I wrote a new article about Streamlit and Headless BI. Basically, it is about capabilities, which I think Headless BI platform should provide through their APIs/SDKs. And obviously, it is also about Streamlit itself 😉 The demo is open source, as usually. https://medium.com/gooddata-developers/streamlit-meets-headless-bi-cb6196b69671

cool beans. can it be done with stlite? https://github.com/whitphx/stlite

What a pity I didn't know about this! 😉 I suppose you have experiences with it. How could I mount my app in stline, if it consists of many files spread across multiple folders and subfolders?

I don't really have experience with it, but that's the only way I'd consider deploying demo streamlit apps 🙂

I see. I complain about it in the article - when running on server, the full page reload happens more often, may be slow, and is sometimes buggy. Running it in browser this way seems to promising. I am curious if things like

@st.cache_data

work in this case. Will try to make it work, will let you know.

it just uses pyodide for the Py parts

I successfully started a small app. My issue is that my streamlit app consists of many files importing each other, stored in multiple folders. The official doc in repo shows, how to mount even multiple files, but the source code is always inlined. Should I implement a small javascript function reading the content of all files and inject it to stline.mount command? Or do I understand it wrong?

not sure. I only briefly skimmed your post and figured making it work in that mode might be more accessible to the wider audience

Putting this to my backlog. Will try to make it fully work in the future. Will let you know if I succeed.