wukong-gitlog-cli 1.0.34 → 1.0.35

package/CHANGELOG.md

@@ -2,6 +2,13 @@
 
 All notable changes to this project will be documented in this file. See [standard-version](https://github.com/conventional-changelog/standard-version) for commit guidelines.
 
+### [1.0.35](https://github.com/tomatobybike/wukong-gitlog-cli/compare/v1.0.34...v1.0.35) (2025-12-12)
+
+
+### Features
+
+* 🎸 createAuthorNormalizer ([a23a80e](https://github.com/tomatobybike/wukong-gitlog-cli/commit/a23a80e6d664cc5db89b40a63456d8e8d4b2c321))
+
 ### [1.0.34](https://github.com/tomatobybike/wukong-gitlog-cli/compare/v1.0.33...v1.0.34) (2025-12-12)
 
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "wukong-gitlog-cli",
-  "version": "1.0.34",
+  "version": "1.0.35",
   "description": "Advanced Git commit log exporter with Excel/JSON/TXT output, grouping, stats and CLI.",
   "keywords": [
     "git",

package/src/utils/authorNormalizer.mjs

@@ -1,42 +1,69 @@
 // authorNormalizer.js
-
 /**
- * 一个按邮箱自动归一化作者名字的工具。
+ * 按邮箱自动归一化作者名字
  * - 优先使用中文姓名
  * - 如果同邮箱出现多个非中文名，保持第一个
- * - 保留原始 author 以防 debug
+ * - 自动清理空格与不可见字符
+ * - 保留原始 author 以便 debug
  */
-
 export function createAuthorNormalizer() {
   const map = {} // email -> canonical name
+  const originalMap = {} // email -> 原始所有 author 名
 
   function isChinese(str) {
     return /[\u4e00-\u9fa5]/.test(str)
   }
 
+  function cleanName(str) {
+    if (!str) return ''
+    return str
+      // eslint-disable-next-line no-misleading-character-class
+      .replace(/[\u200B\u200C\u200D\uFEFF]/g, '') // 去零宽空格
+      .replace(/\s+/g, ' ') // 多空格 -> 单空格
+      .trim()
+  }
+
+  function cleanEmail(str) {
+    return (str || '').trim()
+  }
+
   function getAuthor(name, email) {
-    if (!email) return name
+    const cleanedName = cleanName(name)
+    const cleanedEmail = cleanEmail(email)
+
+    if (!cleanedEmail) return cleanedName || 'Unknown'
 
-    const canonical = map[email]
+    // 记录原始 author 便于 debug
+    if (!originalMap[cleanedEmail]) originalMap[cleanedEmail] = new Set()
+    if (cleanedName) originalMap[cleanedEmail].add(cleanedName)
+
+    const canonical = map[cleanedEmail]
 
     // 首次遇到这个邮箱 → 记录当前作者名
     if (!canonical) {
-      map[email] = name
-      return name
+      map[cleanedEmail] = cleanedName || cleanedEmail
+      return map[cleanedEmail]
     }
 
-    // 如果新的作者名是中文 → 覆盖旧的
-    if (isChinese(name)) {
-      map[email] = name
-      return name
+    // 新名是中文 → 覆盖旧的
+    if (isChinese(cleanedName)) {
+      map[cleanedEmail] = cleanedName
+      return map[cleanedEmail]
     }
 
-    // 新名不是中文 → 返回已有中文（或旧名）
+    // 新名非中文 → 保留已有中文或旧名
     return canonical
   }
 
   return {
     getAuthor,
-    getMap: () => map
+    getMap: () => map,
+    getOriginalMap: () => {
+      const res = {}
+      for (const [email, names] of Object.entries(originalMap)) {
+        res[email] = Array.from(names)
+      }
+      return res
+    }
   }
 }