From 9b956c449fa720e0a7d541888776bf38f80bd26d Mon Sep 17 00:00:00 2001 From: CaptainB Date: Wed, 17 Jun 2026 10:55:26 +0800 Subject: [PATCH] fix: handle multiple entries in zip file extraction to prevent data loss MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --bug=1071129@tapd-62980211 --user=刘瑞斌 【github#6238】从旧版本（2.3.1）升级到最新版（2.10.1-lts），下载知识库的word文档，文件损坏 https://www.tapd.cn/62980211/s/1962254 --- apps/knowledge/models/knowledge.py | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/apps/knowledge/models/knowledge.py b/apps/knowledge/models/knowledge.py index 8beefdb9d3b..3a74ff2d0b8 100644 --- a/apps/knowledge/models/knowledge.py +++ b/apps/knowledge/models/knowledge.py @@ -435,15 +435,19 @@ def get_bytes(self): buffer = io.BytesIO() for chunk in self.get_bytes_stream(): buffer.write(chunk) + data = buffer.getvalue() try: # 解压数据 with zipfile.ZipFile(buffer) as zip_file: + names = [name for name in zip_file.namelist() if not name.endswith("/")] + if len(names) != 1: + return data # 用 zip 内实际存储的条目名，避免文件名不匹配 - name = zip_file.namelist()[0] + name = names[0] return zip_file.read(name) - except Exception as e: + except zipfile.BadZipFile: # 如果数据不是zip格式，直接返回原始数据 - return buffer.getvalue() + return data def get_bytes_stream(self, start=0, end=None, chunk_size=64 * 1024): def _read_with_offset():