From 1a61b6a67760d3640eedb3c1e421c6605c9c6b80 Mon Sep 17 00:00:00 2001
From: HuangHai <10402852@qq.com>
Date: Tue, 24 Jun 2025 19:19:53 +0800
Subject: [PATCH] 'commit'

---
 dsRag/Start.py                                |  66 +++++++++++++++++-
 .../T9_TestReadPptx.cpython-310.pyc           | Bin 0 -> 864 bytes
 .../Util/__pycache__/WordUtil.cpython-310.pyc | Bin 946 -> 946 bytes
 3 files changed, 64 insertions(+), 2 deletions(-)
 create mode 100644 dsRag/Test/__pycache__/T9_TestReadPptx.cpython-310.pyc

diff --git a/dsRag/Start.py b/dsRag/Start.py
index 34cf1968..e5b85b70 100644
--- a/dsRag/Start.py
+++ b/dsRag/Start.py
@@ -1,19 +1,73 @@
+import threading
+import time
+from contextlib import asynccontextmanager
 from pathlib import Path
 
 import uvicorn
 from fastapi import FastAPI, UploadFile, File, HTTPException
+# 在文件开头添加导入
 from pymysql.cursors import DictCursor
 
 from Dao.KbDao import KbDao
 from Model.KbModel import KbModel, KbFileModel
+from Test.T9_TestReadPptx import extract_text_from_pptx
+from Util import PdfUtil, WordUtil
 from Util.MySQLUtil import init_mysql_pool
-from contextlib import asynccontextmanager
+
+import logging
+from logging.handlers import RotatingFileHandler
+
+# 确保logger已在文件开头正确初始化
+logger = logging.getLogger(__name__)
+logger.setLevel(logging.INFO)
+handler = RotatingFileHandler('Logs/document_processor.log', maxBytes=1024*1024, backupCount=5)
+handler.setFormatter(logging.Formatter('%(asctime)s - %(levelname)s - %(message)s'))
+logger.addHandler(handler)
 
 @asynccontextmanager
 async def lifespan(app: FastAPI):
     # 启动时初始化数据库连接池
     app.state.kb_dao = KbDao(await init_mysql_pool())
+    
+    # 启动文档处理线程
+    # 修改函数定义
+    async def document_processor():
+        while True:
+            try:
+                # 获取未处理文档
+                # 处理文档
+                # 保存到ES
+                await asyncio.sleep(10)
+            except Exception as e:
+                logger.error(f"文档处理出错: {e}")
+                await asyncio.sleep(10)
+            
+            time.sleep(10)  # 每10秒检查一次
+    
+    # 修改线程启动部分
+    # 修改线程启动方式
+    import asyncio
+    
+    def run_async_in_thread():
+        loop = asyncio.new_event_loop()
+        asyncio.set_event_loop(loop)
+        try:
+            loop.run_until_complete(document_processor())
+        finally:
+            loop.close()
+    
+    processor_thread = threading.Thread(
+        target=run_async_in_thread,
+        daemon=True
+    )
+    processor_thread.start()
+    
+    # 启动文档处理任务
+    task = asyncio.create_task(document_processor())
     yield
+    # 关闭时取消任务
+    task.cancel()
+    
     # 关闭时清理资源
     await app.state.kb_dao.mysql_pool.close()
 
@@ -192,4 +246,12 @@ async def upload_file(
             }
 
 if __name__ == "__main__":
-    uvicorn.run(app, host="0.0.0.0", port=8000)
\ No newline at end of file
+    uvicorn.run(app, host="0.0.0.0", port=8000)
+
+
+# 确保logger已在文件开头正确初始化
+logger = logging.getLogger(__name__)
+logger.setLevel(logging.INFO)
+handler = RotatingFileHandler('Logs/document_processor.log', maxBytes=1024*1024, backupCount=5)
+handler.setFormatter(logging.Formatter('%(asctime)s - %(levelname)s - %(message)s'))
+logger.addHandler(handler)
\ No newline at end of file
diff --git a/dsRag/Test/__pycache__/T9_TestReadPptx.cpython-310.pyc b/dsRag/Test/__pycache__/T9_TestReadPptx.cpython-310.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..64d7ffcd2d68b2c977f122a15880b0986879bcd0
GIT binary patch
literal 864
zcmYjPO=}ZD7@nE^$Yz@sL@D+bZ$>ah5J3?U>{&uB2uo#IGSg<$&F;E0fmAn8NLwvZ
z+e0n3Mo2DN!S*6ZZTuyBHQC_FtB42RiB<Z-yk9f(%rnnSFg|V~u&1$A=_v?(lf_w6
zp|}f!d*CRt5XF>uxJE33Wv+ONrBIb>tH@HBM)f5UU8fmKXG%5GF__-P(G{jLy@jnT
zGpW(S7%`J&Wu(dj$Z3}8ZxA)#D*t-2Ul@BbWrl8;cYQ7#&&!9)q7wL5!%#HL??46>
zaFQ!Uj5A+2!VUcUWh~pLrbE<V+=IbO-~@_M3pep1YLXaN@eF!2)l_1oN+3JmRD}|&
z6t_rpDORg$N9&@d7Hd$Ejvi}Wp!;~@ah8%!CMI1B1eN+_fbI8QN=oVOvwrV$|M)Q7
z+8%7}rW-41djpzj`{Q7JZE*A@Q3Y#=1e*mTBCyZWx#v>G6P;I_5Mrj{a7TzJf@eU9
z#zo|YiSC3U^J$`11J^I;Np`{Ym>oKzVku$7t(*k4%YiRI1xp)2!d{6u!uf!r9Kij$
zHgjv9@`pjR1fdiKRaO>|%{q(obBv4mxtq3(v&^9-xJ@x!P9_04a>~LU9kCDvHCtZP
z060Dg&LAo_6_c3Q#E@aKLVlZiG$Gj=wq0{v-?r;Fpez=r=Ne+Vf7~7%KOb(dr>#%v
z-buRlK7F+}ID9ue>OymHvNPN{==WX>dz(dhQ2ioo{j271fAy@CKfdnQGjejIdA=@x
zaK3-MIXHYhXde{Dl8kNBplsWTDuG6FCnVtix%p0w*|wFHv$j+e7J{US&KfQk{PYFF
s$o0iYzB6td&Pt5iwSd+=c1M0VE_tdbrx@5!jB_f<VaRX}Um)twKeodUVgLXD

literal 0
HcmV?d00001

diff --git a/dsRag/Util/__pycache__/WordUtil.cpython-310.pyc b/dsRag/Util/__pycache__/WordUtil.cpython-310.pyc
index 9f7445b8edcb70bc9c59832b79ac882f9531083a..3094c6c78a240f09b5154e293ceb4844d5aadcf3 100644
GIT binary patch
delta 122
zcmdnQzKNYHpO=@50SNjkH*%#iG2WV7z$D2SHMxt)7|7noBn2dIGRZPZPyWtS>>&$O
zsL67RB_%((qKFB|DPjQ;AmJiuR6+*Gy~SaZo1apelWNBZ<Q2042_6OxMjmDWItLl_

delta 122
zcmdnQzKNYHpO=@50SNZ`Y~)I1VvL$xz$D3dYjPKpF_68FNeW2bWRhi+ocx`s*h3nq
zP?IH!B_%((qKFB|DPjQ;AmJiOR6+{Ky~SaZo1apelWNBZ<Q2042_6OxMjmDWy-gTs