Infer the config from the URL instead using the STUDY_NAME/STUDY_CONFIG environment variable

[shankari]

At an even more basic level, in the public dashboard, the viz_scripts are run automatically in the background, so they cannot infer the study name from the URL (because they are not invoked via a URL). However, this is a webapp, so you have the URL, and should be able to infer the study name from the URL like the frontend of the public dashboard does.

https://github.com/e-mission/em-public-dashboard/blob/main/frontend/index.html#L354

This ensures that the code "just works" and avoids requiring a lot of configuration. You should be able to do something similar using https://dash.plotly.com/dash-core-components/location, which you do already get in
https://github.com/e-mission/op-admin-dashboard/pull/7/files#diff-9f23c8b13c9e75b67a44235089ab45e194bb00a6339db3295e9ff6a0646f2cc7R154

I agree that automatically inferring the values is less flexible, and I am open to including the CONFIG_PATH in the environment variables. But making the configuration uber flexible also opens us to subtle bugs caused by mismatched config settings. So I think the study name should be automatically inferred.

[shankari]

Response from @AlirezaRa94

Now that we have STUDY_NAME in the docker-compose, we have the dynamic config before loading the dashboard layouts. But if we want to infer it from the URL, we need to get it in a request, and this happens after loading the dashboard, which will break everything. To solve this problem, we can add a welcome page and set the dynamic config at this time, and then we can load the dashboard. I'm also open to hearing your ideas on this.