// ==UserScript==
// @name         自定义格式化文献
// @namespace    https://bbs.tampermonkey.net.cn/
// @version      0.2.3
// @description  目前支持sciencedirect,rsc和springer, onlinelibrary.wiley.com
// @author       tfsn20
// @match        https://www.sciencedirect.com/*
// @match        https://pubs.rsc.org/*
// @match        https://link.springer.com/*
// @match        https://onlinelibrary.wiley.com/*
// @match        https://webvpn.zzuli.edu.cn/https/*
// @grant        GM_getValue
// @grant        GM_setValue
// ==/UserScript==

/* ==UserConfig==
自定义格式化参数:
  figAnnotation: 
    title: 是否保留图片注释
    description: 是否保留图片注释, 勾选时会保留
    type: checkbox
    default: false
  tableRelated: 
    title: 是否保留表格相关
    description: 是否保留表格相关, 勾选时会保留
    type: checkbox
    default: false
  periodToOneDotLeader: 
    title: 涉及图指, 公式指, 表格指, 标题句点转换为一点前导符, 在标题和关键词末尾加句点; 删除表注和表格
    description: 此选项用于深度学习时分割句子的区别, 勾选时请将1, 2, 4, 5 取消勾选
    type: checkbox
    default: false
  title: 
    title: 是否保留序列标题
    description: 不是论文题目，而是每一部分的小标题
    type: checkbox
    default: true
 ==/UserConfig== */

 (function () {
    'use strict';
    let figAnnotation = GM_getValue('自定义格式化参数.figAnnotation');
    let periodToOneDotLeader = GM_getValue('自定义格式化参数.periodToOneDotLeader');
    let tableRelated = GM_getValue('自定义格式化参数.tableRelated');
    let title = GM_getValue('自定义格式化参数.title');
    let match = window.location.pathname.match(/(?:\/[^\/]*){1}\/([^\/]*)/);
    let webvpnPara = match ? match[1] : null;
    // 使用 JavaScript 动态添加 <script> 标签
    var script = document.createElement('script');
    script.src = 'https://cdn.jsdelivr.net/npm/docx@8.2.4/build/index.umd.min.js';
    // 将 <script> 标签添加到文档的头部
    document.head.appendChild(script);

    function formatInnerText(element) {
        /* 去除 [1,2,,,[3,4],[3]]; 替换波浪线运算符 \u223c ∼为~ 
         替换only负号 \u2212 −为-; 替换en-dash \u2013 –为-; 替换only连字符 \u2010 ‐为-; 
         替换微间隔thin space \u2009  为 ; 替换微间隔hair space \u200a  为 ; 替换空格\u2005  为 ; 替换无中断空格 \u00a0  为 ; 
         去除wt.%, at.%各自与%之间的空格; 
         不替换\u2014 —为-
        */
        return document.querySelector(element).innerText.replace(/ ?\[[\d\[\], ]*\]/g,'').replace(/∼/g, '~')
        .replace(/−/g, '-').replace(/–/g,'-').replace(/‐/g,'-')
        .replace(/ /g,' ').replace(/ /g,' ').replace(/ /g,'').replace(/ /g, ' ')
        .replace(/wt\. /g,'wt.').replace(/at\. /g,'at.')
    }
    function replaceNewlinesWithCondition(text) {
        const lines = text.split('\n');
        let newString = "";
    
        for (let i = 0; i < lines.length - 1; i++) {
            // Add the current line and the conditional newline or '1'
            newString += lines[i];
            // Check if the current or next line is less than 8 characters long
            if ((lines[i].length < 8 && lines[i].length != 0) || (lines[i + 1].length < 8 && lines[i + 1].length != 0)) {
                newString += '';
            } else {
                newString += '\n';
            }
        }
        // Add the last line which doesn't have a newline character after it
        newString += lines[lines.length - 1];
        return newString;
    }
    function createButton(fontColor = '(235, 101, 0, 1)', backColor = '(228, 229, 229, 0.5)') {
        var button = document.createElement('button');
        button.id = 'blinkingButton';
        button.innerHTML = 'Format this document';
        button.style.cssText = `position: fixed; top: 5px; left: 50%; transform: translateX(-50%); padding: 5px 5px; background-color: rgba${backColor}; color: rgba${fontColor}; border: none; border-radius: 10px; cursor: pointer; z-index: 9999; font-family: "Times New Roman", Times, serif; font-size: 16px; `;
        document.body.appendChild(button);
        return button
    }
    function downDoc(text, fileName = 'doc') {
        /* txt */
        // 创建一个Blob对象（二进制大对象
        var blob = new Blob([text], { type: 'text/plain' });
        // 创建一个a标签
        var a = document.createElement('a');
        // 创建一个文件名
        a.href = window.URL.createObjectURL(blob);
        a.download = `${fileName}.txt`;
        // 将a标签附加到DOM中
        document.body.appendChild(a);
        // 模拟点击a标签，触发下载
        a.click();
        // 从DOM中移除a标签
        document.body.removeChild(a);

        /* html */
        var blob = new Blob([`<!DOCTYPE html>
  <html lang="en">
  
  <head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Document</title>
    <style type="text/css">
    #text-container {
        font-family:'Times New Roman';
        font-size: 20px
    }
    </style>
  </head>
  
  <body>
    <div id="text-container"></div>
  </body>
  <script>
    var textWithNewline = \`${text}\`;
    var formattedText = textWithNewline.replace(/\\n/g, '<br>');
    document.getElementById('text-container').innerHTML = formattedText;
  </script>
  
  </html>`], { type: 'text/plain' });
        var a = document.createElement('a');
        a.href = window.URL.createObjectURL(blob);
        a.download = `${fileName}.html`;
        document.body.appendChild(a);
        a.click();
        document.body.removeChild(a);

        /* doc */
        var blob = new Blob(['<!DOCTYPE html><html><head><meta charset="UTF-8"></head><body style="font-family: "Times New Roman", Times, serif; ">'
            + text.replace(/(.*)/mg, '<p>$1</p>') +
            '</body></html>'],
            { type: 'application/msword;charset=utf-8' });
        var a = document.createElement('a');
        a.href = window.URL.createObjectURL(blob);
        a.download = `${fileName}.doc`;
        document.body.appendChild(a);
        a.click();
        document.body.removeChild(a);

        /* docx */
        var text_ = text.match(/(^.*$)/mg);
        var _ = []
        for (var i = 0, len = text_.length; i < len; i++) {
            _.push(new docx.Paragraph({
                children: [
                    new docx.TextRun(text_[i])]
            }))
        }

        const doc = new docx.Document({
            sections: [{
                properties: {},
                children: _,
            }]
        });

        docx.Packer.toBlob(doc).then(blob => {
            var a = document.createElement('a');
            a.href = window.URL.createObjectURL(blob);
            a.download = `${fileName}.docx`;
            document.body.appendChild(a);
            a.click();
            document.body.removeChild(a);
        });
    }

    if (window.location.host == 'www.sciencedirect.com'||webvpnPara=='77726476706e69737468656265737421e7e056d234336155700b8ca891472636a6d29e640e') {
        button = createButton('(235, 101, 0, 1)', '(228, 229, 229, 0.5)')
        button.addEventListener('click', function () {
            // 获取下载文献的名称
            let fileName = document.querySelector('article').querySelector('h1').innerText;
            // 获取所有classname为'download-link'的元素
            document.querySelectorAll('.download-link').forEach((e)=>{e.parentNode.removeChild(e)});
            if (!figAnnotation) document.querySelectorAll("Figure").forEach((e)=>{e.parentNode.removeChild(e)});
            //   为标题末尾增加句点并替换原句点为一点前导符
            if (periodToOneDotLeader) {
                document.querySelector('article').querySelectorAll('h1, h2, h3, h4, h5').forEach(e => {e.innerText = e.innerText.replace(/\./g, '․') + '.'});
                //为图指, 方程指，表指替换一点前导符, 下面的方法不通用
                // var elements = document.querySelector('article').querySelectorAll('a');
                // for (var i = 0; i < elements.length; i++) {
                //     if (elements[i].href.includes('#fig') || elements[i].href.includes('#eqn') || elements[i].href.includes('#tbl')) {
                //         elements[i].innerText = elements[i].innerText.replace(/\./g, '․');
                //     }
                // }
            }
            //删除表注和表格, 存在仅仅字母t开头的列表
            if (!tableRelated) document.querySelectorAll('[id^="t"]').forEach((e)=>{if(e.id.match(/^t.*\d+$/)) e.parentNode.removeChild(e)});
            // 替换img alt 'radical dot'为\u2022 •
            document.querySelectorAll("img").forEach((e)=>{
                if (e.alt == 'radical dot') e.parentNode.replaceChild(document.createTextNode('•'), e)
            })
            if(!title) document.querySelector("#body>div").querySelectorAll('h1, h2, h3, h4, h5').forEach((e)=>{e.parentNode.removeChild(e)});
            var text = `${formatInnerText('article h1')}\n
${formatInnerText("#abstracts")}\n
${(() => {// 获取包含关键词的父元素
                    var keywordsContainer = document.querySelector('.Keywords');
                    // 获取关键词节点列表
                    var keywordNodes = keywordsContainer.querySelectorAll('.keyword');
                    // 初始化一个空字符串，用于存储结果文本
                    //   var resultText = 'Keywords\n\n';
                    var resultText = keywordsContainer.querySelector('h2').innerText + '\n\n';
                    // 遍历关键词节点列表
                    for (var i = 0, len = keywordNodes.length; i < len; i++) {
                        var keywordText = keywordNodes[i].querySelector('span').innerText;
                        // 拼接到结果字符串，并添加换行符
                        if (i == len - 1) {
                            resultText += keywordText;
                            break
                        }
                        resultText += keywordText + '; ';
                    }
                    if (periodToOneDotLeader) {
                        return resultText + '.'
                    } else {
                        return resultText
                    }
                })()}\n
${replaceNewlinesWithCondition(formatInnerText("#body>div"))}`;
if(periodToOneDotLeader) text=text.replace(/(Figs?)\./g,'$1․').replace(/(E|eqs?)\./g,'$1․').replace(/(et al)\./g,'$1․').replace(/e\.g\./g,'e․g․');
            downDoc(text, fileName);
        });
    } else if (window.location.host == 'pubs.rsc.org') {
        // '(238,220,0,1)'黄色 '(72,169,197)'淡蓝 '(0,73,118)'深蓝 '(151,215,0)'草绿
        var button = createButton('(255,255,255,1)', '(239,51,64,0.9)');
        button.addEventListener('click', function () {
            // 去除论文作者
            document.querySelectorAll('.article__authors').forEach((e)=>{e.parentNode.removeChild(e)});
            // 去除a标签（主要去除引用）
            document.querySelectorAll('a').forEach((e)=>{e.parentNode.removeChild(e)});
            var text =formatInnerText(document.querySelector("article").innerText.replace(/\n\nConflicts of interest[\s\S]*/g, ''));
            if(periodToOneDotLeader) text=text.replace(/(Figs?)\./g,'$1․').replace(/(E|eqs?)\./g,'$1․').replace(/(et al)\./g,'$1․').replace(/e\.g\./g,'e․g․');
            downDoc(text);
        })
    } else if (window.location.host == 'link.springer.com' || webvpnPara == '77726476706e69737468656265737421fcfe4f976923784277068ea98a1b203a54') {
        var button = createButton('(27,48,81,0.9)', '(206,219,224,0.8)');
        button.addEventListener('click', function () {
            var fileName = document.querySelector('h1').innerText;
            //清空结论后的元素
            var flag=false;
            var elements = document.querySelector('.main-content').children;
            for (let i = 0; i < elements.length; i++) {
                if(!flag){
                    if(elements[i].getAttribute('data-title')=='Conclusion'){
                        flag=!flag
                    }
                }else{
                    elements[i].innerHTML=''
                }
            }
            //图片注释
            if(!figAnnotation) document.querySelectorAll('[data-test="figure"]').forEach((e)=>{e.parentNode.removeChild(e)});
            //表格
            if(!tableRelated) document.querySelectorAll('[id*=table-]').forEach((e)=>{e.parentNode.removeChild(e)});
            // 小标题加句点
            if(periodToOneDotLeader) document.querySelector('#main').querySelectorAll('h1, h2, h3, h4, h5').forEach((e)=>{e.innerText = e.innerText.replace(/\./g, '․') + '.'});
            // 小标题
            if(!title) document.querySelector('#main').querySelectorAll('h1, h2, h3, h4, h5').forEach((e)=>{e.parentNode.removeChild(e)});
            var text = `${formatInnerText('h1[data-test="article-title"]')}\n
${formatInnerText('section[data-title="Abstract"]')}\n
${formatInnerText('.main-content')}`
if(periodToOneDotLeader) text=text.replace(/(Figs?)\./g,'$1․').replace(/(E|eqs?)\./g,'$1․').replace(/(et al)\./g,'$1․').replace(/e\.g\./g,'e․g․');
            downDoc(text,fileName);
        });
    } else if (window.location.host == 'onlinelibrary.wiley.com'){
        var button = createButton('(27,48,81,0.9)', '(206,219,224,0.8)');
        button.addEventListener('click', function () {
            const fileName=document.querySelector('.citation__title').innerText;
            var sons=document.querySelector('section[class*="article-section__full"]').children;
            var flag=false
            // 遍历并使一些子元素innerHtml为空
            for (var i = 0; i < sons.length; i++) {
                if(!flag){
                    if(sons[i].tagName=='SECTION'){
                        // console.log(i)
                    }else{
                        flag=true
                        sons[i].innerHTML=''
                    }
                }else{
                    sons[i].innerHTML=''
                }
            }
            if(periodToOneDotLeader) document.querySelectorAll('h1, h2, h3, h4, h5').forEach(e => {e.innerText = e.innerText.replace(/\./g, '․') + '.'});
            if(!figAnnotation) document.querySelectorAll('.article-section__inline-figure').forEach((e)=>{e.parentNode.removeChild(e)});
            if(!tableRelated) document.querySelectorAll('.article-table-content').forEach((e)=>{e.parentNode.removeChild(e)});
            // 删除span标签, 去除引用
            document.querySelector('section[class*="article-section__full"]').querySelectorAll('span').forEach((e)=>{e.parentNode.removeChild(e)});
            if(!title) document.querySelector('section[class*="article-section__full"]').querySelectorAll('h1, h2, h3, h4, h5').forEach((e)=>{e.parentNode.removeChild(e)});
            var text = `${formatInnerText('.citation__title')}\n
${formatInnerText('div[class*="abstract-group"]')}\n
${formatInnerText('section[class*="article-section__full"]')}`
if(periodToOneDotLeader) text=text.replace(/(Figs?)\./g,'$1․').replace(/(E|eqs?)\./g,'$1․').replace(/(et al)\./g,'$1․').replace(/e\.g\./g,'e․g․');
            downDoc(text,fileName);
        });
    }

    var style = document.createElement('style');
    style.textContent = `
    @keyframes blink {
      0% {
        opacity: 1;
      }
      50% {
        opacity: 0;
      }
      100% {
        opacity: 1;
      }
    }
  
    #blinkingButton {
      animation: blink 1s infinite;
    }
  `;
    document.head.appendChild(style);
})();