koishi-plugin-douyin-local-parser 0.1.0

package/README.md

@@ -0,0 +1,41 @@
+# koishi-plugin-douyin-parser
+
+抖音链接解析 Koishi 插件，支持普通链接、短链和聊天平台卡片消息中的链接提取，可选合并转发。
+
+解析逻辑参考了 [fllesser/nonebot-plugin-parser](https://github.com/fllesser/nonebot-plugin-parser) 的抖音解析实现，并按本地 `xhs-parser` 的 Koishi 插件结构改写。
+
+## 功能
+
+- 解析 `https://v.douyin.com/...`、`https://jx.douyin.com/...` 短链
+- 解析 `https://www.douyin.com/video/...`
+- 解析 `https://www.douyin.com/note/...`
+- 解析 `https://www.iesdouyin.com/share/video|note|slides/...`
+- 解析 `https://m.douyin.com/share/video|note|slides/...`
+- 解析 `https://jingxuan.douyin.com/m/video|note|slides/...`
+- 从 Koishi 卡片元素的 `data` 字段、转义 JSON、普通文本中提取抖音链接
+- 支持图片、动图、视频、原文链接和作者信息返回
+- 支持 OneBot / Red 适配器的合并转发元素
+
+## 本地测试
+
+```bash
+npm install
+npm test
+npm run build
+```
+
+也可以直接跑一次真实解析：
+
+```bash
+npm run dev -- "https://v.douyin.com/_2ljF4AmKL8/"
+```
+
+## Koishi 使用
+
+构建后在 Koishi 配置中加载本地插件：
+
+```yaml
+plugins:
+  /absolute/path/to/douyin-parser/lib:
+    enabled: true
+```

package/lib/index.d.ts

@@ -0,0 +1,27 @@
+import { Context, Schema } from 'koishi';
+export declare const name = "douyin-parser";
+export interface Config {
+    enabled: boolean;
+    parseMode: ('link' | 'card')[];
+    waitTip?: string | null;
+    useForward: boolean;
+    quote: boolean;
+    middleware: boolean;
+    parseLimit: number;
+    minimumInterval: number;
+    userAgent: string;
+    cookie?: string;
+    timeout: number;
+    showImages: boolean;
+    maxImages: number;
+    maxDescLength: number;
+    descTruncateSuffix: string;
+    showVideo: boolean;
+    showAuthor: boolean;
+    showLink: boolean;
+    showError: boolean;
+    loggerinfo: boolean;
+}
+export declare const Config: Schema<Config>;
+export declare const usage = "\n\u53D1\u9001\u6296\u97F3\u94FE\u63A5\u6216\u5E73\u53F0\u5361\u7247\u5373\u53EF\u81EA\u52A8\u89E3\u6790\u3002\n\n\u652F\u6301\u793A\u4F8B\uFF1A\n\n- https://v.douyin.com/_2ljF4AmKL8/\n- https://www.douyin.com/video/7521023890996514083\n- https://www.douyin.com/note/7469411074119322899\n- https://www.iesdouyin.com/share/video/7521023890996514083\n";
+export declare function apply(ctx: Context, config: Config): void;

package/lib/index.js

@@ -0,0 +1,128 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.usage = exports.Config = exports.name = void 0;
+exports.apply = apply;
+const koishi_1 = require("koishi");
+const parser_1 = require("./parser");
+exports.name = 'douyin-parser';
+const logger = new koishi_1.Logger(exports.name);
+exports.Config = koishi_1.Schema.intersect([
+    koishi_1.Schema.object({
+        enabled: koishi_1.Schema.boolean().default(true).description('开启抖音链接/卡片解析。'),
+        parseMode: koishi_1.Schema.array(koishi_1.Schema.union([
+            koishi_1.Schema.const('link').description('普通链接'),
+            koishi_1.Schema.const('card').description('卡片消息'),
+        ])).role('checkbox').default(['link', 'card']).description('选择解析来源。'),
+        waitTip: koishi_1.Schema.union([
+            koishi_1.Schema.const(null).description('不发送提示'),
+            koishi_1.Schema.string().description('解析前发送提示语').default('正在解析抖音链接...'),
+        ]).default(null).description('等待提示。'),
+    }).description('基础设置'),
+    koishi_1.Schema.object({
+        useForward: koishi_1.Schema.boolean().default(false).description('开启合并转发。主要适用于 onebot / red 适配器。').experimental(),
+        quote: koishi_1.Schema.boolean().default(true).description('普通发送时引用原消息。'),
+        middleware: koishi_1.Schema.boolean().default(false).description('以前置中间件模式捕获消息。').experimental(),
+        parseLimit: koishi_1.Schema.number().min(1).max(10).step(1).default(3).description('单条消息最多解析的链接数量。'),
+        minimumInterval: koishi_1.Schema.number().min(0).max(3600).step(1).default(180).description('同频道同链接去重间隔，单位秒。0 表示不去重。'),
+    }).description('发送设置'),
+    koishi_1.Schema.object({
+        showImages: koishi_1.Schema.boolean().default(true).description('返回图片/动图。'),
+        maxImages: koishi_1.Schema.number().min(0).max(18).step(1).default(9).description('单个作品最多发送图片数。'),
+        maxDescLength: koishi_1.Schema.number().min(0).max(2000).step(10).default(160).description('描述最大字数。设为 0 时不展示描述。'),
+        descTruncateSuffix: koishi_1.Schema.string().default('...(已截断)').description('描述超出最大字数时追加的截断标志。'),
+        showVideo: koishi_1.Schema.boolean().default(true).description('返回视频元素。'),
+        showAuthor: koishi_1.Schema.boolean().default(true).description('展示作者。'),
+        showLink: koishi_1.Schema.boolean().default(true).description('展示原文链接。'),
+    }).description('内容设置'),
+    koishi_1.Schema.object({
+        userAgent: koishi_1.Schema.string().default('Mozilla/5.0 (iPhone; CPU iPhone OS 18_5 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/18.5 Mobile/15E148 Safari/604.1').description('请求抖音页面时使用的 User-Agent。'),
+        cookie: koishi_1.Schema.string().role('textarea').default('').description('可选 Cookie。遇到风控或无法读取页面数据时可填写。'),
+        timeout: koishi_1.Schema.number().min(3).max(60).step(1).default(15).description('请求超时时间，单位秒。'),
+        showError: koishi_1.Schema.boolean().default(false).description('解析失败时向聊天发送错误提示。'),
+        loggerinfo: koishi_1.Schema.boolean().default(false).description('输出调试日志。').experimental(),
+    }).description('网络与调试'),
+]);
+exports.usage = `
+发送抖音链接或平台卡片即可自动解析。
+
+支持示例：
+
+- https://v.douyin.com/_2ljF4AmKL8/
+- https://www.douyin.com/video/7521023890996514083
+- https://www.douyin.com/note/7469411074119322899
+- https://www.iesdouyin.com/share/video/7521023890996514083
+`;
+function apply(ctx, config) {
+    if (!config.enabled)
+        return;
+    const recent = new Map();
+    ctx.middleware(async (session, next) => {
+        const content = session.content || session.stripped?.content || '';
+        const isCard = /^<\w+\s/i.test(content) || content.includes('data=');
+        if (isCard && !config.parseMode.includes('card'))
+            return next();
+        if (!isCard && !config.parseMode.includes('link'))
+            return next();
+        const links = (0, parser_1.extractDouyinLinks)(content).slice(0, config.parseLimit);
+        if (!links.length)
+            return next();
+        const targets = links.filter((link) => shouldProcess(recent, session.channelId || session.guildId || 'private', link, config.minimumInterval));
+        if (!targets.length)
+            return next();
+        handleLinks(session, targets, config).catch((error) => {
+            logger.warn(error);
+        });
+        return next();
+    }, config.middleware);
+}
+async function handleLinks(session, links, config) {
+    let waitTipMessageId;
+    if (config.waitTip) {
+        const result = await session.send(`${koishi_1.h.quote(session.messageId)}${config.waitTip}`);
+        waitTipMessageId = Array.isArray(result) ? result[0] : result;
+    }
+    try {
+        const allMessages = [];
+        for (const link of links) {
+            if (config.loggerinfo)
+                logger.info(`parse ${link}`);
+            const post = await (0, parser_1.fetchDouyinPost)(link, config);
+            allMessages.push(...(0, parser_1.buildDouyinMessages)(post, config, session));
+        }
+        if (!allMessages.length)
+            return;
+        if (config.useForward && (session.platform === 'onebot' || session.platform === 'red')) {
+            await session.send((0, koishi_1.h)('figure', { children: allMessages }));
+            return;
+        }
+        if (config.quote) {
+            await session.send((0, koishi_1.h)('message', koishi_1.h.quote(session.messageId), allMessages[0].children));
+            for (const message of allMessages.slice(1))
+                await session.send(message);
+            return;
+        }
+        for (const message of allMessages)
+            await session.send(message);
+    }
+    catch (error) {
+        logger.warn(error);
+        if (config.showError)
+            await session.send(`抖音解析失败：${error instanceof Error ? error.message : String(error)}`);
+    }
+    finally {
+        if (waitTipMessageId) {
+            await session.bot?.deleteMessage?.(session.channelId, waitTipMessageId).catch?.(() => undefined);
+        }
+    }
+}
+function shouldProcess(recent, channelId, link, seconds) {
+    if (seconds <= 0)
+        return true;
+    const key = `${channelId}:${link}`;
+    const now = Date.now();
+    const last = recent.get(key);
+    if (last && now - last < seconds * 1000)
+        return false;
+    recent.set(key, now);
+    return true;
+}

package/lib/parser.d.ts

@@ -0,0 +1,45 @@
+import { h } from 'koishi';
+export interface DouyinConfigLike {
+    userAgent: string;
+    timeout: number;
+    showVideo: boolean;
+    showImages: boolean;
+    maxImages: number;
+    maxDescLength: number;
+    descTruncateSuffix: string;
+    showAuthor: boolean;
+    showLink: boolean;
+    cookie?: string;
+}
+export interface DouyinPost {
+    id: string;
+    url: string;
+    title: string;
+    desc: string;
+    type: 'video' | 'note' | 'slides' | 'unknown';
+    authorName: string;
+    authorAvatar?: string;
+    createTime?: number;
+    duration?: number;
+    coverUrl?: string;
+    imageUrls: string[];
+    dynamicImageUrls: string[];
+    videoUrls: string[];
+}
+interface ParsedDouyinUrl {
+    url: string;
+    id: string;
+    type: 'video' | 'note' | 'slides' | 'unknown';
+}
+export declare function extractDouyinLinks(content: string): string[];
+export declare function resolveDouyinLink(rawUrl: string, config: DouyinConfigLike): Promise<string>;
+export declare function fetchDouyinPost(rawUrl: string, config: DouyinConfigLike): Promise<DouyinPost>;
+export declare function buildDouyinMessages(post: DouyinPost, config: DouyinConfigLike, session?: {
+    userId?: string;
+    username?: string;
+    author?: {
+        nickname?: string;
+    };
+}): h[];
+export declare function extractRouterDataPost(html: string, parsed?: ParsedDouyinUrl, sourceUrl?: string): DouyinPost | null;
+export {};

package/lib/parser.js

@@ -0,0 +1,335 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.extractDouyinLinks = extractDouyinLinks;
+exports.resolveDouyinLink = resolveDouyinLink;
+exports.fetchDouyinPost = fetchDouyinPost;
+exports.buildDouyinMessages = buildDouyinMessages;
+exports.extractRouterDataPost = extractRouterDataPost;
+const koishi_1 = require("koishi");
+const URL_BOUNDARY = '[^\\s"\'<>\\\\^`{|}，。；！？、【】《》]+';
+const LINK_PATTERNS = [
+    new RegExp(`https?://v\\.douyin\\.com/${URL_BOUNDARY}`, 'gi'),
+    new RegExp(`https?://jx\\.douyin\\.com/${URL_BOUNDARY}`, 'gi'),
+    new RegExp(`https?://(?:www\\.)?douyin\\.com/(?:video|note)/${URL_BOUNDARY}`, 'gi'),
+    new RegExp(`https?://(?:www\\.)?iesdouyin\\.com/share/(?:slides|video|note)/${URL_BOUNDARY}`, 'gi'),
+    new RegExp(`https?://m\\.douyin\\.com/share/(?:slides|video|note)/${URL_BOUNDARY}`, 'gi'),
+    new RegExp(`https?://jingxuan\\.douyin\\.com/m/(?:slides|video|note)/${URL_BOUNDARY}`, 'gi'),
+];
+const IOS_HEADERS = {
+    accept: 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
+    'accept-language': 'zh-CN,zh;q=0.9,en;q=0.8',
+    referer: 'https://www.douyin.com/',
+};
+const ANDROID_HEADERS = {
+    accept: 'application/json,text/plain,*/*',
+    referer: 'https://www.douyin.com/',
+};
+function extractDouyinLinks(content) {
+    const candidates = expandTextCandidates(content);
+    const links = [];
+    for (const candidate of candidates) {
+        const normalized = normalizeText(candidate);
+        for (const pattern of LINK_PATTERNS) {
+            pattern.lastIndex = 0;
+            let match;
+            while ((match = pattern.exec(normalized))) {
+                links.push(cleanUrl(match[0]));
+            }
+        }
+    }
+    return [...new Set(links)];
+}
+async function resolveDouyinLink(rawUrl, config) {
+    const url = ensureProtocol(rawUrl);
+    if (!/(?:v|jx)\.douyin\.com/i.test(url))
+        return url;
+    const response = await fetchWithTimeout(url, config, { redirect: 'manual', headers: IOS_HEADERS });
+    const location = response.headers.get('location');
+    if (location)
+        return new URL(location, url).toString();
+    if (response.status >= 300 && response.status < 400)
+        return url;
+    const followed = await fetchWithTimeout(url, config, { redirect: 'follow', headers: IOS_HEADERS });
+    return followed.url || url;
+}
+async function fetchDouyinPost(rawUrl, config) {
+    const resolvedUrl = await resolveDouyinLink(rawUrl, config);
+    const parsed = parseDouyinUrl(resolvedUrl);
+    if (!parsed.id)
+        throw new Error('未能识别抖音作品 ID。');
+    if (parsed.type === 'slides')
+        return fetchSlidesPost(parsed, config);
+    const errors = [];
+    for (const url of buildPageCandidates(parsed)) {
+        try {
+            const html = await fetchText(url, config, IOS_HEADERS);
+            const post = extractRouterDataPost(html, parsed, url);
+            if (post)
+                return post;
+            errors.push(`${url}: 未找到 _ROUTER_DATA`);
+        }
+        catch (error) {
+            errors.push(`${url}: ${error instanceof Error ? error.message : String(error)}`);
+        }
+    }
+    try {
+        return await fetchSlidesPost(parsed, config);
+    }
+    catch (error) {
+        errors.push(`slidesinfo: ${error instanceof Error ? error.message : String(error)}`);
+    }
+    throw new Error(`抖音解析失败：${errors.join('；')}`);
+}
+function buildDouyinMessages(post, config, session) {
+    const messages = [];
+    const attrs = {
+        userId: session?.userId,
+        nickname: session?.author?.nickname || session?.username,
+    };
+    messages.push((0, koishi_1.h)('message', attrs, koishi_1.h.text(formatPostText(post, config))));
+    if (config.showImages) {
+        for (const imageUrl of post.imageUrls.slice(0, config.maxImages)) {
+            messages.push((0, koishi_1.h)('message', attrs, koishi_1.h.image(imageUrl)));
+        }
+    }
+    if (config.showVideo) {
+        for (const videoUrl of [...post.dynamicImageUrls, ...post.videoUrls].slice(0, 1)) {
+            messages.push((0, koishi_1.h)('message', attrs, koishi_1.h.video(videoUrl)));
+        }
+    }
+    return messages;
+}
+function extractRouterDataPost(html, parsed = { id: '', type: 'unknown', url: '' }, sourceUrl = parsed.url) {
+    const match = html.match(/window\._ROUTER_DATA\s*=\s*([\s\S]*?)<\/script>/);
+    if (!match?.[1])
+        return null;
+    const routerData = parseJsonLike(match[1].trim().replace(/;+\s*$/, ''));
+    const item = firstDefined(deepGet(routerData, ['loaderData', 'video_(id)/page', 'videoInfoRes', 'item_list', 0]), deepGet(routerData, ['loaderData', 'note_(id)/page', 'videoInfoRes', 'item_list', 0]));
+    if (!item)
+        return null;
+    return buildPostFromAweme(item, parsed, sourceUrl);
+}
+function parseDouyinUrl(url) {
+    const normalized = ensureProtocol(url);
+    const match = normalized.match(/(?:douyin\.com\/(?:video|note)|(?:iesdouyin|m\.douyin)\.com\/share\/(?:slides|video|note)|jingxuan\.douyin\.com\/m\/(?:slides|video|note))\/(\d+)/i);
+    const typeMatch = normalized.match(/douyin\.com\/(video|note)\/\d+|(?:iesdouyin|m\.douyin)\.com\/share\/(slides|video|note)\/\d+|jingxuan\.douyin\.com\/m\/(slides|video|note)\/\d+/i);
+    return {
+        url: normalized,
+        id: match?.[1] || '',
+        type: (typeMatch?.[1] || typeMatch?.[2] || typeMatch?.[3] || 'unknown'),
+    };
+}
+function buildPageCandidates(parsed) {
+    const type = parsed.type === 'unknown' ? 'video' : parsed.type;
+    return [
+        `https://m.douyin.com/share/${type}/${parsed.id}`,
+        `https://www.iesdouyin.com/share/${type}/${parsed.id}`,
+    ];
+}
+async function fetchSlidesPost(parsed, config) {
+    const api = new URL('https://www.iesdouyin.com/web/api/v2/aweme/slidesinfo/');
+    api.searchParams.set('aweme_ids', `[${parsed.id}]`);
+    api.searchParams.set('request_source', '200');
+    const response = await fetchWithTimeout(api.toString(), config, { redirect: 'follow', headers: ANDROID_HEADERS });
+    if (!response.ok)
+        throw new Error(`请求抖音图集接口失败：HTTP ${response.status}`);
+    const data = await response.json();
+    const item = data?.aweme_details?.[0];
+    if (!item)
+        throw new Error('图集接口未返回作品数据。');
+    return buildPostFromAweme(item, { ...parsed, type: 'slides' }, parsed.url);
+}
+function buildPostFromAweme(item, parsed, sourceUrl) {
+    const video = item?.video;
+    const images = Array.isArray(item?.images) ? item.images : [];
+    const imageUrls = [];
+    const dynamicImageUrls = [];
+    for (const image of images) {
+        const dynamicUrl = pickUrl(image?.video?.play_addr?.url_list);
+        if (dynamicUrl) {
+            dynamicImageUrls.push(removeWatermark(dynamicUrl));
+            continue;
+        }
+        const imageUrl = pickUrl(image?.url_list);
+        if (imageUrl)
+            imageUrls.push(imageUrl);
+    }
+    const hasImageContent = imageUrls.length > 0 || dynamicImageUrls.length > 0;
+    const videoUrl = hasImageContent ? undefined : pickUrl(video?.play_addr?.url_list);
+    const coverUrl = pickUrl(video?.cover?.url_list);
+    const avatarUrl = pickUrl(item?.author?.avatar_thumb?.url_list) || pickUrl(item?.author?.avatar_medium?.url_list);
+    return {
+        id: parsed.id || String(item?.aweme_id || ''),
+        url: canonicalUrl(parsed, sourceUrl),
+        title: String(item?.desc || '抖音作品'),
+        desc: String(item?.desc || ''),
+        type: parsed.type,
+        authorName: String(item?.author?.nickname || '未知作者'),
+        authorAvatar: avatarUrl,
+        createTime: numberOrUndefined(item?.create_time),
+        duration: typeof video?.duration === 'number' ? Math.round(video.duration / 1000) : undefined,
+        coverUrl,
+        imageUrls: unique(imageUrls),
+        dynamicImageUrls: unique(dynamicImageUrls),
+        videoUrls: videoUrl ? [removeWatermark(videoUrl)] : [],
+    };
+}
+function formatPostText(post, config) {
+    const lines = [`抖音：${post.title || '抖音作品'}`];
+    if (config.showAuthor)
+        lines.push(`作者：${post.authorName}`);
+    if (post.desc && post.desc !== post.title && config.maxDescLength > 0) {
+        lines.push(trimText(post.desc, config.maxDescLength, config.descTruncateSuffix));
+    }
+    if (config.showLink)
+        lines.push(post.url);
+    return lines.join('\n');
+}
+function expandTextCandidates(content) {
+    const values = new Set([content]);
+    try {
+        for (const element of koishi_1.h.parse(content))
+            collectElementText(element, values);
+    }
+    catch {
+        // Some adapters deliver partial XML snippets; regex extraction below still handles them.
+    }
+    for (const match of content.matchAll(/\bdata=(?:"([^"]*)"|'([^']*)')/gi)) {
+        values.add(match[1] || match[2] || '');
+    }
+    for (const value of [...values]) {
+        const decoded = decodeHtmlEntities(value);
+        values.add(decoded);
+        maybeCollectJsonValues(decoded, values);
+    }
+    return [...values];
+}
+function collectElementText(element, values) {
+    if (typeof element === 'string') {
+        values.add(element);
+        return;
+    }
+    for (const value of Object.values(element.attrs || {})) {
+        if (typeof value === 'string')
+            values.add(value);
+    }
+    for (const child of element.children || []) {
+        collectElementText(child, values);
+    }
+}
+function maybeCollectJsonValues(text, values) {
+    try {
+        const json = JSON.parse(text);
+        walkJson(json, values);
+    }
+    catch {
+        const unescaped = text.replace(/\\"/g, '"').replace(/\\\//g, '/');
+        if (unescaped !== text)
+            values.add(unescaped);
+    }
+}
+function walkJson(value, values) {
+    if (typeof value === 'string') {
+        values.add(value);
+        return;
+    }
+    if (Array.isArray(value)) {
+        for (const item of value)
+            walkJson(item, values);
+        return;
+    }
+    if (value && typeof value === 'object') {
+        for (const item of Object.values(value))
+            walkJson(item, values);
+    }
+}
+function normalizeText(text) {
+    let value = decodeHtmlEntities(text).replace(/\\\//g, '/');
+    try {
+        value = decodeURIComponent(value);
+    }
+    catch {
+        // Keep the original if it is only partly percent-encoded.
+    }
+    return value;
+}
+function decodeHtmlEntities(text) {
+    return text
+        .replace(/&quot;/g, '"')
+        .replace(/&#34;/g, '"')
+        .replace(/&#x22;/gi, '"')
+        .replace(/&apos;/g, "'")
+        .replace(/&#39;/g, "'")
+        .replace(/&#x27;/gi, "'")
+        .replace(/&amp;/g, '&')
+        .replace(/&lt;/g, '<')
+        .replace(/&gt;/g, '>');
+}
+function cleanUrl(url) {
+    return ensureProtocol(url)
+        .replace(/[),，。；;！？!]+$/g, '')
+        .replace(/&amp;/g, '&');
+}
+function ensureProtocol(url) {
+    return /^https?:\/\//i.test(url) ? url : `https://${url}`;
+}
+async function fetchText(url, config, headers) {
+    const response = await fetchWithTimeout(url, config, { redirect: 'follow', headers });
+    if (!response.ok)
+        throw new Error(`请求抖音页面失败：HTTP ${response.status}`);
+    return response.text();
+}
+async function fetchWithTimeout(url, config, init) {
+    const controller = new AbortController();
+    const timer = setTimeout(() => controller.abort(), config.timeout * 1000);
+    try {
+        return await fetch(url, {
+            ...init,
+            signal: controller.signal,
+            headers: {
+                'user-agent': config.userAgent,
+                ...(config.cookie ? { cookie: config.cookie } : {}),
+                ...(init.headers || {}),
+            },
+        });
+    }
+    finally {
+        clearTimeout(timer);
+    }
+}
+function parseJsonLike(payload) {
+    return JSON.parse(payload.replace(/[\u0000-\u0008\u000b\u000c\u000e-\u001f\u007f]/g, ''));
+}
+function deepGet(source, path) {
+    let cursor = source;
+    for (const key of path) {
+        if (cursor == null)
+            return undefined;
+        cursor = cursor[key];
+    }
+    return cursor;
+}
+function firstDefined(...values) {
+    return values.find((value) => value != null);
+}
+function pickUrl(value) {
+    return Array.isArray(value) && typeof value[0] === 'string' ? value[0] : undefined;
+}
+function removeWatermark(url) {
+    return url.replace('playwm', 'play');
+}
+function canonicalUrl(parsed, fallback) {
+    if (!parsed.id)
+        return fallback;
+    const type = parsed.type === 'slides' ? 'note' : parsed.type === 'unknown' ? 'video' : parsed.type;
+    return `https://www.douyin.com/${type}/${parsed.id}`;
+}
+function numberOrUndefined(value) {
+    return typeof value === 'number' ? value : undefined;
+}
+function unique(values) {
+    return [...new Set(values.filter(Boolean))];
+}
+function trimText(text, maxLength, suffix) {
+    return text.length > maxLength ? `${text.slice(0, maxLength)}${suffix}` : text;
+}

package/package.json

@@ -0,0 +1,34 @@
+{
+  "name": "koishi-plugin-douyin-local-parser",
+  "version": "0.1.0",
+  "description": "Parse Douyin links and cards for Koishi.",
+  "main": "lib/index.js",
+  "typings": "lib/index.d.ts",
+  "files": [
+    "lib",
+    "src"
+  ],
+  "scripts": {
+    "build": "tsc -p tsconfig.json",
+    "test": "vitest run",
+    "dev": "tsx scripts/local-test.ts"
+  },
+  "keywords": [
+    "chatbot",
+    "koishi",
+    "plugin",
+    "douyin",
+    "parser"
+  ],
+  "license": "MIT",
+  "peerDependencies": {
+    "koishi": "^4.16.8"
+  },
+  "devDependencies": {
+    "@types/node": "^20.14.10",
+    "koishi": "^4.18.8",
+    "tsx": "^4.20.3",
+    "typescript": "^5.5.4",
+    "vitest": "^1.6.0"
+  }
+}

package/src/index.ts

@@ -0,0 +1,156 @@
+import { Context, Logger, Schema, h } from 'koishi'
+import { buildDouyinMessages, extractDouyinLinks, fetchDouyinPost } from './parser'
+
+export const name = 'douyin-parser'
+
+const logger = new Logger(name)
+
+export interface Config {
+  enabled: boolean
+  parseMode: ('link' | 'card')[]
+  waitTip?: string | null
+  useForward: boolean
+  quote: boolean
+  middleware: boolean
+  parseLimit: number
+  minimumInterval: number
+  userAgent: string
+  cookie?: string
+  timeout: number
+  showImages: boolean
+  maxImages: number
+  maxDescLength: number
+  descTruncateSuffix: string
+  showVideo: boolean
+  showAuthor: boolean
+  showLink: boolean
+  showError: boolean
+  loggerinfo: boolean
+}
+
+export const Config: Schema<Config> = Schema.intersect([
+  Schema.object({
+    enabled: Schema.boolean().default(true).description('开启抖音链接/卡片解析。'),
+    parseMode: Schema.array(Schema.union([
+      Schema.const('link').description('普通链接'),
+      Schema.const('card').description('卡片消息'),
+    ])).role('checkbox').default(['link', 'card']).description('选择解析来源。'),
+    waitTip: Schema.union([
+      Schema.const(null).description('不发送提示'),
+      Schema.string().description('解析前发送提示语').default('正在解析抖音链接...'),
+    ]).default(null).description('等待提示。'),
+  }).description('基础设置'),
+  Schema.object({
+    useForward: Schema.boolean().default(false).description('开启合并转发。主要适用于 onebot / red 适配器。').experimental(),
+    quote: Schema.boolean().default(true).description('普通发送时引用原消息。'),
+    middleware: Schema.boolean().default(false).description('以前置中间件模式捕获消息。').experimental(),
+    parseLimit: Schema.number().min(1).max(10).step(1).default(3).description('单条消息最多解析的链接数量。'),
+    minimumInterval: Schema.number().min(0).max(3600).step(1).default(180).description('同频道同链接去重间隔，单位秒。0 表示不去重。'),
+  }).description('发送设置'),
+  Schema.object({
+    showImages: Schema.boolean().default(true).description('返回图片/动图。'),
+    maxImages: Schema.number().min(0).max(18).step(1).default(9).description('单个作品最多发送图片数。'),
+    maxDescLength: Schema.number().min(0).max(2000).step(10).default(160).description('描述最大字数。设为 0 时不展示描述。'),
+    descTruncateSuffix: Schema.string().default('...(已截断)').description('描述超出最大字数时追加的截断标志。'),
+    showVideo: Schema.boolean().default(true).description('返回视频元素。'),
+    showAuthor: Schema.boolean().default(true).description('展示作者。'),
+    showLink: Schema.boolean().default(true).description('展示原文链接。'),
+  }).description('内容设置'),
+  Schema.object({
+    userAgent: Schema.string().default('Mozilla/5.0 (iPhone; CPU iPhone OS 18_5 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/18.5 Mobile/15E148 Safari/604.1').description('请求抖音页面时使用的 User-Agent。'),
+    cookie: Schema.string().role('textarea').default('').description('可选 Cookie。遇到风控或无法读取页面数据时可填写。'),
+    timeout: Schema.number().min(3).max(60).step(1).default(15).description('请求超时时间，单位秒。'),
+    showError: Schema.boolean().default(false).description('解析失败时向聊天发送错误提示。'),
+    loggerinfo: Schema.boolean().default(false).description('输出调试日志。').experimental(),
+  }).description('网络与调试'),
+])
+
+export const usage = `
+发送抖音链接或平台卡片即可自动解析。
+
+支持示例：
+
+- https://v.douyin.com/_2ljF4AmKL8/
+- https://www.douyin.com/video/7521023890996514083
+- https://www.douyin.com/note/7469411074119322899
+- https://www.iesdouyin.com/share/video/7521023890996514083
+`
+
+export function apply(ctx: Context, config: Config) {
+  if (!config.enabled) return
+
+  const recent = new Map<string, number>()
+
+  ctx.middleware(async (session, next) => {
+    const content = session.content || session.stripped?.content || ''
+    const isCard = /^<\w+\s/i.test(content) || content.includes('data=')
+
+    if (isCard && !config.parseMode.includes('card')) return next()
+    if (!isCard && !config.parseMode.includes('link')) return next()
+
+    const links = extractDouyinLinks(content).slice(0, config.parseLimit)
+    if (!links.length) return next()
+
+    const targets = links.filter((link) => shouldProcess(recent, session.channelId || session.guildId || 'private', link, config.minimumInterval))
+    if (!targets.length) return next()
+
+    handleLinks(session, targets, config).catch((error) => {
+      logger.warn(error)
+    })
+
+    return next()
+  }, config.middleware)
+}
+
+async function handleLinks(session: any, links: string[], config: Config) {
+  let waitTipMessageId: string | undefined
+
+  if (config.waitTip) {
+    const result = await session.send(`${h.quote(session.messageId)}${config.waitTip}`)
+    waitTipMessageId = Array.isArray(result) ? result[0] : result
+  }
+
+  try {
+    const allMessages: h[] = []
+
+    for (const link of links) {
+      if (config.loggerinfo) logger.info(`parse ${link}`)
+      const post = await fetchDouyinPost(link, config)
+      allMessages.push(...buildDouyinMessages(post, config, session))
+    }
+
+    if (!allMessages.length) return
+
+    if (config.useForward && (session.platform === 'onebot' || session.platform === 'red')) {
+      await session.send(h('figure', { children: allMessages }))
+      return
+    }
+
+    if (config.quote) {
+      await session.send(h('message', h.quote(session.messageId), allMessages[0].children))
+      for (const message of allMessages.slice(1)) await session.send(message)
+      return
+    }
+
+    for (const message of allMessages) await session.send(message)
+  } catch (error) {
+    logger.warn(error)
+    if (config.showError) await session.send(`抖音解析失败：${error instanceof Error ? error.message : String(error)}`)
+  } finally {
+    if (waitTipMessageId) {
+      await session.bot?.deleteMessage?.(session.channelId, waitTipMessageId).catch?.(() => undefined)
+    }
+  }
+}
+
+function shouldProcess(recent: Map<string, number>, channelId: string, link: string, seconds: number) {
+  if (seconds <= 0) return true
+
+  const key = `${channelId}:${link}`
+  const now = Date.now()
+  const last = recent.get(key)
+  if (last && now - last < seconds * 1000) return false
+
+  recent.set(key, now)
+  return true
+}

package/src/parser.ts

@@ -0,0 +1,401 @@
+import { h } from 'koishi'
+
+export interface DouyinConfigLike {
+  userAgent: string
+  timeout: number
+  showVideo: boolean
+  showImages: boolean
+  maxImages: number
+  maxDescLength: number
+  descTruncateSuffix: string
+  showAuthor: boolean
+  showLink: boolean
+  cookie?: string
+}
+
+export interface DouyinPost {
+  id: string
+  url: string
+  title: string
+  desc: string
+  type: 'video' | 'note' | 'slides' | 'unknown'
+  authorName: string
+  authorAvatar?: string
+  createTime?: number
+  duration?: number
+  coverUrl?: string
+  imageUrls: string[]
+  dynamicImageUrls: string[]
+  videoUrls: string[]
+}
+
+interface ParsedDouyinUrl {
+  url: string
+  id: string
+  type: 'video' | 'note' | 'slides' | 'unknown'
+}
+
+const URL_BOUNDARY = '[^\\s"\'<>\\\\^`{|}，。；！？、【】《》]+'
+const LINK_PATTERNS = [
+  new RegExp(`https?://v\\.douyin\\.com/${URL_BOUNDARY}`, 'gi'),
+  new RegExp(`https?://jx\\.douyin\\.com/${URL_BOUNDARY}`, 'gi'),
+  new RegExp(`https?://(?:www\\.)?douyin\\.com/(?:video|note)/${URL_BOUNDARY}`, 'gi'),
+  new RegExp(`https?://(?:www\\.)?iesdouyin\\.com/share/(?:slides|video|note)/${URL_BOUNDARY}`, 'gi'),
+  new RegExp(`https?://m\\.douyin\\.com/share/(?:slides|video|note)/${URL_BOUNDARY}`, 'gi'),
+  new RegExp(`https?://jingxuan\\.douyin\\.com/m/(?:slides|video|note)/${URL_BOUNDARY}`, 'gi'),
+]
+
+const IOS_HEADERS = {
+  accept: 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
+  'accept-language': 'zh-CN,zh;q=0.9,en;q=0.8',
+  referer: 'https://www.douyin.com/',
+}
+
+const ANDROID_HEADERS = {
+  accept: 'application/json,text/plain,*/*',
+  referer: 'https://www.douyin.com/',
+}
+
+export function extractDouyinLinks(content: string): string[] {
+  const candidates = expandTextCandidates(content)
+  const links: string[] = []
+
+  for (const candidate of candidates) {
+    const normalized = normalizeText(candidate)
+    for (const pattern of LINK_PATTERNS) {
+      pattern.lastIndex = 0
+      let match: RegExpExecArray | null
+      while ((match = pattern.exec(normalized))) {
+        links.push(cleanUrl(match[0]))
+      }
+    }
+  }
+
+  return [...new Set(links)]
+}
+
+export async function resolveDouyinLink(rawUrl: string, config: DouyinConfigLike): Promise<string> {
+  const url = ensureProtocol(rawUrl)
+  if (!/(?:v|jx)\.douyin\.com/i.test(url)) return url
+
+  const response = await fetchWithTimeout(url, config, { redirect: 'manual', headers: IOS_HEADERS })
+  const location = response.headers.get('location')
+  if (location) return new URL(location, url).toString()
+
+  if (response.status >= 300 && response.status < 400) return url
+
+  const followed = await fetchWithTimeout(url, config, { redirect: 'follow', headers: IOS_HEADERS })
+  return followed.url || url
+}
+
+export async function fetchDouyinPost(rawUrl: string, config: DouyinConfigLike): Promise<DouyinPost> {
+  const resolvedUrl = await resolveDouyinLink(rawUrl, config)
+  const parsed = parseDouyinUrl(resolvedUrl)
+  if (!parsed.id) throw new Error('未能识别抖音作品 ID。')
+
+  if (parsed.type === 'slides') return fetchSlidesPost(parsed, config)
+
+  const errors: string[] = []
+  for (const url of buildPageCandidates(parsed)) {
+    try {
+      const html = await fetchText(url, config, IOS_HEADERS)
+      const post = extractRouterDataPost(html, parsed, url)
+      if (post) return post
+      errors.push(`${url}: 未找到 _ROUTER_DATA`)
+    } catch (error) {
+      errors.push(`${url}: ${error instanceof Error ? error.message : String(error)}`)
+    }
+  }
+
+  try {
+    return await fetchSlidesPost(parsed, config)
+  } catch (error) {
+    errors.push(`slidesinfo: ${error instanceof Error ? error.message : String(error)}`)
+  }
+
+  throw new Error(`抖音解析失败：${errors.join('；')}`)
+}
+
+export function buildDouyinMessages(post: DouyinPost, config: DouyinConfigLike, session?: { userId?: string, username?: string, author?: { nickname?: string } }) {
+  const messages: h[] = []
+  const attrs = {
+    userId: session?.userId,
+    nickname: session?.author?.nickname || session?.username,
+  }
+
+  messages.push(h('message', attrs, h.text(formatPostText(post, config))))
+
+  if (config.showImages) {
+    for (const imageUrl of post.imageUrls.slice(0, config.maxImages)) {
+      messages.push(h('message', attrs, h.image(imageUrl)))
+    }
+  }
+
+  if (config.showVideo) {
+    for (const videoUrl of [...post.dynamicImageUrls, ...post.videoUrls].slice(0, 1)) {
+      messages.push(h('message', attrs, h.video(videoUrl)))
+    }
+  }
+
+  return messages
+}
+
+export function extractRouterDataPost(html: string, parsed: ParsedDouyinUrl = { id: '', type: 'unknown', url: '' }, sourceUrl = parsed.url): DouyinPost | null {
+  const match = html.match(/window\._ROUTER_DATA\s*=\s*([\s\S]*?)<\/script>/)
+  if (!match?.[1]) return null
+
+  const routerData = parseJsonLike(match[1].trim().replace(/;+\s*$/, ''))
+  const item = firstDefined(
+    deepGet(routerData, ['loaderData', 'video_(id)/page', 'videoInfoRes', 'item_list', 0]),
+    deepGet(routerData, ['loaderData', 'note_(id)/page', 'videoInfoRes', 'item_list', 0]),
+  )
+  if (!item) return null
+
+  return buildPostFromAweme(item, parsed, sourceUrl)
+}
+
+function parseDouyinUrl(url: string): ParsedDouyinUrl {
+  const normalized = ensureProtocol(url)
+  const match = normalized.match(/(?:douyin\.com\/(?:video|note)|(?:iesdouyin|m\.douyin)\.com\/share\/(?:slides|video|note)|jingxuan\.douyin\.com\/m\/(?:slides|video|note))\/(\d+)/i)
+  const typeMatch = normalized.match(/douyin\.com\/(video|note)\/\d+|(?:iesdouyin|m\.douyin)\.com\/share\/(slides|video|note)\/\d+|jingxuan\.douyin\.com\/m\/(slides|video|note)\/\d+/i)
+  return {
+    url: normalized,
+    id: match?.[1] || '',
+    type: (typeMatch?.[1] || typeMatch?.[2] || typeMatch?.[3] || 'unknown') as ParsedDouyinUrl['type'],
+  }
+}
+
+function buildPageCandidates(parsed: ParsedDouyinUrl) {
+  const type = parsed.type === 'unknown' ? 'video' : parsed.type
+  return [
+    `https://m.douyin.com/share/${type}/${parsed.id}`,
+    `https://www.iesdouyin.com/share/${type}/${parsed.id}`,
+  ]
+}
+
+async function fetchSlidesPost(parsed: ParsedDouyinUrl, config: DouyinConfigLike): Promise<DouyinPost> {
+  const api = new URL('https://www.iesdouyin.com/web/api/v2/aweme/slidesinfo/')
+  api.searchParams.set('aweme_ids', `[${parsed.id}]`)
+  api.searchParams.set('request_source', '200')
+
+  const response = await fetchWithTimeout(api.toString(), config, { redirect: 'follow', headers: ANDROID_HEADERS })
+  if (!response.ok) throw new Error(`请求抖音图集接口失败：HTTP ${response.status}`)
+
+  const data = await response.json() as any
+  const item = data?.aweme_details?.[0]
+  if (!item) throw new Error('图集接口未返回作品数据。')
+
+  return buildPostFromAweme(item, { ...parsed, type: 'slides' }, parsed.url)
+}
+
+function buildPostFromAweme(item: any, parsed: ParsedDouyinUrl, sourceUrl: string): DouyinPost {
+  const video = item?.video
+  const images = Array.isArray(item?.images) ? item.images : []
+  const imageUrls: string[] = []
+  const dynamicImageUrls: string[] = []
+
+  for (const image of images) {
+    const dynamicUrl = pickUrl(image?.video?.play_addr?.url_list)
+    if (dynamicUrl) {
+      dynamicImageUrls.push(removeWatermark(dynamicUrl))
+      continue
+    }
+    const imageUrl = pickUrl(image?.url_list)
+    if (imageUrl) imageUrls.push(imageUrl)
+  }
+
+  const hasImageContent = imageUrls.length > 0 || dynamicImageUrls.length > 0
+  const videoUrl = hasImageContent ? undefined : pickUrl(video?.play_addr?.url_list)
+  const coverUrl = pickUrl(video?.cover?.url_list)
+  const avatarUrl = pickUrl(item?.author?.avatar_thumb?.url_list) || pickUrl(item?.author?.avatar_medium?.url_list)
+
+  return {
+    id: parsed.id || String(item?.aweme_id || ''),
+    url: canonicalUrl(parsed, sourceUrl),
+    title: String(item?.desc || '抖音作品'),
+    desc: String(item?.desc || ''),
+    type: parsed.type,
+    authorName: String(item?.author?.nickname || '未知作者'),
+    authorAvatar: avatarUrl,
+    createTime: numberOrUndefined(item?.create_time),
+    duration: typeof video?.duration === 'number' ? Math.round(video.duration / 1000) : undefined,
+    coverUrl,
+    imageUrls: unique(imageUrls),
+    dynamicImageUrls: unique(dynamicImageUrls),
+    videoUrls: videoUrl ? [removeWatermark(videoUrl)] : [],
+  }
+}
+
+function formatPostText(post: DouyinPost, config: DouyinConfigLike) {
+  const lines = [`抖音：${post.title || '抖音作品'}`]
+
+  if (config.showAuthor) lines.push(`作者：${post.authorName}`)
+  if (post.desc && post.desc !== post.title && config.maxDescLength > 0) {
+    lines.push(trimText(post.desc, config.maxDescLength, config.descTruncateSuffix))
+  }
+  if (config.showLink) lines.push(post.url)
+
+  return lines.join('\n')
+}
+
+function expandTextCandidates(content: string): string[] {
+  const values = new Set<string>([content])
+
+  try {
+    for (const element of h.parse(content)) collectElementText(element, values)
+  } catch {
+    // Some adapters deliver partial XML snippets; regex extraction below still handles them.
+  }
+
+  for (const match of content.matchAll(/\bdata=(?:"([^"]*)"|'([^']*)')/gi)) {
+    values.add(match[1] || match[2] || '')
+  }
+
+  for (const value of [...values]) {
+    const decoded = decodeHtmlEntities(value)
+    values.add(decoded)
+    maybeCollectJsonValues(decoded, values)
+  }
+
+  return [...values]
+}
+
+function collectElementText(element: h, values: Set<string>) {
+  if (typeof element === 'string') {
+    values.add(element)
+    return
+  }
+
+  for (const value of Object.values(element.attrs || {})) {
+    if (typeof value === 'string') values.add(value)
+  }
+
+  for (const child of element.children || []) {
+    collectElementText(child as h, values)
+  }
+}
+
+function maybeCollectJsonValues(text: string, values: Set<string>) {
+  try {
+    const json = JSON.parse(text)
+    walkJson(json, values)
+  } catch {
+    const unescaped = text.replace(/\\"/g, '"').replace(/\\\//g, '/')
+    if (unescaped !== text) values.add(unescaped)
+  }
+}
+
+function walkJson(value: unknown, values: Set<string>) {
+  if (typeof value === 'string') {
+    values.add(value)
+    return
+  }
+  if (Array.isArray(value)) {
+    for (const item of value) walkJson(item, values)
+    return
+  }
+  if (value && typeof value === 'object') {
+    for (const item of Object.values(value)) walkJson(item, values)
+  }
+}
+
+function normalizeText(text: string) {
+  let value = decodeHtmlEntities(text).replace(/\\\//g, '/')
+  try {
+    value = decodeURIComponent(value)
+  } catch {
+    // Keep the original if it is only partly percent-encoded.
+  }
+  return value
+}
+
+function decodeHtmlEntities(text: string) {
+  return text
+    .replace(/&quot;/g, '"')
+    .replace(/&#34;/g, '"')
+    .replace(/&#x22;/gi, '"')
+    .replace(/&apos;/g, "'")
+    .replace(/&#39;/g, "'")
+    .replace(/&#x27;/gi, "'")
+    .replace(/&amp;/g, '&')
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+}
+
+function cleanUrl(url: string) {
+  return ensureProtocol(url)
+    .replace(/[),，。；;！？!]+$/g, '')
+    .replace(/&amp;/g, '&')
+}
+
+function ensureProtocol(url: string) {
+  return /^https?:\/\//i.test(url) ? url : `https://${url}`
+}
+
+async function fetchText(url: string, config: DouyinConfigLike, headers: Record<string, string>) {
+  const response = await fetchWithTimeout(url, config, { redirect: 'follow', headers })
+  if (!response.ok) throw new Error(`请求抖音页面失败：HTTP ${response.status}`)
+  return response.text()
+}
+
+async function fetchWithTimeout(url: string, config: DouyinConfigLike, init: RequestInit) {
+  const controller = new AbortController()
+  const timer = setTimeout(() => controller.abort(), config.timeout * 1000)
+
+  try {
+    return await fetch(url, {
+      ...init,
+      signal: controller.signal,
+      headers: {
+        'user-agent': config.userAgent,
+        ...(config.cookie ? { cookie: config.cookie } : {}),
+        ...(init.headers || {}),
+      },
+    })
+  } finally {
+    clearTimeout(timer)
+  }
+}
+
+function parseJsonLike(payload: string) {
+  return JSON.parse(payload.replace(/[\u0000-\u0008\u000b\u000c\u000e-\u001f\u007f]/g, ''))
+}
+
+function deepGet(source: any, path: Array<string | number>) {
+  let cursor = source
+  for (const key of path) {
+    if (cursor == null) return undefined
+    cursor = cursor[key]
+  }
+  return cursor
+}
+
+function firstDefined<T>(...values: T[]) {
+  return values.find((value) => value != null)
+}
+
+function pickUrl(value: unknown) {
+  return Array.isArray(value) && typeof value[0] === 'string' ? value[0] : undefined
+}
+
+function removeWatermark(url: string) {
+  return url.replace('playwm', 'play')
+}
+
+function canonicalUrl(parsed: ParsedDouyinUrl, fallback: string) {
+  if (!parsed.id) return fallback
+  const type = parsed.type === 'slides' ? 'note' : parsed.type === 'unknown' ? 'video' : parsed.type
+  return `https://www.douyin.com/${type}/${parsed.id}`
+}
+
+function numberOrUndefined(value: unknown) {
+  return typeof value === 'number' ? value : undefined
+}
+
+function unique(values: string[]) {
+  return [...new Set(values.filter(Boolean))]
+}
+
+function trimText(text: string, maxLength: number, suffix: string) {
+  return text.length > maxLength ? `${text.slice(0, maxLength)}${suffix}` : text
+}