From c9afcd27592ed43e76914673f922a6bc113ecbc0 Mon Sep 17 00:00:00 2001
From: Nell Hardcastle <nell@dev-nell.com>
Date: Wed, 19 Nov 2025 12:22:20 -0800
Subject: [PATCH] fix(worker): Skip re-export on drop if remote contains all
 files for a branch or tag

---
 .../datalad/datalad_service/common/asyncio.py |  1 +
 .../datalad/datalad_service/tasks/publish.py  | 21 +++++++++++++++++--
 2 files changed, 20 insertions(+), 2 deletions(-)

diff --git a/services/datalad/datalad_service/common/asyncio.py b/services/datalad/datalad_service/common/asyncio.py
index 91bdc292a..a13acb995 100644
--- a/services/datalad/datalad_service/common/asyncio.py
+++ b/services/datalad/datalad_service/common/asyncio.py
@@ -22,3 +22,4 @@ async def run_check(command, dataset_path, env=None):
     stdout, stderr = await process.communicate()
     if process.returncode != 0:
         raise subprocess.CalledProcessError(process.returncode, command, stdout, stderr)
+    return stdout.decode('utf-8')
diff --git a/services/datalad/datalad_service/tasks/publish.py b/services/datalad/datalad_service/tasks/publish.py
index d37a63110..401271ed0 100644
--- a/services/datalad/datalad_service/tasks/publish.py
+++ b/services/datalad/datalad_service/tasks/publish.py
@@ -105,8 +105,12 @@ async def export_backup_and_drop(dataset_path):
         await s3_backup_push(dataset_path)
         for tag in tags:
             logger.info(f'Exporting/dropping tag {dataset_id}@{tag.name}')
-            # Private datasets need to export each tag for this step
-            if not public_dataset:
+            export_ran = False
+            if await find_in_remote(dataset_path, tag.name, get_s3_remote()):
+                export_ran = True
+                await s3_export(dataset_path, get_s3_remote(), tag.name)
+            if tag == tags[-1] and export_ran:
+                # Always export the most recent tag again if any export ran
                 await s3_export(dataset_path, get_s3_remote(), tag.name)
             await fsck_and_drop(dataset_path, tag.name)
             logger.info(f'Exporting/dropping tag {dataset_id}@{tag.name} complete')
@@ -116,6 +120,19 @@ async def export_backup_and_drop(dataset_path):
     logger.info(f'{dataset_id} export_backup_and_drop complete')
 
 
+async def find_in_remote(dataset_path, tag, remote):
+    """Check if any git-annex objects available locally for a branch are not present in a remote."""
+    output = await run_check(
+        ['git-annex', 'find', f'--branch={tag}', '--not', f'--in={remote}'],
+        dataset_path,
+    )
+    if len(output) > 0:
+        # Some keys are missing
+        return False
+    # All keys are present
+    return True
+
+
 @broker.task
 async def export_dataset(
     dataset_path,