message),
|
|
88
89
|
)
|
|
@@ -134,11 +135,11 @@ class ProtobufFrameSerializer(FrameSerializer):
|
|
|
134
135
|
if "pts" in args_dict:
|
|
135
136
|
del args_dict["pts"]
|
|
136
137
|
|
|
137
|
-
# Special handling for MessageFrame ->
|
|
138
|
+
# Special handling for MessageFrame -> OutputTransportMessageUrgentFrame
|
|
138
139
|
if class_name == MessageFrame:
|
|
139
140
|
try:
|
|
140
141
|
msg = json.loads(args_dict["data"])
|
|
141
|
-
instance =
|
|
142
|
+
instance = InputTransportMessageFrame(message=msg)
|
|
142
143
|
logger.debug(f"ProtobufFrameSerializer: Transport message {instance}")
|
|
143
144
|
except Exception as e:
|
|
144
145
|
logger.error(f"Error parsing MessageFrame data: {e}")
|
pipecat/serializers/telnyx.py
CHANGED
|
@@ -29,8 +29,8 @@ from pipecat.frames.frames import (
|
|
|
29
29
|
Frame,
|
|
30
30
|
InputAudioRawFrame,
|
|
31
31
|
InputDTMFFrame,
|
|
32
|
+
InterruptionFrame,
|
|
32
33
|
StartFrame,
|
|
33
|
-
StartInterruptionFrame,
|
|
34
34
|
)
|
|
35
35
|
from pipecat.serializers.base_serializer import FrameSerializer, FrameSerializerType
|
|
36
36
|
|
|
@@ -137,7 +137,7 @@ class TelnyxFrameSerializer(FrameSerializer):
|
|
|
137
137
|
self._hangup_attempted = True
|
|
138
138
|
await self._hang_up_call()
|
|
139
139
|
return None
|
|
140
|
-
elif isinstance(frame,
|
|
140
|
+
elif isinstance(frame, InterruptionFrame):
|
|
141
141
|
answer = {"event": "clear"}
|
|
142
142
|
return json.dumps(answer)
|
|
143
143
|
elif isinstance(frame, AudioRawFrame):
|
pipecat/serializers/twilio.py
CHANGED
|
@@ -22,10 +22,10 @@ from pipecat.frames.frames import (
|
|
|
22
22
|
Frame,
|
|
23
23
|
InputAudioRawFrame,
|
|
24
24
|
InputDTMFFrame,
|
|
25
|
+
InterruptionFrame,
|
|
26
|
+
OutputTransportMessageFrame,
|
|
27
|
+
OutputTransportMessageUrgentFrame,
|
|
25
28
|
StartFrame,
|
|
26
|
-
StartInterruptionFrame,
|
|
27
|
-
TransportMessageFrame,
|
|
28
|
-
TransportMessageUrgentFrame,
|
|
29
29
|
)
|
|
30
30
|
from pipecat.serializers.base_serializer import FrameSerializer, FrameSerializerType
|
|
31
31
|
|
|
@@ -61,6 +61,8 @@ class TwilioFrameSerializer(FrameSerializer):
|
|
|
61
61
|
call_sid: Optional[str] = None,
|
|
62
62
|
account_sid: Optional[str] = None,
|
|
63
63
|
auth_token: Optional[str] = None,
|
|
64
|
+
region: Optional[str] = None,
|
|
65
|
+
edge: Optional[str] = None,
|
|
64
66
|
params: Optional[InputParams] = None,
|
|
65
67
|
):
|
|
66
68
|
"""Initialize the TwilioFrameSerializer.
|
|
@@ -70,13 +72,42 @@ class TwilioFrameSerializer(FrameSerializer):
|
|
|
70
72
|
call_sid: The associated Twilio Call SID (optional, but required for auto hang-up).
|
|
71
73
|
account_sid: Twilio account SID (required for auto hang-up).
|
|
72
74
|
auth_token: Twilio auth token (required for auto hang-up).
|
|
75
|
+
region: Twilio region (e.g., "au1", "ie1"). Must be specified with edge.
|
|
76
|
+
edge: Twilio edge location (e.g., "sydney", "dublin"). Must be specified with region.
|
|
73
77
|
params: Configuration parameters.
|
|
74
78
|
"""
|
|
79
|
+
self._params = params or TwilioFrameSerializer.InputParams()
|
|
80
|
+
|
|
81
|
+
# Validate hangup-related parameters if auto_hang_up is enabled
|
|
82
|
+
if self._params.auto_hang_up:
|
|
83
|
+
# Validate required credentials
|
|
84
|
+
missing_credentials = []
|
|
85
|
+
if not call_sid:
|
|
86
|
+
missing_credentials.append("call_sid")
|
|
87
|
+
if not account_sid:
|
|
88
|
+
missing_credentials.append("account_sid")
|
|
89
|
+
if not auth_token:
|
|
90
|
+
missing_credentials.append("auth_token")
|
|
91
|
+
|
|
92
|
+
if missing_credentials:
|
|
93
|
+
raise ValueError(
|
|
94
|
+
f"auto_hang_up is enabled but missing required parameters: {', '.join(missing_credentials)}"
|
|
95
|
+
)
|
|
96
|
+
|
|
97
|
+
# Validate region and edge are both provided if either is specified
|
|
98
|
+
if (region and not edge) or (edge and not region):
|
|
99
|
+
raise ValueError(
|
|
100
|
+
"Both edge and region parameters are required if one is set. "
|
|
101
|
+
f"Twilio's FQDN format requires both: api.{{edge}}.{{region}}.twilio.com. "
|
|
102
|
+
f"Got: region='{region}', edge='{edge}'"
|
|
103
|
+
)
|
|
104
|
+
|
|
75
105
|
self._stream_sid = stream_sid
|
|
76
106
|
self._call_sid = call_sid
|
|
77
107
|
self._account_sid = account_sid
|
|
78
108
|
self._auth_token = auth_token
|
|
79
|
-
self.
|
|
109
|
+
self._region = region
|
|
110
|
+
self._edge = edge
|
|
80
111
|
|
|
81
112
|
self._twilio_sample_rate = self._params.twilio_sample_rate
|
|
82
113
|
self._sample_rate = 0 # Pipeline input rate
|
|
@@ -122,7 +153,7 @@ class TwilioFrameSerializer(FrameSerializer):
|
|
|
122
153
|
self._hangup_attempted = True
|
|
123
154
|
await self._hang_up_call()
|
|
124
155
|
return None
|
|
125
|
-
elif isinstance(frame,
|
|
156
|
+
elif isinstance(frame, InterruptionFrame):
|
|
126
157
|
answer = {"event": "clear", "streamSid": self._stream_sid}
|
|
127
158
|
return json.dumps(answer)
|
|
128
159
|
elif isinstance(frame, AudioRawFrame):
|
|
@@ -135,7 +166,7 @@ class TwilioFrameSerializer(FrameSerializer):
|
|
|
135
166
|
if serialized_data is None or len(serialized_data) == 0:
|
|
136
167
|
# Ignoring in case we don't have audio
|
|
137
168
|
return None
|
|
138
|
-
|
|
169
|
+
|
|
139
170
|
payload = base64.b64encode(serialized_data).decode("utf-8")
|
|
140
171
|
answer = {
|
|
141
172
|
"event": "media",
|
|
@@ -144,7 +175,7 @@ class TwilioFrameSerializer(FrameSerializer):
|
|
|
144
175
|
}
|
|
145
176
|
|
|
146
177
|
return json.dumps(answer)
|
|
147
|
-
elif isinstance(frame, (
|
|
178
|
+
elif isinstance(frame, (OutputTransportMessageFrame, OutputTransportMessageUrgentFrame)):
|
|
148
179
|
return json.dumps(frame.message)
|
|
149
180
|
|
|
150
181
|
# Return None for unhandled frames
|
|
@@ -158,25 +189,14 @@ class TwilioFrameSerializer(FrameSerializer):
|
|
|
158
189
|
account_sid = self._account_sid
|
|
159
190
|
auth_token = self._auth_token
|
|
160
191
|
call_sid = self._call_sid
|
|
192
|
+
region = self._region
|
|
193
|
+
edge = self._edge
|
|
161
194
|
|
|
162
|
-
|
|
163
|
-
|
|
164
|
-
if not call_sid:
|
|
165
|
-
missing.append("call_sid")
|
|
166
|
-
if not account_sid:
|
|
167
|
-
missing.append("account_sid")
|
|
168
|
-
if not auth_token:
|
|
169
|
-
missing.append("auth_token")
|
|
170
|
-
|
|
171
|
-
logger.warning(
|
|
172
|
-
f"Cannot hang up Twilio call: missing required parameters: {', '.join(missing)}"
|
|
173
|
-
)
|
|
174
|
-
return
|
|
195
|
+
region_prefix = f"{region}." if region else ""
|
|
196
|
+
edge_prefix = f"{edge}." if edge else ""
|
|
175
197
|
|
|
176
198
|
# Twilio API endpoint for updating calls
|
|
177
|
-
endpoint =
|
|
178
|
-
f"https://api.twilio.com/2010-04-01/Accounts/{account_sid}/Calls/{call_sid}.json"
|
|
179
|
-
)
|
|
199
|
+
endpoint = f"https://api.{edge_prefix}{region_prefix}twilio.com/2010-04-01/Accounts/{account_sid}/Calls/{call_sid}.json"
|
|
180
200
|
|
|
181
201
|
# Create basic auth from account_sid and auth_token
|
|
182
202
|
auth = aiohttp.BasicAuth(account_sid, auth_token)
|
pipecat/services/ai_service.py
CHANGED
|
@@ -97,9 +97,7 @@ class AIService(FrameProcessor):
|
|
|
97
97
|
pass
|
|
98
98
|
|
|
99
99
|
async def _update_settings(self, settings: Mapping[str, Any]):
|
|
100
|
-
from pipecat.services.
|
|
101
|
-
SessionProperties,
|
|
102
|
-
)
|
|
100
|
+
from pipecat.services.openai.realtime.events import SessionProperties
|
|
103
101
|
|
|
104
102
|
for key, value in settings.items():
|
|
105
103
|
logger.debug("Update request for:", key, value)
|
|
@@ -111,9 +109,7 @@ class AIService(FrameProcessor):
|
|
|
111
109
|
logger.debug("Attempting to update", key, value)
|
|
112
110
|
|
|
113
111
|
try:
|
|
114
|
-
from pipecat.services.
|
|
115
|
-
TurnDetection,
|
|
116
|
-
)
|
|
112
|
+
from pipecat.services.openai.realtime.events import TurnDetection
|
|
117
113
|
|
|
118
114
|
if isinstance(self._session_properties, SessionProperties):
|
|
119
115
|
current_properties = self._session_properties
|
|
@@ -42,7 +42,6 @@ from pipecat.frames.frames import (
|
|
|
42
42
|
LLMTextFrame,
|
|
43
43
|
LLMUpdateSettingsFrame,
|
|
44
44
|
UserImageRawFrame,
|
|
45
|
-
VisionImageRawFrame,
|
|
46
45
|
)
|
|
47
46
|
from pipecat.metrics.metrics import LLMTokenUsage
|
|
48
47
|
from pipecat.processors.aggregators.llm_context import LLMContext
|
|
@@ -152,7 +151,7 @@ class AnthropicLLMService(LLMService):
|
|
|
152
151
|
self,
|
|
153
152
|
*,
|
|
154
153
|
api_key: str,
|
|
155
|
-
model: str = "claude-sonnet-4-
|
|
154
|
+
model: str = "claude-sonnet-4-5-20250929",
|
|
156
155
|
params: Optional[InputParams] = None,
|
|
157
156
|
client=None,
|
|
158
157
|
retry_timeout_secs: Optional[float] = 5.0,
|
|
@@ -163,7 +162,7 @@ class AnthropicLLMService(LLMService):
|
|
|
163
162
|
|
|
164
163
|
Args:
|
|
165
164
|
api_key: Anthropic API key for authentication.
|
|
166
|
-
model: Model name to use. Defaults to "claude-sonnet-4-
|
|
165
|
+
model: Model name to use. Defaults to "claude-sonnet-4-5-20250929".
|
|
167
166
|
params: Optional model parameters for inference.
|
|
168
167
|
client: Optional custom Anthropic client instance.
|
|
169
168
|
retry_timeout_secs: Request timeout in seconds for retry logic.
|
|
@@ -495,12 +494,6 @@ class AnthropicLLMService(LLMService):
|
|
|
495
494
|
context = frame.context
|
|
496
495
|
elif isinstance(frame, LLMMessagesFrame):
|
|
497
496
|
context = AnthropicLLMContext.from_messages(frame.messages)
|
|
498
|
-
elif isinstance(frame, VisionImageRawFrame):
|
|
499
|
-
# This is only useful in very simple pipelines because it creates
|
|
500
|
-
# a new context. Generally we want a context manager to catch
|
|
501
|
-
# UserImageRawFrames coming through the pipeline and add them
|
|
502
|
-
# to the context.
|
|
503
|
-
context = AnthropicLLMContext.from_image_frame(frame)
|
|
504
497
|
elif isinstance(frame, LLMUpdateSettingsFrame):
|
|
505
498
|
await self._update_settings(frame.settings)
|
|
506
499
|
elif isinstance(frame, LLMEnablePromptCachingFrame):
|
|
@@ -626,22 +619,6 @@ class AnthropicLLMContext(OpenAILLMContext):
|
|
|
626
619
|
self._restructure_from_openai_messages()
|
|
627
620
|
return self
|
|
628
621
|
|
|
629
|
-
@classmethod
|
|
630
|
-
def from_image_frame(cls, frame: VisionImageRawFrame) -> "AnthropicLLMContext":
|
|
631
|
-
"""Create context from a vision image frame.
|
|
632
|
-
|
|
633
|
-
Args:
|
|
634
|
-
frame: The vision image frame to process.
|
|
635
|
-
|
|
636
|
-
Returns:
|
|
637
|
-
New Anthropic context with the image message.
|
|
638
|
-
"""
|
|
639
|
-
context = cls()
|
|
640
|
-
context.add_image_frame_message(
|
|
641
|
-
format=frame.format, size=frame.size, image=frame.image, text=frame.text
|
|
642
|
-
)
|
|
643
|
-
return context
|
|
644
|
-
|
|
645
622
|
def set_messages(self, messages: List):
|
|
646
623
|
"""Set the messages list and reset cache tracking.
|
|
647
624
|
|
pipecat/services/asyncai/tts.py
CHANGED
|
@@ -20,8 +20,8 @@ from pipecat.frames.frames import (
|
|
|
20
20
|
EndFrame,
|
|
21
21
|
ErrorFrame,
|
|
22
22
|
Frame,
|
|
23
|
+
InterruptionFrame,
|
|
23
24
|
StartFrame,
|
|
24
|
-
StartInterruptionFrame,
|
|
25
25
|
TTSAudioRawFrame,
|
|
26
26
|
TTSStartedFrame,
|
|
27
27
|
TTSStoppedFrame,
|
|
@@ -119,7 +119,6 @@ class AsyncAITTSService(InterruptibleTTSService):
|
|
|
119
119
|
"""
|
|
120
120
|
super().__init__(
|
|
121
121
|
aggregate_sentences=aggregate_sentences,
|
|
122
|
-
push_text_frames=False,
|
|
123
122
|
pause_frame_processing=True,
|
|
124
123
|
push_stop_frames=True,
|
|
125
124
|
sample_rate=sample_rate,
|
|
@@ -275,7 +274,7 @@ class AsyncAITTSService(InterruptibleTTSService):
|
|
|
275
274
|
direction: The direction to push the frame.
|
|
276
275
|
"""
|
|
277
276
|
await super().push_frame(frame, direction)
|
|
278
|
-
if isinstance(frame, (TTSStoppedFrame,
|
|
277
|
+
if isinstance(frame, (TTSStoppedFrame, InterruptionFrame)):
|
|
279
278
|
self._started = False
|
|
280
279
|
|
|
281
280
|
async def _receive_messages(self):
|