jsdbjksdbvf

app.py

@@ -1,24 +1,33 @@
 import streamlit as st
 import os
+import zipfile
 import torch
 import torch.nn as nn
 import torchvision.transforms as transforms
 import torchvision.models as models
 import matplotlib.pyplot as plt
 import seaborn as sns
+import pandas as pd
 from PIL import Image
 from torchvision import datasets
 from torch.utils.data import DataLoader
 from sklearn.metrics import accuracy_score, classification_report, confusion_matrix
 import random
 
-# ✅ Sidebar Navigation
-st.sidebar.title("Navigation")
-page = st.sidebar.radio("Go to", ["Dataset", "Visualizations", "Model Metrics", "Disease Predictor"])
+# ✅ Automatically unzip `train.zip` if `train/` folder is missing
+DATASET_PATH = "train"
+ZIP_FILE = "train.zip"
+
+if not os.path.exists(DATASET_PATH):
+    if os.path.exists(ZIP_FILE):
+        with zipfile.ZipFile(ZIP_FILE, 'r') as zip_ref:
+            zip_ref.extractall(".")  # Extract to current directory
 
-# ✅ Dataset Path
-DATASET_PATH = "train"  # Update if needed
-CLASS_NAMES = os.listdir(DATASET_PATH)  # Get class names from folder structure
+# ✅ Load Class Names from Dataset
+if os.path.exists(DATASET_PATH):
+    CLASS_NAMES = sorted(os.listdir(DATASET_PATH))
+else:
+    CLASS_NAMES = []
 
 # ✅ Load Model
 @st.cache_resource
@@ -31,71 +40,80 @@ def load_model():
 
 model = load_model()
 
+# ✅ Sidebar Navigation
+st.sidebar.title("Navigation")
+page = st.sidebar.radio("Go to", ["Dataset", "Visualizations", "Model Metrics", "Disease Predictor"])
+
 # ✅ Dataset Page – Show Sample Images
 if page == "Dataset":
     st.title("📊 Dataset Preview")
     st.write(f"### Classes: {CLASS_NAMES}")
 
-    # Show sample images from each class
-    cols = st.columns(4)
-    for i, class_name in enumerate(CLASS_NAMES[:4]):  # Show 4 classes
-        class_path = os.path.join(DATASET_PATH, class_name)
-        image_name = random.choice(os.listdir(class_path))
-        image_path = os.path.join(class_path, image_name)
-        image = Image.open(image_path)
-        cols[i].image(image, caption=class_name, use_column_width=True)
+    if CLASS_NAMES:
+        cols = st.columns(4)
+        for i, class_name in enumerate(CLASS_NAMES[:4]):  # Show 4 classes
+            class_path = os.path.join(DATASET_PATH, class_name)
+            if os.path.exists(class_path):
+                image_name = random.choice(os.listdir(class_path))
+                image_path = os.path.join(class_path, image_name)
+                image = Image.open(image_path)
+                cols[i].image(image, caption=class_name, use_column_width=True)
 
 # ✅ Visualizations Page – Show Class Distribution
 elif page == "Visualizations":
     st.title("📈 Dataset Visualizations")
 
-    # Count images per class
-    class_counts = {cls: len(os.listdir(os.path.join(DATASET_PATH, cls))) for cls in CLASS_NAMES}
-    
-    # Pie Chart
-    st.write("### Disease Distribution")
-    fig, ax = plt.subplots()
-    ax.pie(class_counts.values(), labels=class_counts.keys(), autopct='%1.1f%%', colors=plt.cm.viridis.colors)
-    st.pyplot(fig)
-
-    # Bar Chart
-    st.write("### Class Count")
-    fig, ax = plt.subplots()
-    sns.barplot(x=list(class_counts.keys()), y=list(class_counts.values()), palette="viridis")
-    plt.xticks(rotation=45)
-    st.pyplot(fig)
+    if CLASS_NAMES:
+        class_counts = {cls: len(os.listdir(os.path.join(DATASET_PATH, cls))) for cls in CLASS_NAMES}
+
+        # Pie Chart with Proper Colors
+        st.write("### Disease Distribution")
+        fig, ax = plt.subplots()
+        colors = sns.color_palette("husl", len(CLASS_NAMES))  # Generate unique colors
+        ax.pie(class_counts.values(), labels=class_counts.keys(), autopct='%1.1f%%', colors=colors)
+        st.pyplot(fig)
+
+        # Bar Chart
+        st.write("### Class Count")
+        fig, ax = plt.subplots()
+        sns.barplot(x=list(class_counts.keys()), y=list(class_counts.values()), palette="husl")
+        plt.xticks(rotation=45)
+        st.pyplot(fig)
 
 # ✅ Model Metrics Page
 elif page == "Model Metrics":
     st.title("📊 Model Performance")
 
-    # Load True Labels and Predictions
-    y_true = torch.load("y_true.pth")
-    y_pred = torch.load("y_pred.pth")
-
-    # Accuracy
-    accuracy = accuracy_score(y_true, y_pred)
-    st.write(f"### ✅ Accuracy: {accuracy:.2f}")
-
-    # Classification Report
-    st.write("### 📋 Classification Report")
-    report = classification_report(y_true, y_pred, target_names=CLASS_NAMES, output_dict=True)
-    st.write(pd.DataFrame(report).T)
-
-    # Confusion Matrix
-    st.write("### 🔀 Confusion Matrix")
-    cm = confusion_matrix(y_true, y_pred)
-    fig, ax = plt.subplots()
-    sns.heatmap(cm, annot=True, fmt="d", cmap="Blues", xticklabels=CLASS_NAMES, yticklabels=CLASS_NAMES)
-    st.pyplot(fig)
+    try:
+        # Load True Labels and Predictions
+        y_true = torch.load("y_true.pth")
+        y_pred = torch.load("y_pred.pth")
+
+        # Accuracy
+        accuracy = accuracy_score(y_true, y_pred)
+        st.write(f"### ✅ Accuracy: {accuracy:.2f}")
+
+        # Classification Report
+        st.write("### 📋 Classification Report")
+        report = classification_report(y_true, y_pred, target_names=CLASS_NAMES, output_dict=True)
+        st.write(pd.DataFrame(report).T)
+
+        # Confusion Matrix
+        st.write("### 🔀 Confusion Matrix")
+        cm = confusion_matrix(y_true, y_pred)
+        fig, ax = plt.subplots()
+        sns.heatmap(cm, annot=True, fmt="d", cmap="Blues", xticklabels=CLASS_NAMES, yticklabels=CLASS_NAMES)
+        st.pyplot(fig)
+    except:
+        st.error("🚨 Model metrics files (`y_true.pth` and `y_pred.pth`) not found!")
 
 # ✅ Disease Predictor Page
 elif page == "Disease Predictor":
     st.title("🌿 Plant Disease Classifier")
-    
+
     # File Upload
     uploaded_file = st.file_uploader("Upload a plant leaf image", type=["jpg", "png", "jpeg"])
-    
+
     if uploaded_file is not None:
         image = Image.open(uploaded_file)
         st.image(image, caption="Uploaded Image", use_column_width=True)