wangliang/crm-ai
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(bridge): 商品链接兜底与强制检索

Browse Source
This commit is contained in:
wangliang
2026-10-23 13:15:37 +08:00
parent 4d7b89492b
commit 22d96edee6
9 changed files with 414 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
bridge/src/output-guard.js
+11 -2
View File
@@ -120,14 +120,23 @@ function escapeRe(s) {
return String(s).replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
}
// 链接被删后不允许留下“点不动的链接文字”：通用 CTA（查看详情/View details…）整段去掉。
const CTA_LABEL_RE = /^(?:[👉➡️→]\s*)?(?:查看详情|查看商品|查看产品|点击查看|点此查看|立即查看|了解更多|详情|view\s+details?|see\s+details?|view\s+product|learn\s+more|shop\s+now|click\s+here)\s*[:：]?\s*$/i;
const CTA_LABEL_SOURCE = '(?:[👉➡️→]\\s*)?(?:查看详情|查看商品|查看产品|点击查看|点此查看|立即查看|了解更多|view\\s+details?|see\\s+details?|view\\s+product|learn\\s+more|shop\\s+now|click\\s+here)\\s*[:：]\\s*';
function removeUrls(content, urls) {
let out = String(content || '');
const labelOrNothing = (text) => (CTA_LABEL_RE.test(String(text).trim()) ? '' : text);
for (const url of urls) {
const e = escapeRe(url);
out = out.replace(new RegExp('\\[([^\\]]*)\\]\\(' + e + '\\)', 'g'), '$1');
out = out.replace(new RegExp('<a\\b[^>]*href=["\\\']' + e + '["\\\'][^>]*>([\\s\\S]*?)<\\/a>', 'gi'), '$1');
out = out.replace(new RegExp('\\[([^\\]]*)\\]\\(' + e + '\\)', 'g'), (_, text) => labelOrNothing(text));
out = out.replace(new RegExp('<a\\b[^>]*href=["\\\']' + e + '["\\\'][^>]*>([\\s\\S]*?)<\\/a>', 'gi'), (_, text) => labelOrNothing(text));
out = out.split(url).join('');
}
// 删除链接后残留的 “→ 查看详情：” 片段（连同前面的 <br>，避免出现空行）
out = out.replace(new RegExp('(?:<br\\s*\\/?>)?\\s*' + CTA_LABEL_SOURCE + '(?=<br\\s*\\/?>|<\\/p>|\\n|$)', 'gi'), '');
out = out.replace(/(?:<br\s*\/?>)?\s*[👉➡️→]\s*(?=<br\s*\/?>|<\/p>|\n|$)/gi, '');
out = out.replace(/<p>\s*<\/p>/gi, '').replace(/<p>\s*(?:<br\s*\/?>\s*)+/gi, '<p>');
return out.replace(/[ \t]{2,}/g, ' ').replace(/\n{3,}/g, '\n\n').trim();
}