fix: 知识库文档上传(尾斜杠+FormData+真文件接收); AI回复强制Markdown; 平台全功能知识库注入系统提示

backend/python/app/api/knowledge.py

@@ -851,9 +851,12 @@ def get_documents(
 
 
 @router.post("/bases/{base_id}/documents")
-def upload_document(
+async def upload_document(
     base_id: str,
-    req: CreateDocumentRequest,
+    file: Optional[UploadFile] = File(None),
+    filename: str = Form(...),
+    file_type: Optional[str] = Form(None),
+    file_size: Optional[int] = Form(None),
     db: Session = Depends(get_db),
     current_user: User = Depends(get_current_user),
 ):
@@ -864,7 +867,7 @@ def upload_document(
         raise HTTPException(status_code=404, detail="知识库不存在")
 
     # 确定文件扩展名和存储路径
-    file_ext = (req.file_type or "txt").lower()
+    file_ext = (file_type or "txt").lower()
     if file_ext.startswith("."):
         file_ext = file_ext[1:]
 
@@ -877,28 +880,69 @@ def upload_document(
     stored_filename = f"{doc_id}.{file_ext}"
     stored_path = os.path.join(storage_dir, stored_filename)
 
-    # 文本解析：根据文件类型生成模拟文档内容
-    # 实际生产环境中应使用 PyPDF2/pdfplumber（PDF）、python-docx（Word）、openpyxl（Excel）等库
-    file_size = req.file_size or len(req.filename.encode('utf-8')) * 100
-    content = _generate_document_content(req.filename, file_ext, file_size)
+    # 文件内容解析
+    actual_size = file_size or 0
+    content = ""
+    uploaded_content = b""
 
-    # 将内容写入存储文件
-    try:
-        with open(stored_path, "w", encoding="utf-8") as f:
-            f.write(content)
-        logger.info(f"知识库文档已写入: {stored_path}, size={len(content)} chars")
-    except Exception as e:
-        logger.error(f"知识库文档写入失败: {stored_path}, error={e}")
-        raise HTTPException(status_code=500, detail=f"文件存储失败: {str(e)}")
+    # 如果有真实文件上传，先保存并读取内容
+    if file and file.filename:
+        try:
+            uploaded_content = await file.read()
+            actual_size = len(uploaded_content)
+            # 先写入原始文件
+            with open(stored_path, "wb") as f:
+                f.write(uploaded_content)
+            # 对于文本类格式，尝试直接解析
+            text_exts = {"txt", "md", "csv", "json", "yaml", "yml", "xml", "py", "js", "ts", "html", "css", "log"}
+            if file_ext in text_exts:
+                try:
+                    content = uploaded_content.decode("utf-8")
+                except UnicodeDecodeError:
+                    try:
+                        content = uploaded_content.decode("gbk")
+                    except UnicodeDecodeError:
+                        content = f"[二进制文件: {filename}, 大小: {actual_size} bytes]"
+            else:
+                # 非文本格式：记录文件但内容用摘要代替
+                content = _generate_document_content(filename, file_ext, actual_size)
+                logger.info(f"非文本文件 {filename}({file_ext}) 已保存到 {stored_path}, 使用模拟内容用于检索")
+        except Exception as e:
+            logger.error(f"文件读取失败: {e}")
+            content = _generate_document_content(filename, file_ext, actual_size)
+    else:
+        # 无文件上传（兼容旧 JSON 模式）：生成模拟内容
+        actual_size = file_size or len(filename.encode('utf-8')) * 100
+        content = _generate_document_content(filename, file_ext, actual_size)
+        # 写入模拟内容
+        try:
+            with open(stored_path, "w", encoding="utf-8") as f:
+                f.write(content)
+        except Exception as e:
+            logger.error(f"知识库文档写入失败: {stored_path}, error={e}")
+            raise HTTPException(status_code=500, detail=f"文件存储失败: {str(e)}")
 
-    # 文本分块：将内容按段落分割为chunks
+    # 确保文件已写入（防止重复写入）
+    if uploaded_content and file_ext not in {"txt", "md", "csv", "json", "yaml", "yml", "xml", "py", "js", "ts", "html", "css", "log"}:
+        pass  # 已在上面 wb 写入
+    elif not os.path.exists(stored_path):
+        try:
+            with open(stored_path, "w", encoding="utf-8") as f:
+                f.write(content)
+        except Exception as e:
+            logger.error(f"知识库文档写入失败: {stored_path}, error={e}")
+            raise HTTPException(status_code=500, detail=f"文件存储失败: {str(e)}")
+
+    logger.info(f"知识库文档已处理: {stored_path}, ext={file_ext}, size={actual_size}, content_len={len(content)}")
+
+    # 文本分块
     chunks = _split_content_into_chunks(content, chunk_size=500)
 
     doc = KnowledgeDocument(
         id=doc_id,
         knowledge_base_id=base_id,
-        filename=req.filename,
-        file_size=file_size,
+        filename=filename,
+        file_size=actual_size,
         file_type=file_ext,
         status="completed",
         chunk_count=len(chunks),
@@ -911,8 +955,8 @@ def upload_document(
     kb.document_count = (kb.document_count or 0) + 1
     kb.updated_at = datetime.now()
 
-    # 同步创建Asset记录（问题9：资产关联知识库文档用量）
-    _sync_asset_for_knowledge_doc(db, current_user.id, doc_id, req.filename, file_ext, file_size)
+    # 同步创建Asset记录
+    _sync_asset_for_knowledge_doc(db, current_user.id, doc_id, filename, file_ext, actual_size)
 
     db.commit()
     db.refresh(doc)