Update src/streamlit_app.py

src/streamlit_app.py

@@ -7,13 +7,19 @@ forums](https://discuss.streamlit.io).
 
 In the meantime, below is an example of what you can do with just a few lines of code:
 """
+
 import os
 
-# Fix for Hugging Face: use safe temp path and disable file watcher
-os.environ["STREAMLIT_HOME"] = "/tmp/.streamlit"
-os.environ["STREAMLIT_WATCH_DISABLE"] = "true"
+# Redirect Streamlit and Matplotlib config to temporary, writable directories
+os.environ["STREAMLIT_CONFIG_DIR"] = "/tmp/.streamlit"
+os.environ["MPLCONFIGDIR"] = "/tmp"
 
 import streamlit as st
+
+# Disable Streamlit usage stats and file watcher to prevent config writes
+st._config.set_option("browser.gatherUsageStats", False)
+st._config.set_option("server.fileWatcherType", "none")
+
 import torch
 import torch.nn as nn
 import torch.nn.functional as F
@@ -91,6 +97,7 @@ model.eval()
 
 # Label dictionary
 label_dict = {0: "Meningioma", 1: "Glioma", 2: "No Tumor", 3: "Pituitary"}
+
 # Preprocessing
 def preprocess_image(image):
     transform = transforms.Compose([
@@ -99,6 +106,7 @@ def preprocess_image(image):
         transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
     ])
     return transform(image).unsqueeze(0).to(device)
+
 # Grad-CAM
 def visualize_grad_cam(image, model, target_layer, label):
     img_np = np.array(image) / 255.0
@@ -112,6 +120,7 @@ def visualize_grad_cam(image, model, target_layer, label):
     grayscale_cam_resized = cv2.resize(grayscale_cam, (224, 224))
     visualization = show_cam_on_image(img_np, grayscale_cam_resized, use_rgb=True)
     return visualization
+
 # LIME
 def model_predict(images):
     preprocessed_images = [preprocess_image(Img.fromarray(img)) for img in images]
@@ -120,6 +129,7 @@ def model_predict(images):
         logits = model(images_tensor)
         probabilities = F.softmax(logits, dim=1)
     return probabilities.cpu().numpy()
+
 def visualize_lime(image):
     explainer = LimeImageExplainer()
     original_image = np.array(image)
@@ -127,6 +137,7 @@ def visualize_lime(image):
     top_label = explanation.top_labels[0]
     temp, mask = explanation.get_image_and_mask(label=top_label, positive_only=True, num_features=10, hide_rest=False)
     return mark_boundaries(temp / 255.0, mask)
+
 # SHAP
 def visualize_shap(image):
     img_tensor = preprocess_image(image).to(device)
@@ -147,6 +158,7 @@ def visualize_shap(image):
     ax.axis('off')
     plt.tight_layout()
     return fig
+
 # Streamlit UI
 st.title("Brain Tumor Classification with Grad-CAM, LIME, and SHAP")
 uploaded_file = st.file_uploader("Upload an MRI Image", type=["jpg", "png", "jpeg"])