app.py

import gradio as gr
import os
import pandas as pd
import plotly.graph_objects as go
import asyncio
from datetime import datetime
import re
import pathlib

# --- Configuration ---
# Update frequency: how many files to process before updating the UI
UPDATE_INTERVAL = 250 

def parse_filename_words(filename):
    """
    Extracts contiguous groups of letters from a filename, ignoring numbers and symbols.
    Example: "Aarons123File-482.md" -> "Aarons, File, md"
    """
    # Find all sequences of letters
    words = re.findall('[A-Za-z]+', filename)
    return ", ".join(words) if words else "N/A"

def get_file_info(path, root_path):
    """
    Gathers required information for a single file.
    Returns a dictionary or None if path is not a file or is inaccessible.
    """
    try:
        if not os.path.isfile(path):
            return None
            
        stat = os.stat(path)
        size = stat.st_size
        
        # Skip empty files
        if size == 0:
            return None

        # Determine the top-level directory for color grouping
        try:
            relative_path = os.path.relpath(path, root_path)
            top_level_dir = relative_path.split(os.sep)[0]
        except ValueError:
            top_level_dir = os.path.basename(root_path)

        # Get parent directory relative to the root for treemap structure
        parent_path = str(pathlib.Path(*pathlib.Path(relative_path).parts[:-1]))
        if parent_path == ".":
            parent_path = top_level_dir

        return {
            'path': path,
            'label': os.path.basename(path),
            'parent': parent_path,
            'size': size,
            'color_group': top_level_dir,
            'created': datetime.fromtimestamp(stat.st_ctime).strftime('%Y-%m-%d %H:%M:%S'),
            'modified': datetime.fromtimestamp(stat.st_mtime).strftime('%Y-%m-%d %H:%M:%S'),
            'keywords': parse_filename_words(os.path.basename(path))
        }
    except (OSError, FileNotFoundError):
        return None

def create_treemap_figure(df):
    """
    Generates the Plotly treemap figure from a DataFrame of file info.
    """
    if df.empty:
        return go.Figure(go.Treemap(
            labels=["Your scan will appear here."],
            parents=[""],
            values=[1]
        ))
    
    # Ensure the root of the treemap is visible
    root_label = os.path.basename(df.iloc[0]['path'])
    
    fig = go.Figure(go.Treemap(
        ids=df['path'],
        labels=df['label'],
        parents=df['parent'],
        values=df['size'],
        marker_colors=df['color_group'], # Color by top-level folder
        tiling_method='squarified', # Use the squarified algorithm
        root_label=root_label,
        customdata=df[['size', 'modified', 'created', 'keywords']],
        hovertemplate=(
            "<b>%{label}</b><br>"
            "Size: %{customdata[0]:.2s}B<br>"
            "Modified: %{customdata[1]}<br>"
            "Created: %{customdata[2]}<br>"
            "Keywords: %{customdata[3]}<br>"
            "Path: %{id}<extra></extra>"
        ),
        pathbar={'visible': True} # Show breadcrumb trail
    ))
    
    fig.update_layout(
        margin=dict(t=50, l=25, r=25, b=25),
        title="File System Treemap"
    )
    return fig

async def scan_directory(directory, stop_flag_state, progress=gr.Progress(track_tqdm=True)):
    """
    Asynchronously scans a directory, yielding updates to the UI.
    """
    if not directory or not os.path.isdir(directory):
        yield create_treemap_figure(pd.DataFrame()), "Invalid directory path.", pd.DataFrame()
        return

    file_list = []
    processed_count = 0
    
    # Reset stop flag at the beginning of a new scan
    stop_flag_state['stop'] = False

    # Create a DataFrame to hold results
    df = pd.DataFrame()

    progress(0, desc="Starting scan...")

    for root, _, files in os.walk(directory, topdown=True):
        if stop_flag_state['stop']:
            progress(1.0, "Scan stopped by user.")
            break

        for name in files:
            file_path = os.path.join(root, name)
            info = get_file_info(file_path, directory)
            if info:
                file_list.append(info)
                processed_count += 1
            
            # Yield updates periodically to keep the UI responsive
            if processed_count % UPDATE_INTERVAL == 0:
                df = pd.DataFrame(file_list)
                yield create_treemap_figure(df), f"Scanned {processed_count} files...", df
                await asyncio.sleep(0.01) # Allow other tasks to run

    # Final update after loop finishes or is stopped
    df = pd.DataFrame(file_list)
    final_status = f"Scan complete. Found {len(df)} files."
    if stop_flag_state['stop']:
        final_status = f"Scan stopped. Displaying {len(df)} found files."
    
    yield create_treemap_figure(df), final_status, df

def stop_scan(stop_flag_state):
    """Sets the stop flag to True."""
    stop_flag_state['stop'] = True
    return stop_flag_state, "Stopping scan..."

# --- Gradio UI ---
with gr.Blocks(theme=gr.themes.Soft(), title="File System Treemap") as app:
    
    stop_flag = gr.State({'stop': False})

    gr.Markdown("# 📁 File System Treemap Visualizer")
    gr.Markdown("Enter a directory path to generate a squarified treemap. The visualization will build in real-time.")

    with gr.Row():
        path_input = gr.Textbox(
            label="Directory Path", 
            placeholder="e.g., C:\\Users\\YourUser\\Documents",
            scale=3
        )
        start_button = gr.Button("Start Scan", variant="primary", scale=1)
        stop_button = gr.Button("Stop Scan", variant="stop", scale=1)
    
    status_label = gr.Label("Status: Ready")
    
    with gr.Tabs():
        with gr.TabItem("Treemap Visualization"):
            plot_output = gr.Plot(interactive=True)
        with gr.TabItem("Data Table"):
            data_output = gr.DataFrame(wrap=True)

    # Event Handlers
    start_button.click(
        fn=scan_directory,
        inputs=[path_input, stop_flag],
        outputs=[plot_output, status_label, data_output]
    )
    
    stop_button.click(
        fn=stop_scan,
        inputs=[stop_flag],
        outputs=[stop_flag, status_label],
        cancels=[start_button.click] # This cancels the running 'scan_directory' event
    )

if __name__ == "__main__":
    app.launch()