Request: st.stop()

Problem

I frequently want to verify some condition before proceeding with my Streamlit script. If that condition isn’t met, I’d like the script to stop right there.

Right now, I my work-around is to wrap the logic in a function so that I can use return for this purpose:

def get_some_input_which_might_fail():
    input = st.text_input('input')
    if not condition(input):
        st.error('Please update input so that blah, blah...')
        return
    do_something(input)

get_some_input_which_might_fail()

but I’d rather not define a function just to get this functionality.

Solution

I would suggest that we define a StopExecution Exception with the unique property that it is not rendered by the Streamlit runtime.

This would allow us to rewrite the code above more elegantly:

input = st.text_input('input')
if not condition(input):
    st.error('Please update input so that blah, blah...')
    raise StopExecution
do_something(input)

An Even Cooler Use Case

StopExecution would also allow us to essentially define async/await syntax in Streamlit, as described in this gist and reproduced here:

@cache_on_button_press('Authenticate')
def authenticate(username, password):
    return username == "buddha" and password == "s4msara"

username = st.text_input('username')
password = st.text_input('password')

if authenticate(username, password):
    st.success('Logged in.')
else:
    st.error('Incorrect username or password')

To understand what this use case does, just run it:

streamlit run https://gist.githubusercontent.com/treuille/bc4eacbb00bfc846b73eec2984869645/raw/734a2522d68920dfd1c875f321df67b3168afbf5/confirm_button_hack.py

Note that cache_on_button_press can be completely implemetnd in Streamlit but for StopExecution.

[kmcgrady]

st.stop() has been implemented, so I will close this ticket and open up a new issue for st.rerun().

[tvst]

Just dropping this quickly here: we decided to move ahead with st.stop but not st.rerun just yet.