search

HumanSignal / label-studio

Label Studio is a multi-type data labeling and annotation tool with standardized output format
https://labelstud.io
Apache License 2.0
19.34k stars 2.4k forks source link

Google Cloud Storage URIs are not resolved to URLs #4005

Open elarbikonta opened 1 year ago

elarbikonta commented 1 year ago

I have a project with source storage connected to a GCS bucket with "Treat every bucket object as a source file" option checked. When I hit "Sync Storage" I get a new task and I can see it in my task list. But when I click on it the UI is unable to get the file. I see a GET request to /projects/3/undefined$ with a 404 response !

the server is logging this message:

[2023-04-10 14:01:32,222] [django.request::log_response::224] [WARNING] Not Found: /projects/3/undefined$

Is the resolution working by default or I should set a parameter somewhere ?

We are using this image: heartexlabs/label-studio:1.7.2

Thanks

makseq commented 1 year ago

what files do you store in the bucket? Are there media files (images, audio, etc)? or jsons?

this one looks absolutely weird - /projects/3/undefined$

ncgovindans commented 1 year ago

+1 to this. Using images, the resolver doesn't seem to convert gs:// to the signed URL

makseq commented 1 year ago

1 Have you enabled pre-signed urls in LS cloud storage settings? 2 if (1) yes and it doesn't help, try running LS this way:

fflag_feat_back_dev_4166_google_project_id_11012023_long=0 label-studio
ncgovindans commented 1 year ago

@makseq - Thanks for the suggestion, I reverted to 1.7 or 1.6. Can't quite remember. I've upgraded to 1.8.0 to take advantage of the LLM module and the problem has returned.

I want to test your suggestion with docker - can I use the feature_flag.json to do that? Set "on" to false? Anything else?