From 4212ec4790df1fb8b20a5e371c37ce091b7ed258 Mon Sep 17 00:00:00 2001
From: berkecanrizai <63911408+berkecanrizai@users.noreply.github.com>
Date: Tue, 23 Jul 2024 11:28:23 +0300
Subject: [PATCH] fix: slide-search rm dumped files if they are rmed in the
 source (#6947)

GitOrigin-RevId: 5cae9b0388e3bf121f00bf6331e13abec07e65ae
---
 examples/pipelines/slides_ai_search/app.py | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/examples/pipelines/slides_ai_search/app.py b/examples/pipelines/slides_ai_search/app.py
index 47b67df..887082d 100644
--- a/examples/pipelines/slides_ai_search/app.py
+++ b/examples/pipelines/slides_ai_search/app.py
@@ -53,8 +53,16 @@ def dump_img_callback(key, row, time, is_addition):
 def dump_file_callback(key, row, time, is_addition):
     # save parsed files
     file_name = row["path"].value.split("/")[-1]
-    with open(f"{FILE_DUMP_FOLDER}/{file_name}", "wb") as f:
-        f.write(row["data"])
+    file_path = f"{FILE_DUMP_FOLDER}/{file_name}"
+    if is_addition:
+        with open(file_path, "wb") as f:
+            f.write(row["data"])
+    else:
+        try:
+            os.remove(file_path)
+            logging.info(f"Removed file: {file_path}")
+        except Exception as e:
+            logging.info("Error removing %s: %s", file_name, e)
 
 
 if __name__ == "__main__":