Usage of curl_multi function set in php-PHP Tutorial-php.cn

Usage of curl_multi function set in php

WBOY

Release： 2016-07-28 08:26:23

Original

1062 people have browsed it

1. Introduction

I’ve been quite busy during this period, and I haven’t written a blog for a long time. Today I will talk about my experience in using the curl_multi_* function set and the issue of http requests.

When our user php initiates an http request. What would we first think of using? That's right, we will create curl to request. What if we need to initiate multiple http requests in one execution. This is simple, make a url request for each URL. Request to play the first one and then request the second one... Is this the end? What else can we say?

Official website link: http://php.net/manual/zh/book.curl.php

2. Disadvantages of multiple simple curl requests

php中curl_multi函数集的用法

Let’s give a chestnut. There are now three http requests. Each request takes 2s. If you follow a simple curl request (Figure 1-(1)). It took 6 seconds. This is intolerable. The more requests there are, the more time it takes.

Is there a way to reduce query time? Can three http requests be executed at the same time (Figure 1-(1))? There are many ways to solve this problem and reduce the time consumption to 2 seconds. Such as: multi-process, thread, event loop, curl_multi_*, etc. The simplest way is to use curl_multi_* functions. In fact, the internal implementation of curl_multi_* uses an event loop.

3. Simple curl_multi_* application

<code><span><?php</span><span>/**
 *
 * curl_multi_*简单运用
 *
 *<span> @author</span>: rudy
 *<span> @date</span>: 2016/07/12
 */</span><span>/**
 * 根据url,postData获取curl请求对象,这个比较简单,可以看官方文档
 */</span><span><span>function</span><span>getCurlObject</span><span>(<span>$url</span>,<span>$postData</span>=array<span>()</span>,<span>$header</span>=array<span>()</span>)</span>{</span><span>$options</span> = <span>array</span>();
    <span>$url</span> = trim(<span>$url</span>);
    <span>$options</span>[CURLOPT_URL] = <span>$url</span>;
    <span>$options</span>[CURLOPT_TIMEOUT] = <span>10</span>;
    <span>$options</span>[CURLOPT_USERAGENT] = <span>'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/44.0.2403.89 Safari/537.36'</span>;
    <span>$options</span>[CURLOPT_RETURNTRANSFER] = <span>true</span>;
    <span>//    $options[CURLOPT_PROXY] = '127.0.0.1:8888';</span><span>foreach</span>(<span>$header</span><span>as</span><span>$key</span>=><span>$value</span>){
        <span>$options</span>[<span>$key</span>] =<span>$value</span>;
    }
    <span>if</span>(!<span>empty</span>(<span>$postData</span>) && is_array(<span>$postData</span>)){
        <span>$options</span>[CURLOPT_POST] = <span>true</span>;
        <span>$options</span>[CURLOPT_POSTFIELDS] = http_build_query(<span>$postData</span>);
    }
    <span>if</span>(stripos(<span>$url</span>,<span>'https'</span>) === <span>0</span>){
        <span>$options</span>[CURLOPT_SSL_VERIFYPEER] = <span>false</span>;
    }
    <span>$ch</span> = curl_init();
    curl_setopt_array(<span>$ch</span>,<span>$options</span>);

    <span>return</span><span>$ch</span>;
}

<span>// 创建三个待请求的url对象</span><span>$chList</span> = <span>array</span>();
<span>$chList</span>[] = getCurlObject(<span>'https://www.baidu.com'</span>);
<span>$chList</span>[] = getCurlObject(<span>'http://www.jd.com'</span>);
<span>$chList</span>[] = getCurlObject(<span>'http://www.jianshu.com/'</span>);

<span>// 创建多请求执行对象</span><span>$downloader</span> = curl_multi_init();

<span>// 将三个待请求对象放入下载器中</span><span>foreach</span> (<span>$chList</span><span>as</span><span>$ch</span>){
    curl_multi_add_handle(<span>$downloader</span>,<span>$ch</span>);
}

<span>// 轮询</span><span>do</span> {
    <span>while</span> ((<span>$execrun</span> = curl_multi_exec(<span>$downloader</span>, <span>$running</span>)) == CURLM_CALL_MULTI_PERFORM) ;
    <span>if</span> (<span>$execrun</span> != CURLM_OK) {
        <span>break</span>;
    }

    <span>// 一旦有一个请求完成，找出来，处理,因为curl底层是select，所以最大受限于1024</span><span>while</span> (<span>$done</span> = curl_multi_info_read(<span>$downloader</span>))
    {
        <span>// 从请求中获取信息、内容、错误</span><span>$info</span> = curl_getinfo(<span>$done</span>[<span>'handle'</span>]);
        <span>$output</span> = curl_multi_getcontent(<span>$done</span>[<span>'handle'</span>]);
        <span>$error</span> = curl_error(<span>$done</span>[<span>'handle'</span>]);

        <span>// 将请求结果保存,我这里是打印出来</span><span>print</span><span>$output</span>;
<span>//        print "一个请求下载完成!\n";</span><span>// 把请求已经完成了得 curl handle 删除</span>
        curl_multi_remove_handle(<span>$downloader</span>, <span>$done</span>[<span>'handle'</span>]);
    }

    <span>// 当没有数据的时候进行堵塞，把 CPU 使用权交出来，避免上面 do 死循环空跑数据导致 CPU 100%</span><span>if</span> (<span>$running</span>) {
        <span>$rel</span> = curl_multi_select(<span>$downloader</span>, <span>1</span>);
        <span>if</span>(<span>$rel</span> == -<span>1</span>){
            usleep(<span>1000</span>);
        }
    }

    <span>if</span>( <span>$running</span> == <span>false</span>){
        <span>break</span>;
    }
} <span>while</span> (<span>true</span>);

<span>// 下载完毕,关闭下载器</span>
curl_multi_close(<span>$downloader</span>);
<span>echo</span><span>"所有请求下载完成!"</span>;</span></code>

Copy after login

In this example, first create three or more url request objects to be requested. Create a downloader through curl_multi_* functions. Write the request to the downloader. Last poll. Waiting for three requests to complete now. Do processing.

4. Complex curl_multi_* usage

Is this how curl_multi_* is used? too yong too simple! In the above example. The requests in the downloader $downloader are added from the beginning. Can we dynamically add requests to the downloader. Dynamically retrieve completed requests from the downloader. Think about it. What's this? Isn't this the core part of the crawler - the dynamic downloader. How to add it dynamically? We can use multi-process to add via IPC. We can add waits through queues through coroutines.

I implemented a crawler framework through coroutine + curl_multi_*.
Tspider: https://github.com/hirudy/Tspider.
A single process can handle requests 2000-5000/min.

').addClass('pre-numbering').hide(); $(this).addClass('has-numbering').parent().append($numbering); for (i = 1; i ').text(i)); }; $numbering.fadeIn(1700); }); });

The above introduces the usage of the curl_multi function set in PHP, including the relevant content. I hope it will be helpful to friends who are interested in PHP tutorials.