fix: FIT-1654: Exporting YOLO with Images using local storage results in no images

GitOrigin-RevId: d1ff9fb4d8f01ed8b17855b8142288c49b76b592

Author: matt-bernstein

src/label_studio_sdk/_extensions/label_studio_tools/core/utils/io.py

@@ -1,14 +1,14 @@
+import base64
 import hashlib
 import io
 import logging
 import os
 import shutil
-import base64
 from contextlib import contextmanager
 from tempfile import mkdtemp
 from urllib.parse import parse_qs, urlparse
-import jwt
 
+import jwt
 import requests
 from appdirs import user_cache_dir, user_data_dir
 
@@ -71,6 +71,10 @@ def _build_headers(target_url, hostname, access_token):
     return headers
 
 
+def _build_cache_path(cache_dir, target_url, filename):
+    return os.path.join(cache_dir, hashlib.md5(target_url.encode(), usedforsecurity=False).hexdigest()[:8] + "__" + filename)
+
+
 def get_local_path(
     url,
     cache_dir=None,
@@ -178,6 +182,12 @@ def get_local_path(
             logger.debug(
                 f"Local Storage file path exists locally, use it as a local file: {filepath}"
             )
+            if cache_dir and download_resources:
+                os.makedirs(cache_dir, exist_ok=True)
+                cache_path = _build_cache_path(cache_dir, url, os.path.basename(filepath))
+                if not os.path.exists(cache_path):
+                    shutil.copy(filepath, cache_path)
+                return cache_path
             return filepath
 
     # try to get local directories
@@ -310,7 +320,7 @@ def _filename_for(target_url, storage_fp=None):
         return os.path.basename(parsed.path)
 
     def _cache_path(target_url, fname):
-        return os.path.join(cache_dir, hashlib.md5(target_url.encode(), usedforsecurity=False).hexdigest()[:8] + "__" + fname)
+        return _build_cache_path(cache_dir, target_url, fname)
 
     cache_dir = cache_dir or get_cache_dir()
     current_filename = _filename_for(url, storage_filepath)

tests/custom/label_studio_tools/test_get_local_path.py

@@ -1,17 +1,15 @@
 import base64
 import hashlib
-import os
-import tempfile
-import requests
 from types import SimpleNamespace
-from unittest.mock import patch, MagicMock
+from unittest.mock import MagicMock, patch
 
 import pytest
+import requests
 
 from label_studio_sdk._extensions.label_studio_tools.core.utils.io import (
-    get_local_path,
-    get_base64_content,
     _DIR_APP_NAME,
+    get_base64_content,
+    get_local_path,
 )
 
 
@@ -105,6 +103,40 @@ def test_get_local_path_safe_build(mock_data_dir, mock_cache_dir, url, raises_er
             assert "my_dir/1.jpg" in local_path
 
 
+@patch("label_studio_sdk._extensions.label_studio_tools.core.utils.io.LOCAL_FILES_DOCUMENT_ROOT", "")
+def test_get_local_path_local_storage_downloads_to_cache(tmp_path):
+    """Ensure local storage files are copied into cache when download_resources is enabled."""
+    local_root = tmp_path / "local-storage-root"
+    image_dir = local_root / "my_dir"
+    image_dir.mkdir(parents=True)
+    source_file = image_dir / "1.jpg"
+    source_content = b"local-image-bytes"
+    source_file.write_bytes(source_content)
+
+    cache_dir = tmp_path / "cache"
+    cache_dir.mkdir()
+    local_url = "/data/local-files?d=my_dir/1.jpg"
+
+    with patch(
+        "label_studio_sdk._extensions.label_studio_tools.core.utils.io.LOCAL_FILES_DOCUMENT_ROOT", str(local_root)
+    ):
+        cached_path = get_local_path(
+            url=local_url,
+            cache_dir=str(cache_dir),
+            hostname="http://app.heartex.com",
+            access_token="secret",
+            download_resources=True,
+            task_id=1,
+        )
+
+    expected_hash = hashlib.md5(local_url.encode(), usedforsecurity=False).hexdigest()[:8]
+    expected_path = cache_dir / f"{expected_hash}__1.jpg"
+
+    assert cached_path == str(expected_path)
+    assert expected_path.exists()
+    assert expected_path.read_bytes() == source_content
+
+
 def test_get_local_path_storage_proxy_filename_and_auth(monkeypatch, tmp_path):
     """
     Ensure storage proxy URLs use `filepath` for filename, rewrite to hostname,