app.py

import os
import gradio as gr
import requests
import json
import speech_recognition as sr
from tempfile import NamedTemporaryFile
import logging
from dotenv import load_dotenv

# Load environment variables from Hugging Face secrets or .env
load_dotenv()
HUGGINGFACE_API_TOKEN = os.getenv("HUGGING_FACE_API")
GROQ_API_KEY = os.getenv("GROQ_API_KEY")
GROQ_MODEL = os.getenv("GROQ_MODEL", "llama3-8b") #mixtral-8x7b-32768")

# Logging setup
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# API headers
groq_headers = {
    "Authorization": f"Bearer {GROQ_API_KEY}",
    "Content-Type": "application/json"
}
tts_headers = {
    "Authorization": f"Bearer {HUGGINGFACE_API_TOKEN}"
}

# API endpoints
GROQ_API_URL = "https://api.groq.com/openai/v1/chat/completions"
TTS_API_URL = "https://api-inference.huggingface.co/models/microsoft/speecht5_tts"

# Emotion dictionary
emotion_options = {
    "neutral": "Neutral or balanced mood",
    "positive": "Generally positive or optimistic",
    "happy": "Feeling joy or happiness",
    "excited": "Feeling enthusiastic or energetic",
    "sad": "Feeling down or unhappy",
    "angry": "Feeling frustrated or irritated",
    "negative": "Generally negative or pessimistic",
    "anxious": "Feeling worried or nervous"
}

# Recognizer for audio input
def transcribe_audio(audio_path):
    recognizer = sr.Recognizer()
    try:
        with sr.AudioFile(audio_path) as source:
            audio_data = recognizer.record(source)
            return recognizer.recognize_google(audio_data)
    except Exception as e:
        logger.error(f"Audio transcription failed: {e}")
        return ""

# Groq response handler
def get_groq_response(prompt, chat_history):
    messages = [{"role": "system", "content": prompt}]
    for msg in chat_history:
        role, content = ("user", msg[0]) if msg else ("assistant", msg[1])
        messages.append({"role": role, "content": content})

    data = {
        "model": GROQ_MODEL,
        "messages": messages[-20:],  # Limit history
        "temperature": 0.7,
        "max_tokens": 1024
    }

    try:
        res = requests.post(GROQ_API_URL, headers=groq_headers, json=data)
        res.raise_for_status()
        return res.json()["choices"][0]["message"]["content"]
    except Exception as e:
        logger.error(f"Groq API error: {e}")
        return "Sorry, I couldn't generate a response right now."

# Hugging Face TTS
def generate_audio(text):
    response = requests.post(TTS_API_URL, headers=tts_headers, json={"inputs": text})
    if response.status_code == 200:
        with NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
            tmp.write(response.content)
            return tmp.name
    else:
        logger.error(f"TTS generation failed: {response.text}")
        return None

# Conversation state
conversation_history = []

# Main chat logic
def chat_with_ai(audio, text_input, emotion, history):
    global conversation_history
    user_input = text_input.strip() if text_input else ""

    if audio:
        transcription = transcribe_audio(audio)
        if transcription:
            user_input = transcription

    if not user_input:
        return "Please provide a message or audio.", None, history

    conversation_history.append((user_input, None))

    prompt = f"""You are an empathetic AI assistant. The user feels {emotion} ({emotion_options[emotion]}).
Respond supportively and helpfully in a concise manner."""

    ai_response = get_groq_response(prompt, conversation_history)
    conversation_history[-1] = (user_input, ai_response)

    audio_output_path = generate_audio(ai_response)

    return ai_response, audio_output_path, conversation_history

def clear_conversation():
    global conversation_history
    conversation_history = []
    return [], None, None, "Conversation cleared."

# Gradio Interface
with gr.Blocks(title="Mind AID AI Assistant") as iface:
    gr.Markdown("# Mind AID: Emotion-Aware Conversational AI")
    gr.Markdown("AI assistant with emotion-awareness, powered by Groq and Hugging Face TTS.")

    with gr.Row():
        with gr.Column(scale=3):
            emotion = gr.Dropdown(
                label="How are you feeling?",
                choices=list(emotion_options.keys()),
                value="neutral"
            )
            emotion_description = gr.Markdown("**Current mood:** Neutral or balanced mood")
            emotion.change(
                fn=lambda e: f"**Current mood:** {emotion_options[e]}",
                inputs=emotion,
                outputs=emotion_description
            )
        with gr.Column(scale=1):
            clear_btn = gr.Button("Clear Conversation")
            status_box = gr.Textbox(label="Status", interactive=False)

    with gr.Row():
        chat_history = gr.Chatbot(label="Conversation", height=400)

    with gr.Row():
        with gr.Column(scale=4):
            text_input = gr.Textbox(label="Type your message here", lines=2)
        with gr.Column(scale=1):
            audio_input = gr.Audio(type="filepath", label="Or speak")

    with gr.Row():
        submit_btn = gr.Button("Send", variant="primary")
    with gr.Row():
        output_audio = gr.Audio(label="AI Voice")

    submit_btn.click(
        fn=chat_with_ai,
        inputs=[audio_input, text_input, emotion, chat_history],
        outputs=[status_box, output_audio, chat_history]
    )
    text_input.submit(
        fn=chat_with_ai,
        inputs=[audio_input, text_input, emotion, chat_history],
        outputs=[status_box, output_audio, chat_history]
    )
    clear_btn.click(
        fn=clear_conversation,
        inputs=[],
        outputs=[chat_history, audio_input, text_input, status_box]
    )

iface.launch(share=True)