app.py

import streamlit as st
import random
from gradio_client import Client, file

# ====================================
# Example Ayahs
# ====================================

ayahs = {
    "sad": [
        {
            "ayah": "2:286",
            "arabic": "لَا يُكَلِّفُ اللَّهُ نَفْسًا إِلَّا وُسْعَهَا",
            "translation": "Allah does not burden a soul beyond that it can bear.",
            "tafsir": "Every test is within your capacity, with Allah’s help."
        }
    ],
    "happy": [
        {
            "ayah": "94:5-6",
            "arabic": "فَإِنَّ مَعَ الْعُسْرِ يُسْرًا إِنَّ مَعَ الْعُسْرِ يُسْرًا",
            "translation": "Indeed, with hardship comes ease.",
            "tafsir": "Your happiness is part of Allah’s ease."
        }
    ],
    "angry": [
        {
            "ayah": "3:134",
            "arabic": "وَالْكَاظِمِينَ الْغَيْظَ",
            "translation": "Those who restrain anger...",
            "tafsir": "Patience and controlling anger are rewarded."
        }
    ]
}

# ====================================
# Streamlit UI
# ====================================

st.set_page_config(page_title="Qur’an Healing Soul - Emotion from Image", page_icon="🌙")
st.title("🌙 Qur’an Healing Soul - FER via API")

uploaded_file = st.file_uploader("Upload your selfie", type=["jpg", "png", "jpeg"])

if uploaded_file is not None:
    # Show uploaded image
    st.image(uploaded_file, caption="Uploaded Image", use_container_width=True)
    st.info("Detecting emotion... please wait...")

    # Save file locally
    with open("temp_image.png", "wb") as f:
        f.write(uploaded_file.read())

    # Call Gradio API
    client = Client("ElenaRyumina/face_emotion_recognition")
    result = client.predict(
        inp=file("temp_image.png"),
        api_name="/preprocess_image_and_predict"
    )

    _, _, confidences = result

    # Use top label directly
    dominant = confidences["label"]
    st.success(f"Dominant Emotion: **{dominant}**")

    # Map to emotion group
    dominant_lower = dominant.lower()
    if dominant_lower in ["sad", "fear"]:
        key = "sad"
    elif dominant_lower in ["happy", "happiness", "surprise"]:
        key = "happy"
    elif dominant_lower in ["angry", "disgust"]:
        key = "angry"
    else:
        key = "sad"

    # Show ayah
    ayah = random.choice(ayahs[key])
    st.markdown(f"""
    ### 📖 Ayah ({ayah['ayah']})

    **Arabic:** *{ayah['arabic']}*

    **Translation:** {ayah['translation']}

    **Tafsir:** {ayah['tafsir']}
    """)