From f9025ff4152760f6ea1bbc0c6bf72dfaee7e71bd Mon Sep 17 00:00:00 2001
From: wan <nishikawa0228@sj9.so-net.ne.jp>
Date: Mon, 10 Mar 2025 16:23:56 +0900
Subject: [PATCH] fix: skip audio and video files during directory scan in RAG
 Service. (#1509)

---
 py/rag-service/src/main.py | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)

diff --git a/py/rag-service/src/main.py b/py/rag-service/src/main.py
index c1e235d..d5477ee 100644
--- a/py/rag-service/src/main.py
+++ b/py/rag-service/src/main.py
@@ -660,14 +660,33 @@ def scan_directory(directory: Path) -> list[str]:
     """Scan directory and return a list of matched files."""
     spec = get_pathspec(directory)
 
+    audio_video_exts = [
+        ".mp3",
+        ".wav",
+        ".ogg",
+        ".flac",
+        ".aac",
+        ".m4a",
+        ".wma",
+        ".mp4",
+        ".avi",
+        ".mov",
+        ".wmv",
+        ".flv",
+        ".mkv",
+        ".webm",
+    ]
+
     matched_files = []
 
     for root, _, files in os.walk(directory):
         file_paths = [str(Path(root) / file) for file in files]
-        if not spec:
-            matched_files.extend(file_paths)
-            continue
         for file in file_paths:
+            file_ext = Path(file).suffix.lower()
+            if file_ext in audio_video_exts:
+                logger.info("Skipping audio/video file: %s", file)
+                continue
+
             if spec and spec.match_file(os.path.relpath(file, directory)):
                 logger.info("Ignoring file: %s", file)
             else: