import os
import shutil
import datetime
import asyncio
import aiofiles
from collections import defaultdict
import tkinter as tk
from tkinter import filedialog, ttk, messagebox
import threading
import time
class AdvancedFileOrganizer:
def __init__(self, master):
self.master = master
self.master.title("Advanced File Organizer")
self.master.geometry("800x600")
self.search_dir = tk.StringVar()
self.output_dir = tk.StringVar()
self.dry_run = tk.BooleanVar(value=True)
self.log_text = tk.StringVar()
self.create_widgets()
self.extension_mapping = {
# Existing mappings
'.jpg': 'Images', '.jpeg': 'Images', '.png': 'Images', '.gif': 'Images',
'.bmp': 'Images', '.tiff': 'Images', '.eps': 'Images', '.psd': 'Images',
'.webp': 'Images', '.svg': 'Images', '.ico': 'Images', '.heic': 'Images',
'.raw': 'Images', '.psd': 'Images', '.xcf': 'Images',
'.lnk': 'Shortcuts',
'.pdf': 'Documents',
'.txt': 'Documents', '.md': 'Documents', '.rtf': 'Documents',
'.doc': 'Documents', '.docx': 'Documents', '.odt': 'Documents',
'.csv': 'Spreadsheets', '.xlsx': 'Spreadsheets', '.xls': 'Spreadsheets', '.ods': 'Spreadsheets',
'.wav': 'Audio', '.mp3': 'Audio', '.aiff': 'Audio', '.aac': 'Audio',
'.wma': 'Audio', '.flac': 'Audio', '.mid': 'Audio', '.midi': 'Audio',
'.m4a': 'Audio', '.ogg': 'Audio', '.m4p': 'Audio',
'.mp4': 'Videos', '.avi': 'Videos', '.mov': 'Videos', '.wmv': 'Videos',
'.flv': 'Videos', '.mkv': 'Videos', '.webm': 'Videos', '.mpg': 'Videos',
'.mpeg': 'Videos', '.aep': 'Videos', '.prproj': 'Videos',
'.stl': '3D Prints', '.ply': '3D Prints', '.3mf': '3D Prints',
'.gcode': '3D Prints', '.obj': '3D Prints',
'.ttc': 'Fonts', '.otf': 'Fonts', '.eot': 'Fonts', '.fon': 'Fonts',
'.ttf': 'Fonts', '.woff': 'Fonts', '.woff2': 'Fonts',
'.zip': 'Compressed', '.rar': 'Compressed', '.7z': 'Compressed',
'.tar': 'Compressed', '.gz': 'Compressed',
'.exe': 'Executables', '.msi': 'Executables', '.app': 'Executables',
'.py': 'Code', '.java': 'Code', '.cpp': 'Code', '.h': 'Code',
'.js': 'Code', '.html': 'Code', '.css': 'Code', '.php': 'Code',
'.sh': 'Code', '.bat': 'Code',
'.ppt': 'Presentations', '.pptx': 'Presentations', '.key': 'Presentations', '.odp': 'Presentations',
'.epub': 'eBooks', '.mobi': 'eBooks', '.azw': 'eBooks',
'.db': 'Databases', '.sqlite': 'Databases', '.sql': 'Databases',
'.vdi': 'Virtual Machines', '.vmdk': 'Virtual Machines', '.ova': 'Virtual Machines',
'.indd': 'Design', '.qxd': 'Design', '.ai': 'Design',
'.cdr': 'Design', '.dwg': 'Design', '.dxf': 'Design',
'.iso': 'Disk Images', '.dmg': 'Disk Images',
'.json': 'Data Files', '.xml': 'Data Files', '.yaml': 'Data Files',
# New mappings
'.step': 'Design', '.stp': 'Design', '.iges': 'Design', '.igs': 'Design',
'.eps': 'Design', '.svg': 'Design',
'.bak': 'Backups', '.old': 'Backups', '.tmp': 'Backups',
'.ini': 'ConfigFiles', '.cfg': 'ConfigFiles', '.conf': 'ConfigFiles', '.config': 'ConfigFiles',
'.ps1': 'Code', '.vbs': 'Code', '.bash': 'Code',
'.log': 'Logs', '.syslog': 'Logs'
}
def create_widgets(self):
# Search directory selection
ttk.Label(self.master, text="Search Directory:").grid(row=0, column=0, padx=5, pady=5, sticky="w")
ttk.Entry(self.master, textvariable=self.search_dir, width=50).grid(row=0, column=1, padx=5, pady=5)
ttk.Button(self.master, text="Browse", command=self.browse_search_dir).grid(row=0, column=2, padx=5, pady=5)
# Output directory selection
ttk.Label(self.master, text="Output Directory:").grid(row=1, column=0, padx=5, pady=5, sticky="w")
ttk.Entry(self.master, textvariable=self.output_dir, width=50).grid(row=1, column=1, padx=5, pady=5)
ttk.Button(self.master, text="Browse", command=self.browse_output_dir).grid(row=1, column=2, padx=5, pady=5)
# Dry run checkbox
ttk.Checkbutton(self.master, text="Dry Run", variable=self.dry_run).grid(row=2, column=0, padx=5, pady=5, sticky="w")
# Run button
ttk.Button(self.master, text="Run", command=self.run_organizer).grid(row=2, column=1, padx=5, pady=5)
# Log display
self.log_display = tk.Text(self.master, wrap=tk.WORD, width=80, height=20)
self.log_display.grid(row=3, column=0, columnspan=3, padx=5, pady=5)
# Scrollbar for log display
scrollbar = ttk.Scrollbar(self.master, orient="vertical", command=self.log_display.yview)
scrollbar.grid(row=3, column=3, sticky="ns")
self.log_display.configure(yscrollcommand=scrollbar.set)
# Commit changes button (initially disabled)
self.commit_button = ttk.Button(self.master, text="Commit Changes", command=self.run_commit_changes, state="disabled")
self.commit_button.grid(row=4, column=1, padx=5, pady=5)
# Progress bar
self.progress_var = tk.DoubleVar()
self.progress_bar = ttk.Progressbar(self.master, variable=self.progress_var, maximum=100)
self.progress_bar.grid(row=5, column=0, columnspan=3, padx=5, pady=5, sticky="ew")
def browse_search_dir(self):
directory = filedialog.askdirectory()
self.search_dir.set(directory)
def browse_output_dir(self):
directory = filedialog.askdirectory()
self.output_dir.set(directory)
def run_organizer(self):
self.log_display.delete(1.0, tk.END)
self.commit_button.config(state="disabled")
if not self.search_dir.get() or not self.output_dir.get():
self.log("Please select both search and output directories.")
return
# Run the organizer in a separate thread to keep the UI responsive
threading.Thread(target=self._run_organizer_thread, daemon=True).start()
def _run_organizer_thread(self):
search_directory = self.search_dir.get()
output_directory = self.output_dir.get()
is_dry_run = self.dry_run.get()
date_code = datetime.datetime.now().strftime("%m%d%Y")
self.log(f"Searching directory: {search_directory}")
self.log(f"Output directory: {output_directory}")
self.log(f"Dry run: {'Yes' if is_dry_run else 'No'}")
categorized_files = defaultdict(list)
# Scan the directory and categorize files
total_files = 0
for root, _, files in os.walk(search_directory):
for file in files:
total_files += 1
file_path = os.path.join(root, file)
_, ext = os.path.splitext(file)
category = self.extension_mapping.get(ext.lower(), 'Misc')
categorized_files[category].append(file_path)
# Process files
self.changes = []
processed_files = 0
for category, files in categorized_files.items():
if files:
target_folder = os.path.join(output_directory, f'{category}_{date_code}')
self.log(f"Category: {category}")
for file_path in files:
target_path = os.path.join(target_folder, os.path.basename(file_path))
self.changes.append((file_path, target_path))
# Extract the filename from the file path
filename = os.path.basename(file_path)
# Extract the target folder name from the target path
target_folder_name = os.path.basename(target_folder)
# Format the log message with aligned filenames and target folders
self.log(f" {filename:<30} → {target_folder_name}")
processed_files += 1
self.update_progress(processed_files, total_files)
if not is_dry_run:
self.run_commit_changes()
else:
self.log("\nDry run completed. Review the log and click 'Commit Changes' to apply.")
self.master.after(0, lambda: self.commit_button.config(state="normal"))
def update_progress(self, current, total):
progress = (current / total) * 100
self.progress_var.set(progress)
self.master.update_idletasks()
def run_commit_changes(self):
threading.Thread(target=self._commit_changes_thread, daemon=True).start()
def _commit_changes_thread(self):
asyncio.run(self.commit_changes())
async def commit_changes(self):
if not self.changes:
self.log("No changes to commit.")
return
total_files = len(self.changes)
self.progress_var.set(0)
self.progress_bar["maximum"] = total_files
async def process_batch(batch):
tasks = [self.async_move_file(source, target) for source, target in batch]
results = await asyncio.gather(*tasks)
for result in results:
self.log(result)
self.progress_var.set(self.progress_var.get() + 1)
self.master.update_idletasks()
# Process files in batches of 10
batch_size = 10
for i in range(0, len(self.changes), batch_size):
batch = self.changes[i:i+batch_size]
await process_batch(batch)
self.log("All changes committed.")
self.changes = []
self.commit_button.config(state="disabled")
async def async_move_file(self, source, target):
try:
# Create target directory if it doesn't exist
os.makedirs(os.path.dirname(target), exist_ok=True)
# Check if the file already exists in the target location
if os.path.exists(target):
base, extension = os.path.splitext(target)
counter = 1
while os.path.exists(f"{base}_{counter}{extension}"):
counter += 1
target = f"{base}_{counter}{extension}"
# Check file size
file_size = os.path.getsize(source)
if file_size > 100 * 1024 * 1024: # 100 MB
# For large files, use shutil.move with a callback for progress
await asyncio.to_thread(self.move_large_file, source, target, file_size)
else:
# For smaller files, use aiofiles
async with aiofiles.open(source, 'rb') as src, aiofiles.open(target, 'wb') as dst:
await dst.write(await src.read())
await asyncio.to_thread(os.remove, source)
return f"Moved: '{source}' to '{target}'"
except PermissionError:
return f"Permission denied: Unable to move '{source}' to '{target}'"
except FileNotFoundError:
return f"File not found: '{source}'"
except OSError as e:
if "on a different device" in str(e):
# File is on a different device (e.g., network drive), use shutil.copy2 and then remove original
await asyncio.to_thread(shutil.copy2, source, target)
await asyncio.to_thread(os.remove, source)
return f"Copied and removed: '{source}' to '{target}' (different device)"
else:
return f"Error moving '{source}' to '{target}': {str(e)}"
except Exception as e:
return f"Unexpected error moving '{source}' to '{target}': {str(e)}"
def move_large_file(self, source, target, total_size):
with open(source, 'rb') as src, open(target, 'wb') as dst:
copied = 0
while True:
buf = src.read(8388608) # 8MB chunks
if not buf:
break
dst.write(buf)
copied += len(buf)
progress = (copied / total_size) * 100
self.master.after(0, lambda: self.progress_var.set(progress))
os.remove(source)
def log(self, message):
self.log_display.insert(tk.END, message + "\n")
self.log_display.see(tk.END)
if __name__ == "__main__":
root = tk.Tk()
app = AdvancedFileOrganizer(root)
root.mainloop()