app.py

import streamlit as st
import os
from pathlib import Path
from PIL import Image
import numpy as np
from sklearn.ensemble import RandomForestClassifier
from sklearn.preprocessing import LabelEncoder


DATASET_PATH = "PlantVillage"

#  Loading image data from dataset
@st.cache_resource
def load_data(data_dir, img_size=(64, 64)):
    X, y = [], []
    for class_name in os.listdir(data_dir):
        class_path = os.path.join(data_dir, class_name)
        if os.path.isdir(class_path):
            for file in os.listdir(class_path):
                if file.lower().endswith((".jpg", ".jpeg", ".png")):
                    img_path = os.path.join(class_path, file)
                    try:
                        img = Image.open(img_path).convert("RGB").resize(img_size)
                        X.append(np.array(img).flatten())
                        y.append(class_name)
                    except Exception as e:
                        print(f"Error loading image {img_path}: {e}")
    return np.array(X) / 255.0, np.array(y)

# ✅ Training model using RandomForest
@st.cache_resource
def train_model():
    X, y = load_data(DATASET_PATH)
    le = LabelEncoder()
    y_encoded = le.fit_transform(y)
    model = RandomForestClassifier()
    model.fit(X, y_encoded)
    return model, le

# ✅ Getting all images from dataset folders
def get_all_images(data_dir):
    image_paths = []
    valid_extensions = (".jpg", ".jpeg", ".png")
    for class_name in os.listdir(data_dir):
        class_path = Path(data_dir) / class_name
        if class_path.is_dir():
            for file in class_path.iterdir():
                if file.suffix.lower() in valid_extensions:
                    image_paths.append((file, class_name))
    return image_paths

# ✅ Streamlit UI
st.set_page_config(page_title="🥔 Potato Leaf Viewer", layout="centered")
st.title("🥔 Potato Leaf Image Viewer + Classifier")

all_images = get_all_images(DATASET_PATH)

if all_images:
    st.success(f"✅ Found {len(all_images)} images in dataset.")

    # Show filenames for selection
    image_display_names = [f"{cls} - {img.name}" for img, cls in all_images]
    selected_display_name = st.selectbox("Select an image:", image_display_names)

    # Find selected image
    selected_index = image_display_names.index(selected_display_name)
    selected_image_path, actual_class = all_images[selected_index]

    # Load and show image (small size)
    image = Image.open(selected_image_path).convert("RGB").resize((64, 64))

    # Show image in smaller size
    st.image(image, caption=f"📂 Original Class: {actual_class}", width=200)

    # Predict
    model, le = train_model()
    img_array = np.array(image).flatten().reshape(1, -1) / 255.0
    prediction = model.predict(img_array)
    predicted_class = le.inverse_transform(prediction)[0]

    st.markdown(f"### 🧠 Predicted Class: **{predicted_class}**")

else:
    st.warning("⚠️ No images found in the dataset. Please check the folder structure.")