Update app.py

app.py

@@ -1,9 +1,11 @@
 import streamlit as st
-from gradio_client import Client, file
-import tempfile
 import random
+from gradio_client import Client, file
+
+# ====================================
+# Example Ayahs
+# ====================================
 
-# --- Your Qur’an ayahs ---
 ayahs = {
     "sad": [
         {
@@ -31,50 +33,57 @@ ayahs = {
     ]
 }
 
-# --- Streamlit UI ---
-st.set_page_config(page_title="Qur’an Healing Soul - Image Emotion", page_icon="🌙")
-
-st.title("🌙 Qur’an Healing Soul - Emotion from Image")
-
-uploaded_file = st.file_uploader("Upload your selfie", type=["jpg", "jpeg", "png"])
+# ====================================
+# Streamlit UI
+# ====================================
 
-if uploaded_file:
-    st.image(uploaded_file, caption="Uploaded Image")
+st.set_page_config(page_title="Qur’an Healing Soul - Emotion from Image", page_icon="🌙")
+st.title("🌙 Qur’an Healing Soul - FER via API")
 
-    with tempfile.NamedTemporaryFile(delete=False) as tmp_file:
-        tmp_file.write(uploaded_file.read())
-        tmp_file_path = tmp_file.name
+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_column_width=True)
     st.info("Detecting emotion... please wait...")
 
-    # Call the remote FER model
-    client = Client("ElenaRyumina/Facial_Expression_Recognition")
+    # Save file locally
+    with open("temp_image.png", "wb") as f:
+        f.write(uploaded_file.read())
 
+    # Call Gradio API
+    client = Client("ElenaRyumina/Facial_Expression_Recognition")
     result = client.predict(
-        inp=file(tmp_file_path),
+        inp=file("temp_image.png"),
         api_name="/preprocess_image_and_predict"
     )
 
-    # Unpack returned values
-    _, _, confidences = result
+    face_image, heatmap_image, confidences = result
 
+    # Show face & heatmap
+    st.image(face_image, caption="Detected Face", use_column_width=True)
+    st.image(heatmap_image, caption="Heatmap", use_column_width=True)
+
+    # Show probabilities
     st.subheader("Emotion Probabilities")
     st.json(confidences)
 
-    # Find dominant emotion
-    dominant = max(confidences, key=confidences.get)
+    # Use top label directly
+    dominant = confidences["label"]
     st.success(f"Dominant Emotion: **{dominant}**")
 
-    # Map emotion
-    if dominant.lower() in ["sad", "fear"]:
+    # Map to emotion group
+    dominant_lower = dominant.lower()
+    if dominant_lower in ["sad", "fear"]:
         key = "sad"
-    elif dominant.lower() in ["happy", "surprise"]:
+    elif dominant_lower in ["happy", "happiness", "surprise"]:
         key = "happy"
-    elif dominant.lower() in ["angry", "disgust"]:
+    elif dominant_lower in ["angry", "disgust"]:
         key = "angry"
     else:
         key = "sad"
 
+    # Show ayah
     ayah = random.choice(ayahs[key])
     st.markdown(f"""
     ### 📖 Ayah ({ayah['ayah']})
@@ -85,3 +94,4 @@ if uploaded_file:
 
     **Tafsir:** {ayah['tafsir']}
     """)
+