php pthreads v3多线程抓取新浪新闻 php使用pthreads v3多线程实现抓取新浪新闻信息操作示例

CREATE TABLE `tb_sina` (
 `id` int(11) unsigned NOT NULL AUTO_INCREMENT COMMENT 'ID',
 `url` varchar(256) DEFAULT '' COMMENT 'url地址',
 `title` varchar(128) DEFAULT '' COMMENT '标题',
 `time` datetime DEFAULT NULL ON UPDATE CURRENT_TIMESTAMP COMMENT '时间',
 PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COMMENT='sina新闻';

<?php

class DB extends Worker
{
  private static $db;
  private $dsn;
  private $root;
  private $pwd;

  public function __construct($dsn, $root, $pwd)
  {
    $this->dsn = $dsn;
    $this->root = $root;
    $this->pwd = $pwd;
  }

  public function run()
  {
    //创建连接对象
    self::$db = new PDO($this->dsn, $this->root, $this->pwd);

    //把require放到worker线程中，不要放到主线程中，不然会报错找不到类
    require './vendor/autoload.php';
  }

  //返回一个连接资源
  public function getConn()
  {
    return self::$db;
  }
}

class Sina extends Thread
{
  private $name;
  private $url;

  public function __construct($name, $url)
  {
    $this->name = $name;
    $this->url = $url;
  }

  public function run()
  {
    $db = $this->worker->getConn();

    if (empty($db) || empty($this->url)) {
      return false;
    }

    $content = file_get_contents($this->url);
    if (!empty($content)) {
      //获取标题，地址，时间
      $data = QL\QueryList::Query($content, [
        'tit' => ['.c_tit > a', 'text'],
        'url' => ['.c_tit > a', 'href'],
        'time' => ['.c_time', 'text'],
      ], '', 'UTF-8', 'GB2312')->getData();

      //把获取的数据插入数据库
      if (!empty($data)) {
        $sql = 'INSERT INTO tb_sina(`url`, `title`, `time`) VALUES';
        foreach ($data as $row) {
          //修改下时间，新浪的时间格式是这样的04-23 15:30
          $time = date('Y') . '-' . $row['time'] . ':00';
          $sql .= "('{$row['url']}', '{$row['tit']}', '{$time}'),";
        }
        $sql = rtrim($sql, ',');
        $ret = $db->exec($sql);

        if ($ret !== false) {
          echo "线程{$this->name}成功插入{$ret}条数据\n";
        } else {
          var_dump($db->errorInfo());
        }
      }
    }
  }
}

//抓取页面地址
$url = 'http://roll.news.sina.com.cn/s/channel.php?ch=01#col=89&spec=&type=&ch=01&k=&offset_page=0&offset_num=0&num=60&asc=&page=';
//创建pool池
$pool = new Pool(5, 'DB', ['mysql:dbname=test;host=192.168.33.226', 'root', '']);

//获取100个分页数据
for ($ix = 1; $ix <= 100; $ix++) {
  $pool->submit(new Sina($ix, $url . $ix));
}

//循环收集垃圾，阻塞主线程，等待子线程结束
while ($pool->collect()) ;
$pool->shutdown();

composer require jaeger/querylist