如何获取任意屏幕位置的最前端窗口并聚焦（macOS）

下面分别提供AppleScript、Swift、Python三种语言的实现方式，均无需模拟鼠标点击，直接通过系统API定位并激活目标窗口。

AppleScript 实现 #

利用System Events的hit test功能定位坐标处的UI元素，向上遍历找到所属窗口后激活对应应用及窗口。

-- 替换为目标坐标（原点为屏幕左上角）
set targetX to 500
set targetY to 300

tell application "System Events"
    set hitItems to hit test {targetX, targetY}
    if hitItems is not {} then
        set currentItem to item 1 of hitItems
        -- 向上遍历父元素，直到找到窗口
        repeat while class of currentItem is not window
            set currentItem to parent of currentItem
            if currentItem is missing value then exit repeat
        end repeat
        if class of currentItem is window then
            set targetApp to application process of currentItem
            -- 激活应用并将窗口置为最前端
            set frontmost of targetApp to true
            set frontmost of currentItem to true
        end if
    end if
end tell

注意：需要给运行脚本的应用（如Script Editor）开启「系统设置-隐私与安全性-辅助功能」权限。

Swift 实现 #

基于AppKit和Quartz框架，获取屏幕窗口列表并按层级排序，筛选包含目标坐标的窗口后激活对应应用。

import AppKit

func activateWindow(at point: CGPoint) {
    guard let windowInfos = CGWindowListCopyWindowInfo([.optionOnScreenOnly, .excludeDesktopElements], kCGNullWindowID) as? [[String: Any]] else {
        return
    }
    
    // 按窗口层级从高到低排序（值越大越靠前）
    let sortedWindows = windowInfos.sorted {
        ($0[kCGWindowOrderingKey as String] as? Int ?? 0) > ($1[kCGWindowOrderingKey as String] as? Int ?? 0)
    }
    
    let screen = NSScreen.main!
    // 转换坐标：输入Y为屏幕左上角原点，系统原生Y以屏幕底部为原点
    let convertedPoint = CGPoint(x: point.x, y: screen.frame.height - point.y)
    
    for windowInfo in sortedWindows {
        guard let bounds = windowInfo[kCGWindowBounds as String] as? [String: CGFloat],
              let x = bounds["X"], let y = bounds["Y"],
              let width = bounds["Width"], let height = bounds["Height"] else {
            continue
        }
        let windowFrame = CGRect(x: x, y: y, width: width, height: height)
        
        if windowFrame.contains(convertedPoint) {
            guard let pid = windowInfo[kCGWindowOwnerPID as String] as? pid_t,
                  let app = NSRunningApplication(processIdentifier: pid) else {
                continue
            }
            app.activate(options: .activateIgnoringOtherApps)
            print("Activated window at \(point) (app: \(app.bundleIdentifier ?? "unknown"))")
            return
        }
    }
}

// 替换为目标坐标
activateWindow(at: CGPoint(x: 500, y: 300))

使用方式：将代码保存为.swift文件，通过swiftc filename.swift -o activateWindow编译为可执行文件，运行前需给可执行文件开启辅助功能权限。

Python 实现 #

通过pyobjc桥接macOS原生框架，实现逻辑与Swift一致，适合Python开发者快速编写脚本。

import objc
from Cocoa import NSRunningApplication, NSScreen
from Quartz import (
    CGWindowListCopyWindowInfo,
    kCGWindowListOptionOnScreenOnly,
    kCGWindowListExcludeDesktopElements,
    kCGNullWindowID,
    kCGWindowBounds,
    kCGWindowOwnerPID,
    kCGWindowOrderingKey
)

def activate_window_at(x, y):
    window_list = CGWindowListCopyWindowInfo(
        kCGWindowListOptionOnScreenOnly | kCGWindowListExcludeDesktopElements,
        kCGNullWindowID
    )
    if not window_list:
        return
    
    # 按窗口层级从高到低排序
    sorted_windows = sorted(window_list, key=lambda w: w[kCGWindowOrderingKey], reverse=True)
    
    screen_height = NSScreen.mainScreen().frame().size.height
    converted_y = screen_height - y
    
    for window in sorted_windows:
        bounds = window[kCGWindowBounds]
        window_x = bounds['X']
        window_y = bounds['Y']
        window_width = bounds['Width']
        window_height = bounds['Height']
        
        if (window_x <= x <= window_x + window_width) and (window_y <= converted_y <= window_y + window_height):
            pid = window[kCGWindowOwnerPID]
            app = NSRunningApplication.runningApplicationWithProcessIdentifier_(pid)
            if app:
                app.activateWithOptions_(1)  # NSApplicationActivateIgnoringOtherApps
                print(f"Activated app: {app.bundleIdentifier()}")
                return

# 替换为目标坐标
activate_window_at(500, 300)

注意：先安装依赖pip install pyobjc，并给Python解释器开启辅助功能权限。

内容的提问来源于stack exchange，提问作者poplitea