2025-07-12 14:19:48,504 - ERROR livekit - failed to emit event response_content_added
Traceback (most recent call last):
  File "D:\AIVoiveAgent\livekit-front\voice-assistant-frontend-warx7u\venv\Lib\site-packages\livekit\rtc\event_emitter.py", line 58, in emit
    callback(*callback_args)
  File "D:\AIVoiveAgent\livekit-front\voice-assistant-frontend-warx7u\venv\Lib\site-packages\livekit\agents\multimodal\multimodal_agent.py", line 277, in _on_content_added
Traceback (most recent call last):
  File "D:\AIVoiveAgent\livekit-front\voice-assistant-frontend-warx7u\venv\Lib\site-packages\livekit\rtc\event_emitter.py", line 58, in emit
    callback(*callback_args)
  File "D:\AIVoiveAgent\livekit-front\voice-assistant-frontend-warx7u\venv\Lib\site-packages\livekit\agents\multimodal\multimodal_agent.py", line 277, in _on_content_added
    tr_fwd = transcription.TTSSegmentsForwarder(
    callback(*callback_args)
  File "D:\AIVoiveAgent\livekit-front\voice-assistant-frontend-warx7u\venv\Lib\site-packages\livekit\agents\multimodal\multimodal_agent.py", line 277, in _on_content_added
    tr_fwd = transcription.TTSSegmentsForwarder(
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    tr_fwd = transcription.TTSSegmentsForwarder(
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\AIVoiveAgent\livekit-front\voice-assistant-frontend-warx7u\venv\Lib\site-packages\livekit\agents\transcription\tts_forwarder.py", line 110, in __init__
  File "D:\AIVoiveAgent\livekit-front\voice-assistant-frontend-warx7u\venv\Lib\site-packages\livekit\agents\transcription\tts_forwarder.py", line 110, in __init__
    track = _utils.find_micro_track_id(room, identity)
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\AIVoiveAgent\livekit-front\voice-assistant-frontend-warx7u\venv\Lib\site-packages\livekit\agents\transcription\_utils.py", line 26, in find_micro_track_id
    raise ValueError(f"participant {identity} does not have a microphone track")
ValueError: participant agent-AJ_bCAkN5MvsBCy does not have a microphone track

from __future__ import annotations

import logging
import os
import asyncio
from dotenv import load_dotenv

from livekit import rtc
from livekit.agents import (
    AutoSubscribe,
    JobContext,
    WorkerOptions,
    cli,
    llm,
)
from livekit.agents.multimodal import MultimodalAgent
from livekit.plugins import openai

# Import clinic data functions
from clinic_data import fetch_specialties
from actions import Appointment

load_dotenv(dotenv_path=".env.local")
logger = logging.getLogger("my-worker")
logger.setLevel(<http://logging.INFO|logging.INFO>)

# Format clinic data for LLM context
def format_clinic_data(specialties) -> str:
    if not specialties or not specialties.specialties:
        return "لا توجد بيانات متاحة للعيادات حالياً."
    
    clinic_info = []
    for specialty in specialties.specialties:
        doctors_info = []
        for doctor in specialty.doctors:
            schedule_info = [f"يوم {s.day}: {s.from_time} - {<http://s.to|s.to>}" for s in doctor.schedule]
            doctors_info.append(
                f"- الدكتور/ة {doctor.name} (المستوى: {doctor.level})\n"
                f"  المواعيد: {', '.join(schedule_info)}"
            )
        
        clinic_info.append(
            f"العيادة: {specialty.name}\n"
            f"الأطباء:\n" + "\n".join(doctors_info)
        )
    
    return (
        "البيانات الحالية للعيادات المتاحة:\n\n" +
        "\n\n".join(clinic_info) +
        "\n\nعند سؤال المستخدم عن العيادات أو الأطباء، استخدم هذه البيانات."
    )

async def entrypoint(ctx: JobContext):
    <http://logger.info|logger.info>(f"connecting to room {ctx.room.name}")
    await ctx.connect(auto_subscribe=AutoSubscribe.AUDIO_ONLY)

    participant = await ctx.wait_for_participant()

    run_multimodal_agent(ctx, participant)

    <http://logger.info|logger.info>("agent started")

def run_multimodal_agent(ctx: JobContext, participant: rtc.RemoteParticipant):
    <http://logger.info|logger.info>("starting multimodal agent")

    # Fetch clinic data from API
    api_url = os.getenv("CLINIC_API_URL", "<https://localhost:63801/api/Raya/medical/specialties>")
    specialties = fetch_specialties(api_url)
    clinic_data = format_clinic_data(specialties)

    # Create background task to refresh data periodically
    async def refresh_data():
        while True:
            await asyncio.sleep(600)  # Refresh every 10 minutes
            new_specialties = fetch_specialties(api_url)
            if new_specialties:
                nonlocal clinic_data
                clinic_data = format_clinic_data(new_specialties)
                <http://logger.info|logger.info>("Clinic data refreshed")

    asyncio.create_task(refresh_data())
    fnc_ctx = Appointment()
    model = openai.realtime.RealtimeModel(
        instructions=(
            "أنت مساعد صوتي عربي متخصص في حجز مواعيد العيادات. "
            "يجب أن تساعد المستخدمين في:"
            "\n1. معرفة العيادات والأطباء المتاحين"
            "\n2. حجز المواعيد بناءً على التخصص والطبيب والوقت"
            "\n3. الإجابة على استفساراتهم حول العيادات"
            "\n\n" + clinic_data + "\n\n"
            "إرشادات هامة:"
            "\n- استخدم أسلوباً ودوداً وبسيطاً"
            "\n- استخدم اللغة العربية"
            "\n- استخدم جمل مباشرة و قصيرة"
            "\n- استفسر عن: الاسم، التخصص المطلوب، الطبيب المفضل، اليوم والوقت"
            "\n- قدم خيارات من البيانات عند السؤال عن العيادات"
            "\n- تجنب استخدام الإنجليزية أو علامات الترقيم غير المنطوقة"
        ),
        modalities=["audio", "text"],
    )

    # Initial chat context
    chat_ctx = llm.ChatContext()
    chat_ctx.append(
        text=(
            "مرحباً! أنا مساعدك الصوتي لحجز مواعيد العيادات. "
            "كيف يمكنني مساعدتك اليوم؟"
        ),
        role="assistant",
    )

    agent = MultimodalAgent(
        model=model,
        chat_ctx=chat_ctx,
        fnc_ctx=fnc_ctx,
    )
    agent.start(ctx.room, participant)
    agent.generate_reply()

if __name__ == "__main__":
    cli.run_app(
        WorkerOptions(
            entrypoint_fnc=entrypoint,
        )
    )