PHP 多进程下载必应壁纸

    public function run()
    {
        // 检查运行环境
        $this->checkEnv();
        // 守护进程
        $this->daemonize();
        // 安装信号处理器
        $this->installSignalHandler();
        // 初始化 Redis
        $this->initRedis();
        // 初始化进程
        $this->initWorkers();
        // 监听子进程状态
        $this->monitor();
    }

    protected function checkEnv()
    {
        if ('//' == \DIRECTORY_SEPARATOR) {
            exit('目前只支持 linux 系统'.PHP_EOL);
        }

        if (!\extension_loaded( 'pcntl') ) {
            exit('缺少 pcntl 扩展'.PHP_EOL);
        }

        if (!\extension_loaded( 'posix') ) {
            exit('缺少 posix 扩展'.PHP_EOL);
        }

        if (version_compare(PHP_VERSION, 7.1, '<')) {
            declare(ticks = 1);
        } else {
            // 启用异步信号处理
            \pcntl_async_signals(true);
        }
    }

    protected function daemonize()
    {
        if (self::$options['daemonize'] !== true) {
            return;
        }

        // 设置当前进程创建的文件权限为 777
        umask(0);

        $pid = \pcntl_fork();
        if ($pid < 0) {
            $this->log('创建守护进程失败');
            exit;
        } else if ($pid > 0) {
            // 主进程退出
            exit(0);
        }

        // 将当前进程作为会话首进程
        if (\posix_setsid() < 0) {
            $this->log('设置会话首进程失败');
            exit;
        }

        // 两次 fork 保证形成的 daemon 进程绝对不会成为会话首进程
        $pid = \pcntl_fork();
        if ($pid < 0) {
            $this->log('创建守护进程失败');
            exit;
        } else if ($pid > 0) {
            // 主进程退出
            exit(0);
        }
    }

    protected function initRedis()
    {
        $this->redisClient = new Client(self::$options['redis']);
    }

    protected function installSignalHandler()
    {
        // 捕获 SIGINT 信号，终端中断
        \pcntl_signal(SIGINT, [$this, 'stopAllWorkers'], false);

        // 捕获 SIGPIPE 信号，忽略掉所有管道事件
        \pcntl_signal(\SIGPIPE, \SIG_IGN, false);
    }

    protected function stopAllWorkers()
    {
        if (self::$maserPid !== \posix_getpid()) {
            // 子进程
            unset(self::$workers[$this->workerId]);
            exit(0);
        }

        // 父进程
        foreach (self::$workers as $pid) {
            // 给 worker 进程发送关闭信号
            \posix_kill($pid, SIGINT);
        }
    }

    protected function initWorkers()
    {
        self::$maserPid = \posix_getpid();

        $this->forkWorker(1, [$this, 'extractWallpaperUrl']);

        $workerNum = (int) self::$options['worker_num'];

        for ($i = 0; $i < $workerNum; $i++) {
            $this->forkWorker($i + 2, [$this, 'downloadWallpaper']);
        }
    }

    protected function forkWorker($workerId, $callback)
    {
        $pid = \pcntl_fork();

        if ($pid > 0) {
            // 父进程记录子进程 PID
            self::$workers[$workerId] = $pid;
        } elseif ($pid === 0) {
            // 子进程处理业务逻辑
            $this->workerId = $workerId;

            if ($callback instanceof \Closure) {
                $callback();
            } else if (isset($callback[1]) && is_object($callback[0])) {
                \call_user_func($callback);
            }

            exit(0);
        } else {
            $this->log('进程创建失败');
            exit;
        }
    }

    protected function extractWallpaperUrl()
    {
        $this->log('提取壁纸地址进程启动...');

        $page = 1;

        do {
            $html = \file_get_contents("http://bing.ioliu.cn/?p={$page}");
            \preg_match_all('/<img([^>]*)\ssrc="([^\s>]+)"/', $html,$matches);

            if (empty($matches[2]) || \count($matches[2]) === 3) {
                $this->log('壁纸地址提取完毕, 当前页码: %s', $page);
                break;
            }

            $urls = \array_unique(\array_filter($matches[2]));

            if (!empty($urls)) {
                // 将壁纸 url 放入队列中
                $this->redisClient->sadd(self::$options['queue_key'], $urls);
            }

            $this->log('提取壁纸数量: %s, 当前页面: %s', count($urls), $page++);
        } while (true);
    }

    protected function downloadWallpaper()
    {
        $this->log('下载壁纸进程启动...');

        while (self::$freeTime < self::$options['max_free_time']) {
            $url = $this->redisClient->spop(self::$options['queue_key']);

            if (empty($url)) {
                $this->log('空闲时间: %s/%ss', self::$freeTime++, self::$options['max_free_time']);
                \sleep(1);
                continue;
            }

            try {
                $result = $this->saveWallpaper($url);
                if (!$result) {
                    $this->redisClient->sadd(self::$options['queue_key'], [$url]);
                }
            } catch (\Exception $e) {
                $result = false;
                $this->log('保存壁纸异常: %s', $e->getMessage());
            }

            $this->log('壁纸下载%s, %s', $result ? '成功' : '失败', $url);
        }
    }

    protected function monitor()
    {
        while (true) {
            $pid = $this->acceptSignal();

            if ($pid > 0) {
                $this->log('子进程退出信号, PID: %s', $pid);
                // 翻转 workers 的键值
                $workers = \array_flip(self::$workers);
                $workerId = $workers[$pid];
                // 删除子进程
                unset(self::$workers[$workerId]);
                // 如果没有在运行的子进程则退出主进程
                count(self::$workers) === 0 && exit(0);
            } else {
                $this->log('其它信号, PID: %s', $pid);
                exit(0);
            }
        }
    }

    protected function acceptSignal()
    {
        if (\version_compare(PHP_VERSION, 7.1, '>=')) {
            return \pcntl_wait($status, WUNTRACED);
        }

        // 调用等待信号的处理器
        \pcntl_signal_dispatch();
        $pid = \pcntl_wait($status, WUNTRACED);
        \pcntl_signal_dispatch();

        return $pid;
    }

$options = [
    'daemonize'     => false,   // 是否 daemon 化
    'worker_num'    => 3,       // 下载壁纸进程数量
    'max_free_time' => 60,      // 最大空闲时间（秒）
    'save_dir'      => __DIR__.'/wallpaper',    // 壁纸保存位置
    'queue_key'     => 'wallpaper_url_queue',   // 壁纸下载地址的 redis key
    'redis' => [                                // redis 配置
        'scheme' => 'tcp',
        'host'   => '127.0.0.1',
        'port'   => 6379,
    ],
];

$wallpaper = new BingWallpaperDownloader($options);

$wallpaper->run();

vagrant@homestead:~/code/her-cat/download_bing_wallpaper$ php index.php
[2020-02-02 10:41:34] [worker-1] 提取壁纸地址进程启动...
[2020-02-02 10:41:34] [worker-3] 下载壁纸进程启动...
[2020-02-02 10:41:34] [worker-2] 下载壁纸进程启动...
[2020-02-02 10:41:34] [worker-4] 下载壁纸进程启动...
[2020-02-02 10:41:35] [worker-1] 提取壁纸数量: 12, 当前页面: 1
[2020-02-02 10:41:35] [worker-2] 壁纸下载成功, http://h1.ioliu.cn/bing/NutcrackerSeason_EN-AU8373379424_1920x1080.jpg
[2020-02-02 10:41:35] [worker-3] 壁纸下载成功, http://h1.ioliu.cn/bing/zhenghe_ZH-CN9628081460_1920x1080.jpg
[2020-02-02 10:41:36] [worker-4] 壁纸下载成功, http://h1.ioliu.cn/bing/MonumentFountain_EN-AU10536043652_1920x1080.jpg
[2020-02-02 10:41:37] [worker-2] 壁纸下载成功, http://h1.ioliu.cn/bing/JeanLafitte_EN-AU11428973003_1920x1080.jpg
[2020-02-02 10:41:37] [worker-1] 提取壁纸数量: 12, 当前页面: 2
[2020-02-02 10:41:37] [worker-3] 壁纸下载成功, http://h1.ioliu.cn/bing/MorondavaBaobab_EN-AU11363642614_1920x1080.jpg
[2020-02-02 10:41:37] [worker-3] 壁纸下载成功, http://h1.ioliu.cn/bing/SnowHare_ZH-CN9767012872_1920x1080.jpg
[2020-02-02 10:41:38] [worker-4] 壁纸下载成功, http://h1.ioliu.cn/bing/ShenandoahAutumn_EN-AU11784755049_1920x1080.jpg
^C[2020-02-02 10:41:38] [worker-0] 其它信号, PID: -1

vagrant@homestead:~/code/her-cat/download_bing_wallpaper/wallpaper$ ls
AbstractSaltBeds_ZH-CN8351691359_1920x1080.jpg         MauiEucalyptus_ZH-CN5616197787_1920x1080.jpg
AcadiaBlueberries_ZH-CN6014510748_1920x1080.jpg        may1_ZH-CN8582006115_1920x1080.jpg
AdelieBreeding_ZH-CN1750945258_1920x1080.jpg           MeerkatHuddle_ZH-CN1358126294_1920x1080.jpg
AdobeSantaFe_EN-AU3063917358_1920x1080.jpg             MeerkatMob_ZH-CN3788674757_1920x1080.jpg
AerialKluaneNP_ZH-CN4080112842_1920x1080.jpg           MeteorCrater_EN-AU9993563603_1920x1080.jpg