@agentrhq/webcmd 0.1.1 → 0.2.0

package/clis/jianyu/search.test.js

@@ -1,153 +0,0 @@
-import { describe, expect, it } from 'vitest';
-import { __test__ } from './search.js';
-describe('jianyu search helpers', () => {
-    it('builds candidate URLs with supsearch as preferred entry', () => {
-        const candidates = __test__.buildSearchCandidates('procurement');
-        expect(candidates[0]).toContain('/jylab/supsearch/index.html');
-        expect(candidates[0]).toContain('keywords=procurement');
-        expect(candidates[0]).toContain('selectType=title');
-        expect(candidates[0]).toContain('searchGroup=1');
-    });
-    it('builds supsearch URL with required query params', () => {
-        const url = __test__.buildSearchUrl('procurement');
-        expect(url).toContain('keywords=procurement');
-        expect(url).toContain('selectType=title');
-        expect(url).toContain('searchGroup=1');
-    });
-    it('normalizes common date formats', () => {
-        expect(__test__.normalizeDate('2026-4-7')).toBe('2026-04-07');
-        expect(__test__.normalizeDate('2026年4月7日')).toBe('2026-04-07');
-        expect(__test__.normalizeDate('发布时间: 2026/04/07 09:00')).toBe('2026-04-07');
-    });
-    it('deduplicates by title and url', () => {
-        const deduped = __test__.dedupeCandidates([
-            { title: 'A', url: 'https://example.com/1', date: '2026-04-07' },
-            { title: 'A', url: 'https://example.com/1', date: '2026-04-07' },
-            { title: 'A', url: 'https://example.com/2', date: '2026-04-07' },
-        ]);
-        expect(deduped).toHaveLength(2);
-    });
-    it('filters obvious navigation rows before quality gate', () => {
-        const filtered = __test__.filterNavigationRows('电梯', [
-            { title: '招标公告', url: 'https://www.jianyu360.cn/list/stype/ZBGG.html', date: '' },
-            { title: '帮助中心', url: 'https://www.jianyu360.cn/helpCenter/index', date: '' },
-            { title: '某项目电梯采购公告', url: 'https://www.jianyu360.cn/notice/detail/123', date: '2026-04-07' },
-        ]);
-        expect(filtered).toHaveLength(1);
-        expect(filtered[0].title).toContain('电梯采购公告');
-    });
-    it('rejects procurement rows that do not contain query evidence', () => {
-        const filtered = __test__.filterNavigationRows('电梯', [
-            {
-                title: '某项目采购公告',
-                url: 'https://www.jianyu360.cn/notice/detail/123',
-                date: '2026-04-07',
-                contextText: '招标公告 项目编号：ABC-123',
-            },
-        ]);
-        expect(filtered).toHaveLength(0);
-    });
-    it('parses search-index markdown headings', () => {
-        const rows = __test__.parseSearchIndexMarkdown(`
-## [标题一](http://duckduckgo.com/l/?uddg=https%3A%2F%2Fbeijing.jianyu360.cn%2Fjybx%2F20260401_26033143187897.html)
-## [标题二](https://www.jianyu360.cn/nologin/content/ABC.html)
-`);
-        expect(rows).toHaveLength(2);
-        expect(rows[0].title).toBe('标题一');
-        expect(rows[1].url).toContain('jianyu360.cn/nologin/content');
-    });
-    it('unwraps duckduckgo redirect links', () => {
-        const direct = __test__.unwrapDuckDuckGoUrl('https://duckduckgo.com/l/?uddg=https%3A%2F%2Fwww.jianyu360.cn%2Fnologin%2Fcontent%2FXYZ.html');
-        expect(direct).toBe('https://www.jianyu360.cn/nologin/content/XYZ.html');
-    });
-    it('extracts publish date from jianyu jybx urls', () => {
-        const date = __test__.extractDateFromJianyuUrl('https://shandong.jianyu360.cn/jybx/20260310_26030938267551.html');
-        expect(date).toBe('2026-03-10');
-    });
-    it('normalizes api payload rows with fallback url/title fields', () => {
-        const normalized = __test__.normalizeApiRow({
-            noticeTitle: '某项目电梯采购公告',
-            detailUrl: '/jybx/20260310_26030938267551.html',
-            publishTime: '2026-03-10 09:00:00',
-            buyer: '测试单位',
-        });
-        expect(normalized).toBeTruthy();
-        expect(normalized?.title).toContain('电梯采购公告');
-        expect(normalized?.url).toContain('/jybx/20260310_26030938267551.html');
-        expect(normalized?.date).toBe('2026-03-10');
-    });
-    it('keeps later bucket unique api rows before final filtering and slice', async () => {
-        const page = {
-            goto: async () => { },
-            wait: async () => { },
-            evaluate: async () => ({
-                challenge: false,
-                responses: [
-                    {
-                        type: 'fType',
-                        ok: true,
-                        status: 200,
-                        payload: {
-                            list: [
-                                {
-                                    noticeTitle: '某项目电梯采购公告',
-                                    detailUrl: '/jybx/20260310_26030938267551.html',
-                                    publishTime: '2026-03-10 09:00:00',
-                                },
-                                {
-                                    noticeTitle: '某项目电梯采购公告',
-                                    detailUrl: '/jybx/20260310_26030938267551.html',
-                                    publishTime: '2026-03-10 09:00:00',
-                                },
-                            ],
-                        },
-                    },
-                    {
-                        type: 'eType',
-                        ok: true,
-                        status: 200,
-                        payload: {
-                            list: [
-                                {
-                                    noticeTitle: '另一条电梯采购公告',
-                                    detailUrl: '/jybx/20260311_26030938267552.html',
-                                    publishTime: '2026-03-11 09:00:00',
-                                },
-                            ],
-                        },
-                    },
-                ],
-            }),
-        };
-        const result = await __test__.fetchJianyuApiRows(page, '电梯', 1);
-        expect(result.challenge).toBe(false);
-        expect(result.rows).toHaveLength(2);
-        expect(result.rows[0].title).toContain('电梯采购公告');
-        expect(result.rows[1].title).toContain('另一条电梯采购公告');
-    });
-    it('classifies nologin links as blocked detail targets', () => {
-        const signal = __test__.classifyDetailStatus('https://www.jianyu360.cn/nologin/content/ABC.html');
-        expect(signal.detail_status).toBe('blocked');
-    });
-    it('classifies accessible detail urls as ok even when they are not jybx paths', () => {
-        const signal = __test__.classifyDetailStatus('https://www.jianyu360.cn/notice/detail/123');
-        expect(signal.detail_status).toBe('ok');
-        expect(signal.detail_reason).toBe('detail_candidate');
-    });
-    it('classifies list pages as entry_only', () => {
-        const signal = __test__.classifyDetailStatus('https://www.jianyu360.cn/list/stype/ZBGG.html');
-        expect(signal.detail_status).toBe('entry_only');
-    });
-    it('extracts stable notice id from jybx urls', () => {
-        const id = __test__.extractNoticeId('https://shandong.jianyu360.cn/jybx/20260310_26030938267551.html');
-        expect(id).toBe('20260310_26030938267551');
-    });
-    it('keeps only rows inside recency window', () => {
-        const within = __test__.isWithinSinceDays('2026-03-20', 30, new Date('2026-04-09T00:00:00Z'));
-        const stale = __test__.isWithinSinceDays('2026-02-01', 30, new Date('2026-04-09T00:00:00Z'));
-        const missing = __test__.isWithinSinceDays('', 30, new Date('2026-04-09T00:00:00Z'));
-        expect(within).toBe(true);
-        expect(stale).toBe(false);
-        expect(missing).toBe(false);
-    });
-});

package/clis/jianyu/shared/china-bid-search.js

@@ -1,165 +0,0 @@
-import { cleanText, normalizeDate, } from './procurement-contract.js';
-export { cleanText, normalizeDate };
-export function dedupeCandidates(items) {
-    const deduped = [];
-    const seen = new Set();
-    for (const item of items) {
-        const key = `${cleanText(item.title)}\t${cleanText(item.url)}`;
-        if (seen.has(key))
-            continue;
-        seen.add(key);
-        deduped.push(item);
-    }
-    return deduped;
-}
-function withQuery(baseUrl, key, query) {
-    try {
-        const url = new URL(baseUrl);
-        url.searchParams.set(key, query);
-        return url.toString();
-    }
-    catch {
-        return null;
-    }
-}
-export function buildSearchCandidates(query, baseEntries, queryKeys = ['keyword', 'keywords', 'q', 'search', 'title']) {
-    const keyword = cleanText(query);
-    const candidates = [];
-    if (keyword) {
-        for (const entry of baseEntries) {
-            for (const key of queryKeys) {
-                const withKeyword = withQuery(entry, key, keyword);
-                if (withKeyword)
-                    candidates.push(withKeyword);
-            }
-        }
-    }
-    candidates.push(...baseEntries);
-    const ordered = [];
-    const seen = new Set();
-    for (const item of candidates) {
-        const value = cleanText(item);
-        if (!value || seen.has(value))
-            continue;
-        seen.add(value);
-        ordered.push(value);
-    }
-    return ordered;
-}
-export async function detectAuthPrompt(page) {
-    const pageText = cleanText(await page.evaluate('document.body ? document.body.innerText : ""'));
-    return /(请先登录|未登录|登录后|验证码|人机验证|权限不足|无权限|请完善信息后访问)/.test(pageText);
-}
-export async function searchRowsFromEntries(page, { query, candidateUrls, allowedHostFragments, limit, }) {
-    const queryText = cleanText(query);
-    const rows = [];
-    for (const targetUrl of candidateUrls) {
-        await page.goto(targetUrl);
-        await page.wait(2);
-        const payload = await page.evaluate(`
-      (() => {
-        const clean = (value) => (value || '').replace(/\\s+/g, ' ').trim();
-        const parseDate = (text) => {
-          const normalized = clean(text);
-          const match = normalized.match(/(20\\d{2})[.\\-/年](\\d{1,2})[.\\-/月](\\d{1,2})/);
-          if (!match) return '';
-          return match[1] + '-' + String(match[2]).padStart(2, '0') + '-' + String(match[3]).padStart(2, '0');
-        };
-        const toAbsolute = (href) => {
-          if (!href) return '';
-          if (href.startsWith('http://') || href.startsWith('https://')) return href;
-          if (href.startsWith('/')) return new URL(href, window.location.origin).toString();
-          return '';
-        };
-
-        const token = ${JSON.stringify(queryText)};
-        const tokenParts = token.split(/\\s+/).filter(Boolean).map((part) => part.toLowerCase());
-        const allowedHosts = ${JSON.stringify(allowedHostFragments.map((item) => item.toLowerCase()))};
-        const procurementHints = ['招标', '采购', '公告', '项目', '中标', '成交', '询价', '竞价', '比选', '投标', 'notice', 'tender', 'procurement', 'bidding'];
-        const rowSelectors = [
-          'table tbody tr',
-          'table tr',
-          'ul li',
-          'ol li',
-          'article',
-          'section',
-          '.list li',
-          '.notice li',
-          '[class*="list"] li',
-          '[class*="notice"] li',
-          '[class*="item"]',
-          '[class*="row"]',
-        ];
-
-        const rowNodes = [];
-        const rowSeen = new Set();
-        for (const selector of rowSelectors) {
-          const nodes = Array.from(document.querySelectorAll(selector));
-          for (const node of nodes) {
-            const text = clean(node.innerText || node.textContent || '');
-            if (!text || text.length < 8) continue;
-            const lowerText = text.toLowerCase();
-            const hasDate = /(20\\d{2})[.\\-/年](\\d{1,2})[.\\-/月](\\d{1,2})/.test(text);
-            const hasHint = procurementHints.some((hint) => lowerText.includes(hint));
-            const hasQuery = tokenParts.length === 0 || tokenParts.some((part) => lowerText.includes(part));
-            if (!hasDate && !hasHint && !hasQuery) continue;
-            if (rowSeen.has(node)) continue;
-            rowSeen.add(node);
-            rowNodes.push(node);
-          }
-        }
-
-        const rows = [];
-        const seen = new Set();
-        for (const node of rowNodes) {
-          const contextText = clean(node.innerText || node.textContent || '');
-          const contextLower = contextText.toLowerCase();
-          const hasHint = procurementHints.some((hint) => contextLower.includes(hint));
-          const hasQuery = tokenParts.length === 0 || tokenParts.some((part) => contextLower.includes(part));
-          if (!hasHint && !hasQuery) continue;
-
-          const anchors = Array.from(node.querySelectorAll('a[href]'));
-          for (const anchor of anchors) {
-            const title = clean(anchor.textContent || '');
-            if (!title || title.length < 4) continue;
-            const url = toAbsolute(anchor.getAttribute('href') || anchor.href || '');
-            if (!url) continue;
-            const lowerUrl = url.toLowerCase();
-            const hostMatched = allowedHosts.length === 0 || allowedHosts.some((item) => lowerUrl.includes(item));
-            if (!hostMatched) continue;
-
-            const key = title + '\\t' + url;
-            if (seen.has(key)) continue;
-            seen.add(key);
-            rows.push({
-              title,
-              url,
-              date: parseDate(contextText),
-              contextText,
-            });
-          }
-        }
-        return rows;
-      })()
-    `);
-        if (Array.isArray(payload)) {
-            for (const item of payload) {
-                if (!item || typeof item !== 'object')
-                    continue;
-                const row = item;
-                const candidate = {
-                    title: cleanText(row.title),
-                    url: cleanText(row.url),
-                    date: normalizeDate(cleanText(row.date)),
-                    contextText: cleanText(row.contextText),
-                };
-                if (!candidate.title || !candidate.url)
-                    continue;
-                rows.push(candidate);
-            }
-        }
-        if (rows.length >= limit)
-            break;
-    }
-    return dedupeCandidates(rows).slice(0, limit);
-}

package/clis/jianyu/shared/procurement-contract.js

@@ -1,324 +0,0 @@
-const PROCUREMENT_HINTS = [
-    '招标',
-    '采购',
-    '公告',
-    '项目',
-    '中标',
-    '成交',
-    '流标',
-    '终止',
-    '询价',
-    '竞价',
-    '比选',
-    '投标',
-    'tender',
-    'procurement',
-    'bidding',
-    'bid',
-    'notice',
-];
-const RESULT_HINTS = [
-    '中标',
-    '成交',
-    '结果',
-    '候选人',
-    '中选',
-    '定标',
-    '评标',
-    'award',
-    'winner',
-];
-const NOTICE_HINTS = [
-    '招标',
-    '采购',
-    '询价',
-    '比选',
-    '公告',
-    '竞争性',
-    '邀请',
-    '投标',
-    'tender',
-    'procurement',
-    'notice',
-];
-const NEWS_HINTS = [
-    '新闻',
-    '资讯',
-    '动态',
-    '政策',
-    '简讯',
-    'news',
-    'article',
-];
-const NAVIGATION_HINTS = [
-    '首页',
-    '官网',
-    '网站地图',
-    '联系我们',
-    '帮助中心',
-    'english',
-    'login',
-    '注册',
-    '导航',
-    '法规',
-    '政策文件',
-    '服务平台',
-    '信用中国',
-];
-const DETAIL_URL_HINTS = [
-    '/detail',
-    '/content',
-    '/jybx/',
-    '/notice',
-    '/article',
-    '/view',
-    '/project',
-    '/bid',
-    'detail=',
-    'id=',
-];
-const LIST_URL_HINTS = [
-    '/search',
-    '/list',
-    '/index',
-    '/home',
-    '/portal',
-    '/channel',
-    'page=',
-];
-const OWNER_PATTERNS = [
-    /(?:招标人|采购人|业主|建设单位|项目单位)\s*[：:]\s*([^\n，。；]{2,60})/i,
-];
-const CODE_PATTERNS = [
-    /(?:项目编号|招标编号|采购编号|项目编码|项目代码|编号)\s*[：:]\s*([A-Za-z0-9\-_/]{4,60})/i,
-];
-const BUDGET_PATTERNS = [
-    /(?:预算(?:金额)?|控制价|最高限价|限价|采购金额|合同估算价)\s*[：:]\s*([^\n，。；]{2,80})/i,
-];
-const DEADLINE_PATTERNS = [
-    /(?:报名截止时间|投标截止时间|开标时间|响应文件递交截止时间|截止时间|开标日期)\s*[：:]\s*([^\n，。；]{2,80})/i,
-];
-const DATE_PATTERN = /(20\d{2})[.\-/年](\d{1,2})[.\-/月](\d{1,2})/;
-export function cleanText(value) {
-    return typeof value === 'string' ? value.replace(/\s+/g, ' ').trim() : '';
-}
-export function normalizeDate(raw) {
-    const normalized = cleanText(raw);
-    const match = normalized.match(DATE_PATTERN);
-    if (!match)
-        return '';
-    const year = match[1];
-    const month = match[2].padStart(2, '0');
-    const day = match[3].padStart(2, '0');
-    return `${year}-${month}-${day}`;
-}
-function uniqueInOrder(values) {
-    const ordered = [];
-    const seen = new Set();
-    for (const value of values) {
-        const text = cleanText(value);
-        if (!text || seen.has(text))
-            continue;
-        seen.add(text);
-        ordered.push(text);
-    }
-    return ordered;
-}
-function containsAny(haystack, needles) {
-    return needles.some((needle) => haystack.includes(needle.toLowerCase()));
-}
-function extractByPatterns(text, patterns) {
-    for (const pattern of patterns) {
-        const matched = text.match(pattern);
-        if (matched?.[1])
-            return cleanText(matched[1]);
-    }
-    return '';
-}
-function deriveSnippet(text) {
-    const normalized = cleanText(text);
-    if (!normalized)
-        return '';
-    return normalized.slice(0, 220);
-}
-function splitEvidenceBlocks(text, query) {
-    const normalized = cleanText(text);
-    if (!normalized)
-        return [];
-    const queryTokens = query
-        .split(/\s+/)
-        .map((item) => item.toLowerCase().trim())
-        .filter(Boolean);
-    const chunks = normalized
-        .split(/[。！？；\n]/)
-        .map((chunk) => cleanText(chunk))
-        .filter(Boolean);
-    const ranked = chunks
-        .map((chunk) => {
-        const lower = chunk.toLowerCase();
-        const tokenScore = queryTokens.length === 0
-            ? 0
-            : queryTokens.reduce((score, token) => (lower.includes(token) ? score + 2 : score), 0);
-        const procurementScore = containsAny(lower, PROCUREMENT_HINTS) ? 1 : 0;
-        return {
-            chunk,
-            score: tokenScore + procurementScore,
-        };
-    })
-        .sort((a, b) => b.score - a.score || b.chunk.length - a.chunk.length)
-        .slice(0, 5)
-        .map((item) => item.chunk);
-    return uniqueInOrder(ranked);
-}
-function classifyContentType(title, url, contextText) {
-    const haystack = `${title} ${contextText} ${url}`.toLowerCase();
-    if (containsAny(haystack, RESULT_HINTS))
-        return 'result';
-    if (containsAny(haystack, NOTICE_HINTS))
-        return 'notice';
-    if (containsAny(haystack, NEWS_HINTS))
-        return 'news';
-    if (containsAny(haystack, NAVIGATION_HINTS))
-        return 'navigation';
-    return 'unknown';
-}
-function isDetailPage(url) {
-    const lower = cleanText(url).toLowerCase();
-    if (!lower)
-        return false;
-    const hasDetailToken = DETAIL_URL_HINTS.some((hint) => lower.includes(hint));
-    if (!hasDetailToken)
-        return false;
-    const hasListToken = LIST_URL_HINTS.some((hint) => lower.includes(hint));
-    return !hasListToken;
-}
-function buildQualityFlags(core) {
-    const flags = [];
-    if (!core.project_owner)
-        flags.push('missing_project_owner');
-    if (!core.project_code)
-        flags.push('missing_project_code');
-    if (!core.budget_or_limit)
-        flags.push('missing_budget');
-    if (!core.deadline_or_open_time)
-        flags.push('missing_deadline');
-    if (core.content_type === 'navigation')
-        flags.push('navigation_risk');
-    if (!core.is_detail_page)
-        flags.push('list_page_url');
-    return flags;
-}
-function queryMatched(text, query) {
-    const tokenParts = query
-        .split(/\s+/)
-        .map((part) => part.toLowerCase().trim())
-        .filter(Boolean);
-    if (tokenParts.length === 0)
-        return true;
-    const lower = text.toLowerCase();
-    return tokenParts.some((part) => lower.includes(part));
-}
-function normalizeCoreRecord(row, { sourceSite, }) {
-    const title = cleanText(row.title);
-    const url = cleanText(row.url);
-    const contextText = cleanText(row.contextText);
-    const date = normalizeDate(cleanText(row.date || contextText));
-    const publishTime = date;
-    const contentType = classifyContentType(title, url, contextText);
-    const projectOwner = extractByPatterns(contextText, OWNER_PATTERNS);
-    const projectCode = extractByPatterns(contextText, CODE_PATTERNS);
-    const budget = extractByPatterns(contextText, BUDGET_PATTERNS);
-    const deadline = extractByPatterns(contextText, DEADLINE_PATTERNS);
-    const snippet = deriveSnippet(contextText || title);
-    const core = {
-        title,
-        url,
-        date,
-        publish_time: publishTime,
-        source_site: sourceSite,
-        is_detail_page: isDetailPage(url),
-        content_type: contentType,
-        project_owner: projectOwner,
-        project_code: projectCode,
-        budget_or_limit: budget,
-        deadline_or_open_time: deadline,
-        snippet,
-        summary: snippet,
-        quality_flags: [],
-    };
-    core.quality_flags = buildQualityFlags(core);
-    return core;
-}
-function qualityRejectReason(core, query) {
-    if (!core.title || !core.url)
-        return 'missing_identity';
-    if (core.content_type === 'navigation')
-        return 'navigation_only';
-    const searchable = `${core.title} ${core.snippet} ${core.url}`.toLowerCase();
-    const hasQuery = queryMatched(searchable, query);
-    if (!hasQuery)
-        return 'query_mismatch';
-    return null;
-}
-function dedupeByTitleUrl(items) {
-    const deduped = [];
-    const seen = new Set();
-    for (const item of items) {
-        const key = `${item.title}\t${item.url}`;
-        if (seen.has(key))
-            continue;
-        seen.add(key);
-        deduped.push(item);
-    }
-    return deduped;
-}
-export function formatTaxonomyError(taxonomy, { site, command, detail, }) {
-    return `[taxonomy=${taxonomy}] site=${site} command=${command} ${cleanText(detail)}`;
-}
-export function taxonomyError(taxonomy, context) {
-    return new Error(formatTaxonomyError(taxonomy, context));
-}
-export function toProcurementSearchRecords(rows, { site, query, limit, }) {
-    const normalizedRows = dedupeByTitleUrl(rows.map((row) => normalizeCoreRecord(row, { sourceSite: site })));
-    const accepted = [];
-    for (const row of normalizedRows) {
-        const rejectReason = qualityRejectReason(row, query);
-        if (rejectReason)
-            continue;
-        accepted.push(row);
-    }
-    if (normalizedRows.length > 0 && accepted.length === 0) {
-        throw taxonomyError('extraction_drift', {
-            site,
-            command: 'search',
-            detail: `all rows rejected by quality gate (raw=${normalizedRows.length})`,
-        });
-    }
-    return accepted
-        .slice(0, Math.max(1, limit))
-        .map((row, index) => ({
-        rank: index + 1,
-        ...row,
-    }));
-}
-export function toProcurementDetailRecord({ title, url, contextText, publishTime, }, { site, query = '', }) {
-    const core = normalizeCoreRecord({
-        title,
-        url,
-        date: publishTime,
-        contextText,
-    }, { sourceSite: site });
-    const detailText = cleanText(contextText).slice(0, 6000);
-    const evidenceBlocks = splitEvidenceBlocks(detailText, query);
-    return {
-        ...core,
-        detail_text: detailText,
-        evidence_blocks: evidenceBlocks,
-    };
-}
-export const __test__ = {
-    classifyContentType,
-    isDetailPage,
-    splitEvidenceBlocks,
-    qualityRejectReason,
-};