Context

The streamlit app wasn't useable. Lots of core features simply don't work with the Elasticsearch backend. This fixes as many bugs as I could find.

While #185 closes #184 , this definitely super-closes it.

Changes proposed in this pull request

• Allows use of basic Elasticsearch tier, turning off hybrid by default (closes Elasticsearch retriever in streamlip-app doesn't work without subscription #203 )
  • Adds ELASTIC_SUBSCRIPTION_LEVEL environment variable
• Mounts streamlit code for quicker dev loop
• Parses string UUIDs returned from Elasticsearch and turns them into UUIDs: Pydantic was throwing tonnes of errors
  • This results in a bit of datatype wrangling in the various Streamlit pages
• Changes the change_selected_model() function so it doesn't print an API key for all to see
• Refactors get_files_by_uuid() into utils to reduce redundancy
• Fixes multiple calls to st.session_state.storage_handler that used the old item_uuid=file.uuid, model_type="File" arguments rather than the PersistableModel object directly
• Fixes issue where streamlit was making new buckets instead of reading from the environment
• Ran pre-commit on streamlit_app/ and redbox/

Guidance to review

I don't know how to run integration tests and would appreciate help if that needs doing.

• docker compose down
• rm -rf data/*
• docker compose up core-api embedder ingester streamlit-app --build
• Play with the app, enjoy it working

Relevant links

• Inspiration for the AfterValidator code

Things to check

• I have added any new ENV vars in all deployed environments
• I have tested any code added or changed
• I have run integration tests
• I have manually tested the streamlit app