Update src/streamlit_app.py

src/streamlit_app.py

@@ -9,14 +9,14 @@ import tensorflow as tf
 # CONFIG
 # ======================
 st.set_page_config(page_title="Tomato Detector 🍅")
-st.title("🍅 Tomato Disease Detection (Local Image Mode)")
+st.title("🍅 Tomato Disease Detection")
 
 # ======================
 # PATHS
 # ======================
 MODEL_PATH = "/app/src/best_tomato_model.h5"
 JSON_PATH  = "/app/src/class_indices.json"
-IMAGE_PATH = "/app/src/test.jpg"   # 🔥 LOCAL IMAGE
+DEFAULT_IMAGE_PATH = "/app/src/test.jpg"  # fallback image
 
 # ======================
 # DEBUG FILES
@@ -34,10 +34,6 @@ if not os.path.exists(JSON_PATH):
     st.error("class_indices.json not found")
     st.stop()
 
-if not os.path.exists(IMAGE_PATH):
-    st.error("test.png not found in /app/src")
-    st.stop()
-
 # ======================
 # LOAD MODEL
 # ======================
@@ -49,12 +45,13 @@ model = load_model()
 st.success("Model loaded")
 
 # ======================
-# LOAD CLASSES (FIXED ORDER)
+# LOAD CLASSES
 # ======================
 with open(JSON_PATH) as f:
     class_indices = json.load(f)
 
 class_names = [None] * len(class_indices)
+
 for k, v in class_indices.items():
     class_names[v] = k
 
@@ -72,16 +69,37 @@ def preprocess(img):
     return np.expand_dims(arr, axis=0)
 
 # ======================
-# LOAD LOCAL IMAGE
+# FILE UPLOADER UI
+# ======================
+uploaded_file = st.file_uploader(
+    "Upload tomato leaf image 🍅",
+    type=["jpg", "jpeg", "png"]
+)
+
+# ======================
+# IMAGE SOURCE SELECT
 # ======================
-image = Image.open(IMAGE_PATH)
+if uploaded_file is not None:
+    image = Image.open(uploaded_file)
+    st.success("Using uploaded image")
+else:
+    if os.path.exists(DEFAULT_IMAGE_PATH):
+        image = Image.open(DEFAULT_IMAGE_PATH)
+        st.warning("Using default test image")
+    else:
+        st.error("No image available")
+        st.stop()
 
-st.image(image, caption="Local test image", use_column_width=True)
+# ======================
+# SHOW IMAGE
+# ======================
+st.image(image, caption="Input image", use_column_width=True)
 
 # ======================
 # PREDICT
 # ======================
 try:
+
     img = preprocess(image)
 
     preds = model.predict(img, verbose=0)