You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I would like to see a change in how remote-files are saved locally, preferably using a checksum over the file content or similar.
I am uploading files from a remote where the basenames sometimes are the same and with the currently implementation of _download_data in cvat/apps/engine/task.py the upload fails with a filename collision even though the files themselves are different.
Using a checksum over the file would additionally allow reuse of the uploaded data over different tasks/jobs if desired.
The text was updated successfully, but these errors were encountered:
It was already requested by another user to save such images with full url path or something like that (http://example.com/a/b/mypic.png -> /example.com/a/b/mypic.png). Probably we can provide a parameter into REST API to follow the approach.
I would like to see a change in how remote-files are saved locally, preferably using a checksum over the file content or similar.
I am uploading files from a remote where the basenames sometimes are the same and with the currently implementation of _download_data in cvat/apps/engine/task.py the upload fails with a filename collision even though the files themselves are different.
Using a checksum over the file would additionally allow reuse of the uploaded data over different tasks/jobs if desired.
The text was updated successfully, but these errors were encountered: