drpy.js 71.3 KB
Newer Older
1
// import 'https://gitcode.net/qq_32394351/dr_py/-/raw/master/libs/es6py.js';
H
hjdhnx 已提交
2
// import {是否正版,urlDeal,setResult,setResult2,setHomeResult,maoss,urlencode} from 'http://192.168.10.103:5705/libs/es6py.js';
3
// import 'http://192.168.1.124:5705/libs/es6py.js';
H
hjdhnx 已提交
4 5
import cheerio from 'https://gitcode.net/qq_32394351/dr_py/-/raw/master/libs/cheerio.min.js';
// import cheerio from 'http://192.168.10.103:5705/libs/cheerio.min.js';
H
hjdhnx 已提交
6
import 'https://gitcode.net/qq_32394351/dr_py/-/raw/master/libs/crypto-js.js';
H
hjdhnx 已提交
7 8
import 'https://gitcode.net/qq_32394351/dr_py/-/raw/master/libs/drT.js';
// import 'http://192.168.10.103:5705/libs/drT.js';
H
hjdhnx 已提交
9
// import muban from 'https://gitcode.net/qq_32394351/dr_py/-/raw/master/js/模板.js';
H
hjdhnx 已提交
10
// import muban from 'http://192.168.10.103:5705/admin/view/模板.js';
H
hjdhnx 已提交
11 12

// const key = 'drpy_zbk';
13
// eval(req('http://192.168.1.124:5705/libs/es6py.js').content);
H
hjdhnx 已提交
14
function init_test(){
15
    // console.log(typeof(CryptoJS));
H
hjdhnx 已提交
16
    console.log("init_test_start");
H
hjdhnx 已提交
17
    console.log("当前版本号:"+VERSION);
H
hjdhnx 已提交
18
    console.log(RKEY);
H
hjdhnx 已提交
19
    console.log(JSON.stringify(rule));
20
    // console.log('123456的md5值是:'+md5('123456'));
H
hjdhnx 已提交
21 22 23 24
    // let aa = base64Encode('编码测试一下')
    // log(aa);
    // let bb = base64Decode(aa);
    // log('bb:'+bb);
H
hjdhnx 已提交
25 26 27 28 29 30 31
    // clearItem(RULE_CK);
    // console.log(JSON.stringify(rule));
    // console.log(request('https://www.baidu.com',{withHeaders:true}));
    // console.log(request('https://www.baidu.com/favicon.ico',{toBase64:true}));
    // require('http://192.168.10.99:5705/txt/pluto/drT.js');
    // console.log(typeof(drT));
    // console.log(drT.renderText('{{fl.cate}},hi, {{fl}}哈哈.{{fl}}',{sort: 1,cate:'movie'},'fl'));
H
hjdhnx 已提交
32
    console.log("init_test_end");
H
hjdhnx 已提交
33 34
}

H
hjdhnx 已提交
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54
/**
 * 执行预处理代码
 */
function pre(){
    if(typeof(rule.预处理) === 'string' && rule.预处理 && rule.预处理.trim()){
        let code = rule.预处理.trim();
        console.log("执行预处理代码:"+code);
        if(code.startsWith('js:')){
            code = code.replace('js:','');
        }
        try {
            // code里可以进行get 或者 post请求cookie并改变rule.headers 里的cookie
            //  直接操作 rule_fetch_params 这个变量 .headers.Cookie
            eval(code);
        }catch (e) {
            console.log('预处理执行失败:'+e.message);
        }
    }
}

H
hjdhnx 已提交
55
let rule = {};
H
hjdhnx 已提交
56
const VERSION = '3.9.18';
H
hjdhnx 已提交
57
/** 已知问题记录
58
 * 1.影魔的jinjia2引擎不支持 {{fl}}对象直接渲染 (有能力解决的话尽量解决下,支持对象直接渲染字符串转义,如果加了|safe就不转义)[影魔牛逼,最新的文件发现这问题已经解决了]
H
hjdhnx 已提交
59 60
 * Array.prototype.append = Array.prototype.push; 这种js执行后有毛病,for in 循环列表会把属性给打印出来 (这个大毛病需要重点排除一下)
 * 2.import es6py.js但是里面的函数没有被装载进来.比如drpy规则报错setResult2 is undefiend(合并文件了可以不管了)
H
hjdhnx 已提交
61
 * 3.无法重复导入cheerio(怎么解决drpy和parseTag里都需要导入cheerio的问题) 无法在副文件导入cheerio (现在是全部放在drpy一个文件里了,凑合解决?)
62
 * 4.有个错误不知道哪儿来的 executeScript: com.quickjs.JSObject$Undefined cannot be cast to java.lang.String 在 点击选集播放打印init_test_end后面打印(貌似不影响使用)
H
hjdhnx 已提交
63
 * 5.需要实现 stringify 函数,比起JSON.stringify函数,它会原封不动保留中文不会编码unicode
64 65
 * 6.base64Encode,base64Decode,md5函数还没有实现 (抄影魔代码实现了)
 * 7.eval(getCryptoJS());还没有实现 (可以空实现了,以后遇到能忽略)
H
hjdhnx 已提交
66
 * done:  jsp:{pdfa,pdfh,pd},json:{pdfa,pdfh,pd},jq:{pdfa,pdfh,pd}
67 68 69 70 71 72
 *  * 电脑看日志调试
 adb tcpip 5555
 adb connect 192.168.10.192
 adb devices -l
 adb logcat -c
 adb logcat | grep -i QuickJS
H
hjdhnx 已提交
73
 * **/
H
hjdhnx 已提交
74 75 76 77 78 79 80 81 82


/*** 以下是内置变量和解析方法 **/
const MOBILE_UA = 'Mozilla/5.0 (Linux; Android 11; M2007J3SC Build/RKQ1.200826.002; wv) AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 Chrome/77.0.3865.120 MQQBrowser/6.2 TBS/045714 Mobile Safari/537.36';
const PC_UA = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/95.0.4638.54 Safari/537.36';
const UA = 'Mozilla/5.0';
const UC_UA = 'Mozilla/5.0 (Linux; U; Android 9; zh-CN; MI 9 Build/PKQ1.181121.001) AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 Chrome/57.0.2987.108 UCBrowser/12.5.5.1035 Mobile Safari/537.36';
const IOS_UA = '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';
const RULE_CK = 'cookie'; // 源cookie的key值
H
hjdhnx 已提交
83
// const KEY = typeof(key)!=='undefined'&&key?key:'drpy_' + (rule.title || rule.host); // 源的唯一标识
H
hjdhnx 已提交
84 85 86 87
const CATE_EXCLUDE = '首页|留言|APP|下载|资讯|新闻|动态';
const TAB_EXCLUDE = '猜你|喜欢|APP|下载|剧情|热播';
const OCR_RETRY = 3;//ocr验证重试次数
// const OCR_API = 'http://dm.mudery.com:10000';//ocr在线识别接口
H
hjdhnx 已提交
88
// const OCR_API = 'http://192.168.3.239:5705/parse/ocr';//ocr在线识别接口
H
hjdhnx 已提交
89 90
// const OCR_API = 'http://cms.nokia.press/parse/ocr';//ocr在线识别接口
const OCR_API = 'http://cms.nokia.press:5706/parse/ocr';//ocr在线识别接口
H
hjdhnx 已提交
91 92 93
if(typeof(MY_URL)==='undefined'){
    var MY_URL; // 全局注入变量,pd函数需要
}
H
hjdhnx 已提交
94
var RKEY; // 源的唯一标识
H
hjdhnx 已提交
95 96 97
var fetch;
var print;
var log;
H
hjdhnx 已提交
98 99
var rule_fetch_params;
var fetch_params; // 每个位置单独的
H
hjdhnx 已提交
100
var oheaders;
H
hjdhnx 已提交
101
// var play_url; // 二级详情页注入变量,为了适配js模式0 (不在这里定义了,直接二级里定义了个空字符串)
H
hjdhnx 已提交
102 103 104
var _pdfh;
var _pdfa;
var _pd;
H
hjdhnx 已提交
105
// const DOM_CHECK_ATTR = ['url', 'src', 'href', 'data-original', 'data-src'];
H
hjdhnx 已提交
106
const DOM_CHECK_ATTR = /(url|src|href|-original|-src|-play|-url)$/;
H
hjdhnx 已提交
107 108 109
const SELECT_REGEX = /:eq|:lt|:gt|#/g;
const SELECT_REGEX_A = /:eq|:lt|:gt/g;

110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251
/**
es6py扩展
 */
if (typeof Object.assign != 'function') {
    Object.assign = function () {
        var target = arguments[0];
        for (var i = 1; i < arguments.length; i++) {
            var source = arguments[i];
            for (var key in source) {
                if (Object.prototype.hasOwnProperty.call(source, key)) {
                    target[key] = source[key];
                }
            }
        }
        return target;
    };
}
if (!String.prototype.includes) {
    String.prototype.includes = function (search, start) {
        if (typeof start !== 'number') {
            start = 0;
        }

        if (start + search.length > this.length) {
            return false;
        } else {
            return this.indexOf(search, start) !== -1;
        }
    };
}

if (!Array.prototype.includes) {
    Object.defineProperty(Array.prototype, 'includes', {
        value: function (searchElement, fromIndex) {

            if (this == null) {//this是空或者未定义,抛出错误
                throw new TypeError('"this" is null or not defined');
            }

            var o = Object(this);//将this转变成对象
            var len = o.length >>> 0;//无符号右移0位,获取对象length属性,如果未定义就会变成0

            if (len === 0) {//length为0直接返回false未找到目标值
                return false;
            }

            var n = fromIndex | 0;//查找起始索引
            var k = Math.max(n >= 0 ? n : len - Math.abs(n), 0);//计算正确起始索引,因为有可能是负值

            while (k < len) {//从起始索引处开始循环
                if (o[k] === searchElement) {//如果某一位置与寻找目标相等,返回true,找到了
                    return true;
                }
                k++;
            }
            return false;//未找到,返回false
        }
    });
}
if (typeof String.prototype.startsWith != 'function') {
    String.prototype.startsWith = function (prefix){
        return this.slice(0, prefix.length) === prefix;
    };
}
if (typeof String.prototype.endsWith != 'function') {
    String.prototype.endsWith = function(suffix) {
        return this.indexOf(suffix, this.length - suffix.length) !== -1;
    };
}
Object.prototype.myValues=function(obj){
    if(obj ==null) {
        throw new TypeError("Cannot convert undefined or null to object");
    }
    var res=[]
    for(var k in obj){
        if(obj.hasOwnProperty(k)){//需判断是否是本身的属性
            res.push(obj[k]);
        }
    }
    return res;
}
if (typeof Object.prototype.values != 'function') {
    Object.prototype.values=function(obj){
        if(obj ==null) {
            throw new TypeError("Cannot convert undefined or null to object");
        }
        var res=[]
        for(var k in obj){
            if(obj.hasOwnProperty(k)){//需判断是否是本身的属性
                res.push(obj[k]);
            }
        }
        return res;
    }
}
if (typeof Array.prototype.join != 'function') {
    Array.prototype.join = function (emoji) {
        // emoji = emoji||',';
        emoji = emoji||'';
        let self = this;
        let str = "";
        let i = 0;
        if (!Array.isArray(self)) {throw String(self)+'is not Array'}
        if(self.length===0){return ''}
        if (self.length === 1){return String(self[0])}
        i = 1;
        str = this[0];
        for (; i < self.length; i++) {
            str += String(emoji)+String(self[i]);
        }
        return str;
    };
}

String.prototype.rstrip = function (chars) {
    let regex = new RegExp(chars + "$");
    return this.replace(regex, "");
};

Array.prototype.append = Array.prototype.push;
String.prototype.strip = String.prototype.trim;
function 是否正版(vipUrl){
    let flag = new RegExp('qq\.com|iqiyi\.com|youku\.com|mgtv\.com|bilibili\.com|sohu\.com|ixigua\.com|pptv\.com|miguvideo\.com|le\.com|1905\.com|fun\.tv');
    return  flag.test(vipUrl);
}
function urlDeal(vipUrl){
    if(!vipUrl){
        return ''
    }
    if(!是否正版(vipUrl)){
        return vipUrl
    }
    if(!/miguvideo/.test(vipUrl)){
        vipUrl=vipUrl.split('#')[0].split('?')[0];
    }
    return vipUrl
}
function setResult(d){
    if(!Array.isArray(d)){
        return []
    }
    VODS = [];
H
hjdhnx 已提交
252
    // print(d);
253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293
    d.forEach(function (it){
        let obj = {
            vod_id:it.url||'',
            vod_name: it.title||'',
            vod_remarks: it.desc||'',
            vod_content: it.content||'',
            vod_pic: it.pic_url||it.img||'',
        };
        let keys = Object.keys(it);
        if(keys.includes('tname')){
            obj.type_name = it.tname||'';
        }
        if(keys.includes('tid')){
            obj.type_id = it.tid||'';
        }
        if(keys.includes('year')){
            obj.vod_year = it.year||'';
        }
        if(keys.includes('actor')){
            obj.vod_actor = it.actor||'';
        }
        if(keys.includes('director')){
            obj.vod_director = it.director||'';
        }
        if(keys.includes('area')){
            obj.vod_area = it.area||'';
        }
        VODS.push(obj);
    });
    return VODS
}
function setResult2(res){
    VODS = res.list||[];
    return VODS
}
function setHomeResult(res){
    if(!res||typeof(res)!=='object'){
        return []
    }
    return setResult(res.list);
}
H
hjdhnx 已提交
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322
// 猫了个咪
function rc(js) {
    if (js === 'maomi_aes.js') {
        var a = CryptoJS.enc.Utf8.parse("625222f9149e961d");
        var t = CryptoJS.enc.Utf8.parse("5efdtf6060e2o330");
        return {
            De: function (word) {
                word = CryptoJS.enc.Hex.parse(word)
                return CryptoJS.AES.decrypt(CryptoJS.enc.Base64.stringify(word), a, {
                    iv: t,
                    mode: CryptoJS.mode.CBC,
                    padding: CryptoJS.pad.Pkcs7
                }).toString(CryptoJS.enc.Utf8)
            },
            En: function (word) {
                // print(a);
                // print(word);
                var Encrypted = CryptoJS.AES.encrypt(word, a, {
                    iv: t,
                    mode: CryptoJS.mode.CBC,
                    padding: CryptoJS.pad.Pkcs7
                });
                return Encrypted.ciphertext.toString();
            }
        };
    }
    return {};
}

323 324 325
// 千万不要用for in 推荐 forEach (for in 会打乱顺序)
//猫函数
function maoss(jxurl, ref, key) {
326
    fetch_params = JSON.parse(JSON.stringify(rule_fetch_params));
327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377
    eval(getCryptoJS());
    try {
        var getVideoInfo = function (text) {
            return CryptoJS.AES.decrypt(text, key, {iv: iv, padding: CryptoJS.pad.Pkcs7}).toString(CryptoJS.enc.Utf8);
        };
        var token_key = key == undefined ? 'dvyYRQlnPRCMdQSe' : key;
        if (ref) {
            var html = request(jxurl, {
                headers: {
                    'Referer': ref
                }
            });
        } else {
            var html = request(jxurl);
        }
        // print(html);
        if (html.indexOf('&btwaf=') != -1) {
            html = request(jxurl + '&btwaf' + html.match(/&btwaf(.*?)"/)[1], {
                headers: {
                    'Referer': ref
                }
            })
        }
        var token_iv = html.split('_token = "')[1].split('"')[0];
        var key = CryptoJS.enc.Utf8.parse(token_key);
        var iv = CryptoJS.enc.Utf8.parse(token_iv);
        // log("iv:"+iv);
        //  log(html);
        // print(key);
        // print(iv);
        eval(html.match(/var config = {[\s\S]*?}/)[0] + '');
        // config.url = config.url.replace(/,/g,'');
        // print(config.url);
        if (!config.url.startsWith('http')) {
            //config.url = decodeURIComponent(AES(config.url, key, iv));
            config.url = CryptoJS.AES.decrypt(config.url, key, {
                iv: iv,
                padding: CryptoJS.pad.Pkcs7
            }).toString(CryptoJS.enc.Utf8)
        }
        return config.url;
    } catch (e) {
        return '';
    }
}

function urlencode (str) {
    str = (str + '').toString();
    return encodeURIComponent(str).replace(/!/g, '%21').replace(/'/g, '%27').replace(/\(/g, '%28').
    replace(/\)/g, '%29').replace(/\*/g, '%2A').replace(/%20/g, '+');
}
H
hjdhnx 已提交
378 379

function base64Encode(text){
380 381
    return CryptoJS.enc.Base64.stringify(CryptoJS.enc.Utf8.parse(text));
    // return text
H
hjdhnx 已提交
382 383
}

H
hjdhnx 已提交
384
function base64Decode(text){
385 386 387 388 389 390 391 392 393 394 395
    return CryptoJS.enc.Utf8.stringify(CryptoJS.enc.Base64.parse(text));
    // return text
}

function md5(text) {
    return CryptoJS.MD5(text).toString();
}

function getCryptoJS(){
    // return request('https://gitcode.net/qq_32394351/dr_py/-/raw/master/libs/crypto-hiker.js');
    return 'console.log("CryptoJS已装载");'
H
hjdhnx 已提交
396 397
}

H
hjdhnx 已提交
398 399
let VODS = [];// 一级或者搜索需要的数据列表
let VOD = {};// 二级的单个数据
H
hjdhnx 已提交
400 401
let TABS = [];// 二级的自定义线路列表 如: TABS=['道长在线','道长在线2']
let LISTS = [];// 二级的自定义选集播放列表 如: LISTS=[['第1集$http://1.mp4','第2集$http://2.mp4'],['第3集$http://1.mp4','第4集$http://2.mp4']]
H
hjdhnx 已提交
402 403 404
globalThis.encodeUrl = urlencode;
globalThis.urlencode = urlencode;

405 406


H
hjdhnx 已提交
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
/**
 *  url拼接
 * @param fromPath 初始当前页面url
 * @param nowPath 相对当前页面url
 * @returns {*}
 */
function urljoin(fromPath, nowPath) {
    fromPath = fromPath||'';
    nowPath = nowPath||'';
    return joinUrl(fromPath, nowPath);
    // try {
    //     // import Uri from './uri.min.js';
    //     // var Uri = require('./uri.min.js');
    //     // eval(request('https://cdn.bootcdn.net/ajax/libs/URI.js/1.19.11/URI.min.js'));
    //     // let new_uri = URI(nowPath, fromPath);

    //     let new_uri = Uri(nowPath, fromPath);
    //     new_uri = new_uri.toString();
    //     // console.log(new_uri);
    //     // return fromPath + nowPath
    //     return new_uri
    // }
    // catch (e) {
    //     console.log('urljoin发生错误:'+e.message);
    //     if(nowPath.startsWith('http')){
    //         return nowPath
    //     }if(nowPath.startsWith('/')){
    //         return getHome(fromPath)+nowPath
    //     }
    //     return fromPath+nowPath
    // }
}
439
var urljoin2 = urljoin;
H
hjdhnx 已提交
440

H
hjdhnx 已提交
441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460
// 内置 pdfh,pdfa,pd
const defaultParser = {
    pdfh:pdfh,
    pdfa:pdfa,
    pd(html,parse,uri){
        let ret = this.pdfh(html,parse);
        if(typeof(uri)==='undefined'||!uri){
            uri = '';
        }
        if(DOM_CHECK_ATTR.test(parse)){
            if(/http/.test(ret)){
                ret = ret.substr(ret.indexOf('http'));
            }else{
                ret = urljoin(MY_URL,ret)
            }
        }
        return ret
    },
};

H
hjdhnx 已提交
461

H
hjdhnx 已提交
462 463 464 465 466 467 468
/**
 *  pdfh原版优化,能取style属性里的图片链接
 * @param html 源码
 * @param parse 解析表达式
 * @returns {string|*}
 */
function pdfh2(html,parse){
469 470 471 472 473 474 475 476 477
    let html2 = html;
    try {
        if(typeof(html)!=='string'){
            html2 = html.rr(html.ele).toString();
        }
    }catch (e) {
        print('html对象转文本发生了错误:'+e.message);
    }
    let result = defaultParser.pdfh(html2,parse);
H
hjdhnx 已提交
478 479 480 481 482 483 484 485 486
    let option = parse.includes('&&')?parse.split('&&').slice(-1)[0]:parse.split(' ').slice(-1)[0];
    if(/style/.test(option.toLowerCase())&&/url\(/.test(result)){
        try {
            result =  result.match(/url\((.*?)\)/)[1];
        }catch (e) {}
    }
    return result
}

487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504
/**
 * pdfa原版优化,可以转换jq的html对象
 * @param html
 * @param parse
 * @returns {*}
 */
function pdfa2(html,parse){
    let html2 = html;
    try {
        if(typeof(html)!=='string'){
            html2 = html.rr(html.ele).toString();
        }
    }catch (e) {
        print('html对象转文本发生了错误:'+e.message);
    }
    return defaultParser.pdfa(html2,parse);
}

H
hjdhnx 已提交
505
/**
H
hjdhnx 已提交
506
 * pd原版方法重写-增加自动urljoin
H
hjdhnx 已提交
507 508 509 510 511
 * @param html
 * @param parse
 * @param uri
 * @returns {*}
 */
H
hjdhnx 已提交
512
function pd2(html,parse,uri){
H
hjdhnx 已提交
513
    let ret = pdfh2(html,parse);
H
hjdhnx 已提交
514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530
    if(typeof(uri)==='undefined'||!uri){
        uri = '';
    }
    if(DOM_CHECK_ATTR.test(parse)){
        if(/http/.test(ret)){
            ret = ret.substr(ret.indexOf('http'));
        }else{
            ret = urljoin(MY_URL,ret)
        }
    }
    // MY_URL = getItem('MY_URL',MY_URL);
    // console.log(`规则${RKEY}打印MY_URL:${MY_URL},uri:${uri}`);
    return ret
}

const parseTags = {
    jsp:{
H
hjdhnx 已提交
531
        pdfh:pdfh2,
532
        pdfa:pdfa2,
H
hjdhnx 已提交
533 534
        // pd:pd2,
        pd:pd,
H
hjdhnx 已提交
535 536 537 538 539 540
    },
    json:{
        pdfh(html, parse) {
            if (!parse || !parse.trim()){
                return '';
            }
H
hjdhnx 已提交
541 542
            if (typeof(html) === 'string'){
                // print('jsonpath:pdfh字符串转dict');
H
hjdhnx 已提交
543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569
                html = JSON.parse(html);
            }
            parse = parse.trim();
            if (!parse.startsWith('$.')){
                parse = '$.' + parse;
            }
            parse = parse.split('||');
            for (let ps of parse) {
                let ret = cheerio.jp(ps, html);
                if (Array.isArray(ret)){
                    ret = ret[0] || '';
                } else{
                    ret = ret || ''
                }
                if (ret && typeof (ret) !== 'string'){
                    ret = ret.toString();
                }
                if(ret){
                    return ret
                }
            }
            return '';
        },
        pdfa(html, parse) {
            if (!parse || !parse.trim()){
                return '';
            }
H
hjdhnx 已提交
570 571
            if (typeof(html) === 'string'){
                // print('jsonpath:pdfa字符串转dict');
H
hjdhnx 已提交
572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593
                html = JSON.parse(html);
            }
            parse = parse.trim()
            if (!parse.startsWith('$.')){
                parse = '$.' + parse;
            }
            let ret = cheerio.jp(parse, html);
            if (Array.isArray(ret) && Array.isArray(ret[0]) && ret.length === 1){
                return ret[0] || []
            }
            return ret || []
        },
        pd(html,parse){
            let ret = parseTags.json.pdfh(html,parse);
            if(ret){
                return urljoin(MY_URL,ret);
            }
            return ret
        },
    },
    jq:{
        pdfh(html, parse, base_url) {
H
hjdhnx 已提交
594
            if (!parse || !parse.trim()) {
H
hjdhnx 已提交
595 596
                return ''
            }
H
hjdhnx 已提交
597
            let eleFind = typeof html === 'object';
H
hjdhnx 已提交
598
            let option = undefined;
H
hjdhnx 已提交
599 600 601 602 603 604 605
            if (eleFind && parse.startsWith('body&&')) {
                parse = parse.substr(6);
                if (parse.indexOf('&&') < 0) {
                    option = parse.trim();
                    parse = '*=*';
                }
            }
H
hjdhnx 已提交
606 607 608 609 610 611
            if (parse.indexOf('&&') > -1) {
                let sp = parse.split('&&');
                option = sp[sp.length - 1];
                sp.splice(sp.length - 1);
                if (sp.length > 1) {
                    for (let i in sp) {
H
hjdhnx 已提交
612 613 614 615 616
                        //Javascript自定义Array.prototype干扰for-in循环
                        if(sp.hasOwnProperty(i)){
                            if (!SELECT_REGEX.test(sp[i])) {
                                sp[i] = sp[i] + ':eq(0)';
                            }
H
hjdhnx 已提交
617 618 619 620 621 622 623 624 625 626
                        }
                    }
                } else {
                    if (!SELECT_REGEX.test(sp[0])) {
                        sp[0] = sp[0] + ':eq(0)';
                    }
                }
                parse = sp.join(' ');
            }
            let result = '';
H
hjdhnx 已提交
627 628
            const $ = eleFind ? html.rr : cheerio.load(html);
            let ret = eleFind ? ((parse === '*=*' || $(html.ele).is(parse)) ? html.ele : $(html.ele).find(parse)) : $(parse);
H
hjdhnx 已提交
629
            if (option) {
H
hjdhnx 已提交
630
                if (option === 'Text') {
H
hjdhnx 已提交
631 632
                    result = $(ret).text();
                }
H
hjdhnx 已提交
633
                else if (option === 'Html') {
H
hjdhnx 已提交
634 635
                    result = $(ret).html();
                }
H
hjdhnx 已提交
636
                else {
H
hjdhnx 已提交
637
                    result = $(ret).attr(option);
H
hjdhnx 已提交
638 639 640 641 642
                    if(/style/.test(option.toLowerCase())&&/url\(/.test(result)){
                        try {
                            result =  result.match(/url\((.*?)\)/)[1];
                        }catch (e) {}
                    }
H
hjdhnx 已提交
643 644
                }
                if (result && base_url && DOM_CHECK_ATTR.test(option)) {
H
hjdhnx 已提交
645
                    if (/http/.test(result)) {
H
hjdhnx 已提交
646
                        result = result.substr(result.indexOf('http'));
H
hjdhnx 已提交
647 648
                    } else {
                        result = urljoin(base_url, result)
H
hjdhnx 已提交
649 650 651 652 653 654 655 656
                    }
                }
            } else {
                result = $(ret).toString();
            }
            return result;
        },
        pdfa(html, parse) {
H
hjdhnx 已提交
657
            if (!parse || !parse.trim()) {
H
hjdhnx 已提交
658
                print('!parse');
H
hjdhnx 已提交
659 660
                return [];
            }
H
hjdhnx 已提交
661
            let eleFind = typeof html === 'object';
H
hjdhnx 已提交
662
            // print('parse前:'+parse);
H
hjdhnx 已提交
663 664 665
            if (parse.indexOf('&&') > -1) {
                let sp = parse.split('&&');
                for (let i in sp) {
H
hjdhnx 已提交
666 667 668 669 670 671
                    if(sp.hasOwnProperty(i)){
                        if (!SELECT_REGEX_A.test(sp[i]) && i < sp.length - 1) {
                            if(sp[i]!=='body'){
                                // sp[i] = sp[i] + ':eq(0)';
                                sp[i] = sp[i] + ':first';
                            }
H
hjdhnx 已提交
672
                        }
H
hjdhnx 已提交
673 674 675 676
                    }
                }
                parse = sp.join(' ');
            }
H
hjdhnx 已提交
677
            // print('parse后:'+parse);
H
hjdhnx 已提交
678 679
            const $ = eleFind ? html.rr : cheerio.load(html);
            let ret = eleFind ? ($(html.ele).is(parse) ? html.ele : $(html.ele).find(parse)) : $(parse);
H
hjdhnx 已提交
680
            let result = [];
H
hjdhnx 已提交
681 682
            // print('outerHTML:');
            // print($(ret[0]).prop("outerHTML"));
H
hjdhnx 已提交
683 684
            if (ret) {
                ret.each(function (idx, ele) {
H
hjdhnx 已提交
685
                    result.push({ rr: $, ele: ele });
H
hjdhnx 已提交
686
                    // result.push({ rr: $, ele: $(ele).prop("outerHTML")}); // 性能贼差
H
hjdhnx 已提交
687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708
                });
            }
            return result;
        },
        pd(html,parse,uri){
            return parseTags.jq.pdfh(html, parse, MY_URL);
        },
    },
    getParse(p0){//非js开头的情况自动获取解析标签
        if(p0.startsWith('jsp:')){
            return this.jsp
        }else if(p0.startsWith('json:')){
            return this.json
        }else if(p0.startsWith('jq:')){
            return this.jq
        }else {
            return this.jq
        }
    }
};

const stringify = JSON.stringify;
H
hjdhnx 已提交
709
const jsp = parseTags.jsp;
710
const jq = parseTags.jq;
H
hjdhnx 已提交
711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729

/*** 后台需要实现的java方法并注入到js中 ***/

/**
 * 读取本地文件->应用程序目录
 * @param filePath
 * @returns {string}
 */
function readFile(filePath){
    filePath = filePath||'./uri.min.js';
    var fd = os.open(filePath);
    var buffer = new ArrayBuffer(1024);
    var len = os.read(fd, buffer, 0, 1024);
    console.log(len);
    let text = String.fromCharCode.apply(null, new Uint8Array(buffer));
    console.log(text);
    return text
}

H
hjdhnx 已提交
730 731 732 733 734 735 736
/**
 * 处理返回的json数据
 * @param html
 * @returns {*}
 */
function dealJson(html) {
    try {
737
        // html = html.match(/[\w|\W|\s|\S]*?(\{[\w|\W|\s|\S]*\})/).group[1];
H
hjdhnx 已提交
738 739 740 741
        html = html.trim();
        if(!((html.startsWith('{') && html.endsWith('}'))||(html.startsWith('[') && html.endsWith(']')))){
            html = '{'+html.match(/.*?\{(.*)\}/m)[1]+'}';
        }
H
hjdhnx 已提交
742 743
    } catch (e) {
    }
H
hjdhnx 已提交
744 745 746 747
    try {
        html = JSON.parse(html);
    }catch (e) {}
    // console.log(typeof(html));
H
hjdhnx 已提交
748 749 750
    return html;
}

H
hjdhnx 已提交
751 752 753 754 755 756 757 758 759
/**
 * 验证码识别逻辑,需要java实现(js没有bytes类型,无法调用后端的传递图片二进制获取验证码文本的接口)
 * @type {{api: string, classification: (function(*=): string)}}
 */
var OcrApi={
    api:OCR_API,
    classification:function (img){ // img是byte类型,这里不方便搞啊
        let code = '';
        try {
760
            let html = request(this.api,{data:{img:img},headers:{'User-Agent':PC_UA},'method':'POST'},true);
H
hjdhnx 已提交
761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780
            html = JSON.parse(html);
            code = html.url||'';
        }catch (e) {}
        return code
    }
};
/**
 * 验证码识别,暂未实现
 * @param url 验证码图片链接
 * @returns {string} 验证成功后的cookie
 */
function verifyCode(url){
    let cnt = 0;
    let host = getHome(url);
    let cookie = '';
    while (cnt < OCR_RETRY){
        try{
            // let obj = {headers:headers,timeout:timeout};
            let yzm_url = `${host}/index.php/verify/index.html`;
            console.log(`验证码链接:${yzm_url}`);
781
            let hhtml = request(yzm_url,{withHeaders:true,toBase64:true},true);
H
hjdhnx 已提交
782 783
            let json = JSON.parse(hhtml);
            if(!cookie){
784 785 786 787
                // print(json);
                let setCk = Object.keys(json).find(it=>it.toLowerCase()==='set-cookie');
                // cookie = json['set-cookie']?json['set-cookie'].split(';')[0]:'';
                cookie = setCk?json[setCk].split(';')[0]:'';
H
hjdhnx 已提交
788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822
            }
            // console.log(hhtml);
            console.log('cookie:'+cookie);
            let img = json.body;
            // console.log(img);
            let code = OcrApi.classification(img);
            console.log(`第${cnt+1}次验证码识别结果:${code}`);
            let submit_url = `${host}/index.php/ajax/verify_check?type=search&verify=${code}`;
            console.log(submit_url);
            let html = request(submit_url,{headers:{Cookie:cookie,'User-Agent':MOBILE_UA},'method':'POST'});
            // console.log(html);
            html = JSON.parse(html);
            if(html.msg === 'ok'){
                console.log(`第${cnt+1}次验证码提交成功`);
                return cookie // 需要返回cookie
            }else if(html.msg!=='ok'&&cnt+1>=OCR_RETRY){
                cookie = ''; // 需要清空返回cookie
            }
        }catch (e) {
            console.log(`第${cnt+1}次验证码提交失败:${e.message}`);
            if(cnt+1>=OCR_RETRY){
                cookie = '';
            }
        }
        cnt+=1
    }
    return cookie
}

/**
 * 存在数据库配置表里, key字段对应值value,没有就新增,有就更新,调用此方法会清除key对应的内存缓存
 * @param k 键
 * @param v 值
 */
function setItem(k,v){
H
hjdhnx 已提交
823 824
    local.set(RKEY,k,v);
    console.log(`规则${RKEY}设置${k} => ${v}`)
H
hjdhnx 已提交
825 826 827 828 829 830 831 832 833
}

/**
 *  获取数据库配置表对应的key字段的value,没有这个key就返回value默认传参.需要有缓存,第一次获取后会存在内存里
 * @param k 键
 * @param v 值
 * @returns {*}
 */
function getItem(k,v){
H
hjdhnx 已提交
834
    return local.get(RKEY,k) || v;
H
hjdhnx 已提交
835 836 837 838 839 840 841
}

/**
 *  删除数据库key对应的一条数据,并清除此key对应的内存缓存
 * @param k
 */
function clearItem(k){
H
hjdhnx 已提交
842
    local.delete(RKEY,k);
H
hjdhnx 已提交
843 844 845 846 847 848 849 850 851 852 853 854 855 856 857
}

/*** js自封装的方法 ***/

/**
 * 获取链接的host(带http协议的完整链接)
 * @param url 任意一个正常完整的Url,自动提取根
 * @returns {string}
 */
function getHome(url){
    if(!url){
        return ''
    }
    let tmp = url.split('//');
    url = tmp[0] + '//' + tmp[1].split('/')[0];
H
hjdhnx 已提交
858 859 860
    try {
        url = decodeURIComponent(url);
    }catch (e) {}
H
hjdhnx 已提交
861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898
    return url
}

/**
 * get参数编译链接,类似python params字典自动拼接
 * @param url 访问链接
 * @param obj 参数字典
 * @returns {*}
 */
function buildUrl(url,obj){
    obj = obj||{};
    if(url.indexOf('?')<0){
        url += '?'
    }
    let param_list = [];
    let keys = Object.keys(obj);
    keys.forEach(it=>{
        param_list.push(it+'='+obj[it])
    });
    let prs = param_list.join('&');
    if(keys.length > 0 && !url.endsWith('?')){
        url += '&'
    }
    url+=prs;
    return url
}

/**
 * 远程依赖执行函数
 * @param url 远程js地址
 */
function require(url){
    eval(request(url));
}
/**
 * 海阔网页请求函数完整封装
 * @param url 请求链接
 * @param obj 请求对象 {headers:{},method:'',timeout:5000,body:'',withHeaders:false}
899
 * @param ocr_flag 标识此flag是用于请求ocr识别的,自动过滤content-type指定编码
H
hjdhnx 已提交
900 901
 * @returns {string|string|DocumentFragment|*}
 */
902 903
function request(url,obj,ocr_flag){
    ocr_flag = ocr_flag||false;
H
hjdhnx 已提交
904
    if(typeof(obj)==='undefined'||!obj||obj==={}){
H
hjdhnx 已提交
905 906 907 908 909 910 911 912
        if(!fetch_params||!fetch_params.headers){
            let headers = {
                'User-Agent':MOBILE_UA,
            };
            if(rule.headers){
                Object.assign(headers,rule.headers);
            }
            fetch_params.headers = headers;
H
hjdhnx 已提交
913
        }
H
hjdhnx 已提交
914 915
        if(!fetch_params.headers.Referer){
            fetch_params.headers.Referer = getHome(url)
H
hjdhnx 已提交
916
        }
H
hjdhnx 已提交
917
        obj = fetch_params;
H
hjdhnx 已提交
918 919 920 921 922 923 924 925 926 927
    }else{
        let headers = obj.headers||{};
        let keys = Object.keys(headers).map(it=>it.toLowerCase());
        if(!keys.includes('user-agent')){
            headers['User-Agent'] = MOBILE_UA;
        }if(!keys.includes('referer')){
            headers['Referer'] = getHome(url);
        }
        obj.headers = headers;
    }
928
    if(rule.encoding&&rule.encoding!=='utf-8'&&!ocr_flag){
H
hjdhnx 已提交
929 930 931
        if(!obj.headers.hasOwnProperty('Content-Type')){ // 手动指定了就不管
            obj.headers["Content-Type"] = 'text/html; charset='+rule.encoding;
        }
932
    }
H
hjdhnx 已提交
933
    if(typeof(obj.body)!='undefined'&&obj.body&&typeof (obj.body)==='string'){
H
hjdhnx 已提交
934
        let data = {};
H
hjdhnx 已提交
935
        obj.body.split('&').forEach(it=>{
H
hjdhnx 已提交
936 937 938
            data[it.split('=')[0]] = it.split('=')[1]
        });
        obj.data = data;
H
hjdhnx 已提交
939 940 941 942
        delete obj.body
    }else if(typeof(obj.body)!='undefined'&&obj.body&&typeof (obj.body)==='object'){
        obj.data = obj.body;
        delete obj.body
H
hjdhnx 已提交
943 944 945 946 947 948 949 950
    }
    if(!url){
        return obj.withHeaders?'{}':''
    }
    if(obj.toBase64){ // 返回base64,用于请求图片
        obj.buffer = 2;
        delete obj.toBase64
    }
H
hjdhnx 已提交
951 952
    console.log(JSON.stringify(obj.headers));
    // console.log('request:'+url+' obj:'+JSON.stringify(obj));
H
hjdhnx 已提交
953
    console.log('request:'+url);
H
hjdhnx 已提交
954 955
    let res = req(url, obj);
    let html = res.content||'';
H
hjdhnx 已提交
956
    // console.log(html);
H
hjdhnx 已提交
957 958 959 960 961 962 963 964 965
    if(obj.withHeaders){
        let htmlWithHeaders = res.headers;
        htmlWithHeaders.body = html;
        return JSON.stringify(htmlWithHeaders);
    }else{
        return html
    }
}

H
hjdhnx 已提交
966 967 968 969 970 971 972 973 974 975 976
/**
 *  快捷post请求
 * @param url 地址
 * @param obj 对象
 * @returns {string|DocumentFragment|*}
 */
function post(url,obj){
    obj.method = 'POST';
    return request(url,obj);
}

H
hjdhnx 已提交
977 978 979 980 981 982
fetch = request;
print = function (data){
    data = data||'';
    if(typeof(data)!=='string'){
        try {
            data = JSON.stringify(data);
H
hjdhnx 已提交
983 984 985
        }catch (e) {
            console.log('print:'+e.message)
        }
H
hjdhnx 已提交
986
    }
H
hjdhnx 已提交
987 988 989 990 991
    if(typeof(data)!=='string'){
        console.log(typeof(data)+':'+data.length);
    }else{
        console.log(data);
    }
H
hjdhnx 已提交
992
}
H
md  
hjdhnx 已提交
993
log = print;
H
hjdhnx 已提交
994 995 996 997 998 999 1000 1001 1002 1003 1004
/**
 * 检查宝塔验证并自动跳过获取正确源码
 * @param html 之前获取的html
 * @param url 之前的来源url
 * @param obj 来源obj
 * @returns {string|DocumentFragment|*}
 */
function checkHtml(html,url,obj){
    if(/\?btwaf=/.test(html)){
        let btwaf = html.match(/btwaf(.*?)"/)[1];
        url = url.split('#')[0]+'?btwaf'+btwaf;
H
hjdhnx 已提交
1005
        print('宝塔验证访问链接:'+url);
H
hjdhnx 已提交
1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050
        html = request(url,obj);
    }
    return html
}

/**
 *  带一次宝塔验证的源码获取
 * @param url 请求链接
 * @param obj 请求参数
 * @returns {string|DocumentFragment}
 */
function getCode(url,obj){
    let html = request(url,obj);
    html = checkHtml(html,url,obj);
    return html
}

/**
 * 源rule专用的请求方法,自动注入cookie
 * @param url 请求链接
 * @returns {string|DocumentFragment}
 */
function getHtml(url){
    let obj = {};
    if(rule.headers){
        obj.headers = rule.headers;
    }
    let cookie = getItem(RULE_CK,'');
    if(cookie){
        if(obj.headers && ! Object.keys(obj.headers).map(it=>it.toLowerCase()).includes('cookie')){
            obj.headers['Cookie'] = cookie;
        }else if(!obj.headers){
            obj.headers = {Cookie:cookie};
        }
    }
    let html = getCode(url,obj);
    return html
}

/**
 * 首页分类解析,筛选暂未实现
 * @param homeObj 首页传参对象
 * @returns {string}
 */
function homeParse(homeObj) {
H
hjdhnx 已提交
1051
    fetch_params = JSON.parse(JSON.stringify(rule_fetch_params));
H
hjdhnx 已提交
1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066
    let classes = [];
    if (homeObj.class_name && homeObj.class_url) {
        let names = homeObj.class_name.split('&');
        let urls = homeObj.class_url.split('&');
        let cnt = Math.min(names.length, urls.length);
        for (let i = 0; i < cnt; i++) {
            classes.push({
                'type_id': urls[i],
                'type_name': names[i]
            });
        }
    }

    if (homeObj.class_parse) {
        let p = homeObj.class_parse.split(';');
1067 1068 1069 1070 1071 1072
        let _ps = parseTags.getParse(p[0]);
        _pdfa = _ps.pdfa;
        _pdfh = _ps.pdfh;
        _pd = _ps.pd;
        MY_URL = rule.url;
        if (p.length >= 3) { // 可以不写正则
H
hjdhnx 已提交
1073 1074 1075
            try {
                let html = getHtml(homeObj.MY_URL);
                if (html) {
H
hjdhnx 已提交
1076
                    homeHtmlCache = html;
1077
                    let list = _pdfa(html, p[0]);
H
hjdhnx 已提交
1078
                    if (list && list.length > 0) {
H
hjdhnx 已提交
1079
                        list.forEach((it,idex) => {
H
hjdhnx 已提交
1080
                            try {
1081
                                let name = _pdfh(it, p[1]);
H
hjdhnx 已提交
1082 1083 1084
                                if (homeObj.cate_exclude && (new RegExp(homeObj.cate_exclude).test(name))) {
                                    return;
                                }
1085 1086 1087
                                // let url = pdfh(it, p[2]);
                                let url = _pd(it, p[2]);
                                if (p.length > 3 && p[3]) {
H
hjdhnx 已提交
1088 1089 1090 1091 1092
                                    let exp = new RegExp(p[3]);
                                    url = url.match(exp)[1];
                                }

                                classes.push({
H
hjdhnx 已提交
1093 1094
                                    'type_id': url.trim(),
                                    'type_name': name.trim()
H
hjdhnx 已提交
1095 1096
                                });
                            } catch (e) {
H
hjdhnx 已提交
1097
                                console.log(`分类列表定位第${idex}个元素正常报错:${e.message}`);
H
hjdhnx 已提交
1098 1099 1100 1101 1102 1103 1104 1105 1106 1107
                            }
                        });
                    }
                }
            } catch (e) {
                console.log(e.message);
            }

        }
    }
H
hjdhnx 已提交
1108 1109
    // 排除分类
    classes = classes.filter(it=>!homeObj.cate_exclude || !(new RegExp(homeObj.cate_exclude).test(it.type_name)));
H
hjdhnx 已提交
1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120
    let resp = {
        'class': classes
    };
    if(homeObj.filter){
        resp.filters = homeObj.filter;
    }
    console.log(JSON.stringify(resp));
    return JSON.stringify(resp);

}

H
hjdhnx 已提交
1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133
/**
 * 推荐和搜索单字段继承一级
 * @param p 推荐或搜索的解析分割;列表
 * @param pn 自身列表序号
 * @param pp  一级解析分割;列表
 * @param ppn 继承一级序号
 * @returns {*}
 */
function getPP(p, pn, pp, ppn){
    let ps = p[pn] === '*' && pp.length > ppn ?pp[ppn]:p[pn]
    return ps
}

H
hjdhnx 已提交
1134 1135 1136 1137 1138 1139
/**
 *  首页推荐列表解析
 * @param homeVodObj
 * @returns {string}
 */
function homeVodParse(homeVodObj){
H
hjdhnx 已提交
1140
    fetch_params = JSON.parse(JSON.stringify(rule_fetch_params));
H
hjdhnx 已提交
1141 1142 1143 1144
    let d = [];
    MY_URL = homeVodObj.homeUrl;
    // setItem('MY_URL',MY_URL);
    console.log(MY_URL);
H
hjdhnx 已提交
1145
    let t1 = (new Date()).getTime();
H
hjdhnx 已提交
1146
    let p = homeVodObj.推荐;
H
hjdhnx 已提交
1147
    print('p:'+p);
H
hjdhnx 已提交
1148 1149 1150 1151
    if(p==='*' && rule.一级){
        p = rule.一级;
        homeVodObj.double = false;
    }
H
hjdhnx 已提交
1152
    if(!p||typeof(p)!=='string'){
H
hjdhnx 已提交
1153 1154
        return '{}'
    }
H
hjdhnx 已提交
1155
    p = p.trim();
H
hjdhnx 已提交
1156
    let pp = rule.一级.split(';');
H
hjdhnx 已提交
1157
    if(p.startsWith('js:')){
H
hjdhnx 已提交
1158 1159 1160
        const TYPE = 'home';
        var input = MY_URL;
        const HOST = rule.host;
H
hjdhnx 已提交
1161
        eval(p.replace('js:',''));
H
hjdhnx 已提交
1162 1163 1164 1165 1166 1167 1168 1169
        d = VODS;
    }else {
        p = p.split(';');
        if (!homeVodObj.double && p.length < 5) {
            return '{}'
        } else if (homeVodObj.double && p.length < 6) {
            return '{}'
        }
H
hjdhnx 已提交
1170 1171
        let p0 = getPP(p,0,pp,0)
        let _ps = parseTags.getParse(p0);
H
hjdhnx 已提交
1172 1173 1174
        _pdfa = _ps.pdfa;
        _pdfh = _ps.pdfh;
        _pd = _ps.pd;
H
hjdhnx 已提交
1175 1176
        let is_json = p0.startsWith('json:');
        p0 = p0.replace(/^(jsp:|json:|jq:)/,'');
H
hjdhnx 已提交
1177
        // print(p[0]);
H
hjdhnx 已提交
1178 1179
        let html = homeHtmlCache || getHtml(MY_URL);
        homeHtmlCache = undefined;
H
hjdhnx 已提交
1180
        if(is_json){
H
hjdhnx 已提交
1181
            // print('是json,开始处理');
H
hjdhnx 已提交
1182 1183
            html = dealJson(html);
        }
H
hjdhnx 已提交
1184 1185 1186
        try {
            console.log('double:' + homeVodObj.double);
            if (homeVodObj.double) {
H
hjdhnx 已提交
1187
                let items = _pdfa(html, p0);
H
hjdhnx 已提交
1188 1189 1190
                // console.log(items.length);
                for (let item of items) {
                    // console.log(p[1]);
H
hjdhnx 已提交
1191
                    let items2 = _pdfa(item, p[1]);
H
hjdhnx 已提交
1192 1193 1194
                    // console.log(items2.length);
                    for (let item2 of items2) {
                        try {
H
hjdhnx 已提交
1195 1196
                            let p2 = getPP(p,2,pp,1);
                            let title = _pdfh(item2, p2);
H
hjdhnx 已提交
1197 1198
                            let img = '';
                            try {
H
hjdhnx 已提交
1199 1200 1201 1202 1203 1204 1205 1206 1207
                                let p3 = getPP(p,3,pp,2);
                                img = _pd(item2, p3);
                            } catch (e) {}
                            let desc = '';
                            try {
                                let p4 = getPP(p,4,pp,3);
                                desc = _pdfh(item2, p4);
                            }catch (e) {}
                            let p5 = getPP(p,5,pp,4);
H
hjdhnx 已提交
1208
                            let links = [];
H
hjdhnx 已提交
1209 1210
                            for (let _p5 of p5.split('+')) {
                                let link = !homeVodObj.detailUrl ? _pd(item2, _p5, MY_URL) : _pdfh(item2, _p5);
H
hjdhnx 已提交
1211 1212
                                links.push(link);
                            }
H
hjdhnx 已提交
1213 1214 1215 1216 1217 1218 1219
                            let content;
                            if(p.length > 6 && p[6]){
                                let p6 = getPP(p,6,pp,5);
                                content = _pdfh(item2, p6);
                            } else{
                                content = '';
                            }
H
hjdhnx 已提交
1220 1221 1222 1223
                            let vid = links.join('$');
                            if(rule.二级==='*'){
                                vid = vid+'@@'+title+'@@'+img;
                            }
H
hjdhnx 已提交
1224 1225 1226 1227
                            let vod = {
                                vod_name: title,
                                vod_pic: img,
                                vod_remarks: desc,
H
hjdhnx 已提交
1228
                                vod_content: content,
H
hjdhnx 已提交
1229
                                vod_id: vid
H
hjdhnx 已提交
1230
                            };
H
hjdhnx 已提交
1231
                            // print(vod);
H
hjdhnx 已提交
1232 1233
                            d.push(vod);
                        } catch (e) {
H
hjdhnx 已提交
1234
                            console.log('首页列表处理发生错误:'+e.message);
H
hjdhnx 已提交
1235 1236 1237 1238 1239 1240 1241 1242 1243
                        }

                    }


                }


            } else {
H
hjdhnx 已提交
1244
                let items = _pdfa(html, p0);
H
hjdhnx 已提交
1245 1246
                for (let item of items) {
                    try {
H
hjdhnx 已提交
1247 1248
                        let p1 = getPP(p,1,pp,1);
                        let title = _pdfh(item, p1);
H
hjdhnx 已提交
1249 1250
                        let img = '';
                        try {
H
hjdhnx 已提交
1251 1252 1253 1254 1255 1256 1257 1258 1259
                            let p2 = getPP(p,2,pp,2);
                            img = _pd(item, p2, MY_URL);
                        } catch (e) {}
                        let desc = '';
                        try {
                            let p3 = getPP(p,3,pp,3);
                            desc = _pdfh(item, p3);
                        }catch (e) {}
                        let p4 = getPP(p,4,pp,4);
H
hjdhnx 已提交
1260
                        let links = [];
H
hjdhnx 已提交
1261 1262
                        for (let _p5 of p4.split('+')) {
                            let link = !homeVodObj.detailUrl ? _pd(item, _p5, MY_URL) : _pdfh(item, _p5);
H
hjdhnx 已提交
1263 1264
                            links.push(link);
                        }
H
hjdhnx 已提交
1265 1266 1267 1268 1269 1270 1271
                        let content;
                        if(p.length > 5 && p[5]){
                            let p5 = getPP(p,5,pp,5);
                            content = _pdfh(item, p5);
                        }else{
                            content = ''
                        }
H
hjdhnx 已提交
1272 1273 1274 1275
                        let vid = links.join('$');
                        if(rule.二级==='*'){
                            vid = vid+'@@'+title+'@@'+img;
                        }
H
hjdhnx 已提交
1276 1277 1278 1279
                        let vod = {
                            vod_name: title,
                            vod_pic: img,
                            vod_remarks: desc,
H
hjdhnx 已提交
1280
                            vod_content: content,
H
hjdhnx 已提交
1281
                            vod_id: vid
H
hjdhnx 已提交
1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296
                        };
                        d.push(vod);

                    } catch (e) {

                    }

                }

            }

        } catch (e) {

        }
    }
H
hjdhnx 已提交
1297 1298
    let t2 = (new Date()).getTime();
    console.log('加载首页推荐耗时:'+(t2-t1)+'毫秒');
H
hjdhnx 已提交
1299
    // console.log(JSON.stringify(d));
H
hjdhnx 已提交
1300 1301 1302 1303 1304 1305 1306
    if(rule.图片来源){
        d.forEach(it=>{
            if(it.vod_pic&&it.vod_pic.startsWith('http')){
                it.vod_pic = it.vod_pic + rule.图片来源;
            }
        });
    }
H
hjdhnx 已提交
1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317
    return JSON.stringify({
        list:d
    })
}

/**
 * 一级分类页数据解析
 * @param cateObj
 * @returns {string}
 */
function categoryParse(cateObj) {
H
hjdhnx 已提交
1318
    fetch_params = JSON.parse(JSON.stringify(rule_fetch_params));
H
hjdhnx 已提交
1319
    let p = cateObj.一级;
H
hjdhnx 已提交
1320 1321 1322
    if(!p||typeof(p)!=='string'){
        return '{}'
    }
H
hjdhnx 已提交
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336
    let d = [];
    // let url = cateObj.url.replaceAll('fyclass', cateObj.tid).replaceAll('fypage', cateObj.pg);
    let url = cateObj.url.replaceAll('fyclass', cateObj.tid);
    if(rule.filter_url){
        if(!/fyfilter/.test(url)){
            if(!url.endsWith('&')&&!rule.filter_url.startsWith('&')){
                url+='&'
            }
            url+=rule.filter_url;
        }else{
            url = url.replace('fyfilter', rule.filter_url);
        }
        // console.log('filter:'+cateObj.filter);
        let fl = cateObj.filter?cateObj.extend:{};
H
hjdhnx 已提交
1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351
        // 自动合并 不同分类对应的默认筛选
        if(rule.filter_def && typeof(rule.filter_def)==='object'){
            try {
                if(Object.keys(rule.filter_def).length>0 && rule.filter_def.hasOwnProperty(cateObj.tid)){
                    let self_fl_def = rule.filter_def[cateObj.tid];
                    if(self_fl_def && typeof(self_fl_def)==='object'){
                        // 引用传递转值传递,避免污染self变量
                        let fl_def = JSON.parse(JSON.stringify(self_fl_def));
                        fl = Object.assign(fl_def,fl);
                    }
                }
            }catch (e) {
                print('合并不同分类对应的默认筛选出错:'+e.message);
            }
        }
H
hjdhnx 已提交
1352 1353
        let new_url;
        new_url = cheerio.jinja2(url,{fl:fl});
1354
        // console.log('jinjia2执行后的new_url类型为:'+typeof(new_url));
H
hjdhnx 已提交
1355 1356 1357 1358
        if(/object Object/.test(new_url)){
            new_url = drT.renderText(url,fl);
        }
        url = new_url;
H
hjdhnx 已提交
1359 1360 1361 1362
    }
    if(/fypage/.test(url)){
        if(url.includes('(')&&url.includes(')')){
            let url_rep = url.match(/.*?\((.*)\)/)[1];
H
hjdhnx 已提交
1363
            // console.log(url_rep);
H
hjdhnx 已提交
1364
            let cnt_page = url_rep.replaceAll('fypage', cateObj.pg);
H
hjdhnx 已提交
1365 1366 1367
            // console.log(cnt_page);
            let cnt_pg = eval(cnt_page);
            // console.log(cnt_pg);
H
hjdhnx 已提交
1368 1369 1370 1371 1372 1373 1374
            url = url.replaceAll(url_rep,cnt_pg).replaceAll('(','').replaceAll(')','');
        }else{
            url = url.replaceAll('fypage',cateObj.pg);
        }
    }
    if(cateObj.pg === 1 && url.includes('[')&&url.includes(']')){
        url = url.split('[')[1].split(']')[0];
H
hjdhnx 已提交
1375 1376
    }else if(cateObj.pg > 1 && url.includes('[')&&url.includes(']')){
        url = url.split('[')[0];
H
hjdhnx 已提交
1377 1378 1379 1380
    }
    MY_URL = url;
    // setItem('MY_URL',MY_URL);
    console.log(MY_URL);
H
hjdhnx 已提交
1381
    p = p.trim();
1382
    const MY_CATE = cateObj.tid;
H
hjdhnx 已提交
1383
    if(p.startsWith('js:')){
H
hjdhnx 已提交
1384
        var MY_FL = cateObj.extend;
H
hjdhnx 已提交
1385 1386 1387
        const TYPE = 'cate';
        var input = MY_URL;
        const MY_PAGE = cateObj.pg;
H
hjdhnx 已提交
1388
        var desc = '';
H
hjdhnx 已提交
1389 1390 1391 1392 1393 1394 1395
        eval(p.trim().replace('js:',''));
        d = VODS;
    }else {
        p = p.split(';');
        if (p.length < 5) {
            return '{}'
        }
H
hjdhnx 已提交
1396 1397 1398 1399
        let _ps = parseTags.getParse(p[0]);
        _pdfa = _ps.pdfa;
        _pdfh = _ps.pdfh;
        _pd = _ps.pd;
H
hjdhnx 已提交
1400
        let is_json = p[0].startsWith('json:');
H
hjdhnx 已提交
1401
        p[0] = p[0].replace(/^(jsp:|json:|jq:)/,'');
H
hjdhnx 已提交
1402 1403 1404
        try {
            let html = getHtml(MY_URL);
            if (html) {
H
hjdhnx 已提交
1405 1406 1407
                if(is_json){
                    html = dealJson(html);
                }
H
hjdhnx 已提交
1408
                let list = _pdfa(html, p[0]);
H
hjdhnx 已提交
1409
                list.forEach(it => {
1410
                    let links = p[4].split('+').map(p4=>{
1411
                        return !rule.detailUrl?_pd(it, p4,MY_URL):_pdfh(it, p4);
1412 1413 1414
                    });
                    let link = links.join('$');
                    let vod_id = rule.detailUrl?MY_CATE+'$'+link:link;
H
hjdhnx 已提交
1415 1416 1417 1418 1419 1420 1421

                    let vod_name = _pdfh(it, p[1]).replace(/\n|\t/g,'').trim();
                    let vod_pic = _pd(it, p[2],MY_URL);

                    if(rule.二级==='*'){
                        vod_id = vod_id+'@@'+vod_name+'@@'+vod_pic;
                    }
H
hjdhnx 已提交
1422
                    d.push({
1423
                        'vod_id': vod_id,
H
hjdhnx 已提交
1424 1425
                        'vod_name': vod_name,
                        'vod_pic': vod_pic,
H
hjdhnx 已提交
1426
                        'vod_remarks': _pdfh(it, p[3]).replace(/\n|\t/g,'').trim(),
H
hjdhnx 已提交
1427 1428 1429 1430 1431 1432 1433
                    });
                });
            }
        } catch (e) {
            console.log(e.message);
        }
    }
H
hjdhnx 已提交
1434 1435 1436 1437 1438 1439 1440
    if(rule.图片来源){
        d.forEach(it=>{
            if(it.vod_pic&&it.vod_pic.startsWith('http')){
                it.vod_pic = it.vod_pic + rule.图片来源;
            }
        });
    }
1441
    // print(d);
H
hjdhnx 已提交
1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456
    return d.length<1?'{}':JSON.stringify({
        'page': parseInt(cateObj.pg),
        'pagecount': 999,
        'limit': 20,
        'total': 999,
        'list': d,
    });
}

/**
 * 搜索列表数据解析
 * @param searchObj
 * @returns {string}
 */
function searchParse(searchObj) {
H
hjdhnx 已提交
1457
    fetch_params = JSON.parse(JSON.stringify(rule_fetch_params));
H
hjdhnx 已提交
1458 1459 1460 1461 1462
    let d = [];
    if(!searchObj.searchUrl){
        return '{}'
    }
    let p = searchObj.搜索==='*'&&rule.一级 ? rule.一级 : searchObj.搜索;
H
hjdhnx 已提交
1463 1464 1465 1466
    if(!p||typeof(p)!=='string'){
        return '{}'
    }
    p = p.trim();
H
hjdhnx 已提交
1467
    let pp = rule.一级.split(';');
H
hjdhnx 已提交
1468 1469 1470
    let url = searchObj.searchUrl.replaceAll('**', searchObj.wd).replaceAll('fypage', searchObj.pg);
    MY_URL = url;
    console.log(MY_URL);
H
hjdhnx 已提交
1471
    // log(searchObj.搜索);
H
hjdhnx 已提交
1472
    // setItem('MY_URL',MY_URL);
H
hjdhnx 已提交
1473
    if(p.startsWith('js:')){
H
hjdhnx 已提交
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485
        const TYPE = 'search';
        const MY_PAGE = searchObj.pg;
        const KEY = searchObj.wd;
        var input = MY_URL;
        var detailUrl = rule.detailUrl||'';
        eval(p.trim().replace('js:',''));
        d = VODS;
    }else{
        p = p.split(';');
        if (p.length < 5) {
            return '{}'
        }
H
hjdhnx 已提交
1486 1487
        let p0 = getPP(p,0,pp,0);
        let _ps = parseTags.getParse(p0);
H
hjdhnx 已提交
1488 1489 1490
        _pdfa = _ps.pdfa;
        _pdfh = _ps.pdfh;
        _pd = _ps.pd;
H
hjdhnx 已提交
1491 1492
        let is_json = p0.startsWith('json:');
        p0 = p0.replace(/^(jsp:|json:|jq:)/,'');
H
hjdhnx 已提交
1493
        // print('1381 p0:'+p0);
H
hjdhnx 已提交
1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511
        try {
            let html = getHtml(MY_URL);
            if (html) {
                if(/系统安全验证|输入验证码/.test(html)){
                    let cookie = verifyCode(MY_URL);
                    if(cookie){
                        console.log(`本次成功过验证,cookie:${cookie}`);
                        setItem(RULE_CK,cookie);
                    }else{
                        console.log(`本次自动过搜索验证失败,cookie:${cookie}`);
                    }
                    // obj.headers['Cookie'] = cookie;
                    html = getHtml(MY_URL);
                }
                if(!html.includes(searchObj.wd)){
                    console.log('搜索结果源码未包含关键字,疑似搜索失败,正为您打印结果源码');
                    console.log(html);
                }
H
hjdhnx 已提交
1512
                if(is_json){
H
hjdhnx 已提交
1513
                    // console.log(html);
H
hjdhnx 已提交
1514 1515
                    html = dealJson(html);
                }
H
hjdhnx 已提交
1516 1517
                console.log(JSON.stringify(html));
                console.log(html);
H
hjdhnx 已提交
1518
                let list = _pdfa(html, p0);
H
hjdhnx 已提交
1519 1520
                // print(list.length);
                // print(list);
H
hjdhnx 已提交
1521
                list.forEach(it => {
H
hjdhnx 已提交
1522 1523 1524 1525 1526 1527
                    let p1 = getPP(p, 1, pp, 1);
                    let p2 = getPP(p, 2, pp, 2);
                    let p3 = getPP(p, 3, pp, 3);
                    let p4 = getPP(p, 4, pp, 4);
                    let links = p4.split('+').map(_p4=>{
                        return !rule.detailUrl?_pd(it, _p4,MY_URL):_pdfh(it, _p4)
1528 1529
                    });
                    let link = links.join('$');
H
hjdhnx 已提交
1530 1531 1532
                    let content;
                    if(p.length > 5 && p[5]){
                        let p5 = getPP(p,5,pp,5);
H
hjdhnx 已提交
1533
                        content = _pdfh(it, p5);
H
hjdhnx 已提交
1534 1535 1536
                    }else{
                        content = '';
                    }
H
hjdhnx 已提交
1537 1538 1539 1540 1541 1542
                    let vod_id = link;
                    let vod_name = _pdfh(it, p1).replace(/\n|\t/g,'').trim();
                    let vod_pic = _pd(it, p2,MY_URL);
                    if(rule.二级==='*'){
                        vod_id = vod_id+'@@'+vod_name+'@@'+vod_pic;
                    }
H
hjdhnx 已提交
1543
                    let ob = {
H
hjdhnx 已提交
1544 1545 1546
                        'vod_id': vod_id,
                        'vod_name': vod_name,
                        'vod_pic': vod_pic,
H
hjdhnx 已提交
1547 1548
                        'vod_remarks': _pdfh(it, p3).replace(/\n|\t/g,'').trim(),
                        'vod_content': content.replace(/\n|\t/g,'').trim(),
H
hjdhnx 已提交
1549 1550 1551 1552 1553 1554
                    };
                    d.push(ob);
                });

            }
        } catch (e) {
H
hjdhnx 已提交
1555
            print('搜索发生错误:'+e.message);
H
hjdhnx 已提交
1556 1557 1558
            return '{}'
        }
    }
H
hjdhnx 已提交
1559 1560 1561 1562 1563 1564 1565 1566
    if(rule.图片来源){
        d.forEach(it=>{
            if(it.vod_pic&&it.vod_pic.startsWith('http')){
                it.vod_pic = it.vod_pic + rule.图片来源;
            }
        });
    }
    // print(d);
H
hjdhnx 已提交
1567 1568 1569 1570 1571 1572 1573
    return JSON.stringify({
        'page': parseInt(searchObj.pg),
        'pagecount': 10,
        'limit': 20,
        'total': 100,
        'list': d,
    });
H
hjdhnx 已提交
1574 1575 1576 1577 1578 1579 1580 1581
}

/**
 * 二级详情页数据解析
 * @param detailObj
 * @returns {string}
 */
function detailParse(detailObj){
H
hjdhnx 已提交
1582
    let t1 = (new Date()).getTime();
H
hjdhnx 已提交
1583
    fetch_params = JSON.parse(JSON.stringify(rule_fetch_params));
H
hjdhnx 已提交
1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594
    let orId = detailObj.orId;
    let vod_name = '片名';
    let vod_pic = '';
    let vod_id = orId;
    if(rule.二级==='*'){
        // vod_id = orId.split('@@')[0]; // 千万不能分割
        let extra = orId.split('@@');
        vod_name = extra.length>1?extra[1]:vod_name;
        vod_pic = extra.length>2?extra[2]:vod_pic;
    }
    // print(vod_pic);
H
hjdhnx 已提交
1595
    let vod = {
H
hjdhnx 已提交
1596 1597 1598 1599
        vod_id: vod_id, //"id",
        vod_name: vod_name,
        vod_pic: vod_pic,
        type_name: "类型",
H
hjdhnx 已提交
1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613
        vod_year: "年份",
        vod_area: "地区",
        vod_remarks: "更新信息",
        vod_actor: "主演",
        vod_director: "导演",
        vod_content: "简介"
    };
    let p = detailObj.二级;
    let url = detailObj.url;
    let detailUrl = detailObj.detailUrl;
    let fyclass = detailObj.fyclass;
    let tab_exclude = detailObj.tab_exclude;
    let html = detailObj.html||'';
    MY_URL = url;
H
hjdhnx 已提交
1614
    // console.log(MY_URL);
H
hjdhnx 已提交
1615 1616 1617 1618 1619 1620
    // setItem('MY_URL',MY_URL);
    if(p==='*'){
        vod.vod_play_from = '道长在线';
        vod.vod_remarks = detailUrl;
        vod.vod_actor = '没有二级,只有一级链接直接嗅探播放';
        vod.vod_content = MY_URL;
H
hjdhnx 已提交
1621
        vod.vod_play_url = '嗅探播放$' + MY_URL.split('@@')[0];
H
hjdhnx 已提交
1622 1623 1624
    }else if(typeof(p)==='string'&&p.trim().startsWith('js:')){
        const TYPE = 'detail';
        var input = MY_URL;
H
hjdhnx 已提交
1625
        var play_url = '';
H
hjdhnx 已提交
1626
        eval(p.trim().replace('js:',''));
1627
        vod = VOD;
H
hjdhnx 已提交
1628
        console.log(JSON.stringify(vod));
H
hjdhnx 已提交
1629 1630 1631 1632
    }else if(p&&typeof(p)==='object'){
        if(!html){
            html = getHtml(MY_URL);
        }
H
hjdhnx 已提交
1633
        let _impJQP = true;
H
hjdhnx 已提交
1634 1635 1636
        let _ps;
        if(p.is_json){
            _ps = parseTags.json;
H
hjdhnx 已提交
1637
            html = dealJson(html);
H
hjdhnx 已提交
1638
            _impJQP = false;
H
hjdhnx 已提交
1639 1640 1641 1642 1643 1644 1645
        }else if(p.is_jsp){
            _ps = parseTags.jsp;
        }else if(p.is_jq){
            _ps = parseTags.jq;
        }else{
            _ps = parseTags.jq;
        }
H
hjdhnx 已提交
1646 1647 1648 1649
        if (_impJQP) {
            let c$ = cheerio.load(html);
            html = { rr: c$, ele: c$('html')[0] }
        }
H
hjdhnx 已提交
1650 1651 1652
        _pdfa = _ps.pdfa;
        _pdfh = _ps.pdfh;
        _pd = _ps.pd;
H
hjdhnx 已提交
1653 1654
        if(p.title){
            let p1 = p.title.split(';');
H
hjdhnx 已提交
1655 1656
            vod.vod_name = _pdfh(html, p1[0]).replace(/\n|\t/g,'').trim();
            let type_name = p1.length > 1 ? _pdfh(html, p1[1]).replace(/\n|\t/g,'').replace(/ /g,'').trim():'';
H
hjdhnx 已提交
1657 1658 1659 1660 1661
            vod.type_name = type_name||vod.type_name;
        }
        if(p.desc){
            try{
                let p1 = p.desc.split(';');
H
hjdhnx 已提交
1662 1663 1664 1665 1666 1667
                vod.vod_remarks =  _pdfh(html, p1[0]).replace(/\n|\t/g,'').trim();
                vod.vod_year = p1.length > 1 ? _pdfh(html, p1[1]).replace(/\n|\t/g,'').trim():'';
                vod.vod_area = p1.length > 2 ? _pdfh(html, p1[2]).replace(/\n|\t/g,'').trim():'';
                // vod.vod_actor = p1.length > 3 ? _pdfh(html, p1[3]).replaceAll('\n', ' ').trim():'';
                vod.vod_actor = p1.length > 3 ? _pdfh(html, p1[3]).replace(/\n|\t/g,'').trim():'';
                vod.vod_director = p1.length > 4 ? _pdfh(html, p1[4]).replace(/\n|\t/g,'').trim():'';
H
hjdhnx 已提交
1668 1669 1670 1671 1672 1673 1674 1675
            }
            catch (e) {

            }
        }
        if(p.content){
            try{
                let p1 = p.content.split(';');
H
hjdhnx 已提交
1676
                vod.vod_content =  _pdfh(html, p1[0]).replace(/\n|\t/g,'').trim();
H
hjdhnx 已提交
1677 1678 1679 1680 1681 1682
            }
            catch (e) {}
        }
        if(p.img){
            try{
                let p1 = p.img.split(';');
H
hjdhnx 已提交
1683
                vod.vod_pic =  _pd(html, p1[0],MY_URL);
H
hjdhnx 已提交
1684 1685 1686 1687 1688 1689 1690
            }
            catch (e) {}
        }

        let vod_play_from = '$$$';
        let playFrom = [];
        if(p.重定向&&p.重定向.startsWith('js:')){
1691
            print('开始执行重定向代码:'+p.重定向);
H
hjdhnx 已提交
1692
            html = eval(p.重定向.replace('js:',''));
H
hjdhnx 已提交
1693 1694 1695 1696
            if (_impJQP) {
                let c$ = cheerio.load(html);
                html = { rr: c$, ele: c$('html')[0] }
            }
H
hjdhnx 已提交
1697 1698 1699 1700
        }
        
// console.log(2);
        if(p.tabs){
H
hjdhnx 已提交
1701 1702 1703
            if(p.tabs.startsWith('js:')){
                print('开始执行tabs代码:'+p.tabs);
                if(html&&typeof (html)!=='string'){
H
升级  
hjdhnx 已提交
1704 1705 1706
                    try { // 假装是jq的对象拿来转换一下字符串,try为了防止json的情况报错
                        html = html.rr(html.ele).toString();
                    }catch (e) {}
H
hjdhnx 已提交
1707 1708 1709 1710 1711 1712
                }
                var input = MY_URL;
                eval(p.tabs.replace('js:',''));
                playFrom = TABS;
            }else{
                let p_tab = p.tabs.split(';')[0];
H
hjdhnx 已提交
1713
                // console.log(p_tab);
H
hjdhnx 已提交
1714 1715
                let vHeader = _pdfa(html, p_tab);
                console.log(vHeader.length);
H
hjdhnx 已提交
1716 1717
                let tab_text = p.tab_text||'body&&Text';
                // print('tab_text:'+tab_text);
H
hjdhnx 已提交
1718
                for(let v of vHeader){
H
hjdhnx 已提交
1719
                    let v_title = _pdfh(v,tab_text).trim();
H
hjdhnx 已提交
1720 1721 1722 1723 1724
                    console.log(v_title);
                    if(tab_exclude&& (new RegExp(tab_exclude)).test(v_title)){
                        continue;
                    }
                    playFrom.push(v_title);
H
hjdhnx 已提交
1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736
                }
            }
            console.log(JSON.stringify(playFrom));
        }else{
            playFrom = ['道长在线']
        }
        vod.vod_play_from = playFrom.join(vod_play_from);

// console.log(3);
        let vod_play_url = '$$$';
        let vod_tab_list = [];
        if(p.lists){
H
hjdhnx 已提交
1737 1738
            if(p.lists.startsWith('js:')){
                print('开始执行lists代码:'+p.lists);
H
hjdhnx 已提交
1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761
                try {
                    if(html&&typeof (html)!=='string'){
                        // 假装是jq的对象拿来转换一下字符串,try为了防止json的情况报错
                        try {
                            html = html.rr(html.ele).toString();
                        }catch (e) {}
                    }
                    var input = MY_URL;
                    var play_url = '';
                    eval(p.lists.replace('js:',''));
                    for(let i in LISTS){
                        if(LISTS.hasOwnProperty(i)){
                            // print(i);
                            try {
                                LISTS[i] = LISTS[i].map(it=>it.split('$').slice(0,2).join('$'));
                            }catch (e) {
                                print('格式化LISTS发生错误:'+e.message);
                            }
                        }
                    }
                    vod_play_url = LISTS.map(it=>it.join('#')).join(vod_play_url);
                }catch (e) {
                    print('js执行lists: 发生错误:'+e.message);
H
hjdhnx 已提交
1762
                }
H
hjdhnx 已提交
1763

H
hjdhnx 已提交
1764
            }else{
H
hjdhnx 已提交
1765 1766 1767 1768
                let list_text = p.list_text||'body&&Text';
                let list_url = p.list_url||'a&&href';
                // print('list_text:'+list_text);
                // print('list_url:'+list_url);
H
md  
hjdhnx 已提交
1769
                // print('list_parse:'+p.lists);
H
hjdhnx 已提交
1770
                let is_tab_js = p.tabs.trim().startsWith('js:');
H
hjdhnx 已提交
1771 1772
                for(let i=0;i<playFrom.length;i++){
                    let tab_name = playFrom[i];
H
hjdhnx 已提交
1773
                    let tab_ext =  p.tabs.split(';').length > 1 && !is_tab_js ? p.tabs.split(';')[1] : '';
H
hjdhnx 已提交
1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785
                    let p1 = p.lists.replaceAll('#idv', tab_name).replaceAll('#id', i);
                    tab_ext = tab_ext.replaceAll('#idv', tab_name).replaceAll('#id', i);
                    console.log(p1);
                    // console.log(html);
                    let vodList = [];
                    try {
                        vodList =  _pdfa(html, p1);
                        console.log('len(vodList):'+vodList.length);
                    }catch (e) {
                        // console.log(e.message);
                    }
                    let new_vod_list = [];
H
hjdhnx 已提交
1786
                    // print('tab_ext:'+tab_ext);
H
hjdhnx 已提交
1787 1788 1789
                    let tabName = tab_ext?_pdfh(html, tab_ext):tab_name;
                    console.log(tabName);
                    // console.log('cheerio解析Text');
H
hjdhnx 已提交
1790 1791 1792 1793
                    // 此处存在性能问题: pt版2000集需要650毫秒,俊版1300毫秒 特么的优化不动 主要后面定位url的我拿他没法
                    // 主要性能问题在于 _pd(it, list_url, MY_URL)
                    let tt1 = (new Date()).getTime();
                    vodList.forEach((it,idex)=>{
H
hjdhnx 已提交
1794 1795 1796
                        // 请注意,这里要固定pdfh解析body&&Text,不需要下划线,没写错
                        // new_vod_list.push(pdfh(it,'body&&Text')+'$'+_pd(it,'a&&href',MY_URL));
                        // new_vod_list.push(cheerio.load(it).text()+'$'+_pd(it,'a&&href',MY_URL));
H
hjdhnx 已提交
1797
                        // new_vod_list.push(_pdfh(it, list_text).trim() + '$' + _pd(it, list_url, MY_URL));
H
hjdhnx 已提交
1798 1799 1800
                        // new_vod_list.push(_pdfh(it, list_text).trim() + '$' +idex);
                        // new_vod_list.push(idex + '$' +_pdfh(it, list_url));
                        new_vod_list.push(_pdfh(it, list_text).trim() + '$' +_pd(it, list_url,MY_URL));
H
hjdhnx 已提交
1801
                    });
H
hjdhnx 已提交
1802
                    console.log(`drpy影响性能代码共计列表数循环次数:${vodList.length},耗时:${(new Date()).getTime()-tt1}毫秒`);
H
hjdhnx 已提交
1803 1804 1805 1806
                    let vlist = new_vod_list.join('#');
                    vod_tab_list.push(vlist);
                }
                vod_play_url = vod_tab_list.join(vod_play_url);
H
hjdhnx 已提交
1807 1808
            }
        }
H
hjdhnx 已提交
1809
        vod.vod_play_url = vod_play_url;
H
hjdhnx 已提交
1810
    }
H
hjdhnx 已提交
1811 1812 1813
    if(rule.图片来源 && vod.vod_pic && vod.vod_pic.startsWith('http')){
        vod.vod_pic = vod.vod_pic + rule.图片来源;
    }
H
hjdhnx 已提交
1814
    if(!vod.vod_id){
H
hjdhnx 已提交
1815
        vod.vod_id = vod_id;
H
hjdhnx 已提交
1816
    }
H
hjdhnx 已提交
1817 1818
    let t2 = (new Date()).getTime();
    console.log(`加载二级界面${MY_URL}耗时:${t2-t1}毫秒`);
H
hjdhnx 已提交
1819
    // print(vod);
H
hjdhnx 已提交
1820 1821 1822 1823 1824
    return JSON.stringify({
        list: [vod]
    })
}

1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837
/**
 * 判断是否需要解析
 * @param url
 * @returns {number|number}
 */
function tellIsJx(url){
    try {
        let is_vip = !/\.(m3u8|mp4|m4a)$/.test(url.split('?')[0]) && 是否正版(url);
        return is_vip?1:0
    }catch (e) {
        return 1
    }
}
H
hjdhnx 已提交
1838 1839 1840 1841 1842 1843
/**
 * 选集播放点击事件解析
 * @param playObj
 * @returns {string}
 */
function playParse(playObj){
H
hjdhnx 已提交
1844
    fetch_params = JSON.parse(JSON.stringify(rule_fetch_params));
H
hjdhnx 已提交
1845
    MY_URL = playObj.url;
H
hjdhnx 已提交
1846 1847 1848 1849 1850 1851 1852
    if(!/http/.test(MY_URL)){
        try {
            MY_URL = base64Decode(MY_URL);
        }catch (e) {}
    }
    MY_URL = decodeURIComponent(MY_URL);
    var input = MY_URL;//注入给免嗅js
H
hjdhnx 已提交
1853 1854
    let common_play = {
        parse:1,
1855
        url:input,
H
hjdhnx 已提交
1856
        // url:urlencode(input),
1857
        jx:tellIsJx(input)
H
hjdhnx 已提交
1858 1859 1860 1861 1862 1863
    };
    let lazy_play;
    if(!rule.play_parse||!rule.lazy){
        lazy_play =  common_play;
    }else if(rule.play_parse&&rule.lazy&&typeof(rule.lazy)==='string'){
        try {
H
hjdhnx 已提交
1864 1865 1866
            let lazy_code = rule.lazy.replace('js:','').trim();
            print('开始执行js免嗅=>'+lazy_code);
            eval(lazy_code);
H
hjdhnx 已提交
1867 1868
            lazy_play = typeof(input) === 'object'?input:{
                parse:1,
1869
                jx:tellIsJx(input),
H
hjdhnx 已提交
1870 1871 1872
                url:input
            };
        }catch (e) {
H
hjdhnx 已提交
1873
            print('js免嗅错误:'+e.message);
H
hjdhnx 已提交
1874 1875 1876 1877 1878
            lazy_play =  common_play;
        }
    }else{
        lazy_play =  common_play;
    }
H
升级  
hjdhnx 已提交
1879 1880
    // print('play_json:'+typeof(rule.play_json));
    // console.log(Array.isArray(rule.play_json));
H
hjdhnx 已提交
1881 1882 1883 1884 1885 1886
    if(Array.isArray(rule.play_json) && rule.play_json.length >0){ // 数组情况判断长度大于0
        let web_url = lazy_play.url;
        for(let pjson of rule.play_json){
            if(pjson.re && (pjson.re==='*'||web_url.match(new RegExp(pjson.re)))){
                if(pjson.json && typeof(pjson.json)==='object'){
                    let base_json = pjson.json;
H
升级  
hjdhnx 已提交
1887 1888
                    // print('开始合并:');
                    // print(base_json);
H
hjdhnx 已提交
1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906
                    lazy_play = Object.assign(lazy_play,base_json);
                    break;
                }
            }
        }
    }else if(rule.play_json && !Array.isArray(rule.play_json)){ // 其他情况 非[] 判断true/false
        let base_json = {
            jx:1,
            parse:1,
        };
        lazy_play = Object.assign(lazy_play,base_json);
    }else if(!rule.play_json){ // 不解析传0
        let base_json = {
            jx:0,
            parse:1,
        };
        lazy_play = Object.assign(lazy_play,base_json);
    }
H
hjdhnx 已提交
1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917
    console.log(JSON.stringify(lazy_play));
    return JSON.stringify(lazy_play);
}

/**
 * js源预处理特定返回对象中的函数
 * @param ext
 */
 function init(ext) {
    console.log('init');
    try {
H
hjdhnx 已提交
1918 1919 1920 1921 1922 1923 1924 1925
        // make shared jsContext happy muban不能import,不然会造成换源继承后变量被篡改
        if (typeof (globalThis.mubanJs) === 'undefined') {
            let mubanJs = request('https://gitcode.net/qq_32394351/dr_py/-/raw/master/js/模板.js', { 'User-Agent': MOBILE_UA });
            mubanJs = mubanJs.replace('export default', '(function() {return muban;}()) // export default');
            // console.log(mubanJs);
            globalThis.mubanJs = mubanJs;
        }
        let muban = eval(globalThis.mubanJs);
H
hjdhnx 已提交
1926 1927 1928 1929 1930
        if (typeof ext == 'object'){
            rule = ext;
        } else if (typeof ext == 'string') {
            if (ext.startsWith('http')) {
                let js = request(ext,{'method':'GET'});
H
hjdhnx 已提交
1931 1932 1933 1934
                if (js){
                    eval(js.replace('var rule', 'rule'));
                }
                }
H
hjdhnx 已提交
1935 1936 1937 1938 1939 1940 1941
        } else {
            eval(ext.replace('var rule', 'rule'));
        }
        if (rule.模板 && muban.hasOwnProperty(rule.模板)) {
            print('继承模板:'+rule.模板);
            rule = Object.assign(muban[rule.模板], rule);
        }
H
hjdhnx 已提交
1942 1943 1944 1945 1946 1947 1948 1949
        /** 处理一下 rule规则关键字段没传递的情况 **/
        let rule_cate_excludes = (rule.cate_exclude||'').split('|').filter(it=>it.trim());
        let rule_tab_excludes = (rule.tab_exclude||'').split('|').filter(it=>it.trim());
        rule_cate_excludes = rule_cate_excludes.concat(CATE_EXCLUDE.split('|').filter(it=>it.trim()));
        rule_tab_excludes = rule_tab_excludes.concat(TAB_EXCLUDE.split('|').filter(it=>it.trim()));

        rule.cate_exclude = rule_cate_excludes.join('|');
        rule.tab_exclude = rule_tab_excludes.join('|');
H
hjdhnx 已提交
1950
        rule.host = (rule.host||'').rstrip('/');
H
hjdhnx 已提交
1951
        rule.url = rule.url||'';
H
hjdhnx 已提交
1952
        rule.double = rule.double||false;
H
hjdhnx 已提交
1953
        rule.homeUrl = rule.homeUrl||'';
1954
        rule.detailUrl = rule.detailUrl||'';
H
hjdhnx 已提交
1955
        rule.searchUrl = rule.searchUrl||'';
1956 1957
        rule.homeUrl = rule.host&&rule.homeUrl?urljoin(rule.host,rule.homeUrl):(rule.homeUrl||rule.host);
        rule.detailUrl = rule.host&&rule.detailUrl?urljoin(rule.host,rule.detailUrl):rule.detailUrl;
H
hjdhnx 已提交
1958 1959 1960 1961 1962 1963 1964 1965 1966
        if(rule.url.includes('[')&&rule.url.includes(']')){
            let u1 = rule.url.split('[')[0]
            let u2 = rule.url.split('[')[1].split(']')[0]
            rule.url = rule.host && rule.url?urljoin(rule.host,u1)+'['+urljoin(rule.host,u2)+']':rule.url;
        }else{
            rule.url = rule.host && rule.url ? urljoin(rule.host,rule.url) : rule.url;
        }
        rule.searchUrl = rule.host && rule.searchUrl ? urljoin(rule.host,rule.searchUrl) : rule.searchUrl;

H
hjdhnx 已提交
1967 1968
        rule.timeout = rule.timeout||5000;
        rule.encoding = rule.编码||rule.encoding||'utf-8';
H
hjdhnx 已提交
1969
        rule.图片来源 = rule.图片来源||'';
H
hjdhnx 已提交
1970
        rule.play_json = rule.hasOwnProperty('play_json')?rule.play_json:[];
H
hjdhnx 已提交
1971
        if(rule.headers && typeof(rule.headers) === 'object'){
H
hjdhnx 已提交
1972 1973 1974 1975 1976 1977 1978 1979 1980
            try {
                let header_keys = Object.keys(rule.headers);
                for(let k of header_keys){
                    if(k.toLowerCase() === 'user-agent'){
                        let v = rule.headers[k];
                        console.log(v);
                        if(['MOBILE_UA','PC_UA','UC_UA','IOS_UA','UA'].includes(v)){
                            rule.headers[k] = eval(v);
                        }
H
hjdhnx 已提交
1981 1982
                    }
                }
H
hjdhnx 已提交
1983 1984
            }catch (e) {
                console.log('处理headers发生错误:'+e.message);
H
hjdhnx 已提交
1985 1986
            }
        }
H
hjdhnx 已提交
1987 1988
        // print(rule.headers);
        rule_fetch_params  = {'headers': rule.headers||false, 'timeout': rule.timeout, 'encoding': rule.encoding};
H
hjdhnx 已提交
1989
        oheaders = rule.headers||{};
H
hjdhnx 已提交
1990
        RKEY = typeof(key)!=='undefined'&&key?key:'drpy_' + (rule.title || rule.host);
H
hjdhnx 已提交
1991
        pre(); // 预处理
H
hjdhnx 已提交
1992
        init_test();
H
hjdhnx 已提交
1993 1994 1995 1996 1997
    }catch (e) {
        console.log('init_test发生错误:'+e.message);
    }
}

H
hjdhnx 已提交
1998 1999
let homeHtmlCache = undefined;

H
hjdhnx 已提交
2000 2001 2002 2003 2004 2005 2006 2007 2008
/**
 * js源获取首页分类和筛选特定返回对象中的函数
 * @param filter 筛选条件字典对象
 * @returns {string}
 */
function home(filter) {
    console.log("home");
    let homeObj = {
        filter:rule.filter||false,
H
hjdhnx 已提交
2009
        MY_URL: rule.homeUrl,
H
hjdhnx 已提交
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023
        class_name: rule.class_name || '',
        class_url: rule.class_url || '',
        class_parse: rule.class_parse || '',
        cate_exclude: rule.cate_exclude,
    };
    return homeParse(homeObj);
}

/**
 * js源获取首页推荐数据列表特定返回对象中的函数
 * @param params
 * @returns {string}
 */
function homeVod(params) {
H
hjdhnx 已提交
2024
    console.log("homeVod");
H
hjdhnx 已提交
2025 2026 2027
    let homeVodObj = {
        推荐:rule.推荐,
        double:rule.double,
2028 2029
        homeUrl:rule.homeUrl,
        detailUrl:rule.detailUrl
H
hjdhnx 已提交
2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044
    };
    return homeVodParse(homeVodObj)
    // return "{}";
}

/**
 * js源获取分类页一级数据列表特定返回对象中的函数
 * @param tid 分类id
 * @param pg 页数
 * @param filter 当前选中的筛选条件
 * @param extend 扩展
 * @returns {string}
 */
function category(tid, pg, filter, extend) {
    let cateObj = {
H
hjdhnx 已提交
2045
        url: rule.url,
H
hjdhnx 已提交
2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061
        一级: rule.一级,
        tid: tid,
        pg: parseInt(pg),
        filter: filter,
        extend: extend
    };
    // console.log(JSON.stringify(extend));
    return categoryParse(cateObj)
}

/**
 * js源获取二级详情页数据特定返回对象中的函数
 * @param vod_url 一级列表中的vod_id或者是带分类的自拼接 vod_id 如 fyclass$vod_id
 * @returns {string}
 */
function detail(vod_url) {
H
hjdhnx 已提交
2062
    let orId = vod_url;
H
hjdhnx 已提交
2063 2064 2065 2066 2067 2068
    let fyclass = '';
    if(vod_url.indexOf('$')>-1){
        let tmp = vod_url.split('$');
        fyclass = tmp[0];
        vod_url = tmp[1];
    }
H
hjdhnx 已提交
2069
    let detailUrl = vod_url.split('@@')[0];
H
hjdhnx 已提交
2070 2071 2072 2073 2074 2075 2076 2077 2078
    let url;
    if(!detailUrl.startsWith('http')&&!detailUrl.includes('/')){
        url = rule.detailUrl.replaceAll('fyid', detailUrl).replaceAll('fyclass',fyclass);
    }else if(detailUrl.includes('/')){
        url = urljoin(rule.homeUrl,detailUrl);
    }else{
        url = detailUrl
    }
    let detailObj = {
H
hjdhnx 已提交
2079
        orId: orId,
H
hjdhnx 已提交
2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112
        url:url,
        二级:rule.二级,
        detailUrl:detailUrl,
        fyclass:fyclass,
        tab_exclude:rule.tab_exclude,
    }
    return detailParse(detailObj)
}

/**
 * js源选集按钮播放点击事件特定返回对象中的函数
 * @param flag 线路名
 * @param id 播放按钮的链接
 * @param flags 全局配置的flags是否需要解析的标识列表
 * @returns {string}
 */
function play(flag, id, flags) {
    let playObj = {
        url:id,
        flag:flag,
        flags:flags
    }
    return playParse(playObj);
}

/**
 * js源搜索返回的数据列表特定返回对象中的函数
 * @param wd 搜索关键字
 * @param quick 是否来自快速搜索
 * @returns {string}
 */
function search(wd, quick) {
    let searchObj = {
H
hjdhnx 已提交
2113
        searchUrl: rule.searchUrl,
H
hjdhnx 已提交
2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146
        搜索: rule.搜索,
        wd: wd,
        //pg: pg,
        pg: 1,
        quick: quick,
    };
    // console.log(JSON.stringify(searchObj));
    return searchParse(searchObj)
}

function DRPY(){//导出函数
    return {
        init: init,
        home: home,
        homeVod: homeVod,
        category: category,
        detail: detail,
        play: play,
        search: search,
    }
}

// 导出函数对象
export default {
    init: init,
    home: home,
    homeVod: homeVod,
    category: category,
    detail: detail,
    play: play,
    search: search,
    DRPY:DRPY
}