information-system/application/third_party/htmlcompressor/controllers/index.php

<?php

if (!defined('BASEPATH'))
    exit('No direct script access allowed');

class Index extends CI_Controller {

    public function __construct() {
        parent::__construct();
        $this->load->library('simple_html_dom_lib');
    }

    public function index() {
        $this->load->view('welcome');
    }

    public function optimize() {
        $htmlsource = $this->input->post('htmlsource');
        $websitehost = $this->input->post('websitehost');
        if (empty($htmlsource) || empty($websitehost)) {
            $this->output->set_status_header(500);
            echo 'error:htmlsource or websitehost is empty!';
            return false;
        }
        //域名后面不能有/
        if (substr($websitehost, -1, 1) == '/') {
            $websitehost = substr($websitehost, 0, -1);
        }

        $html_object = str_get_html($htmlsource);
        if (!empty($html_object)) {

//提取和下载所有CSS样式，包括链接文件和页面样式
            $link_css_array = array();
            $css_content = '';
            foreach ($html_object->find('link') as $link_css) {
                if ($link_css->rel == 'stylesheet' && !empty($link_css->href)) {
                    $link_css_array[] = $link_css->href;
                    $link_css->outertext = ''; //删除链接
                }
            }
            //print_r($link_css_array);
            foreach ($link_css_array as $item) {
                $css_content.= GET_HTTP($this->format_url($item, $websitehost));
            }
            foreach ($html_object->find('style') as $style_css) {
                if ($style_css->type == "text/css") {
                    $css_content .= $style_css->innertext;
                }
            }
            // echo $css_content;
            // echo $html_object;die();
            // 
//提取和下载所有JS脚本，包括链接文件和页面脚本
            $link_js_array = array();
            $js_content = '';
            foreach ($html_object->find('script') as $link_script) {
                if (!empty($link_script->src)) {
                    $link_js_array[] = $link_script->src;
                    $link_script->outertext = ''; //删除链接，移动到页底
                } else {
                    //网页内的js不需要提取
                    //$js_content.= $link_script->innertext;//js的内容
                    // $js_content.= $link_script;//js的内容，包含<script>标签

                    //$link_script_content = $link_script;
                    //含有$表示调用了jquery的函数，添加延迟加载defer
                   // if (strpos($link_script_content, '$') !== FALSE) {
                   //     $link_script_content = str_replace('defer', 'defer', $link_script_content);
                   //     $link_script_content = str_replace('<script', '<script defer', $link_script_content);
                   //     $js_content.= $link_script_content;
                   // } else {
                        $js_content.= $link_script; //js的内容，包含<script>标签
                   // }
                    $link_script->outertext = ''; //删除js，移动到页底
                }
            }

            foreach ($link_js_array as $item) {
                //$js_content.= GET_HTTP($this->format_url($item, $websitehost));
            }
            //echo $js_content;
            //把网页内容和css提交到purifycss处理
            $optimize_css = GET_HTTP('http://184.172.113.219:33033/', 'html_source=' . urlencode($htmlsource) . '&html_css=' . urlencode($css_content), 'POST');
            if (empty($optimize_css)) {
                $this->output->set_status_header(500);
                echo 'css精简错误';
                echo '<!-- ' . $css_content . ' -->';
                return FALSE;
            }

            //把精简的css添加到head前面
            $html_object = str_replace('</head>', '<style type="text/css">' . $optimize_css . "</style></head>", $html_object);
            //删除多余空格和换行符
            $html_object = str_replace(array("    ", "\t", "\n", "\r"), " ", $html_object);
            //循环n次，把双空格替换为一个空格
            for ($i = 0; $i <= 4; $i++) {
                $html_object = str_replace("  ", " ", $html_object);
            }

//在最后加载原始css文件和js文件
            //把css移动到页面底部，延迟加载
            $lastload_js = '<noscript id="deferred-styles">';
            foreach ($link_css_array as $item) {
                $lastload_js.='<link rel="stylesheet" type="text/css" href="' . $item . '"/>';
            }
            $lastload_js.='</noscript><script>var loadDeferredStyles=function(){var addStylesNode=document.getElementById("deferred-styles");var replacement=document.createElement("div");replacement.innerHTML=addStylesNode.textContent;document.body.appendChild(replacement);addStylesNode.parentElement.removeChild(addStylesNode)};var raf=requestAnimationFrame||mozRequestAnimationFrame||webkitRequestAnimationFrame||msRequestAnimationFrame;if(raf){raf(function(){window.setTimeout(loadDeferredStyles,0)})}else{window.addEventListener("load",loadDeferredStyles)};</script>';
            //把js移动到页面底部
            foreach ($link_js_array as $item) {
                $lastload_js.='<script  src="' . $item . '"></script>';
            }
            $lastload_js.=$js_content;

            $html_object = str_replace('</body>', $lastload_js . '</body>', $html_object);
        }
        echo $html_object;
    }

//格式化url，保证请求的URL有域名，//更换为对应的域名路径
    function format_url($url, $host = '') {
        if (substr($url, 0, 8) == 'https://' || substr($url, 0, 7) == 'http://') {
            return urldecode($url);
        }

        if (substr($url, 0, 2) == '//') { //https或http
            return urldecode(str_replace('//', 'http://', $url));
        }

        return urldecode($host . $url);
    }

}