From b9d1430736a0f19589065b02edd8f47de564b923 Mon Sep 17 00:00:00 2001
From: John Wilkie <124276291+JBWilkie@users.noreply.github.com>
Date: Thu, 7 Mar 2024 16:50:57 +0000
Subject: [PATCH] [PLA-683][external] 2 Minor fixes for importing of PASCAL VOC
 annotations (#783)

* Fixed mislabelled default arg value

* Added support for importing PASCAL VOC annotations to folders + unit test

* Fixed misconfigured querystring filter name in fetch_remote_files()

* removed test file
---
 darwin/cli_functions.py                            |  2 +-
 darwin/importer/formats/pascal_voc.py              |  5 ++++-
 darwin/importer/importer.py                        |  2 +-
 .../importer/formats/import_pascalvoc_test.py      | 14 ++++++++++++++
 4 files changed, 20 insertions(+), 3 deletions(-)

diff --git a/darwin/cli_functions.py b/darwin/cli_functions.py
index ad8d75268..662280e75 100644
--- a/darwin/cli_functions.py
+++ b/darwin/cli_functions.py
@@ -881,7 +881,7 @@ def dataset_import(
     import_reviewers : bool, default: False
         If ``True`` it will import the reviewers from the files to the dataset, if .
         If ``False`` it will not import the reviewers.
-    use_multi_cpu : bool, default: True
+    use_multi_cpu : bool, default: False
         If ``True`` it will use all multiple CPUs to speed up the import process.
     cpu_limit : Optional[int], default: Core count - 2
         The maximum number of CPUs to use for the import process.
diff --git a/darwin/importer/formats/pascal_voc.py b/darwin/importer/formats/pascal_voc.py
index 4470a6149..e6fadf328 100644
--- a/darwin/importer/formats/pascal_voc.py
+++ b/darwin/importer/formats/pascal_voc.py
@@ -3,6 +3,7 @@
 from typing import List, Optional
 
 import darwin.datatypes as dt
+from darwin.path_utils import deconstruct_full_path
 
 
 def parse_path(path: Path) -> Optional[dt.AnnotationFile]:
@@ -57,8 +58,10 @@ def parse_path(path: Path) -> Optional[dt.AnnotationFile]:
     )
     annotation_classes = {annotation.annotation_class for annotation in annotations}
 
+    remote_path, filename = deconstruct_full_path(filename)
+
     return dt.AnnotationFile(
-        path, filename, annotation_classes, annotations, remote_path="/"
+        path, filename, annotation_classes, annotations, remote_path=remote_path
     )
 
 
diff --git a/darwin/importer/importer.py b/darwin/importer/importer.py
index eac3e295c..1dd0880e4 100644
--- a/darwin/importer/importer.py
+++ b/darwin/importer/importer.py
@@ -221,7 +221,7 @@ def get_remote_files(
     for i in range(0, len(filenames), chunk_size):
         chunk = filenames[i : i + chunk_size]
         for remote_file in dataset.fetch_remote_files(
-            {"types": "image,playback_video,video_frame", "filenames": chunk}
+            {"types": "image,playback_video,video_frame", "item_names": chunk}
         ):
             slot_name = _get_slot_name(remote_file)
             remote_files[remote_file.full_path] = (remote_file.id, slot_name)
diff --git a/tests/darwin/importer/formats/import_pascalvoc_test.py b/tests/darwin/importer/formats/import_pascalvoc_test.py
index 2c50dc24f..84d09aa9f 100644
--- a/tests/darwin/importer/formats/import_pascalvoc_test.py
+++ b/tests/darwin/importer/formats/import_pascalvoc_test.py
@@ -171,3 +171,17 @@ def test_returns_annotation_file_with_correct_annotations_with_float_values(
         assert annotation.subs == []
 
         assert annotation_file.remote_path == "/"
+
+    def test_deconstructs_filepath_properly_if_folder_included_in_filename(
+        self, annotation_path: Path
+    ):
+        annotation_path.write_text(
+            "<root><filename>folder/image.jpg</filename><object><name>Class</name><bndbox><xmin>10</xmin><xmax>10</xmax><ymin>10</ymin><ymax>10</ymax></bndbox></object></root>"
+        )
+
+        annotation_file = parse_path(annotation_path)
+
+        assert annotation_file is not None
+        assert annotation_file.path == annotation_path
+        assert annotation_file.filename == "image.jpg"
+        assert annotation_file.remote_path == "/folder"