Label-studio: Containerization

[davebulaval]

It would be interesting to have a container image of the _app_. This way I would be able to launch a container on a remote desktop and ping the host address.

Something like Docker.

[JStumpp]

I can provide a dockerized version, if you are interested.

[shevchenkonik]

Hello guys! Thank you for your interest!

@davebulaval Yes, we thought about this possibility and expected to implement it in the future versions. But now our current focus is on supporting as many data types as possible in Label Studio.

@JStumpp We are interested in it. If you do PR, we will certainly implement this functionality in the project.

[davebulaval]

@JStumpp I cannot help much but I can at least valide the build of the container and if the functionnality work.

[JStumpp]

@davebulaval no problem, I am on it. Should be ready tomorrow.

[JStumpp]

The dockerization is ready, see the pull request https://github.com/heartexlabs/label-studio/pull/31

Give it a try:

docker run -p 8200:8200 -t -i amaigmbh/label-studio -c config.json -l ../examples/chatbot_analysis/config.xml -i ../examples/chatbot_analysis/tasks.json -o output

or mount a folder with your custom configuration files

docker run -p 8200:8200 -t -i -v C:\docker\custom:/app/examples/audio_regions amaigmbh/label-studio

[davebulaval]

First one work well. Second I have a invalid mode error.

[JStumpp]

@davebulaval can you post the exact error? Are you running on windows and did you create the path C:\docker\custom with files from an example?

[davebulaval]

Linux, no I didn't. I'll try later and let you know.

[shevchenkonik]

@JStumpp @davebulaval First one works well, yes. But I can't reproduce with this instruction the second one.

docker run -p 8200:8200 -t -i -v C:\docker\custom:/app/examples/audio_regions amaigmbh/label-studio

Could u expand the instruction? When I started LS using an example from the local path, Docker did not find the config outside directory. And the LS starts with default config audio_regions

My version (_my environment is macOS_):

docker run -p 8200:8200 -t -i -v /Users/shevchenkonik/named_entity:/app/examples/named_entity amaigmbh/label-studio