Upload folder using huggingface_hub

app.py

@@ -1,57 +1,47 @@
-import requests
 import streamlit as st
 import pydicom
 import numpy as np
-import cv2
+import requests
 
 st.title("Medical Image Prediction App")
 st.write("Upload a DICOM (.dcm) image and get prediction.")
 
-# Upload DICOM file
+# --- Upload DICOM file ---
 uploaded_file = st.file_uploader("Upload DICOM file", type=["dcm"])
 
-# Function ONLY for preview (does not modify actual file)
+# --- Preview function (visual only) ---
 def load_dicom_preview(file):
     dicom = pydicom.dcmread(file, force=True)
     img = dicom.pixel_array.astype(np.float32)
 
-    # Normalize for display
+    # Scale to 0–255 for display only
     img -= np.min(img)
     if np.max(img) != 0:
         img = img / np.max(img)
+    img = (img * 255).astype(np.uint8)
 
-    # Convert grayscale → RGB for Streamlit display
+    # Convert grayscale → RGB
     if len(img.shape) == 2:
-        img = np.stack([img] * 3, axis=-1)
+        img = np.stack([img]*3, axis=-1)
 
-    img = (img * 255).astype(np.uint8)
     return img
 
-
-# Display preview safely
+# --- Display preview ---
 if uploaded_file is not None:
     try:
-        # Reset pointer BEFORE reading for preview
         uploaded_file.seek(0)
         preview_img = load_dicom_preview(uploaded_file)
-
         st.image(preview_img, caption="DICOM Preview", use_container_width=True)
-
-        # Reset pointer AGAIN before sending to backend
         uploaded_file.seek(0)
-
     except Exception as e:
         st.error(f"Error reading DICOM file: {e}")
 
-
-# Prediction request
+# --- Predict button ---
 if st.button("Predict"):
-
     if uploaded_file is None:
         st.error("Please upload a DICOM file before predicting.")
     else:
-        uploaded_file.seek(0)  # ensure pointer at beginning
-
+        uploaded_file.seek(0)
         files = {"file": uploaded_file}
 
         response = requests.post(
@@ -61,7 +51,9 @@ if st.button("Predict"):
 
         if response.status_code == 200:
             result = response.json()
-            pred = result.get("Prediction", "No prediction")
-            st.success(f"Prediction: {pred}")
+            pred_class = result.get("Predicted_Class")
+            prob = result.get("Probability")
+            st.success(f"Predicted Class: {pred_class}")
+            st.info(f"Probability: {prob:.4f}")
         else:
             st.error(f"API Error: {response.text}")

requirements.txt

@@ -4,4 +4,3 @@ pydicom==2.4.4
 opencv-python-headless==4.10.0.84
 pandas==2.3.1
 numpy==1.26.0
-#tensorflow==2.18.0