From 22f370322412074174cde20ecfd14ec03657ab63 Mon Sep 17 00:00:00 2001
From: lyg <1543117173@qq.com>
Date: 星期一, 07 七月 2025 16:20:25 +0800
Subject: [PATCH] 生成数据库
---
main.py | 69 ++++++++++++----------------------
1 files changed, 25 insertions(+), 44 deletions(-)
diff --git a/main.py b/main.py
index 701bf1d..6cc6b78 100644
--- a/main.py
+++ b/main.py
@@ -1,53 +1,34 @@
+import asyncio
import os
-from lang_flow import LangFlow
-from markitdown import MarkItDown
+import sys
-from doc_to_docx import doc_to_docx
+from db_struct_flow import DbStructFlow, tc_data_generate
+from knowledgebase.db.doc_db_helper import doc_dbh
+from knowledgebase.doc.doc_processor import DocProcessor
+from knowledgebase.doc.entity_helper import init_entity_helper
-def process_docs(directory):
- # 閬嶅巻鐩綍涓嬬殑鎵�鏈夋枃浠�
- for filename in os.listdir(directory):
- # 鍒ゆ柇鏄惁涓� doc 鏂囦欢
- if filename.endswith(".doc"):
- # 杞崲涓� docx
- doc_to_docx(directory + filename, directory + filename.replace(".doc", ".docx"))
+def doc_split(project_path):
+ docs_path = f'{project_path}/docs'
+ files = os.listdir(docs_path)
+ files = [f'{docs_path}/{x}' for x in filter(lambda x: x.endswith('.docx'), files)]
+ for file in files:
+ DocProcessor(file).process()
-md = MarkItDown()
-
-def to_markdown(dst_dir: str):
- text = ''
- # 閬嶅巻鏂囦欢澶逛笅鐨勬墍鏈夋枃浠�
- for file in os.listdir(dst_dir):
- # 鍒ゆ柇鏄惁涓� docx 鏂囦欢
- if file.endswith(".docx"):
- # 杞崲涓� md
- result = md.convert(dst_dir + file)
- text += '\n\n' + result.text_content
- out_file = dst_dir + 'docs.md'
- with open(out_file, 'w', encoding='utf-8') as f:
- f.write(text)
- return out_file
-
-
-# 1.瑙ｆ瀽鏂囨。
-# 2.杈撳叆鏂囨。
-# 3.鍚姩LangFlow
def main():
- # doc_dir = "D:\\workspace\\PythonProjects\\KnowledgeBase\\doc\\"
- # 澶勭悊鏂囨。
- # process_docs(doc_dir)
- # 鏂囨。杞崲涓簃arkdown
- # md_file = to_markdown(doc_dir)
-
- md_file = 'D:\\workspace\\PythonProjects\\KnowledgeBase\\doc\\test.md'
+ project_path = sys.argv[1]
+ if not project_path:
+ print("missing project path. eg: python main.py <path/to/project>")
+ return
+ # 鎷嗗垎鏂囨。
+ doc_dbh.set_project_path(project_path)
+ init_entity_helper()
+ # doc_split(project_path)
# 鍚姩澶фā鍨嬪鐞嗘祦绋�
- ret_text = LangFlow([md_file]).run()
- # 淇濆瓨缁撴灉
- # with open('D:\\workspace\\PythonProjects\\KnowledgeBase\\doc\\test.text', 'w', encoding='utf-8') as f:
- # f.write(ret_text)
-
-if __name__ == '__main__':
- main()
+ asyncio.run(DbStructFlow(project_path).run())
+ # 鐢熸垚鎸囦护鏁版嵁琛�
+ tc_data_generate()
+if __name__ == "__main__":
+ main()
\ No newline at end of file
--
Gitblit v1.9.1