Update app.py

app.py

@@ -5,9 +5,9 @@ from PIL import Image
 
 # --- KONFIGURASI ---
 IMG_SIZE = (224, 224)
-MODEL_PATH = "best_model.h5" 
+MODEL_PATH = "best_model.h5"
 
-# Definisi Kelas (Sesuai urutan training model)
+# Definisi Kelas (Sesuai urutan training model Anda)
 class_names = [
     'Abrasions', 'Bruises', 'Burns', 'Cut', 'Diabetic Wounds', 
     'Laceration', 'Normal', 'Pressure Wounds', 'Surgical Wounds', 'Venous Wounds'
@@ -97,87 +97,86 @@ def get_first_aid_recommendation(wound_class):
     }
     return rekomendasi.get(wound_class, "Rekomendasi belum tersedia. Silakan konsultasi dengan petugas medis.")
 
-# --- SCRIPT JAVASCRIPT FINAL (ULTRA STABLE) ---
-get_location_js = """
-async function getLoc() {
-    try {
-        const position = await new Promise((resolve, reject) => {
-            navigator.geolocation.getCurrentPosition(resolve, reject, {timeout: 8000});
-        });
-        return `${position.coords.latitude}, ${position.coords.longitude}`;
-    } catch (e) {
-        return "Lokasi tidak tersedia (Izin ditolak/Timeout)";
-    }
-}
-"""
-
-def predict_image(image_input, location_data):
-    # Validasi input dasar
-    if image_input is None:
-        return {}, "Silakan upload gambar.", "⚠️ Foto belum diunggah."
+# --- FUNGSI PREDIKSI ---
+def predict_image(image_input):
     if best_model is None:
-        return {}, "⚠️ Model gagal dimuat di server.", "N/A"
+        return {}, "⚠️ Model tidak ditemukan. Harap upload file model (.h5)."
 
-    # Status lokasi
-    status_lokasi = f"📍 **Koordinat:** {location_data}" if location_data and "tidak" not in location_data.lower() else f"⚠️ {location_data or 'Gagal mengambil lokasi'}"
+    if image_input is None:
+        return {}, "Silakan upload gambar."
 
-    # --- Preprocessing & Prediksi ---
+    # Preprocessing
     img = image_input.resize(IMG_SIZE)
-    img_array = tf.keras.utils.img_to_array(img) / 255.0
-    img_array = np.expand_dims(img_array, 0)
-    
+    img_array = tf.keras.utils.img_to_array(img)
+    img_array = tf.expand_dims(img_array, 0)
+    img_array = img_array / 255.0
+
+    # Prediksi
     predictions = best_model.predict(img_array)
-    # Gunakan softmax jika model Anda belum memilikinya di layer terakhir
     scores = tf.nn.softmax(predictions[0]).numpy()
 
+    # Mapping hasil untuk UI (Bahasa Indonesia)
     translated_output_dict = {
         translation_map.get(class_names[i], class_names[i]): float(scores[i])
         for i in range(len(class_names))
     }
 
+    # Ambil label tertinggi
     top_idx = np.argmax(scores)
     top_label_en = class_names[top_idx]
     top_confidence = scores[top_idx]
 
-    # Logika Threshold
-    if top_confidence < 0.20:
+    # --- LOGIKA THRESHOLD ---
+    THRESHOLD = 0.20
+
+    if top_confidence < THRESHOLD:
         top_label_id = "Normal (Tidak Terdeteksi)"
-        rekomendasi_teks = "Model tidak yakin. Pastikan luka terlihat jelas."
+        rekomendasi_teks = (
+            "**Mohon Maaf:** Model kurang akurat dalam menganalisis foto ini.\n\n"
+            "**Saran:**\n"
+            "1. Pastikan area luka terlihat jelas dan tidak blur.\n"
+            "2. Gunakan pencahayaan yang cukup (terang).\n"
+            "3. Ambil foto dari sudut tegak lurus ke arah luka.\n\n"
+            "Jika Anda merasa luka ini serius, segera hubungi tenaga medis meskipun hasil analisis tidak muncul."
+        )
     else:
         top_label_id = translation_map.get(top_label_en, top_label_en)
         rekomendasi_teks = get_first_aid_recommendation(top_label_en)
 
+    # Format Markdown untuk Gradio
     formatted_output = (
         f"### Analisis: **{top_label_id}**\n\n"
         f"**Langkah Pertolongan:**\n{rekomendasi_teks}\n\n"
         f"--- \n*Tingkat Keyakinan AI: {top_confidence:.2%}*"
     )
-    
-    return translated_output_dict, formatted_output, status_lokasi
+
+    return translated_output_dict, formatted_output
 
 # --- UI INTERFACE ---
-# Pindahkan theme ke launch() atau hapus dari Blocks untuk menghindari error versi 6.0
-with gr.Blocks() as demo:
+with gr.Blocks(theme=gr.themes.Soft(primary_hue="red", secondary_hue="slate")) as demo:
     gr.Markdown("# 🚨 FirstAidLens")
-    
-    location_hidden = gr.Textbox(visible=False)
-    
+    gr.Markdown("Deteksi jenis luka secara instan dan dapatkan panduan pertolongan pertama yang tepat.")
+
     with gr.Row():
-        with gr.Column():
-            input_img = gr.Image(sources=["upload", "webcam"], type="pil", label="Foto Luka")
-            btn_predict = gr.Button("Analisis Luka & Ambil Lokasi", variant="primary")
-            
-        with gr.Column():
-            output_label = gr.Label(num_top_classes=3, label="Hasil Analisis")
-            output_markdown = gr.Markdown("### Panduan Pertolongan Pertama")
-            output_location = gr.Markdown("") 
-
-    btn_predict.click(
+        with gr.Column(scale=1):
+            input_img = gr.Image(
+                sources=["upload", "webcam"],
+                type="pil",
+                label="Ambil Foto Luka"
+            )
+            gr.Markdown("> **Penting:** Hasil AI ini hanya referensi awal. Jika luka parah atau pendarahan tidak berhenti, segera hubungi **112**.")
+
+        with gr.Column(scale=1):
+            output_label = gr.Label(num_top_classes=3, label="Hasil Analisis Jenis Luka")
+            output_markdown = gr.Markdown("### Panduan Pertolongan Pertama\n_Upload atau ambil foto untuk melihat rekomendasi._")
+
+    # Trigger otomatis saat gambar diupload/diambil
+    input_img.change(
         fn=predict_image,
-        inputs=[input_img, location_hidden],
-        outputs=[output_label, output_markdown, output_location],
-        js=get_location_js
+        inputs=input_img,
+        outputs=[output_label, output_markdown]
     )
 
+# Launch (Server Name 0.0.0.0 wajib untuk Docker)
 if __name__ == "__main__":
     demo.launch(server_name="0.0.0.0", server_port=7860)