diff --git a/schema/dataset.py b/schema/dataset.py
new file mode 100644
index 0000000..c60ffed
--- /dev/null
+++ b/schema/dataset.py
@@ -0,0 +1,28 @@
+from typing import Optional
+from pydantic import BaseModel, Field
+from datetime import datetime, timezone
+
+class doc(BaseModel):
+    id: Optional[int] = Field(default=None, description="文档ID")
+    name: str = Field(default="", description="文档名称")
+    path: str = Field(default="", description="文档路径")
+    markdown_files: list[str] = Field(default_factory=list, description="文档路径列表")
+
+class Q_A(BaseModel):
+    question: str = Field(default="", min_length=1,description="问题")
+    answer: str = Field(default="", min_length=1, description="答案")
+
+class dataset_item(BaseModel):
+    id: Optional[int] = Field(default=None, description="数据集项ID")
+    message: list[Q_A] = Field(description="数据集项内容")
+
+class dataset(BaseModel):
+    id: Optional[int] = Field(default=None, description="数据集ID")
+    name: Optional[str] = Field(default=None, description="数据集名称")
+    model_id: Optional[list[str]] = Field(default=None, description="数据集使用的模型ID")
+    description: Optional[str] = Field(default="", description="数据集描述")
+    created_at: datetime = Field(
+        default_factory=lambda: datetime.now(timezone.utc),
+        description="记录创建时间"
+    )
+    dataset_items: list[dataset_item] = Field(default_factory=list, description="数据集项列表")
\ No newline at end of file
diff --git a/tools/__init__.py b/tools/__init__.py
new file mode 100644
index 0000000..1c1e383
--- /dev/null
+++ b/tools/__init__.py
@@ -0,0 +1,2 @@
+from .parse_markdown import parse_markdown
+from .scan_doc_dir import *
\ No newline at end of file
diff --git a/tools/parse_markdown.py b/tools/parse_markdown.py
index 6eeba01..92dae9e 100644
--- a/tools/parse_markdown.py
+++ b/tools/parse_markdown.py
@@ -6,6 +6,27 @@ from pathlib import Path
 sys.path.append(str(Path(__file__).resolve().parent.parent))
 from schema import MarkdownNode
 
+def process_markdown_file(file_path):
+    with open(file_path, 'r', encoding='utf-8') as f:
+        content = f.read()
+    
+    root = parse_markdown(content)
+    results = []
+    
+    def traverse(node, parent_titles):
+        current_titles = parent_titles.copy()
+        current_titles.append(node.title)
+        
+        if not node.children:  # 叶子节点
+            if node.content:
+                full_text = ' -> '.join(current_titles) + '\n' + node.content
+                results.append(full_text)
+        else:
+            for child in node.children:
+                traverse(child, current_titles)
+    
+    traverse(root, [])
+    return results
 def add_child(parent, child):
     parent.children.append(child)
 
@@ -60,10 +81,13 @@ def parse_markdown(markdown):
     return root
 
 if __name__=="__main__":
-    # 从文件读取 Markdown 内容
-    with open("workdir/example.md", "r", encoding="utf-8") as f:
-        markdown = f.read()
+    # # 从文件读取 Markdown 内容
+    # with open("workdir/example.md", "r", encoding="utf-8") as f:
+    #     markdown = f.read()
 
-    # 解析 Markdown 并打印树结构
-    root = parse_markdown(markdown)
-    print_tree(root)
+    # # 解析 Markdown 并打印树结构
+    # root = parse_markdown(markdown)
+    # print_tree(root)
+    for i in process_markdown_file("workdir/example.md"):
+        print("~"*20)
+        print(i)
\ No newline at end of file
diff --git a/tools/scan_doc_dir.py b/tools/scan_doc_dir.py
new file mode 100644
index 0000000..bd17d1b
--- /dev/null
+++ b/tools/scan_doc_dir.py
@@ -0,0 +1,32 @@
+import sys
+import os
+from pathlib import Path
+
+# 添加项目根目录到sys.path
+sys.path.append(str(Path(__file__).resolve().parent.parent))
+from schema import doc
+
+def scan_docs_directory(workdir: str):
+    docs_dir = os.path.join(workdir, "docs")
+
+    doc_list = os.listdir(docs_dir)
+
+    to_return = []
+
+    for doc_name in doc_list:
+        doc_path = os.path.join(docs_dir, doc_name)
+        if os.path.isdir(doc_path):
+            markdown_files = []
+            for root, dirs, files in os.walk(doc_path):
+                for file in files:
+                    if file.endswith(".md"):
+                        markdown_files.append(os.path.join(root, file))
+            to_return.append(doc(name=doc_name, path=doc_path, markdown_files=markdown_files))
+                    
+    return to_return
+
+# 添加测试代码
+if __name__ == "__main__":
+    workdir = os.path.join(os.path.dirname(__file__), "..", "workdir")
+    docs = scan_docs_directory(workdir)
+    print(docs)