生成分卷和全本ePub文档
目前為
// ==UserScript==
// @name 轻小说文库下载
// @namespace wenku8Haoa
// @version 2.0.2
// @description 生成分卷和全本ePub文档
// @author HaoaW
// @match https://www.wenku8.net/*
// @match http://www.wenku8.net/*
// @connect dl.wenku8.com
// @connect picture.wenku8.com
// @connect img.wenku8.com
// @connect pic.wenku8.com
// @require https://cdn.bootcdn.net/ajax/libs/jquery/3.5.1/jquery.js
// @require https://cdn.bootcdn.net/ajax/libs/jszip/2.6.1/jszip.js
// @require https://cdn.bootcdn.net/ajax/libs/FileSaver.js/2.0.5/FileSaver.js
// @icon https://www.wenku8.net/favicon.ico
// @grant GM_xmlhttpRequest
// ==/UserScript==
(function () {
'use strict';
//epub生成;使用addEventListener绑定；
function EpubBuilder(e) {
let bInfo = {
dlDomain: "dl.wenku8.com",
aid: article_id,//本书编号 article_id
title: document.getElementById("title").childNodes[0].textContent, //标题
creator: document.getElementById('info').innerText,//作者
bookUrl: self.location.href,
XHRArr: [],//下载请求；[{start:false,done:false,url:,loadFun:,data:,bookInfo:bInfo}]
loadVolume: (xhr) => {
xhr.start = true;
let navToc = xhr.bookInfo.nav_toc[xhr.VolumeIndex];
GM_xmlhttpRequest({
method: 'GET',
url: xhr.url,
onload: function (response) {
//xhr.VolumeIndex卷号
let chapterIndex = 0;
let ImagesIndex = 0;
if (response.status == 200) {
//https://developer.mozilla.org/zh-CN/docs/Web/Guide/Parsing_and_serializing_XML
//下载分卷文本，转换为html
let domParser = new DOMParser();
let rspHtml = domParser.parseFromString(
`<html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops">
<head>
<meta charset="utf-8"/>
<title>${navToc.volumeName}</title>
<link href="../Styles/default.css" rel="stylesheet" type="text/css"/>
</head>
<body><div class="volumetitle"><h2>${navToc.volumeName}</h2></div><br /></body>
</html>`, "text/html");
rspHtml.body.innerHTML += response.responseText;
//调用自带简转繁
if (currentEncoding != targetEncoding) {
translateBody(rspHtml.body);
}
let removeChild = [];
//处理章节、插图 和 contentdp
for (let i = 0; i < rspHtml.body.children.length; i++) {
let child = rspHtml.body.children[i];
if ("UL" == child.tagName && "contentdp" == child.id) {
removeChild.push(child);
}
//章节
else if ("DIV" == child.tagName && "chaptertitle" == child.className) {
chapterIndex++;
//章节h3、分卷h2、书名h1（没有做）
//<div class="chaptertitle"><div id="chapter_1" name="xxx"><h3>第一章</h3></div></div>
let cTitle = child.innerText;
let aName = child.firstChild.getAttribute("name");
let divEle = document.createElement("div");
divEle.id = `chapter_${chapterIndex}`;
divEle.setAttribute("name", aName);
divEle.innerHTML = `<h3>${cTitle}</h3>`;
child.innerHTML = divEle.outerHTML;
//添加章节导航
navToc.chapterArr.push({
chapterName: cTitle
, chapterID: divEle.id
, chapterHref: `${navToc.volumeHref}#${divEle.id}`
});
}
//内容
else if ("DIV" == child.tagName && "chaptercontent" == child.className) {
for (let j = 0; j < child.children.length; j++) {
let contentChild = child.children[j];
//插图
if ("DIV" == contentChild.tagName && "divimage" == contentChild.className) {//插图
//取得插图下载地址
let imgASrc = contentChild.getAttribute("title");
let imgUrl = new URL(imgASrc);
let imgPath = `Images${imgUrl.pathname}`;
//在html中加入img标签
let imgEle = document.createElement("img");
imgEle.setAttribute("loading", "lazy");
imgEle.setAttribute("src", `../${imgPath}`);
contentChild.innerHTML = imgEle.outerHTML;
//记录图片信息作为epub资源
ImagesIndex++;
let ImagesID = `${xhr.bookInfo.ImagesFix}_${xhr.VolumeIndex}_${ImagesIndex}`;
let images = { path: `${imgPath}`, content: null, id: ImagesID, idName: contentChild.id };
xhr.bookInfo.Images.push(images);
//添加图片下载xhr请求
let xhrImg = { start: false, done: false, url: imgASrc, loadFun: xhr.bookInfo.loadImg, images: images, bookInfo: xhr.bookInfo };
xhr.bookInfo.XHRArr.push(xhrImg);
xhrImg.loadFun(xhrImg);
}
}
}
}
removeChild.forEach(c => rspHtml.body.removeChild(c));
//转换html为xhtml
let xmlSerializer = new XMLSerializer();
let rspXml = xmlSerializer.serializeToString(rspHtml);
let rspXHtml = domParser.parseFromString(rspXml, "application/xhtml+xml");
rspXHtml.firstChild.setAttribute("xmlns:epub", "http://www.idpf.org/2007/ops");
//保存章节内容，作为epub资源
xhr.bookInfo.Text[xhr.VolumeIndex].content = `<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html>
${rspXHtml.firstChild.outerHTML}`;
//生成epub，还有资源未下载则只更新生成进度
xhr.done = true;
xhr.bookInfo.buildEpub(xhr.bookInfo);
}
else {
//重新下载
xhr.bookInfo.refreshProgress(xhr.bookInfo, `${navToc.volumeName} 下载失败，重新下载;`);
xhr.loadFun(xhr);
}
},
onerror: () => {
//重新下载
xhr.bookInfo.refreshProgress(xhr.bookInfo, `${navToc.volumeName} 下载失败，重新下载;`);
xhr.loadFun(xhr);
}
});
},//分卷下载
loadImg: (xhr) => {
xhr.start = true;
GM_xmlhttpRequest({
method: 'GET',
url: xhr.url,
responseType: "arraybuffer",
onload: function (response) {
if (response.status == 200) {
xhr.images.content = response.response;
xhr.done = true;
xhr.bookInfo.buildEpub(xhr.bookInfo);
} else {
//重新下载
xhr.bookInfo.refreshProgress(xhr.bookInfo, `${xhr.images.idName} 下载失败，重新下载;`);
xhr.loadFun(xhr);
}
},
onerror: () => {
//重新下载
xhr.bookInfo.refreshProgress(xhr.bookInfo, `${xhr.images.idName} 下载失败，重新下载;`);
xhr.loadFun(xhr);
}
});
},//图片下载
nav_toc: [],//导航菜单,第一层分卷，第二层章节{volumeName：,volumeID:,volumeHref:,chapterArr:[{chapterName:,chapterID:,chapterHref:}]}
VolumeFix: "Volume",//分卷文件、ID前缀
Text: [],//下载后生成的XHTML；{path:`Text/${volumeHref}`,content:}
ImagesFix: "Img",//图片文件、ID前缀
Images: [],//下载的图片；{path:`Images/${url.pathname}`,content:}
buildEpub: (info) => {
if (info.XHRArr.every(e => e.done)) {
info.refreshProgress(info, `开始生成ePub;`);
let zip = new JSZip();
//epub固定内容
zip.file("mimetype", info.mimetype);
zip.file("META-INF/container.xml", info.container_xml);
//todo: new DOMParser(); 操作xml
let manifest = document.createElement("manifest");
let spine = document.createElement("spine");
//保存css
{
let item = document.createElement("item");
item.setAttribute("id", info.defaultCSS.id);
item.setAttribute("href", info.defaultCSS.path);
item.setAttribute("media-type", "text/css");
manifest.appendChild(item);
//保存css
zip.file(`OEBPS/${info.defaultCSS.path}`, info.defaultCSS.content);
}
//生成并保存nav.xhtml
//<ol><li><a href="Volume_0.xhtml">第一卷</a><ol><li><a href="Volume_0.xhtml#chapter_1">第一章</a></li></ol></li></ol>
{
//生成nav.xhtml
let domParser = new DOMParser();
let navXhtmlDoc = domParser.parseFromString(info.nav_xhtml.content, "application/xhtml+xml");
let tocEle = navXhtmlDoc.getElementById("toc");
let bOlEle = navXhtmlDoc.createElement("ol");
for (let t of info.nav_toc) {
//分卷
let vAEle = navXhtmlDoc.createElement("a");
vAEle.href = t.volumeHref;
vAEle.innerText = t.volumeName;
let vLiEle = navXhtmlDoc.createElement("li");
vLiEle.appendChild(vAEle);
if (t.chapterArr && 0 < t.chapterArr.length) {
//分卷的章节
let vOlEle = navXhtmlDoc.createElement("ol");
for (let c of t.chapterArr) {
let cAEle = navXhtmlDoc.createElement("a");
cAEle.href = c.chapterHref;
cAEle.innerText = c.chapterName;
let cLiEle = navXhtmlDoc.createElement("li");
cLiEle.appendChild(cAEle);
vOlEle.appendChild(cLiEle);
}
vLiEle.appendChild(vOlEle);
}
bOlEle.appendChild(vLiEle);
}
tocEle.appendChild(bOlEle);
let nav_xhtml = `<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html>
${navXhtmlDoc.firstChild.outerHTML}`;
//保存nav.xhtml信息到content.opf
//manifest节点
let item = document.createElement("item");
item.setAttribute("id", info.nav_xhtml.id);
item.setAttribute("href", info.nav_xhtml.path);
item.setAttribute("media-type", "application/xhtml+xml");
item.setAttribute("properties", "nav");
manifest.appendChild(item);
//spine节点
let itemref = document.createElement("itemref");
itemref.setAttribute("idref", info.nav_xhtml.id);
itemref.setAttribute("linear", "no");
spine.appendChild(itemref);
//保存nav.xhtml
zip.file(`OEBPS/${info.nav_xhtml.path}`, nav_xhtml);
}
//保存分卷内容
for (let t of info.Text) {
let item = document.createElement("item");
item.setAttribute("media-type", "application/xhtml+xml");
item.setAttribute("href", t.path);
item.setAttribute("id", t.id);
manifest.appendChild(item);
let itemref = document.createElement("itemref");
itemref.setAttribute("idref", t.id);
spine.appendChild(itemref);
zip.file(`OEBPS/${t.path}`, t.content);
}
//保存图片
for (let t of info.Images) {
let itemImg = document.createElement("item");
//media-type暂固定jpeg
itemImg.setAttribute("media-type", "image/jpeg");
itemImg.setAttribute("href", t.path);
itemImg.setAttribute("id", t.id);
manifest.appendChild(itemImg);
zip.file(`OEBPS/${t.path}`, t.content, { binary: true });
}
//生成书籍信息
let uuid = self.crypto.randomUUID();
let createTime = new Date();
let content_opf = `<?xml version="1.0" encoding="utf-8"?>
<package version="3.0" unique-identifier="BookId" xmlns="http://www.idpf.org/2007/opf">
<metadata xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf">
<dc:language>zh-CN</dc:language>
<dc:title>${bInfo.title}</dc:title>
<meta property="dcterms:modified">${createTime.toISOString()}</meta>
<dc:identifier id="BookId">urn:uuid:${uuid}</dc:identifier>
<dc:creator>${info.creator}</dc:creator>
<!--第一张插图做封面-->
${(info.Images && 0 < info.Images.length) ? `<meta name="cover" content="${info.Images[0].id}" />` : ''}
</metadata>
<manifest>
${manifest.innerHTML}
</manifest>
<spine>
${spine.innerHTML}
</spine>
</package>`;
zip.file("OEBPS/content.opf", content_opf);
//书名.开始卷-结束卷.epub
let epubName = `${bInfo.title}.${info.nav_toc[0].volumeName}`;
if (1 < info.nav_toc.length) {
epubName = epubName + '-' + info.nav_toc[info.nav_toc.length - 1].volumeName;
}
saveAs(zip.generate({ type: "blob", mimeType:"application/epub+zip" }), `${epubName}.epub`);
info.refreshProgress(info, `ePub生成完成,文件名：${epubName}.epub；`);
}
else {
info.refreshProgress(info);
}
},//生成epub，如果XHRArr已经都完成了
mimetype: 'application/epub+zip',//epub mimetype 文件内容
container_xml: `<?xml version="1.0" ?>
<container version="1.0" xmlns="urn:oasis:names:tc:opendocument:xmlns:container">
<rootfiles>
<rootfile full-path="OEBPS/content.opf" media-type="application/oebps-package+xml" />
</rootfiles>
</container>`,//epub container.xml 文件内容
nav_xhtml: {
content: `
<html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" lang="en" xml:lang="en">
<head>
<title>ePub NAV</title>
<meta charset="utf-8"/>
<link href="../Styles/default.css" rel="stylesheet" type="text/css"/>
</head>
<body epub:type="frontmatter">
<nav epub:type="toc" id="toc" role="doc-toc">
<h2>目录</h2>
</nav>
</body>
</html>`
, path: `Text/nav.xhtml`
, id: `nav_xhtml_id`
},//epub nav.xhtml 文件模板
defaultCSS: {
content: `
nav#landmarks {
display:none;
}
nav#page-list {
display:none;
}
ol {
list-style-type: none;
}
.volumetitle ,
.chaptertitle {
text-align: center;
}
`
, id: "default_css_id"
, path: "Styles/default.css"
},//epub default.css 样式文件
progressEle: null,//进度、日志 元素；{txt:,img:,err:}
refreshProgress: (info, err) => {
if (!info.progressEle) {
//epub生成进度，下载文本进度：7/7；下载图片进度：77/77；日志：开始生成ePub;ePub生成完成;
info.progressEle = {};
info.progressEle.txt = document.createElement("span");
info.progressEle.img = document.createElement("span");
info.progressEle.err = document.createElement("span");
let logDiv = document.createElement('div');
logDiv.appendChild(document.createTextNode("epub生成进度，下载文本进度："));
logDiv.appendChild(info.progressEle.txt);
logDiv.appendChild(document.createTextNode("；下载图片进度："));
logDiv.appendChild(info.progressEle.img);
logDiv.appendChild(document.createTextNode("；日志："));
logDiv.appendChild(info.progressEle.err);
document.body.insertBefore(logDiv, document.getElementById('title'));
}
//日志
if (err) { info.progressEle.err.innerText = err + info.progressEle.err.innerText; }
//文本进度
let txtProgress = info.Text.filter((value, key, arr) => { return value.content; }).length;
info.progressEle.txt.innerText = `${txtProgress}/${info.Text.length}`;
//图片进度，文本下载完成后才能得到图片总数
if (txtProgress == info.Text.length) {
let imgProgress = info.Images.filter((value, key, arr) => { return value.content; }).length;
info.progressEle.img.innerText = `${imgProgress}/${info.Images.length}`;
}
},//显示进度日志
start: (e) => {
//全本分卷
let vcssEle = document.getElementsByClassName("vcss");
if ("bookDownload" != e.target.id) { //全本下载 链接的id是bookDownload
vcssEle = [e.target.parentElement];
}
for (let VolumeIndex = 0; VolumeIndex < vcssEle.length; VolumeIndex++) {
let vcss = vcssEle[VolumeIndex];
let vid = vcss.parentElement.nextElementSibling.getElementsByTagName('a')[0].getAttribute('href').split('.')[0];;
//let vid = vcss.getAttribute("vid");
let vcssText = vcss.childNodes[0].textContent;
let navToc = bInfo.nav_toc[VolumeIndex] = {
volumeName: vcssText
, volumeID: `${bInfo.VolumeFix}_${VolumeIndex}`
, volumeHref: `${bInfo.VolumeFix}_${VolumeIndex}.xhtml`
, chapterArr: []
};
bInfo.Text[VolumeIndex] = {
path: `Text/${navToc.volumeHref}`
, content: ""
, id: navToc.volumeID
};
//分卷下载链接
let dlink = `https://${bInfo.dlDomain}/pack.php?aid=${bInfo.aid}&vid=${vid}`;
let xhr = { start: false, done: false, url: dlink, loadFun: bInfo.loadVolume, VolumeIndex: VolumeIndex, data: { vid: vid, vcssText: vcssText }, bookInfo: bInfo };
bInfo.XHRArr.push(xhr);
xhr.loadFun(xhr);
}
bInfo.buildEpub(bInfo);
}//入口，开始下载文件并生成epub；
};
bInfo.start(e);
};
//替换所有
String.prototype.replaceAll = function (exp, newStr) {
return this.replace(new RegExp(exp, "gm"), newStr);
};
//标记含有html代码的行
var htmlLine = function (txt) {
return 'c1ef6520' + txt + '1a5219e9b2b0';
};
//格式化内容
var contentFormat = function (txt) {
//设置章节名样式
let txtTmp = txt.replaceAll(/^ {2}\S+.*$/,
function (match) {
return htmlLine('<br/><div style="background:#e4e1d8" id="title">' + match.trim() + '</div>');
}
);
txtTmp = $('<div></div>').text(txtTmp).html();
txtTmp = txtTmp.replaceAll(/ /, ' ');
txtTmp = txtTmp.replaceAll(/(c1ef6520).*(1a5219e9b2b0)/,
function (match) {
return $('<div></div>').html(match.replaceAll(/( )/, ' ').replace('c1ef6520', '').replace('1a5219e9b2b0', '')).text();
}
);
txtTmp = txtTmp.replaceAll(/[\r\n]+/, '<br/>');
return txtTmp;
};
//格式化图片
var imgFormat = function (txt) {
let imgResultRoot = $('<div></div>');
let imgRoot = $('<div></div>').html(txt);
$('.divimage', imgRoot).each(function () {
let jqthis = $(this);
let imgURL = jqthis.attr('title');//取得图片url
let itemEle = $('<div class="divimage"><a target="_blank"><img border="0" class="imagecontent" /></a></div>');
$('a', itemEle).attr('href', imgURL);
$('img', itemEle).attr('src', imgURL);
imgResultRoot.append(itemEle);
});
return imgResultRoot.html();
}
//格式化从pack.php下载的带有Html的内容
var htmlFormat = function (htmlStr) {
let resultRoot = $('<div></div>');
let strRoot = $('<div></div>').html(htmlStr);
$('.chaptertitle', strRoot).each(function () {
let jqthis = $(this);
let cTitle = jqthis.text();//取得章节名
let itemEle = $('<div style="background-color:silver" id="title"></div>');
itemEle.text(cTitle);
resultRoot.append(itemEle);
resultRoot.append('<br />');
let nextHtml = jqthis.next().html();
if (-1 != cTitle.indexOf('插图')) {
nextHtml = imgFormat(nextHtml);
}
resultRoot.append(nextHtml);
resultRoot.append('<br /><br />');
});
return resultRoot.html();
}
//目录或内容页面会声明章节变量。
if ('undefined' == typeof chapter_id || undefined === chapter_id) { }
else {
//本书编号 article_id
//目录页面章节id定义为 '0'
if ('0' == chapter_id) {//在章节名之后添加下载链接
//书名
let aname = $('#title').text();
//targetEncoding 1: 繁體中文, 2: 简体中文
let charsetDL = 'utf-8';
let charsetDLAll = 'utf8';
if ('1' == targetEncoding) {
charsetDL = 'big5';
charsetDLAll = 'big5';
}
//添加全本下载链接
let allDLink = `https://dl.wenku8.com/down.php?type=${charsetDLAll}&id=${article_id}&fname=${aname}`;
let allaEle = `<a href="${allDLink}"> 全本文本下载(${charsetDLAll})</a>`;
$('#title').append(allaEle);
$('.vcss').each(function () {
let jqthis = $(this);
let vname = jqthis.text();
let nextChapter = jqthis.parent().next().children().first();
let vid = $('a', nextChapter).attr('href').split('.')[0];
//vid = (Number(vid)-1).toString();//部分文章无法使用章节减1来下载，直接使用章节可以下载(下载内容使用-1时章节名包含小说名，不-1不包含小说名)
let dlink = `https://dl.wenku8.com/packtxt.php?aid=${article_id}&vid=${vid}&aname=${aname}&vname=${vname}&charset=${charsetDL}`;
let aEle = `<a href="${dlink}"> 文本下载(${charsetDL})</a>`;
jqthis.append(aEle);
});
//添加 ePub下载(全本)
let allaEpubEle = document.createElement("a");
allaEpubEle.id = "bookDownload";
allaEpubEle.innerText = " ePub下载(全本)";
allaEpubEle.href = "javascript:void(0);";
document.getElementById('title').append(allaEpubEle);
allaEpubEle.addEventListener("click", EpubBuilder);
//添加 ePub下载(本卷)
for (let v of document.getElementsByClassName("vcss")) {
let allaEle = document.createElement("a");
allaEle.href = "javascript:void(0);";
allaEle.innerText = " ePub下载(本卷)";
v.append(allaEle);
allaEle.addEventListener("click", EpubBuilder);
};
}
else {
if ('0' != chapter_id) {//内容页面
//如果包含一个内容为 'null'的span则判定为版权限制
if ($('#contentmain span').first().text().trim() == 'null') {
//设置下一页和上一页快捷键到目录页
preview_page = next_page = index_page;
$('#content').text('正在下载，请稍候...');
//下载带有Html标记的内容
let dlink = 'http://dl.wenku8.com/pack.php?aid=' + article_id +
'&vid=' + chapter_id;
GM_xmlhttpRequest({
method: 'GET',
url: dlink,
onload: function (response) {
if (response.status == 200) {
var formatHtml = htmlFormat(response.responseText);
$('#content').html(formatHtml);
} else {//html下载可能会失败，使用下载文本的方法来下载
if (-1 != $('#title').text().indexOf('插图')) {
$('#content').text('无法获取插图');
} else {
dlink = 'http://dl.wenku8.com/packtxt.php?aid=' + article_id +
'&vid=' + chapter_id;
GM_xmlhttpRequest({
method: 'GET',
url: dlink,
onload: function (response) {
if (response.status == 200) {
let formatTxt = contentFormat(response.responseText);
$('#content').html(formatTxt);
} else {
$('#content').text('下载失败');
}
},
});
}
}
},
});
}
}
}
}
// Your code here...
})();