@agentrhq/webcmd 0.1.2 → 0.2.0

package/clis/1point3acres/search.js

@@ -1,71 +0,0 @@
-/**
- * 一亩三分地 站内搜索 — /bbs/search.php?mod=forum
- *
- * Guests get a "请登录" alert page, so this command needs the live browser
- * session's cookie. Discuz routes search through a 302 redirect to
- * search.php?searchid=<ID>. Node fetch follows redirects automatically as
- * long as we pass the session cookie along.
- */
-import { cli, Strategy } from '@agentrhq/webcmd/registry';
-import { ArgumentError, EmptyResultError } from '@agentrhq/webcmd/errors';
-import { fetchHtml, parseSearchList, assertNotGuestAlert, getCookie, decodeEntities, normalizeLimit, BASE } from './utils.js';
-
-cli({
-    site: '1point3acres',
-    name: 'search',
-    access: 'read',
-    description: '一亩三分地 站内关键字搜索（需要登录）',
-    domain: 'www.1point3acres.com',
-    strategy: Strategy.COOKIE,
-    browser: true,
-    navigateBefore: false,
-    args: [
-        { name: 'query', required: true, positional: true, help: '搜索关键字' },
-        { name: 'limit', type: 'int', default: 20, help: '返回条数（默认 20，最多 50）' },
-        { name: 'fid', type: 'string', default: '', help: '限定版块 ID（可选）' },
-    ],
-    columns: ['rank', 'tid', 'title', 'forum', 'author', 'replies', 'views', 'postTime', 'url'],
-    func: async (page, args) => {
-        const query = String(args.query || '').trim();
-        if (!query) throw new ArgumentError('query 不能为空');
-        const limit = normalizeLimit(args.limit, 20, 50);
-        const fid = String(args.fid || '').trim();
-
-        const cookie = await getCookie(page);
-        const qs = new URLSearchParams({
-            mod: 'forum',
-            srchtxt: query,
-            searchsubmit: 'yes',
-            ...(fid ? { srchfid: fid } : {}),
-        });
-        const url = `${BASE}/search.php?${qs.toString()}`;
-
-        // Node fetch with the session cookie — Discuz's 302 to search.php?searchid=…
-        // is followed by default.
-        const html = await fetchHtml(url, {
-            cookie,
-            headers: { Referer: `${BASE}/` },
-        });
-        assertNotGuestAlert(html);
-
-        const items = parseSearchList(html);
-        if (items.length === 0) {
-            const hint = html.match(/<p>([^<]*?抱歉[^<]*?)<\/p>/);
-            if (hint) {
-                throw new EmptyResultError('1point3acres search', decodeEntities(hint[1].trim()));
-            }
-            throw new EmptyResultError('1point3acres search', `No results for "${query}"`);
-        }
-        return items.slice(0, limit).map((t, i) => ({
-            rank: i + 1,
-            tid: t.tid,
-            title: t.title,
-            forum: t.forum,
-            author: t.author,
-            replies: t.replies,
-            views: t.views,
-            postTime: t.postTime,
-            url: t.url,
-        }));
-    },
-});

package/clis/1point3acres/thread.js

@@ -1,117 +0,0 @@
-/**
- * 一亩三分地 帖子详情 — /bbs/thread-<tid>-<page>-1.html
- *
- * Returns one row per post on the requested page. First row (floor=1) is the
- * main post; the rest are replies. Columns are shaped so `--limit 1` gives
- * just the main post, and larger limits walk down the thread.
- */
-import { cli, Strategy } from '@agentrhq/webcmd/registry';
-import { ArgumentError, EmptyResultError } from '@agentrhq/webcmd/errors';
-import { fetchHtml, decodeEntities, stripHtml, truncate, normalizePositiveInteger, BASE } from './utils.js';
-
-function extract(html, regex, group = 1) {
-    const m = html.match(regex);
-    return m ? m[group] : '';
-}
-
-cli({
-    site: '1point3acres',
-    name: 'thread',
-    access: 'read',
-    description: '一亩三分地 帖子详情 + 楼层（主楼 + 回复）',
-    domain: 'www.1point3acres.com',
-    strategy: Strategy.PUBLIC,
-    browser: false,
-    args: [
-        { name: 'tid', required: true, positional: true, help: '帖子 ID（数字，见 `hot`/`latest` 返回的 tid）' },
-        { name: 'page', type: 'int', default: 1, help: '楼层分页页码（默认 1）' },
-        { name: 'limit', type: 'int', default: 10, help: '返回楼层条数（默认 10，含主楼）' },
-        { name: 'contentLimit', type: 'int', default: 400, help: '每楼正文截断长度（默认 400 字符，最少 50）' },
-    ],
-    columns: ['floor', 'pid', 'author', 'postTime', 'content', 'url'],
-    func: async (args) => {
-        const tid = String(args.tid || '').trim();
-        if (!/^\d+$/.test(tid)) {
-            throw new ArgumentError('tid must be a numeric thread id');
-        }
-        const page = normalizePositiveInteger(args.page, 1, 'page');
-        const limit = normalizePositiveInteger(args.limit, 10, 'limit');
-        const contentLimit = normalizePositiveInteger(args.contentLimit, 400, 'contentLimit', { min: 50 });
-
-        const url = `${BASE}/thread-${tid}-${page}-1.html`;
-        const html = await fetchHtml(url);
-
-        // Sanity: real thread page will contain postlist + at least one post div.
-        if (!/id="postlist"/.test(html) && !/id="post_\d+"/.test(html)) {
-            throw new EmptyResultError('1point3acres thread', `帖子 ${tid} 不存在或被删除`);
-        }
-
-        // Split posts: each post block is bounded by <div id="post_<PID>">…</div> next post or postlist end.
-        // NOTE: intermediate objects intentionally use postId/body/offset (not pid/html/start) to
-        // avoid being mistaken for row-shaped objects by the silent-column-drop audit.
-        const postBlocks = [];
-        const re = /<div id="post_(\d+)"[^>]*>/g;
-        const offsets = [];
-        let m;
-        while ((m = re.exec(html))) offsets.push({ postId: m[1], offset: m.index });
-        for (let i = 0; i < offsets.length; i++) {
-            const segStart = offsets[i].offset;
-            const segEnd = i + 1 < offsets.length ? offsets[i + 1].offset : html.length;
-            postBlocks.push({ postId: offsets[i].postId, body: html.slice(segStart, segEnd) });
-        }
-
-        const rows = [];
-        for (let i = 0; i < postBlocks.length && rows.length < limit; i++) {
-            const { postId: pid, body: block } = postBlocks[i];
-            // Discuz authi block holds the author link + post time metadata.
-            const authiMatch = block.match(/<div class="authi"[\s\S]*?<\/div>/);
-            const authiBlock = authiMatch ? authiMatch[0] : '';
-            const authorCandidates = [
-                /<a [^>]*class="[^"]*\bxi2\b[^"]*"[^>]*>\s*([^<]+?)\s*<\/a>/,
-                /<a [^>]*href="space-uid-\d+\.html"[^>]*>\s*([^<]+?)\s*<\/a>/,
-                /<a [^>]*class="[^"]*\bxw1\b[^"]*"[^>]*>\s*([^<]+?)\s*<\/a>/,
-            ];
-            let author = '';
-            for (const re of authorCandidates) {
-                const v = decodeEntities(extract(authiBlock || block, re));
-                if (v && !/匿名卡|变色卡|关贴卡/.test(v)) { author = v; break; }
-            }
-            // Time: prefer <span title="YYYY-MM-DD HH:MM:SS"> (per-post, precise).
-            // <meta itemprop="datePublished"> is the *thread* publish time on this site — avoid.
-            const postTime = extract(authiBlock, /<span title="([^"]+)">/) ||
-                extract(block, /id="authorposton\d+"[^>]*>\s*<span title="([^"]+)">/) ||
-                extract(block, /id="authorposton\d+"[^>]*>\s*([^<]+?)\s*</) ||
-                extract(block, /<meta itemprop="datePublished" content="([^"]+)"/);
-            // Floor: first post on page 1 is the 楼主, subsequent posts carry <em>N#</em>.
-            const floorEm = extract(block, /<em>(\d+)<\/em>\s*#?\s*<\/a>/) ||
-                extract(block, /id="postnum\d+"[^>]*>\s*<em>(\d+)<\/em>/);
-            const isMainPost = page === 1 && i === 0;
-            const floor = floorEm ? Number(floorEm) : (isMainPost ? 1 : (page - 1) * 10 + i + 1);
-            const contentMatch = block.match(/id="postmessage_\d+"[^>]*>([\s\S]*?)<\/td>/);
-            const content = truncate(stripHtml(contentMatch ? contentMatch[1] : ''), contentLimit);
-            rows.push({
-                floor,
-                pid,
-                author,
-                postTime: postTime.trim(),
-                content,
-                url: `${BASE}/forum.php?mod=redirect&goto=findpost&ptid=${tid}&pid=${pid}`,
-            });
-        }
-
-        // Attach the thread title + forum name as a leading synthetic row only when rows exist
-        // and only for page 1, so agents get the title without needing a separate call.
-        if (page === 1 && rows.length > 0) {
-            const title = decodeEntities(
-                extract(html, /<span id="thread_subject">([^<]+)<\/span>/).trim() ||
-                extract(html, /<title>([^<]+?)\s*[-|]/).trim()
-            );
-            rows[0].content = title ? `【${title}】\n${rows[0].content}` : rows[0].content;
-        }
-
-        if (!rows.length) {
-            throw new EmptyResultError('1point3acres thread', `帖子 ${tid} 第 ${page} 页没有可读取楼层`);
-        }
-        return rows;
-    },
-});

package/clis/1point3acres/user.js

@@ -1,77 +0,0 @@
-/**
- * 一亩三分地 用户资料 — /bbs/space-uid-<uid>.html  or  /bbs/space-username-<name>.html
- *
- * Guest-visible fields: username, uid, user group, register/last-access times,
- * post/thread/digest counts, credits, rice (大米 — site currency), profile URL.
- * Users can be queried by numeric uid or by username (both routes are public).
- */
-import { cli, Strategy } from '@agentrhq/webcmd/registry';
-import { ArgumentError, EmptyResultError } from '@agentrhq/webcmd/errors';
-import { fetchHtml, decodeEntities, BASE } from './utils.js';
-
-cli({
-    site: '1point3acres',
-    name: 'user',
-    access: 'read',
-    description: '一亩三分地 用户空间（用户组 / 积分 / 大米 / 帖子数 等）',
-    domain: 'www.1point3acres.com',
-    strategy: Strategy.PUBLIC,
-    browser: false,
-    args: [
-        { name: 'who', required: true, positional: true, help: '用户名或 uid（纯数字按 uid 查，否则按用户名）' },
-    ],
-    columns: [
-        'uid', 'username', 'group', 'credits', 'rice',
-        'posts', 'threads', 'digests', 'registerTime', 'lastAccess', 'profileUrl',
-    ],
-    func: async (args) => {
-        const who = String(args.who || '').trim();
-        if (!who) throw new ArgumentError('who 不能为空', '传用户名或数字 uid');
-        const url = /^\d+$/.test(who)
-            ? `${BASE}/space-uid-${who}.html`
-            : `${BASE}/space-username-${encodeURIComponent(who)}.html`;
-
-        const html = await fetchHtml(url);
-        if (/<title>提示信息/.test(html) && /(没有找到|不存在)/.test(html)) {
-            throw new EmptyResultError('1point3acres user', `用户 "${who}" 不存在`);
-        }
-
-        const pick = (re) => {
-            const m = html.match(re);
-            return m ? decodeEntities(m[1].trim()) : '';
-        };
-        // <li>KEY: VAL</li>   — tolerant of optional <span>, colons fullwidth/半角, 颗/根/粒 suffixes.
-        const pickLi = (label) => {
-            const re = new RegExp(`<li>\\s*${label}[：:\\s]*(?:<[^>]+>)?\\s*([^<]+?)\\s*(?:<|$)`);
-            const m = html.match(re);
-            return m ? decodeEntities(m[1].trim()) : '';
-        };
-
-        const username =
-            pick(/<p class="mtm[^"]*"[^>]*>\s*<a [^>]*>([^<]+?)<\/a>/) ||
-            pick(/<title>([^<]+?)的个人资料/);
-        const uid = pick(/uid=(\d+)/) || pick(/space-uid-(\d+)\.html/);
-        const group = pickLi('用户组');
-        const credits = pickLi('积分');
-        const rice = pickLi('大米');
-        const posts = pickLi('帖子数');
-        const threads = pickLi('主题数');
-        const digests = pickLi('精华数');
-        const registerTime = pickLi('注册时间');
-        const lastAccess = pickLi('最后访问');
-
-        return [{
-            uid,
-            username,
-            group,
-            credits,
-            rice,
-            posts,
-            threads,
-            digests,
-            registerTime,
-            lastAccess,
-            profileUrl: uid ? `${BASE}/space-uid-${uid}.html` : url,
-        }];
-    },
-});

package/clis/1point3acres/utils.js

@@ -1,247 +0,0 @@
-/**
- * Shared helpers for 一亩三分地 (1point3acres.com) adapters.
- *
- * Site is a Discuz!X PHP BBS that serves GBK-encoded HTML.
- * - Thread listings:  /bbs/forum.php?mod=guide&view={hot|new|digest|newthread}
- * - Forum:            /bbs/forum-<fid>-<page>.html
- * - Thread detail:    /bbs/thread-<tid>-<page>-1.html
- * - User profile:     /bbs/space-uid-<uid>.html  or  /bbs/space-username-<name>.html
- * - Search:           /bbs/search.php?mod=forum  (COOKIE — guests get an alert page)
- */
-import { AuthRequiredError, ArgumentError, CommandExecutionError } from '@agentrhq/webcmd/errors';
-
-export const BASE = 'https://www.1point3acres.com/bbs';
-
-/**
- * Validate `limit` per typed-fail-fast convention (no silent clamp).
- * Throws ArgumentError on non-positive / non-integer / out-of-range input.
- */
-export function normalizeLimit(value, defaultValue, maxValue, label = 'limit') {
-    const limit = normalizePositiveInteger(value, defaultValue, label);
-    if (limit > maxValue) {
-        throw new ArgumentError(`${label} must be <= ${maxValue}`);
-    }
-    return limit;
-}
-
-/** Validate a positive integer argument without silently flooring/clamping. */
-export function normalizePositiveInteger(value, defaultValue, label = 'value', { min = 1 } = {}) {
-    const raw = value ?? defaultValue;
-    const limit = Number(raw);
-    if (!Number.isInteger(limit) || limit <= 0) {
-        throw new ArgumentError(`${label} must be a positive integer`);
-    }
-    if (limit < min) {
-        throw new ArgumentError(`${label} must be >= ${min}`);
-    }
-    return limit;
-}
-
-const UA = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/146.0 Safari/537.36';
-
-/** Fetch a GBK-encoded Discuz page and return decoded UTF-8 HTML. */
-export async function fetchHtml(url, { headers = {}, cookie = '' } = {}) {
-    let res;
-    try {
-        res = await fetch(url, {
-            headers: {
-                'User-Agent': UA,
-                'Accept': 'text/html,application/xhtml+xml',
-                'Accept-Language': 'zh-CN,zh;q=0.9,en;q=0.8',
-                ...(cookie ? { Cookie: cookie } : {}),
-                ...headers,
-            },
-            redirect: 'follow',
-        });
-    } catch (error) {
-        throw new CommandExecutionError(`1point3acres request failed: ${error?.message || error}`);
-    }
-    if (!res.ok) {
-        throw new CommandExecutionError(`1point3acres request failed: HTTP ${res.status} ${res.statusText} from ${url}`);
-    }
-    const buf = await res.arrayBuffer();
-    return new TextDecoder('gbk').decode(buf);
-}
-
-/** Pull cookie string from the live browser session for this domain.
- *  Discuz auth cookies (4Oaf_61d6_*, session) are HttpOnly and set on the
- *  root domain `.1point3acres.com`, so we need `getCookies` (not document.cookie)
- *  AND we need to query both host + root domain and merge.
- */
-export async function getCookie(page) {
-    if (!page) return '';
-    const seen = new Map();
-    if (typeof page.getCookies === 'function') {
-        for (const opts of [{ domain: 'www.1point3acres.com' }, { domain: '.1point3acres.com' }]) {
-            try {
-                const cookies = await page.getCookies(opts);
-                for (const c of cookies || []) {
-                    if (!seen.has(c.name)) seen.set(c.name, c.value);
-                }
-            } catch { /* try next */ }
-        }
-    }
-    if (seen.size > 0) {
-        return [...seen].map(([k, v]) => `${k}=${v}`).join('; ');
-    }
-    try {
-        const result = await page.evaluate('document.cookie');
-        return typeof result === 'string' ? result : '';
-    } catch {
-        return '';
-    }
-}
-
-/** Detect the "you are a guest" alert page that Discuz returns for protected actions. */
-export function assertNotGuestAlert(html, domain = 'www.1point3acres.com') {
-    if (/<title>提示信息 \| 一亩三分地<\/title>/.test(html) && /无法进行此操作/.test(html)) {
-        throw new AuthRequiredError(domain, '需要登录一亩三分地后再使用该命令');
-    }
-}
-
-const ENTITY_MAP = {
-    '&nbsp;': ' ', '&amp;': '&', '&lt;': '<', '&gt;': '>',
-    '&quot;': '"', '&#39;': "'", '&apos;': "'",
-};
-
-/** Decode HTML entities (numeric + common named). */
-export function decodeEntities(s) {
-    if (!s) return '';
-    return s
-        .replace(/&#(\d+);/g, (_, n) => String.fromCodePoint(Number(n)))
-        .replace(/&#[xX]([0-9a-fA-F]+);/g, (_, n) => String.fromCodePoint(parseInt(n, 16)))
-        .replace(/&(nbsp|amp|lt|gt|quot|#39|apos);/g, m => ENTITY_MAP[m] || m);
-}
-
-/** Strip HTML tags and collapse whitespace, returning plain text. */
-export function stripHtml(html) {
-    if (!html) return '';
-    return decodeEntities(
-        String(html)
-            .replace(/<br\s*\/?>/gi, '\n')
-            .replace(/<\/(p|div|li|tr)>/gi, '\n')
-            .replace(/<[^>]+>/g, '')
-    ).replace(/[ \t]+\n/g, '\n').replace(/\n{3,}/g, '\n\n').trim();
-}
-
-/** Truncate text to n characters with ellipsis. */
-export function truncate(s, n = 300) {
-    if (!s) return '';
-    return s.length > n ? s.slice(0, n) + '…' : s;
-}
-
-/** Extract all <tbody id="normalthread_*"> blocks from a forum/guide page. */
-export function parseThreadRows(html) {
-    const rows = [];
-    const re = /<tbody id="(normalthread|stickthread)_(\d+)"[^>]*>([\s\S]*?)<\/tbody>/g;
-    let m;
-    while ((m = re.exec(html))) {
-        const [, kind, tid, inner] = m;
-        rows.push({ kind, tid, inner });
-    }
-    return rows;
-}
-
-/** Parse a single Discuz thread row (inner HTML of the tbody). */
-export function parseThreadRow({ kind, tid, inner }) {
-    const titleMatches = [...inner.matchAll(/<a [^>]*class="[^"]*\bxst\b[^"]*"[^>]*>([^<]+)<\/a>/g)];
-    const title = titleMatches.length
-        ? decodeEntities(titleMatches[titleMatches.length - 1][1].trim())
-        : '';
-
-    const forumMatch = inner.match(/<a href="forum-(\d+)-1\.html"[^>]*target="_blank"[^>]*>([^<]+)<\/a>/);
-    const fid = forumMatch ? forumMatch[1] : '';
-    const forumName = forumMatch ? decodeEntities(forumMatch[2].trim()) : '';
-
-    // <td class="by"> blocks; first with <cite> = author, last with <cite> = last reply
-    const byBlocks = [...inner.matchAll(/<td class="by"[^>]*>([\s\S]*?)<\/td>/g)].map(m => m[1]);
-    const readCite = (block) => {
-        const m = block.match(/<cite[^>]*>([\s\S]*?)<\/cite>/);
-        if (!m) return '';
-        return decodeEntities(m[1].replace(/<[^>]+>/g, '').trim());
-    };
-    const readTime = (block) => {
-        const titleM = block.match(/<span [^>]*title="([^"]+)"[^>]*>/);
-        if (titleM) return titleM[1].trim();
-        const plainA = block.match(/<em>[\s\S]*?<a [^>]*>\s*([^<]+?)\s*<\/a>/);
-        if (plainA) return decodeEntities(plainA[1].trim());
-        const plainSpan = block.match(/<em>[\s\S]*?<span[^>]*>\s*([^<]+?)\s*<\/span>/);
-        if (plainSpan) return decodeEntities(plainSpan[1].trim());
-        const bare = block.match(/<em>\s*([^<]+?)\s*<\/em>/);
-        return bare ? decodeEntities(bare[1].trim()) : '';
-    };
-    let authorBlock = '';
-    let lastBlock = '';
-    for (const b of byBlocks) {
-        if (/<cite/.test(b)) {
-            if (!authorBlock) authorBlock = b;
-            lastBlock = b;
-        }
-    }
-    const author = authorBlock ? readCite(authorBlock) : '';
-    const postTime = authorBlock ? readTime(authorBlock) : '';
-    const lastReplyUser = lastBlock && lastBlock !== authorBlock ? readCite(lastBlock) : '';
-    const lastReplyTime = lastBlock && lastBlock !== authorBlock ? readTime(lastBlock) : '';
-
-    const numMatch = inner.match(/<td class="num"[^>]*>\s*<a[^>]*class="xi2"[^>]*>(\d+)<\/a>(?:\s*<em>(\d+)<\/em>)?/);
-    const replies = numMatch ? Number(numMatch[1]) : 0;
-    const views = numMatch && numMatch[2] ? Number(numMatch[2]) : 0;
-    return {
-        tid,
-        kind,
-        title,
-        author,
-        forum: forumName,
-        fid,
-        replies,
-        views,
-        postTime,
-        lastReplyUser,
-        lastReplyTime,
-        url: `${BASE}/thread-${tid}-1-1.html`,
-    };
-}
-
-/** Quick one-shot listing parser used by hot/latest/digest/forum. */
-export function parseThreadList(html) {
-    return parseThreadRows(html).map(parseThreadRow).filter(t => t.title);
-}
-
-/**
- * Parse Discuz search results page (different HTML shape than forum listings).
- * Each hit is <li class="pbw" id="TID"> containing h3 > a[href*="tid=TID"],
- * <p class="xg1">N 个回复 - M 次查看</p>, and a time/author/forum <p>.
- */
-export function parseSearchList(html) {
-    const items = [];
-    const re = /<li class="pbw" id="(\d+)">([\s\S]*?)<\/li>/g;
-    let m;
-    while ((m = re.exec(html))) {
-        const [, tid, inner] = m;
-        const titleMatch = inner.match(/<h3[^>]*>\s*<a [^>]*>([\s\S]*?)<\/a>/);
-        const titleRaw = titleMatch ? titleMatch[1] : '';
-        const title = decodeEntities(titleRaw.replace(/<[^>]+>/g, '')).trim();
-        if (!title) continue;
-
-        const statsMatch = inner.match(/<p class="xg1">\s*([\d,]+)\s*个回复\s*-\s*([\d,]+)\s*次查看\s*<\/p>/);
-        const replies = statsMatch ? Number(statsMatch[1].replace(/,/g, '')) : 0;
-        const views = statsMatch ? Number(statsMatch[2].replace(/,/g, '')) : 0;
-
-        const metaMatch = inner.match(/<p>\s*<span>([^<]+)<\/span>[\s\S]*?<a [^>]*space-uid-\d+[^>]*>([^<]+?)<\/a>[\s\S]*?<a [^>]*href="forum-(\d+)-[^"]*"[^>]*>([^<]+?)<\/a>/);
-        const postTime = metaMatch ? decodeEntities(metaMatch[1].trim()) : '';
-        const author = metaMatch ? decodeEntities(metaMatch[2].trim()) : '';
-        const fid = metaMatch ? metaMatch[3] : '';
-        const forumName = metaMatch ? decodeEntities(metaMatch[4].trim()) : '';
-
-        items.push({
-            tid, title, author, forum: forumName, fid,
-            replies, views, postTime,
-            // Search pages don't show lastReplyTime separately — surface postTime instead.
-            lastReplyUser: '', lastReplyTime: postTime,
-            url: `${BASE}/thread-${tid}-1-1.html`,
-        });
-    }
-    return items;
-}
-
-export { UA };

package/clis/36kr/article.js

@@ -1,66 +0,0 @@
-/**
- * 36kr article detail — INTERCEPT strategy.
- *
- * Fetches the full content of a 36kr article given its ID or URL.
- */
-import { cli, Strategy } from '@agentrhq/webcmd/registry';
-import { CliError } from '@agentrhq/webcmd/errors';
-/** Extract article ID from a full URL or a bare numeric ID string */
-function parseArticleId(input) {
-    const m = input.match(/\/p\/(\d+)/);
-    return m ? m[1] : input.replace(/\D/g, '');
-}
-cli({
-    site: '36kr',
-    name: 'article',
-    access: 'read',
-    description: '获取36氪文章正文内容',
-    domain: 'www.36kr.com',
-    strategy: Strategy.INTERCEPT,
-    args: [
-        { name: 'id', positional: true, required: true, help: 'Article ID or full 36kr article URL' },
-    ],
-    columns: ['field', 'value'],
-    func: async (page, args) => {
-        const articleId = parseArticleId(String(args.id ?? ''));
-        if (!articleId) {
-            throw new CliError('INVALID_ARGUMENT', 'Invalid article ID or URL');
-        }
-        await page.installInterceptor('36kr.com/api');
-        await page.goto(`https://www.36kr.com/p/${articleId}`);
-        await page.wait(5);
-        const data = await page.evaluate(`
-      (() => {
-        // Title: 36kr uses class "article-title" on h1
-        const title = document.querySelector('.article-title, h1')?.textContent?.trim() || '';
-        // Author: second .author-name (first is empty nav link, second has real name)
-        const authorEls = document.querySelectorAll('.author-name');
-        const author = Array.from(authorEls).map(el => el.textContent?.trim()).filter(Boolean)[0] || '';
-        // Date: 36kr uses class "title-icon-item item-time" for the publish date
-        const dateRaw = document.querySelector('.item-time')?.textContent?.trim() || '';
-        const date = dateRaw.replace(/^[·\s]+/, '').trim();
-        // Article body paragraphs
-        const bodyEls = document.querySelectorAll('[class*="article-content"] p, [class*="rich-text"] p, .article p');
-        const body = Array.from(bodyEls)
-          .map(el => el.textContent?.trim())
-          .filter(t => t && t.length > 10)
-          .join(' ')
-          .slice(0, 800);
-        return { title, author, date, body };
-      })()
-    `);
-        if (!data?.title) {
-            throw new CliError('NOT_FOUND', 'Article not found or failed to load', 'Check the article ID');
-        }
-        if (!data.body) {
-            throw new CliError('PARSE_ERROR', 'Article body not found', '36kr page loaded but no article body paragraphs were extracted');
-        }
-        return [
-            { field: 'title', value: data.title },
-            { field: 'author', value: data.author || '' },
-            { field: 'date', value: data.date || '' },
-            { field: 'url', value: `https://36kr.com/p/${articleId}` },
-            { field: 'body', value: data.body || '' },
-        ];
-    },
-});

package/clis/36kr/article.test.js

@@ -1,46 +0,0 @@
-import { describe, expect, it, vi } from 'vitest';
-import { getRegistry } from '@agentrhq/webcmd/registry';
-import { CliError } from '@agentrhq/webcmd/errors';
-import './article.js';
-
-function makePage(evaluateResult) {
-    return {
-        installInterceptor: vi.fn().mockResolvedValue(undefined),
-        goto: vi.fn().mockResolvedValue(undefined),
-        wait: vi.fn().mockResolvedValue(undefined),
-        evaluate: vi.fn().mockResolvedValue(evaluateResult),
-    };
-}
-
-describe('36kr article', () => {
-    it('emits empty-string for missing optional author / date instead of a sentinel', async () => {
-        const command = getRegistry().get('36kr/article');
-        expect(command?.func).toBeDefined();
-        const page = makePage({ title: 'Real Title', author: '', date: '', body: 'Real article body' });
-        const rows = await command.func(page, { id: '1234567' });
-        const byField = Object.fromEntries(rows.map((r) => [r.field, r.value]));
-        expect(byField.title).toBe('Real Title');
-        expect(byField.author).toBe('');
-        expect(byField.date).toBe('');
-        expect(byField.body).toBe('Real article body');
-        expect(byField.url).toBe('https://36kr.com/p/1234567');
-    });
-
-    it('throws CliError NOT_FOUND when the page exposes no title', async () => {
-        const command = getRegistry().get('36kr/article');
-        const page = makePage({ title: '', author: 'x', date: 'y', body: 'z' });
-        await expect(command.func(page, { id: '1234567' })).rejects.toBeInstanceOf(CliError);
-    });
-
-    it('throws CliError PARSE_ERROR when the page exposes title but no body', async () => {
-        const command = getRegistry().get('36kr/article');
-        const page = makePage({ title: 'Real Title', author: 'x', date: 'y', body: '' });
-        await expect(command.func(page, { id: '1234567' })).rejects.toMatchObject({ code: 'PARSE_ERROR' });
-    });
-
-    it('throws CliError INVALID_ARGUMENT when no numeric id can be parsed', async () => {
-        const command = getRegistry().get('36kr/article');
-        const page = makePage({});
-        await expect(command.func(page, { id: 'not-a-url' })).rejects.toBeInstanceOf(CliError);
-    });
-});