splitdata.py

import os
import shutil
import random

# ✅ Define paths
data_dir = "PlantVillage"
train_dir = "dataset/train"
test_dir = "dataset/test"
split_ratio = 0.8  # 80% train, 20% test

# ✅ Ensure train/test folders exist
os.makedirs(train_dir, exist_ok=True)
os.makedirs(test_dir, exist_ok=True)

# ✅ Split dataset
for category in os.listdir(data_dir):  # Loop through plant disease categories
    category_path = os.path.join(data_dir, category)
    
    if os.path.isdir(category_path):  # Ensure it's a folder
        images = os.listdir(category_path)
        random.shuffle(images)  # Shuffle to ensure randomness
        
        split_index = int(len(images) * split_ratio)
        train_images = images[:split_index]
        test_images = images[split_index:]

        # ✅ Create category folders
        os.makedirs(os.path.join(train_dir, category), exist_ok=True)
        os.makedirs(os.path.join(test_dir, category), exist_ok=True)

        # ✅ Move images
        for img in train_images:
            shutil.move(os.path.join(category_path, img), os.path.join(train_dir, category, img))
        for img in test_images:
            shutil.move(os.path.join(category_path, img), os.path.join(test_dir, category, img))

print("✅ Dataset successfully split into train/test!")