feat: xhs解析视频 & xhs解析速度提升

This commit is contained in:
zhiyu1998 2023-03-20 21:17:15 +08:00
parent 251c382d2b
commit 3ae7c26a3b
2 changed files with 63 additions and 62 deletions

View File

@ -12,7 +12,7 @@ import HttpProxyAgent from "https-proxy-agent";
import { mkdirsSync } from "../utils/file.js"; import { mkdirsSync } from "../utils/file.js";
import { downloadBFile, getDownloadUrl, mergeFileToMp4, getDynamic } from "../utils/bilibili.js"; import { downloadBFile, getDownloadUrl, mergeFileToMp4, getDynamic } from "../utils/bilibili.js";
import { parseUrl, parseM3u8, downloadM3u8Videos, mergeAcFileToMp4 } from "../utils/acfun.js"; import { parseUrl, parseM3u8, downloadM3u8Videos, mergeAcFileToMp4 } from "../utils/acfun.js";
import { transMap, douyinTypeMap, TEN_THOUSAND } from "../utils/constant.js"; import { transMap, douyinTypeMap, TEN_THOUSAND, XHS_CK } from "../utils/constant.js";
import { getIdVideo, generateRandomStr } from "../utils/common.js"; import { getIdVideo, generateRandomStr } from "../utils/common.js";
import config from "../model/index.js"; import config from "../model/index.js";
@ -535,69 +535,65 @@ export class tools extends plugin {
// 小红书解析 // 小红书解析
async redbook(e) { async redbook(e) {
// 解析短号
const msgUrl = /(http:|https:)\/\/(xhslink|xiaohongshu).com\/[A-Za-z\d._?%&+\-=\/#@]*/.exec( const msgUrl = /(http:|https:)\/\/(xhslink|xiaohongshu).com\/[A-Za-z\d._?%&+\-=\/#@]*/.exec(
e.msg, e.msg,
)[0]; )[0];
const url = `https://dlpanda.com/zh-CN/xhs?url=${msgUrl}`; let id;
if (msgUrl.includes('xhslink')) {
await axios await fetch(msgUrl, {
.get(url, { redirect: 'follow',
headers: { }).then(resp => {
"User-Agent": const uri = decodeURIComponent(resp.url)
"Mozilla/5.0 (Linux; Android 5.0; SM-G900P Build/LRX21T) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.25 Mobile Safari/537.36", id = /explore\/(\w+)/.exec(uri)[1];
"Content-Type": "application/json",
"Accept-Encoding": "gzip,deflate,compress",
},
timeout: 10000,
proxy: false,
}) })
.then(async resp => { } else {
const reg = /<img(.*)src="\/\/ci\.xiaohongshu\.com(.*?)"/g; id = /explore\/(\w+)/.exec(msgUrl)[1];
}
const downloadPath = `${this.defaultPath}${this.e.group_id || this.e.user_id}`; const downloadPath = `${this.defaultPath}${this.e.group_id || this.e.user_id}`;
// 创建文件夹(如果没有过这个群) // 获取信息
if (!fs.existsSync(downloadPath)) { fetch(`https://www.xiaohongshu.com/discovery/item/${id}`, {
mkdirsSync(downloadPath); headers: {
} "user-agent": "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 Edg/110.0.0.0",
const res = resp.data.match(reg); "cookie": Buffer.from(XHS_CK, 'base64').toString('utf-8')
const imagesPath = res.map(item => { }
const addr = `https:${item.split('"')[3]}`; }).then(async resp => {
return axios const xhsHtml = await resp.text();
.get(addr, { const reg = /window.__INITIAL_STATE__=(.*?)<\/script>/
headers: { const resJson = xhsHtml.match(reg)[0]
"User-Agent": const res = JSON.parse(resJson.match(reg)[1]);
"Mozilla/5.0 (Linux; Android 5.0; SM-G900P Build/LRX21T) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.25 Mobile Safari/537.36", const noteData = res.noteData.data.noteData
}, const { title, desc, type } = noteData
responseType: "stream", e.reply(`识别:小红书, ${title}\n${desc}`)
}) let imgPromise = []
.then(resp => { if (type === 'video') {
const filepath = `${downloadPath}/${/com\/(.*)\?/.exec(addr)[1]}.jpg`; const url = noteData.video.url;
const writer = fs.createWriteStream(filepath); this.downloadVideo(url).then(path => {
resp.data.pipe(writer); e.reply(segment.video(path + "/temp.mp4"));
return new Promise((resolve, reject) => { })
writer.on("finish", () => resolve(filepath)); return true;
writer.on("error", reject); } else if (type === 'normal') {
}); noteData.imageList.map (async (item, index) => {
}); imgPromise.push(this.downloadImg(item.url, downloadPath, index.toString()))
}); })
let path = []; }
const images = await Promise.all(imagesPath).then(paths => { let path = [];
return paths.map(item => { const images = await Promise.all(imgPromise).then(paths => {
path.push(item); return paths.map(item => {
return { path.push(item);
message: segment.image(fs.readFileSync(item)), return {
nickname: e.sender.card || e.user_id, message: segment.image(fs.readFileSync(item)),
user_id: e.user_id, nickname: e.sender.card || e.user_id,
}; user_id: e.user_id,
}); };
});
await this.reply(await Bot.makeForwardMsg(images));
// 清理文件
path.forEach(item => {
fs.unlinkSync(item);
}); });
}); });
await this.reply(await Bot.makeForwardMsg(images));
// 清理文件
path.forEach(item => {
fs.unlinkSync(item);
});
})
return true; return true;
} }
@ -743,11 +739,14 @@ export class tools extends plugin {
* 下载一张网络图片(自动以url的最后一个为名字) * 下载一张网络图片(自动以url的最后一个为名字)
* @param img * @param img
* @param dir * @param dir
* @param fileName
* @returns {Promise<unknown>} * @returns {Promise<unknown>}
*/ */
async downloadImg(img, dir) { async downloadImg(img, dir, fileName='') {
const filename = img.split("/").pop(); if (fileName === "") {
const filepath = `${dir}/${filename}`; fileName = img.split("/").pop();
}
const filepath = `${dir}/${fileName}`;
const writer = fs.createWriteStream(filepath); const writer = fs.createWriteStream(filepath);
return axios return axios
.get(img, { .get(img, {

View File

@ -24,4 +24,6 @@ export const douyinTypeMap = {
export const TEN_THOUSAND = 10000; export const TEN_THOUSAND = 10000;
export const CAT_LIMIT = 10; export const CAT_LIMIT = 10;
export const XHS_CK = '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'