From ea0bde374237c37d497cbd930f9205dcb51f04fb Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?API=E6=96=87=E6=A1=A3=E7=BB=B4=E6=8A=A4Agent?=
 <api-doc-agent@xiaoxiajianji.com>
Date: Sun, 28 Jun 2026 15:58:07 +0800
Subject: [PATCH] fix: add security validations to duplication upload API

P0-1: Add MIME type whitelist validation (video files only)
- Accept: mp4, mpeg, mov, avi, webm, mkv, 3gp
- Reject non-video files with 415 Unsupported Media Type

P0-2: Add file size limit validation
- Use OSS_DIRECT_UPLOAD_MAX_MB config (default 800MB)
- Check Content-Length header before reading file
- Verify actual file size after reading
- Return 413 Request Entity Too Large if exceeded

Reference: apps/api/app/api/routes/upload.py implementation
---
 ...ables.py => 012_add_duplication_tables.py} |  8 +--
 apps/api/app/api/routes/duplication.py        | 54 ++++++++++++++++++-
 2 files changed, 56 insertions(+), 6 deletions(-)
 rename alembic/versions/{011_add_duplication_tables.py => 012_add_duplication_tables.py} (97%)

diff --git a/alembic/versions/011_add_duplication_tables.py b/alembic/versions/012_add_duplication_tables.py
similarity index 97%
rename from alembic/versions/011_add_duplication_tables.py
rename to alembic/versions/012_add_duplication_tables.py
index cea50acbf..8940b0e7c 100644
--- a/alembic/versions/011_add_duplication_tables.py
+++ b/alembic/versions/012_add_duplication_tables.py
@@ -1,7 +1,7 @@
 """Phase 2 - 查重功能：duplication_records + duplication_segments
 
-Revision ID: 011
-Revises: 010
+Revision ID: 012
+Revises: 011
 Create Date: 2026-06-28
 
 This migration creates two new tables:
@@ -12,8 +12,8 @@ from alembic import op
 import sqlalchemy as sa
 
 # revision identifiers
-revision = "011"
-down_revision = "010"
+revision = "012"
+down_revision = "011"
 branch_labels = None
 depends_on = None
 
diff --git a/apps/api/app/api/routes/duplication.py b/apps/api/app/api/routes/duplication.py
index 2c5c4fcec..c9d8b832c 100644
--- a/apps/api/app/api/routes/duplication.py
+++ b/apps/api/app/api/routes/duplication.py
@@ -30,6 +30,32 @@ logger = logging.getLogger(__name__)
 
 router = APIRouter()
 
+# 查重功能只接受视频文件
+ALLOWED_VIDEO_MIME_TYPES = frozenset({
+    "video/mp4", "video/mpeg", "video/quicktime", "video/x-msvideo",
+    "video/webm", "video/x-matroska", "video/3gpp",
+})
+
+
+def _validate_video_mime_type(content_type: str | None) -> str:
+    """验证视频文件的 MIME 类型，如果无效则抛出异常。"""
+    if not content_type:
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail="Content-Type header is required",
+        )
+    
+    # 处理带参数的类型，如 "video/mp4; charset=utf-8"
+    base_type = content_type.split(";")[0].strip().lower()
+    
+    if base_type not in ALLOWED_VIDEO_MIME_TYPES:
+        raise HTTPException(
+            status_code=status.HTTP_415_UNSUPPORTED_MEDIA_TYPE,
+            detail=f"只支持视频文件。支持的类型: mp4, mpeg, mov, avi, webm, mkv, 3gp",
+        )
+    
+    return base_type
+
 
 def _to_record_response(record: DuplicationRecord) -> DuplicationRecordResponse:
     return DuplicationRecordResponse(
@@ -86,6 +112,21 @@ async def upload_for_duplication(
             detail="文件名不能为空",
         )
 
+    # P0-1: 验证 MIME 类型（只接受视频文件）
+    validated_content_type = _validate_video_mime_type(file.content_type)
+
+    # P0-2: 验证文件大小（参考 OSS_DIRECT_UPLOAD_MAX_MB）
+    from app.config import get_settings
+    settings = get_settings()
+    max_size_bytes = settings.OSS_DIRECT_UPLOAD_MAX_MB * 1024 * 1024
+    
+    # 先检查 Content-Length header（如果可用）
+    if file.size is not None and file.size > max_size_bytes:
+        raise HTTPException(
+            status_code=status.HTTP_413_REQUEST_ENTITY_TOO_LARGE,
+            detail=f"文件超过上传限制 ({settings.OSS_DIRECT_UPLOAD_MAX_MB}MB)",
+        )
+
     # 读取文件内容并上传到 OSS
     file_id = uuid4().hex[:8]
     safe_filename = file.filename.replace("/", "_").replace("\\", "_")
@@ -94,6 +135,15 @@ async def upload_for_duplication(
     try:
         content = await file.read()
         file_size = len(content)
+        
+        # 再次检查实际文件大小
+        if file_size > max_size_bytes:
+            raise HTTPException(
+                status_code=status.HTTP_413_REQUEST_ENTITY_TOO_LARGE,
+                detail=f"文件超过上传限制 ({settings.OSS_DIRECT_UPLOAD_MAX_MB}MB)",
+            )
+    except HTTPException:
+        raise
     except Exception as exc:
         raise HTTPException(
             status_code=status.HTTP_400_BAD_REQUEST,
@@ -104,7 +154,7 @@ async def upload_for_duplication(
         storage_service.upload_file(
             content,
             storage_key,
-            content_type=file.content_type or "video/mp4",
+            content_type=validated_content_type,
         )
     except Exception as exc:
         raise HTTPException(
@@ -176,7 +226,7 @@ def delete_duplication_record(
     duplication_repository: Any = Depends(get_duplication_repository),
 ) -> None:
     """删除查重记录。"""
-    # 先检查记录是否存在且属于当前用户
+    # 检查记录是否存在且属于当前用户
     detail_uc = GetDuplicationDetailUseCase(duplication_repository)
     record = detail_uc.execute(record_id)
     if record is None or record.user_id != authenticated_user.user.id: