Update app.py

app.py

@@ -1,56 +1,132 @@
-
-import streamlit as st
-import pandas as pd
+import gradio as gr
 import pickle
 
-# Load the fitted model
-model = pickle.load(open('model (10).pkl', 'rb'))
+# Load model + scaler (lebih efisien: sekali di awal)
+model = pickle.load(open("model (10).pkl", "rb"))
+scaler = pickle.load(open("scaler.pkl", "rb"))
 
 # Mapping antara kelas dan nama tipe almond
 class_mapping = {
-    0: 'Sanora',
-    1: 'Mamra',
-    2: 'Regular',
+    0: "Sanora",
+    1: "Mamra",
+    2: "Regular",
     # tambahkan kelas lainnya sesuai kebutuhan
 }
 
-st.title('Almond Classification')
-st.write('This web app classifies almonds based on your input features.')
-
-# Input untuk setiap fitur
-length_major_axis = st.number_input('Length (major axis)', min_value=269.356903, max_value=279.879883)
-width_minor_axis = st.number_input('Width (minor axis)', min_value=176.023636, max_value=227.940628)
-thickness_depth = st.number_input('Thickness (depth)', min_value=107.253448, max_value=127.795132)
-area = st.number_input('Area', min_value=18471.5, max_value=36683.0)
-perimeter = st.number_input('Perimeter', min_value=551.688379, max_value=887.310743)
-roundness = st.slider('Roundness', min_value=0.472718, max_value=0.643761, step=0.01)
-solidity = st.slider('Solidity', min_value=0.931800, max_value=0.973384, step=0.01)
-compactness = st.slider('Compactness', min_value=1.383965, max_value=1.764701, step=0.01)
-aspect_ratio = st.slider('Aspect Ratio', min_value=1.530231, max_value=1.705716, step=0.01)
-eccentricity = st.slider('Eccentricity', min_value=0.75693, max_value=0.81012, step=0.01)
-extent = st.slider('Extent', min_value=0.656535, max_value=0.725739, step=0.01)
-convex_area = st.number_input('Convex hull (convex area)', min_value=18068.0, max_value=36683.0, step=0.01)
-
-# Tombol untuk memprediksi
-if st.button('Predict'):
-        # Muat scaler
-    scaler = pickle.load(open('scaler.pkl', 'rb'))
-    
-    # Input dari pengguna
-    input_features = [[length_major_axis, width_minor_axis, thickness_depth, area,
-                       perimeter, roundness, solidity, compactness, aspect_ratio,
-                       eccentricity, extent, convex_area]]
-    
-    # Lakukan scaling pada input
-    input_features_scaled = scaler.transform(input_features)
-    
-    # Prediksi menggunakan model
-    prediction = model.predict(input_features_scaled)
-
-
-
-    # Menggunakan mapping untuk mendapatkan nama tipe almond
-    predicted_class_name = class_mapping[prediction[0]]
-    
-    st.write(f'The predicted class is: {predicted_class_name}')
+def predict_almond(
+    length_major_axis,
+    width_minor_axis,
+    thickness_depth,
+    area,
+    perimeter,
+    roundness,
+    solidity,
+    compactness,
+    aspect_ratio,
+    eccentricity,
+    extent,
+    convex_area,
+):
+    # Susun input sesuai urutan fitur saat training
+    X = [[
+        length_major_axis, width_minor_axis, thickness_depth, area,
+        perimeter, roundness, solidity, compactness, aspect_ratio,
+        eccentricity, extent, convex_area
+    ]]
+
+    # Scaling + prediksi
+    X_scaled = scaler.transform(X)
+    pred = model.predict(X_scaled)[0]
+
+    return class_mapping.get(int(pred), f"Unknown class: {pred}")
+
+with gr.Blocks(title="Almond Classification") as demo:
+    gr.Markdown("# Almond Classification")
+    gr.Markdown("This web app classifies almonds based on your input features.")
+
+    with gr.Row():
+        with gr.Column():
+            length_major_axis = gr.Number(
+                label="Length (major axis)",
+                minimum=269.356903, maximum=279.879883,
+                value=(269.356903 + 279.879883) / 2,
+                step=0.001,
+            )
+            width_minor_axis = gr.Number(
+                label="Width (minor axis)",
+                minimum=176.023636, maximum=227.940628,
+                value=(176.023636 + 227.940628) / 2,
+                step=0.001,
+            )
+            thickness_depth = gr.Number(
+                label="Thickness (depth)",
+                minimum=107.253448, maximum=127.795132,
+                value=(107.253448 + 127.795132) / 2,
+                step=0.001,
+            )
+            area = gr.Number(
+                label="Area",
+                minimum=18471.5, maximum=36683.0,
+                value=(18471.5 + 36683.0) / 2,
+                step=0.1,
+            )
+            perimeter = gr.Number(
+                label="Perimeter",
+                minimum=551.688379, maximum=887.310743,
+                value=(551.688379 + 887.310743) / 2,
+                step=0.001,
+            )
+            convex_area = gr.Number(
+                label="Convex hull (convex area)",
+                minimum=18068.0, maximum=36683.0,
+                value=(18068.0 + 36683.0) / 2,
+                step=0.1,
+            )
+
+        with gr.Column():
+            roundness = gr.Slider(
+                label="Roundness",
+                minimum=0.472718, maximum=0.643761, step=0.01,
+                value=(0.472718 + 0.643761) / 2,
+            )
+            solidity = gr.Slider(
+                label="Solidity",
+                minimum=0.931800, maximum=0.973384, step=0.01,
+                value=(0.931800 + 0.973384) / 2,
+            )
+            compactness = gr.Slider(
+                label="Compactness",
+                minimum=1.383965, maximum=1.764701, step=0.01,
+                value=(1.383965 + 1.764701) / 2,
+            )
+            aspect_ratio = gr.Slider(
+                label="Aspect Ratio",
+                minimum=1.530231, maximum=1.705716, step=0.01,
+                value=(1.530231 + 1.705716) / 2,
+            )
+            eccentricity = gr.Slider(
+                label="Eccentricity",
+                minimum=0.75693, maximum=0.81012, step=0.01,
+                value=(0.75693 + 0.81012) / 2,
+            )
+            extent = gr.Slider(
+                label="Extent",
+                minimum=0.656535, maximum=0.725739, step=0.01,
+                value=(0.656535 + 0.725739) / 2,
+            )
+
+    btn = gr.Button("Predict")
+    out = gr.Textbox(label="The predicted class is:", interactive=False)
+
+    btn.click(
+        fn=predict_almond,
+        inputs=[
+            length_major_axis, width_minor_axis, thickness_depth, area,
+            perimeter, roundness, solidity, compactness, aspect_ratio,
+            eccentricity, extent, convex_area
+        ],
+        outputs=out,
+    )
 
+if __name__ == "__main__":
+    demo.launch()