From e8a38f493e17c70acb2ffa586826ec415c40e010 Mon Sep 17 00:00:00 2001
From: "WIN-C27C5MSVTPO\\16961" <1696136552@qq.com>
Date: Tue, 16 Dec 2025 14:48:18 +0800
Subject: [PATCH] =?UTF-8?q?=E5=AF=BC=E5=85=A5?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 app/admin/controller/SkuController.php | 78 +++++++++++---------------
 1 file changed, 33 insertions(+), 45 deletions(-)

diff --git a/app/admin/controller/SkuController.php b/app/admin/controller/SkuController.php
index bffb8cf..8f420fe 100644
--- a/app/admin/controller/SkuController.php
+++ b/app/admin/controller/SkuController.php
@@ -264,10 +264,10 @@ class SkuController extends BaseController
             return '';
         }
 
-        // 1. 加载 docx
-        $phpWord = WordIOFactory::load($docxPath, 'Word2007');
+        // 1. 读取 docx
+        $phpWord = WordIOFactory::load($docxPath);
 
-        // 2. 临时目录（非常重要）
+        // 2. 导出 HTML
         $tmpDir = runtime_path('tmp/' . uniqid());
         if (!is_dir($tmpDir)) {
             mkdir($tmpDir, 0755, true);
@@ -275,40 +275,40 @@ class SkuController extends BaseController
 
         $tmpHtml = $tmpDir . '/index.html';
 
-        // 3. 导出 HTML
         $writer = WordIOFactory::createWriter($phpWord, 'HTML');
         $writer->save($tmpHtml);
 
         $html = file_get_contents($tmpHtml);
 
-        // 4. 清洗 Word 垃圾 HTML（关键）
+        // 3. 只取 body 内容（关键）
+        if (preg_match('/<body[^>]*>(.*?)<\/body>/is', $html, $m)) {
+            $html = $m[1];
+        }
+
+        // 4. 清洗 Word 垃圾样式（关键）
         $html = $this->cleanWordHtml($html);
 
+        // 5. 处理图片并上传 OSS
         $uploadService = new UploadService();
 
-        // 5. 处理图片
         $html = preg_replace_callback(
             '/<img[^>]+src="([^"]+)"[^>]*>/i',
             function ($matches) use ($tmpDir, $uploadService) {
                 $src = $matches[1];
 
-                // 只处理本地图片
-                if (preg_match('#^(http|https|data:)#i', $src)) {
-                    return $matches[0];
+                // Word 导出的是相对路径
+                if (!str_starts_with($src, 'http')) {
+                    $localImg = $tmpDir . '/' . ltrim($src, '/');
+                    if (is_file($localImg)) {
+                        $ossUrl = $uploadService->uploadLocalFile([
+                            'path'  => $localImg,
+                            'scene' => 'desc'
+                        ]);
+                        return preg_replace('/src="[^"]+"/', 'src="' . $ossUrl . '"', $matches[0]);
+                    }
                 }
 
-                $localImg = realpath($tmpDir . '/' . $src);
-                if (!$localImg || !is_file($localImg)) {
-                    return $matches[0];
-                }
-
-                // 上传 OSS
-                $ossUrl = $uploadService->uploadLocalFile([
-                    'path'  => $localImg,
-                    'scene' => 'desc'
-                ]);
-
-                return str_replace($src, $ossUrl, $matches[0]);
+                return $matches[0];
             },
             $html
         );
@@ -319,44 +319,32 @@ class SkuController extends BaseController
         return trim($html);
     }
 
-    function deleteDir(string $dir)
+    function deleteDir(string $dir): void
     {
         if (!is_dir($dir)) return;
         foreach (scandir($dir) as $file) {
             if ($file === '.' || $file === '..') continue;
             $path = $dir . '/' . $file;
-            is_dir($path) ? $this->deleteDir($path) : unlink($path);
+            is_dir($path) ? deleteDir($path) : @unlink($path);
         }
-        rmdir($dir);
+        @rmdir($dir);
     }
 
     function cleanWordHtml(string $html): string
     {
-        // 删除 Word 注释
-        $html = preg_replace('/<!--(.|\s)*?-->/', '', $html);
-
-        // 删除 style / meta / link
-        $html = preg_replace('/<(style|meta|link)[^>]*>.*?<\/\1>/is', '', $html);
-
-        // 删除 xml 标签
-        $html = preg_replace('/<xml>.*?<\/xml>/is', '', $html);
-
-        // 删除 mso-xxx 样式
+        // 移除 Office 命名空间、样式
         $html = preg_replace('/\s*mso-[^:]+:[^;"]+;?/i', '', $html);
+        $html = preg_replace('/\s*class="[^"]*"/i', '', $html);
+        $html = preg_replace('/\s*style="[^"]*"/i', '', $html);
 
-        // 清理 class="Mso..."
-        $html = preg_replace('/class="Mso[^"]*"/i', '', $html);
+        // 移除多余标签
+        $html = preg_replace('/<(\/?)(html|head|meta|link|span|font|xml|o:p)[^>]*>/i', '', $html);
 
-        // 删除空 span
-        $html = preg_replace('/<span[^>]*>\s*<\/span>/i', '', $html);
+        // 清理空段落
+        $html = preg_replace('/<p>\s*<\/p>/i', '', $html);
 
-        // 删除 Office 条件注释
-        $html = preg_replace('/<!\[if.*?\]>.*?<!\[endif\]>/is', '', $html);
-
-        // 只保留 body 内容
-        if (preg_match('/<body[^>]*>(.*?)<\/body>/is', $html, $m)) {
-            $html = $m[1];
-        }
+        // 规范段落
+        $html = str_replace(['<p>', '</p>'], ["<p>", "</p>\n"], $html);
 
         return $html;
     }