QueryList-Rule-Baidu/Baidu.php

<?php
/**
 * Created by PhpStorm.
 * User: Jaeger <JaegerCode@gmail.com>
 * Date: 2017/10/1
 * Baidu searcher
 */

namespace QL\Ext;

use QL\Contracts\PluginContract;
use QL\QueryList;

class Baidu implements PluginContract
{
    private $is_pc;
    protected $ql;
    protected $keyword;
    protected $pageNumber = 10;
    protected $httpOpt = [
        'PC'=>[ 'headers' => [
            'User-Agent' => 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_3) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/80.0.3987.149 Safari/537.36',
            'Accept-Encoding' => 'gzip, deflate, br',
        ]],
        'M'=>[ 'headers' => [
            'User-Agent' => 'Mozilla/5.0 (iPhone; CPU iPhone OS 13_2_3 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/13.0.3 Mobile/15E148 Safari/604.1',
            'Accept-Encoding' => 'gzip, deflate, br',
        ]],
    ];
    const API = [
        'PC' => 'https://www.baidu.com/s',
        'M' => 'https://m.baidu.com/s',
    ];
    const RULES = [
        'PC' => [
            'title' => ['h3', 'text'],
            'link' => ['h3>a', 'href']
        ],
        'M' => [
            'title' => ['h3', 'text'],
            'link' => ['', 'data-log']
        ],
    ];
    const RANGE = [
        'PC' => '.result',
        'M' => '.result'
    ];
    /**
     * @var string
     */
    private $api;

    public function __construct(QueryList $ql, $pageNumber, $api)
    {
        $this->ql = $ql->rules(self::RULES[$api])->range(self::RANGE[$api]);
        $this->pageNumber = $pageNumber;
        $this->api = self::API[$api];
        $this->is_pc = $api == "PC";
        $this->httpOpt=$this->httpOpt[$api];

    }

    public static function install(QueryList $queryList, ...$opt)
    {
        $name = $opt[0] ?? 'baidu';
        $api = $opt[1] ?? 'PC';
        $queryList->bind($name, function ($pageNumber = 10) use ($api) {
            return new Baidu($this, $pageNumber, $api);
        });
    }

    public function setHttpOpt(array $httpOpt = [])
    {
        $this->httpOpt = $httpOpt;
        return $this;
    }

    public function search($keyword)
    {
        $this->keyword = $keyword;
        return $this;
    }

    public function page($page = 1, $realURL = false)
    {
        return $this->query($page)->query()->getData(function ($item) use ($realURL) {
            if (!$this->is_pc) {
                $item['link'] = json_decode($item['link'], true)['mu'];
            } else {
                $realURL && $item['link'] = $this->getRealURL($item['link']);
            }
            return $item;
        });
    }

    public function getCount()
    {
        if (!$this->is_pc) {
            return 1000;
        }
        $count = 0;
        $text = $this->query(1)->find('.nums')->text();
        if (preg_match('/[\d,]+/', $text, $arr)) {
            $count = str_replace(',', '', $arr[0]);
        }
        return (int)$count;
    }

    public function getCountPage()
    {
        $count = $this->getCount();
        $countPage = $this->ceil($count / $this->pageNumber);
        return $countPage;
    }

    protected function query($page = 1)
    {
        if ($this->is_pc) {
            $this->ql->get($this->api, [
                'wd' => $this->keyword,
                'rn' => $this->pageNumber,
                'pn' => $this->pageNumber * ($page - 1)
            ], $this->httpOpt);
        } else {
            $this->ql->get($this->api, [
                'word' => $this->keyword,
                'usm' => $this->pageNumber,
                'pn' => $this->pageNumber * ($page - 1)
            ], $this->httpOpt);
        }

        return $this->ql;
    }

    /**
     * 得到百度跳转的真正地址
     * @param $url
     * @return mixed
     */
    protected function getRealURL($url)
    {
        if (empty($url)) return $url;
        $header = get_headers($url, 1);
        if (strpos($header[0], '301') || strpos($header[0], '302')) {
            if (is_array($header['Location'])) {
                //return $header['Location'][count($header['Location'])-1];
                return $header['Location'][0];
            } else {
                return $header['Location'];
            }
        } else {
            return $url;
        }
    }

}