app.py

import gradio as gr
import os
import requests
import base64

ORCHESTRATOR_URL = os.getenv("HF_ORCHESTRATOR_URL", "")
WEBHOOK_SECRET = os.getenv("HF_ORCHESTRATOR_WEBHOOK_SECRET", "")
BASIC_USER = os.getenv("N8N_BASIC_AUTH_USER", "")
BASIC_PASS = os.getenv("N8N_BASIC_AUTH_PASSWORD", "")

def _headers():
    headers = {"x-webhook-secret": WEBHOOK_SECRET, "Content-Type": "application/json"}
    if BASIC_USER and BASIC_PASS:
        import base64 as b64
        token = b64.b64encode(f"{BASIC_USER}:{BASIC_PASS}".encode()).decode()
        headers["Authorization"] = f"Basic {token}"
    return headers

def chat_send(message, history, conversation_id):
    payload = {"secret": WEBHOOK_SECRET, "conversation_id": conversation_id, "text": message}
    r = requests.post(f"{ORCHESTRATOR_URL}/voice-agent", json=payload, headers=_headers())
    data = r.json()
    if r.status_code != 200 or data.get("error"):
        return history + [[message, data.get("error", "Error")]], conversation_id, ""
    assistant = data.get("assistant_response_text") or ""
    convo = data.get("conversation_id") or conversation_id
    audio_html = ""
    if data.get("tts_audio_url"):
        audio_html = f"<audio controls src=\"{data['tts_audio_url']}\"></audio>"
    return history + [[message, assistant]], convo, audio_html

def voice_send(audio_path, history, conversation_id):
    audio_b64 = None
    if audio_path:
        with open(audio_path, "rb") as f:
            audio_b64 = base64.b64encode(f.read()).decode("utf-8")
    payload = {"secret": WEBHOOK_SECRET, "conversation_id": conversation_id, "audio": audio_b64}
    r = requests.post(f"{ORCHESTRATOR_URL}/voice-agent", json=payload, headers=_headers())
    data = r.json()
    if r.status_code != 200 or data.get("error"):
        return history + [["", data.get("error", "Error")]], conversation_id, ""
    user_text = data.get("transcript_text", "")
    assistant = data.get("assistant_response_text") or ""
    convo = data.get("conversation_id") or conversation_id
    audio_html = ""
    if data.get("tts_audio_url"):
        audio_html = f"<audio controls src=\"{data['tts_audio_url']}\"></audio>"
    return history + [[user_text, assistant]], convo, audio_html

with gr.Blocks() as demo:
    gr.Markdown("# Voice Agent")
    chatbox = gr.Chatbot()
    msg = gr.Textbox(label="Message")
    send_btn = gr.Button("Send")
    mic = gr.Audio(sources=["microphone"], type="filepath", label="Talk")
    voice_btn = gr.Button("Send Voice")
    audio_out = gr.HTML()
    conversation_state = gr.State(value=None)
    send_btn.click(chat_send, [msg, chatbox, conversation_state], [chatbox, conversation_state, audio_out])
    voice_btn.click(voice_send, [mic, chatbox, conversation_state], [chatbox, conversation_state, audio_out])

if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=7860)