Python用win32api模拟Shift+左键点击的问题及优化咨询

• 应用/系统的输入过滤机制：很多GUI应用会对鼠标点击的按下（LEFTDOWN）和抬起（LEFTUP）事件间隔设置最小阈值，间隔过短会被判定为无效输入（比如误触）。0.055秒（55毫秒）刚好是目标应用或系统的临界值，低于这个值就会被过滤。
• mouse_event API的局限性：这是Windows的旧版输入模拟API，系统对它的优先级低于真实用户输入，且它不会同步等待系统处理事件，短间隔的事件容易被系统输入队列忽略。
• 光标移动的同步问题：SetCursorPos调用后，系统需要极短时间完成光标位置更新，紧接着发送点击事件可能导致点击位置未生效，进而被应用忽略。

1. 替换为现代API：SendInput #

Windows官方推荐使用SendInput替代mouse_event和keybd_event，它能更精准地模拟用户输入，事件优先级更高，不容易被忽略。以下是重写后的核心函数：

import win32api
import win32con
import ctypes
from ctypes import wintypes

user32 = ctypes.WinDLL('user32', use_last_error=True)

INPUT_MOUSE = 0
INPUT_KEYBOARD = 1

KEYEVENTF_KEYUP = 0x0002
MOUSEEVENTF_LEFTDOWN = 0x0002
MOUSEEVENTF_LEFTUP = 0x0004

class MOUSEINPUT(ctypes.Structure):
    _fields_ = (("dx", wintypes.LONG),
                ("dy", wintypes.LONG),
                ("mouseData", wintypes.DWORD),
                ("dwFlags", wintypes.DWORD),
                ("time", wintypes.DWORD),
                ("dwExtraInfo", wintypes.ULONG_PTR))

class KEYBDINPUT(ctypes.Structure):
    _fields_ = (("wVk", wintypes.WORD),
                ("wScan", wintypes.WORD),
                ("dwFlags", wintypes.DWORD),
                ("time", wintypes.DWORD),
                ("dwExtraInfo", wintypes.ULONG_PTR))

class INPUT(ctypes.Structure):
    class _INPUT(ctypes.Union):
        _fields_ = (("mi", MOUSEINPUT),
                    ("ki", KEYBDINPUT))
    _anonymous_ = ("_input",)
    _fields_ = (("type", wintypes.DWORD),
                ("_input", _INPUT))

def send_input(inputs):
    n = len(inputs)
    LPINPUT = INPUT * n
    pInputs = LPINPUT(*inputs)
    cbSize = ctypes.c_int(ctypes.sizeof(INPUT))
    return user32.SendInput(n, pInputs, cbSize)

def set_cursor_pos(x, y):
    user32.SetCursorPos(x, y)

def press_shift():
    ki = KEYBDINPUT(wVk=win32con.VK_LSHIFT, wScan=0, dwFlags=0, time=0, dwExtraInfo=0)
    send_input([INPUT(type=INPUT_KEYBOARD, ki=ki)])

def release_shift():
    ki = KEYBDINPUT(wVk=win32con.VK_LSHIFT, wScan=0, dwFlags=KEYEVENTF_KEYUP, time=0, dwExtraInfo=0)
    send_input([INPUT(type=INPUT_KEYBOARD, ki=ki)])

def mouse_click(x, y):
    set_cursor_pos(x, y)
    # 模拟左键按下+抬起
    mi_down = MOUSEINPUT(dx=0, dy=0, mouseData=0, dwFlags=MOUSEEVENTF_LEFTDOWN, time=0, dwExtraInfo=0)
    mi_up = MOUSEINPUT(dx=0, dy=0, mouseData=0, dwFlags=MOUSEEVENTF_LEFTUP, time=0, dwExtraInfo=0)
    send_input([INPUT(type=INPUT_MOUSE, mi=mi_down), INPUT(type=INPUT_MOUSE, mi=mi_up)])

使用SendInput后，可大幅缩短甚至移除部分sleep，因为它的事件处理更可靠。

2. 优化延迟逻辑 #

• 用Windows原生的win32api.Sleep(ms)替代time.sleep()，前者是毫秒级精准延迟，不受系统调度误差影响。
• 仅在必要处添加最小延迟，比如光标移动后加1ms延迟确保位置生效：

def mouse_click(x, y):
    set_cursor_pos(x, y)
    win32api.Sleep(1)  # 确保光标到位
    mi_down = MOUSEINPUT(dx=0, dy=0, mouseData=0, dwFlags=MOUSEEVENTF_LEFTDOWN, time=0, dwExtraInfo=0)
    mi_up = MOUSEINPUT(dx=0, dy=0, mouseData=0, dwFlags=MOUSEEVENTF_LEFTUP, time=0, dwExtraInfo=0)
    send_input([INPUT(type=INPUT_MOUSE, mi=mi_down), INPUT(type=INPUT_MOUSE, mi=mi_up)])

3. 调整Shift与点击的时序 #

确保Shift按下事件完全被系统处理后再触发点击，可在press_shift()后加1ms延迟避免事件队列混乱：

def clicks():
   for offs in coord:
        x, y = offs
        x += random.randint(-15, 15)
        y += random.randint(-15, 15)
        press_shift()
        win32api.Sleep(1)
        mouse_click(x, y)
        win32api.Sleep(1)
        release_shift()
        print('click on', offs)

• PyAutoGUI：最适合新手的跨平台库，封装了所有输入模拟逻辑，自动处理延迟和同步，代码极简：

  import pyautogui
  def shift_click(x, y):
      pyautogui.moveTo(x, y)
      pyautogui.keyDown('shift')
      pyautogui.click()
      pyautogui.keyUp('shift')
  

• pynput：更灵活的输入控制库，支持监听和模拟输入，适合精细控制场景：

  from pynput.mouse import Controller, Button
  from pynput.keyboard import Controller as KeyboardController, Key
  mouse = Controller()
  keyboard = KeyboardController()
  
  def shift_click(x, y):
      mouse.position = (x, y)
      with keyboard.pressed(Key.shift):
          mouse.click(Button.left, 1)
  

• ctypes：直接调用Windows API，适合进阶用户，完全自定义输入逻辑，无需依赖第三方库。

内容的提问来源于stack exchange，提问作者R3w6