Run Cron Job in streamlit applications

[sasadangelo]

Hi,

I have the following streamlit app deployed on Community Cloud here:

app

The github code is here:

GitHub - sasadangelo/calcolatore-bot

It is an italian application but basically you can see a table that is rendered starting from a CSV. Now I am able to update this csv simply running aa method with this code:

def update_catalog():
catalog = BOTCatalog()
catalog.update()
catalog.save()

The problem is that I need to run this code everyday aat midnight in the Rome timezone. The operation is very quick but in the future it could be longer. I would prefer to have a separate process to do this job instead of the streamlit main one.

It’s not clear to me the best approach to separate my fetching/refreshing data from the main streamlit application. For the moment, I don’t care about possible user interlock. At the moment I am the only person who is using this application.

[MattGuitarMurphy]

I wonder if you could use this timer code (without display) in order to reload every so often.

How to make a timer?

Hi! You could use the metrics and display a countdown: import streamlit as st import time st.set_page_config() ph = st.empty() N = 5*60 for secs in range(N,0,-1): mm, ss = secs//60, secs%60 ph.metric("Countdown", f"{mm:02d}:{ss:02d}") time.sleep(1) [countdown] But there is a huge catch: the app wont be able to do anything else until the countdown ends (It stays on the countdown loop). I’d love to know if there’s a way to avoid this!

[sasadangelo]

I prefer something more easy and robust.
I see, at the moment, two options:

• the use of schedule library
• the use of Cron Job

The former needs a separate thread to run the script in order to separate the main application from the data fetch. The latter requires some system information like user but I don’t know how the Community Cloud works, I do not know under which user runs my application.