From 3b7a6eaa6cbcf4cfe7ba905eae3e4ba464e9a731 Mon Sep 17 00:00:00 2001
From: Magicbook1108 <newyorkupperbay@gmail.com>
Date: Wed, 29 Apr 2026 11:29:17 +0800
Subject: [PATCH] Feat: sync deleted files in Bitbucket (#14450)

### What problem does this PR solve?

Feat: sync deleted files in Bitbucket

### Type of change

- [x] New Feature (non-breaking change which adds functionality)
---
 common/data_source/bitbucket/connector.py                 | 5 +----
 rag/svr/sync_data_source.py                               | 8 +++++++-
 web/src/pages/user-setting/data-source/constant/index.tsx | 5 ++++-
 3 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/common/data_source/bitbucket/connector.py b/common/data_source/bitbucket/connector.py
index 4b0240fa5f..0557d2a503 100644
--- a/common/data_source/bitbucket/connector.py
+++ b/common/data_source/bitbucket/connector.py
@@ -355,10 +355,7 @@ if __name__ == "__main__":
     start_time = datetime.fromtimestamp(0, tz=timezone.utc)
     end_time = datetime.now(timezone.utc)
 
-    for doc_batch in bitbucket.retrieve_all_slim_docs_perm_sync(
-        start=start_time.timestamp(),
-        end=end_time.timestamp(),
-    ):
+    for doc_batch in bitbucket.retrieve_all_slim_docs_perm_sync():
         for doc in doc_batch:
             print(doc)
 
diff --git a/rag/svr/sync_data_source.py b/rag/svr/sync_data_source.py
index 2c6d72cc94..ac0d4d0cb7 100644
--- a/rag/svr/sync_data_source.py
+++ b/rag/svr/sync_data_source.py
@@ -1334,12 +1334,17 @@ class Bitbucket(SyncBase):
             "bitbucket_api_token": self.conf["credentials"].get("bitbucket_api_token"),
             }
         )
+        file_list = None
 
         if task["reindex"] == "1" or not task["poll_range_start"]:
             start_time = datetime.fromtimestamp(0, tz=timezone.utc)
             _begin_info = "totally"
         else:
             start_time = task.get("poll_range_start")
+            if self.conf.get("sync_deleted_files"):
+                file_list = []
+                for slim_batch in self.connector.retrieve_all_slim_docs_perm_sync():
+                    file_list.extend(slim_batch)
             _begin_info = f"from {start_time}"
         
         end_time = datetime.now(timezone.utc)
@@ -1371,7 +1376,8 @@ class Bitbucket(SyncBase):
                 yield batch
 
         self.log_connection("Bitbucket", f"workspace({self.conf.get('workspace')})", task)
-
+        if file_list is not None:
+            return wrapper(), file_list
         return wrapper()
 
 
diff --git a/web/src/pages/user-setting/data-source/constant/index.tsx b/web/src/pages/user-setting/data-source/constant/index.tsx
index 2bb4d267f7..774b5c3f91 100644
--- a/web/src/pages/user-setting/data-source/constant/index.tsx
+++ b/web/src/pages/user-setting/data-source/constant/index.tsx
@@ -58,7 +58,7 @@ export const DataSourceFeatureVisibilityMap = {
   [DataSourceKey.GITHUB]: {
     syncDeletedFiles: true,
   },
-[DataSourceKey.GOOGLE_DRIVE]: {
+  [DataSourceKey.GOOGLE_DRIVE]: {
     syncDeletedFiles: true,
   },
   [DataSourceKey.CONFLUENCE]: {
@@ -85,6 +85,9 @@ export const DataSourceFeatureVisibilityMap = {
   [DataSourceKey.JIRA]: {
     syncDeletedFiles: true,
   },
+  [DataSourceKey.BITBUCKET]: {
+    syncDeletedFiles: true,
+  },
   [DataSourceKey.AIRTABLE]: {
     syncDeletedFiles: true,
   },