Hi -
I am trying to bring up Streamlit locally and add a get_url method so I can call st.get_url() to fetch the current URL from the user’s browser.
I have followed the set up in the contributing guide, however my backend python changes are not reflected, only frontend is!
Here is what I did to test that my python changes are reflected locally…
- Run
yarn start at streamlit/frontend
- Made changes to
__init__.py at streamlit/lib/streamlit (like change spinner default text here )
- Added a new file called
test.py at streamlit/lib containing below
import streamlit as st
import time
with st.spinner():
time.sleep(3)
st.success('Done!')
- Run
streamlit run test.py at streamlit/lib. The spinner text is not updated to reflect my local changes at http://localhost:3000 !
I think streamlit is still pointing my local binary instead of fetching my local changes, how do I fix this?
Thanks in advance!
A Random Streamlit Fan
Hi Random Streamlit Fan 
If your which streamlit is pointing to your local binary instead of the Pipenv dev environment, maybe your environment is not active ? Did you run :
cd lib
pipenv shell
before streamlit run test.py ? I usually run which streamlit to ensure I’m in the correct environment
Hi @andfanilo
Thanks for replying! Yes I did run pipenv shell before.
However my which streamlit always shows the /Users/zhaoooyue/opt/anaconda3/bin/streamlit, which is not my local streamlit path. Any idea how to change this?
Thanks
Hmmm, I’m more of a Windows guy than a Unix guy, and more of a Conda user than Pipenv user
anyway, perhaps in your PATH environment variable, even after you use pipenv shell the path to your local Streamlit is the first one to appear ?
If your pipenv path is the first one in PATH then perhaps Streamlit is actually not installed and you should rerun make all-devel in your environment.
Good luck !
Thanks for your help! It’s due to a clash with anaconda - works fine after I uninstalled anaconda 
1 Like
Ahah glad you found the clash though it meant the removal of Anaconda 
I’m surprised, I’m working in a Unix VM with pipenv installed on my Miniconda and I don’t get your problem…oh well as long as it works 
1 Like