Improve error message when importing invalid archive #5760
Description
A common scenario when importing datasets to Geti is that the user uploads an invalid archive, for several reasons:
- the user accidentally selects the wrong zip file
- the user uploads a dataset in a format that is not supported by Geti
- the user uploads a dataset in a supported data format, but the internal layout of the zip archive is irregular (e.g. files nested inside a folder at the root level, or directories with non-standard names, etc...)
In this case, it is important that the application outputs a user-friendly and descriptive error message.
The current message is not very user-friendly:
The server logs contain valuable information to trace the root cause, but these are indeed logs and such information would be too technical to show in the UI:
ValueError: Could not detect dataset format in 'data/staged_datasets/e96c5689-f324-406c-8acb-236d7f436e08/dataset_import'. Expected one of: Datumaro (metadata.json + data.parquet), COCO (annotations/*.json with 'images' and 'annotations' keys), YOLO (data.yaml, obj.names, or images/ + labels/ directories), or VOC (JPEGImages/ + Annotations/ or ImageSets/ directories)
A proper error message to show in the UI could be:
An error occurred during import: the dataset could not be recognized in any of the supported formats. Please verify that the dataset is well-formed and in a supported format; if the problem persists, report the issue.