PhotoSimRemove.py

import os
import shutil
from imagededup.methods import PHash
import tkinter as tk
from tkinter import ttk
from PIL import Image, ImageTk
import time

def get_image_info(img_path):
    try:
        stat = os.stat(img_path)
        img = Image.open(img_path)
        info = {
            '文件名': os.path.basename(img_path),
            '创建时间': time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(stat.st_ctime)),
            '修改时间': time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(stat.st_mtime)),
            '文件大小': f'{stat.st_size / 1024:.1f} KB',
            '像素': f'{img.width}x{img.height}',
            '模式': img.mode,
            '格式': img.format,
        }
        img.close()
        return info
    except Exception as e:
        return {'文件名': os.path.basename(img_path), '错误': str(e)}

def show_images_with_checkboxes(image_paths, progress_text=''):
    root = tk.Tk()
    root.title(f'请选择要保留的图片 {progress_text}')
    selected = [tk.BooleanVar(master=root, value=True) for _ in image_paths]
    photos = []
    panels = []
    labels = []
    min_sizes = []

    # 创建Canvas和滚动条
    canvas = tk.Canvas(root, width=1200, height=500)
    h_scroll = tk.Scrollbar(root, orient='horizontal', command=canvas.xview)
    canvas.configure(xscrollcommand=h_scroll.set)
    canvas.grid(row=0, column=0, sticky='nsew')
    h_scroll.grid(row=1, column=0, sticky='ew')

    # 创建Frame放到Canvas里
    frame = tk.Frame(canvas)
    canvas.create_window((0, 0), window=frame, anchor='nw')

    # 支持缩放
    def zoom(event, idx, factor):
        img = Image.open(image_paths[idx])
        size = panels[idx].image.width(), panels[idx].image.height()
        min_w, min_h = min_sizes[idx]
        new_w = max(int(size[0] * factor), min_w)
        new_h = max(int(size[1] * factor), min_h)
        try:
            resample = Image.Resampling.LANCZOS
        except AttributeError:
            resample = Image.LANCZOS
        img = img.resize((new_w, new_h), resample)
        photo = ImageTk.PhotoImage(img)
        panels[idx].configure(image=photo)
        panels[idx].image = photo
        img.close()
        # 更新canvas scrollregion
        frame.update_idletasks()
        canvas.config(scrollregion=canvas.bbox("all"))

    for i, img_path in enumerate(image_paths):
        try:
            img = Image.open(img_path)
            img.thumbnail((300, 300))
            min_sizes.append(img.size)
            photo = ImageTk.PhotoImage(img)
            photos.append(photo)
            panel = tk.Label(frame, image=photo)
            panel.image = photo
            panel.grid(row=0, column=i, padx=10, pady=5)
            panel.bind('<MouseWheel>', lambda e, idx=i: zoom(e, idx, 1.2 if e.delta > 0 else 0.8))
            panels.append(panel)
            info = get_image_info(img_path)
            info_text = '\n'.join([f'{k}: {v}' for k, v in info.items()])
            label = tk.Label(frame, text=info_text, justify='left', font=('Consolas', 9))
            label.grid(row=1, column=i)
            labels.append(label)
            chk = tk.Checkbutton(frame, text='保留', variable=selected[i])
            chk.grid(row=2, column=i)
            img.close()
        except Exception as e:
            print(f'无法打开图片: {img_path}, 错误: {e}')

    # 确定按钮放在主窗口下方
    btn = ttk.Button(root, text='确定', command=root.quit)
    btn.grid(row=2, column=0, pady=10)

    # 更新scrollregion
    frame.update_idletasks()
    canvas.config(scrollregion=canvas.bbox("all"))

    root.mainloop()
    root.destroy()
    return [i for i, var in enumerate(selected) if var.get()]

def get_all_image_paths(root_dir, exts={'.jpg', '.jpeg', '.png', '.bmp', '.gif'}):
    image_paths = []
    for dirpath, _, filenames in os.walk(root_dir):
        for fname in filenames:
            if os.path.splitext(fname)[1].lower() in exts:
                image_paths.append(os.path.join(dirpath, fname))
    return image_paths

def build_groups(duplicates, root_dir):
    # 构建无向图
    from collections import defaultdict, deque
    graph = defaultdict(set)
    for key, dupes in duplicates.items():
        for d in dupes:
            graph[key].add(d)
            graph[d].add(key)
    visited = set()
    groups = []
    for node in graph:
        if node not in visited:
            group = []
            queue = deque([node])
            visited.add(node)
            while queue:
                curr = queue.popleft()
                group.append(os.path.join(root_dir, curr))
                for neighbor in graph[curr]:
                    if neighbor not in visited:
                        visited.add(neighbor)
                        queue.append(neighbor)
            if len(group) > 1:
                groups.append(group)
    return groups

if __name__ == '__main__':
    root_dir = r'K:\photo'
    to_delete_dir = os.path.join(root_dir, '待删除')
    os.makedirs(to_delete_dir, exist_ok=True)

    phasher = PHash()
    encodings = phasher.encode_images(image_dir=root_dir)
    print(f'编码图片数量: {len(encodings)}')
    duplicates = phasher.find_duplicates(encoding_map=encodings)

    groups = build_groups(duplicates, root_dir)
    total = len(groups)
    for idx, group in enumerate(groups, 1):
        progress_text = f'({idx}/{total})'
        keep_indices = show_images_with_checkboxes(group, progress_text)
        for i, img_path in enumerate(group):
            if i not in keep_indices:
                try:
                    shutil.move(img_path, os.path.join(to_delete_dir, os.path.basename(img_path)))
                    print(f'已移动: {img_path}')
                except Exception as e:
                    print(f'移动失败: {img_path}, 错误: {e}')