Best way(s) to enhance Pandas/CPU bound performance in Streamlit

[Charly_Wargnier]

Hi guys,

There are usually several ways to boost CPU bound operations in Python, such as:

• Dask (https://dask.org/)
• Numba (https://numba.pydata.org/)
• Ray (https://github.com/ray-project/ray)
• Vaex https://vaex.readthedocs.io/en/latest/
• Modin (https://modin.readthedocs.io/en/latest/)

I was just wondering whether any was preferred when using Streamlit? (especially for data wrangling on large tabular/Pandas-like data).

Thanks,
Charly

[Charly_Wargnier]

@randyzwitch: I remember you mentioned that Numba which was fine on your local machine yet crashing in Streamlit Sharing.

Did you manage to fix it?

Charly

[randyzwitch]

It hasn’t happened for a while, so I have no reason to believe it doesn’t work well on Streamlit sharing

[riyad_parvez]

I prefer using Ray/Modin. Is there any plan to support any multiprocessing/cluster framework on streamlit?