Use cache not fragment

[swo]

To make it so that viewing different per-simulation results did not re-run all the simulations, I used st.fragment, so that the graph viz would be separated from the rest of the app.

Probably what we want instead is to cache the results of the simulations: https://docs.streamlit.io/develop/concepts/architecture/caching

Then we could have other adjustments to the viz that are all downstream of the simulation, without wrapping each one in a fragment.

Conceptually I think it makes more sense to cache than to fragment.

There is a "max cache size" option, which could become important, if someone drags sliders around. Although I don't know how many simulations it would take to produce any meaningful length of output.

[afmagee42]

Fragment is confusing to me, while caching makes sense. So I endorse this.