apunkt/dograh
1
0
Fork 0
mirror of https://github.com/dograh-hq/dograh.git synced 2026-06-22 08:38:13 +02:00
Code Issues Projects Releases Packages Wiki Activity

feat: add recording audio option in tool and node transitions (#232)

Browse source 
* feat: allow uploading recording as part of node transition

* feat: allow recordings in tool transitions

* chore: fix tests
This commit is contained in:
Abhishek 2026-04-10 17:53:42 +05:30 • committed by GitHub
parent 3f19a16e7f
commit 7c245051d2
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
54 changed files with 3575 additions and 640 deletions
Download patch file Download diff file Expand all files Collapse all files

87
api/services/workflow/pipecat_engine_custom_tools.py
View file

@ -16,6 +16,7 @@ from loguru import logger
from api.db import db_client
from api.enums import ToolCategory, WorkflowRunMode
from api.services.pipecat.audio_playback import play_audio, play_audio_loop
from api.services.telephony.call_transfer_manager import get_call_transfer_manager
from api.services.telephony.factory import get_telephony_provider
from api.services.telephony.transfer_event_protocol import TransferContext
@ -27,7 +28,6 @@ from api.services.workflow.tools.custom_tool import (
execute_http_tool,
tool_to_function_schema,
)
from api.utils.hold_audio import play_hold_audio_loop
from pipecat.adapters.schemas.function_schema import FunctionSchema
from pipecat.frames.frames import (
FunctionCallResultProperties,
@ -77,6 +77,45 @@ class CustomToolManager:
self._engine = engine
self._organization_id: Optional[int] = None
async def _play_config_message(
self, config: dict, *, append_to_context: bool = False
) -> bool:
"""Play a message from tool config — text or pre-recorded audio.
Returns True if a message was queued, False otherwise.
"""
message_type = config.get("messageType", "none")
if message_type == "audio":
recording_pk = config.get("audioRecordingId")
if recording_pk and self._engine._fetch_recording_audio:
result = await self._engine._fetch_recording_audio(
recording_pk=int(recording_pk)
)
if result:
await play_audio(
result.audio,
sample_rate=self._engine._audio_config.pipeline_sample_rate
if self._engine._audio_config
else 16000,
queue_frame=self._engine._transport_output.queue_frame,
transcript=result.transcript,
)
return True
else:
logger.warning(f"Failed to fetch recording pk={recording_pk}")
return False
if message_type == "custom":
custom_message = config.get("customMessage", "")
if custom_message:
await self._engine.task.queue_frame(
TTSSpeakFrame(custom_message, append_to_context=append_to_context)
)
return True
return False
async def get_organization_id(self) -> Optional[int]:
"""Get and cache the organization ID from workflow run."""
if self._organization_id is None:
@ -250,9 +289,30 @@ class CustomToolManager:
try:
# Queue custom message before executing the API call
# Queue custom message (text or audio) before executing the API call
config = tool.definition.get("config", {}) if tool.definition else {}
custom_msg_type = config.get("customMessageType", "text")
custom_message = config.get("customMessage", "")
if custom_message:
if custom_msg_type == "audio":
recording_pk = config.get("customMessageRecordingId")
if recording_pk and self._engine._fetch_recording_audio:
logger.info(
f"Playing audio message before HTTP tool: pk={recording_pk}"
)
self._engine._queued_speech_mute_state = "waiting"
result = await self._engine._fetch_recording_audio(
recording_pk=int(recording_pk)
)
if result:
await play_audio(
result.audio,
sample_rate=self._engine._audio_config.pipeline_sample_rate
if self._engine._audio_config
else 16000,
queue_frame=self._engine._transport_output.queue_frame,
transcript=result.transcript,
)
elif custom_message:
logger.info(
f"Playing custom message before HTTP tool: {custom_message}"
)
@ -299,8 +359,6 @@ class CustomToolManager:
try:
# Get the end call configuration
config = tool.definition.get("config", {})
message_type = config.get("messageType", "none")
custom_message = config.get("customMessage", "")
# Handle end call reason if enabled
end_call_reason_enabled = config.get("endCallReason", False)
@ -322,10 +380,8 @@ class CustomToolManager:
properties=properties,
)
if message_type == "custom" and custom_message:
# Queue the custom message to be spoken
logger.info(f"Playing custom goodbye message: {custom_message}")
await self._engine.task.queue_frame(TTSSpeakFrame(custom_message))
played = await self._play_config_message(config)
if played:
# End the call after the message (not immediately)
await self._engine.end_call_with_reason(
EndTaskReason.END_CALL_TOOL_REASON.value,
@ -370,8 +426,6 @@ class CustomToolManager:
# Get the transfer call configuration
config = tool.definition.get("config", {})
destination = config.get("destination", "")
message_type = config.get("messageType", "none")
custom_message = config.get("customMessage", "")
timeout_seconds = config.get(
"timeout", 30
) # Default 30 seconds if not configured
@ -443,10 +497,9 @@ class CustomToolManager:
)
return
if message_type == "custom" and custom_message:
logger.info(f"Playing pre-transfer message: {custom_message}")
played = await self._play_config_message(config)
if played:
self._engine._queued_speech_mute_state = "waiting"
await self._engine.task.queue_frame(TTSSpeakFrame(custom_message))
# Get organization ID for provider configuration
organization_id = await self.get_organization_id()
@ -537,10 +590,10 @@ class CustomToolManager:
# Start hold music as background task
hold_music_task = asyncio.create_task(
play_hold_audio_loop(
self._engine.task,
hold_music_stop_event,
sample_rate,
play_audio_loop(
stop_event=hold_music_stop_event,
sample_rate=sample_rate,
queue_frame=self._engine._transport_output.queue_frame,
)
)