@sns-parse/core 0.3.0-alpha.1 → 0.4.0-alpha.10

package/lib/utils/url.js

@@ -0,0 +1,89 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.cleanUrl = cleanUrl;
+exports.linkTypeParser = linkTypeParser;
+exports.extractAllUrlsFromMessage = extractAllUrlsFromMessage;
+function cleanUrl(url) {
+    url = url.replace(/&/g, '&')
+        .replace(/"/g, '"')
+        .replace(/&lt;/g, '<')
+        .replace(/&gt;/g, '>')
+        .replace(/\\\//g, '/');
+    url = url.replace(/^[\s"'<（(“”‘’]+/, '');
+    url = url.replace(/[\s"'<>\{\}\[\]`,;，。！？：；“”‘’…—～.()）]+$/, '');
+    // 剔除链接后残留的 XML/HTML 标签片段（如卡片消息里的 <op>…（上游 v1.6.7）
+    const tagStart = url.indexOf('<');
+    if (tagStart > 0)
+        url = url.slice(0, tagStart);
+    if (!/^https?:\/\//i.test(url)) {
+        if (/^\/\//.test(url))
+            url = 'https:' + url;
+        else
+            return url;
+    }
+    return url;
+}
+function linkTypeParser(content, rules) {
+    content = content.replace(/\\\//g, '/');
+    const matches = [];
+    const seen = new Set();
+    for (const rule of rules) {
+        let match;
+        rule.pattern.lastIndex = 0;
+        while ((match = rule.pattern.exec(content)) !== null) {
+            let url = match[0];
+            url = cleanUrl(url);
+            if (!url)
+                continue;
+            if (seen.has(url))
+                continue;
+            seen.add(url);
+            // 记录出现位置，多链接按正文出现顺序排列（上游 v1.6.7）
+            matches.push({ type: rule.type, url, id: match[1] || url, pos: match.index });
+        }
+    }
+    return matches.sort((a, b) => a.pos - b.pos).map(({ pos, ...m }) => m);
+}
+function extractAllUrlsFromMessage(session, rules) {
+    const content = session.content?.trim() || '';
+    const matchedLinks = linkTypeParser(content, rules);
+    const cardsContent = [];
+    if (session.elements) {
+        for (const elem of session.elements) {
+            if (elem.type === 'xml' && elem.data)
+                cardsContent.push(elem.data);
+            else if (elem.type === 'json' && elem.data) {
+                try {
+                    const json = JSON.parse(elem.data);
+                    const extract = (obj) => {
+                        if (!obj || typeof obj !== 'object')
+                            return;
+                        for (const val of Object.values(obj)) {
+                            if (typeof val === 'string')
+                                cardsContent.push(val);
+                            else if (typeof val === 'object')
+                                extract(val);
+                        }
+                    };
+                    extract(json);
+                }
+                catch { }
+            }
+        }
+    }
+    for (const cardContent of cardsContent) {
+        matchedLinks.push(...linkTypeParser(cardContent, rules));
+    }
+    const cleanResult = [];
+    const seenUrls = new Set();
+    for (const link of matchedLinks) {
+        const cleaned = cleanUrl(link.url);
+        if (!cleaned || !/^https?:\/\//i.test(cleaned))
+            continue;
+        if (!seenUrls.has(cleaned)) {
+            seenUrls.add(cleaned);
+            cleanResult.push({ ...link, url: cleaned });
+        }
+    }
+    return cleanResult;
+}

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@sns-parse/core",
-  "version": "0.3.0-alpha.1+upstream.1.6.7",
-  "description": "sns-parse 核心契约与宿主抽象：类型、可注入日志、Host、发送层契约、平台定义",
+  "version": "0.4.0-alpha.10+upstream.1.6.7",
+  "description": "sns-parse 核心：契约、宿主抽象、配置 DSL 与解析引擎（fetcher/parser/翻译/X 原生解析/合并/GIF/TLS 指纹客户端）",
   "main": "lib/index.js",
   "types": "lib/index.d.ts",
   "files": [
@@ -11,7 +11,8 @@
   "scripts": {
     "build": "tsc",
     "clean": "rimraf lib",
-    "prepublishOnly": "npm run build"
+    "test": "tsx test/engine.test.ts",
+    "prepublishOnly": "tsc"
   },
   "keywords": [
     "sns-parse",
@@ -21,10 +22,20 @@
     "koishi"
   ],
   "devDependencies": {
+    "@char46/tlsget-rs": "^1.0.0",
     "@types/node": "^20.11.0",
     "rimraf": "^5.0.5",
+    "tsx": "^4.23.15",
     "typescript": "^5.3.3"
   },
+  "dependencies": {
+    "axios": "^1.16.1",
+    "https-proxy-agent": "^5.0.1"
+  },
+  "optionalDependencies": {
+    "@char46/tlsget-rs": "^1.0.0",
+    "ffmpeg-static": "^5.2.0"
+  },
   "publishConfig": {
     "access": "public",
     "registry": "https://registry.npmjs.org"
@@ -35,5 +46,6 @@
   },
   "engines": {
     "node": ">=16.0.0"
-  }
+  },
+  "packageManager": "pnpm@12.3.4"
 }