From 8a2542157fc22fd49daad465704aa1ab78dc87ba Mon Sep 17 00:00:00 2001
From: yihong <zouzou0208@gmail.com>
Date: Mon, 3 Mar 2025 10:26:45 +0800
Subject: [PATCH] Fix: possible memory leaks close #5277 (#5500)

### What problem does this PR solve?

close #5277 by make sure the file close

### Type of change

- [x] Performance Improvement

---------

Signed-off-by: yihong0618 <zouzou0208@gmail.com>
---
 api/utils/file_utils.py      | 1 +
 deepdoc/parser/pdf_parser.py | 7 ++++++-
 deepdoc/vision/__init__.py   | 1 +
 rag/svr/task_executor.py     | 1 +
 4 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/api/utils/file_utils.py b/api/utils/file_utils.py
index ed014ac0e..d76be81c9 100644
--- a/api/utils/file_utils.py
+++ b/api/utils/file_utils.py
@@ -188,6 +188,7 @@ def thumbnail_img(filename, blob):
                 buffered = BytesIO()
             else:
                 break
+        pdf.close()
         return img
 
     elif re.match(r".*\.(jpg|jpeg|png|tif|gif|icon|ico|webp)$", filename):
diff --git a/deepdoc/parser/pdf_parser.py b/deepdoc/parser/pdf_parser.py
index 2282a87fb..d4645f8a9 100644
--- a/deepdoc/parser/pdf_parser.py
+++ b/deepdoc/parser/pdf_parser.py
@@ -950,7 +950,9 @@ class RAGFlowPdfParser:
         try:
             pdf = pdfplumber.open(
                 fnm) if not binary else pdfplumber.open(BytesIO(binary))
-            return len(pdf.pages)
+            total_page = len(pdf.pages)
+            pdf.close()
+            return total_page 
         except Exception:
             logging.exception("total_page_number")
 
@@ -996,8 +998,11 @@ class RAGFlowPdfParser:
             dfs(outlines, 0)
         except Exception as e:
             logging.warning(f"Outlines exception: {e}")
+        finally:
+            self.pdf.close()
         if not self.outlines:
             logging.warning("Miss outlines")
+        
 
         logging.debug("Images converted.")
         self.is_english = [re.search(r"[a-zA-Z0-9,/¸;:'\[\]\(\)!@#$%^&*\"?<>._-]{30,}", "".join(
diff --git a/deepdoc/vision/__init__.py b/deepdoc/vision/__init__.py
index 64afcaf72..c67178c8f 100644
--- a/deepdoc/vision/__init__.py
+++ b/deepdoc/vision/__init__.py
@@ -42,6 +42,7 @@ def init_in_out(args):
 
         for i, page in enumerate(images):
             outputs.append(os.path.split(fnm)[-1] + f"_{i}.jpg")
+        pdf.close()
 
     def images_and_outputs(fnm):
         nonlocal outputs, images
diff --git a/rag/svr/task_executor.py b/rag/svr/task_executor.py
index 3fc69f3f5..05388f8df 100644
--- a/rag/svr/task_executor.py
+++ b/rag/svr/task_executor.py
@@ -18,6 +18,7 @@
 # beartype_all(conf=BeartypeConf(violation_type=UserWarning))    # <-- emit warnings from all code
 import random
 import sys
+
 from api.utils.log_utils import initRootLogger, get_project_base_directory
 from graphrag.general.index import WithCommunity, WithResolution, Dealer
 from graphrag.light.graph_extractor import GraphExtractor as LightKGExt