Navigation Menu

Skip to content

Instantly share code, notes, and snippets.

@treuille
Last active June 5, 2021 16:15
Show Gist options
  • Star 5 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save treuille/f988f78c4610c78322d089eb77f74598 to your computer and use it in GitHub Desktop.
Save treuille/f988f78c4610c78322d089eb77f74598 to your computer and use it in GitHub Desktop.
Workaround: Adding ttl and sesssion uniqueness to @st.cache
import streamlit as st
import streamlit.ReportThread as ReportThread
from streamlit.server.Server import Server
import time
import functools
import random
import string
# Copied from tvst's great gist:
# https://gist.github.com/tvst/6ef6287b2f3363265d51531c62a84f51
def get_session_id():
# Hack to get the session object from Streamlit.
ctx = ReportThread.get_report_ctx()
session = None
session_infos = Server.get_current()._session_infos.values()
for session_info in session_infos:
s = session_info.session
if (
(hasattr(s, '_main_dg') and s._main_dg == ctx.main_dg)
# Streamlit < 0.54.0
or
# Streamlit >= 0.54.0
(not hasattr(s, '_main_dg') and s.enqueue == ctx.enqueue)
):
session = session_info.session
if session is None:
raise RuntimeError(
"Oh noes. Couldn't get your Streamlit Session object"
'Are you doing something fancy with threads?')
return id(session)
def fancy_cache(func=None, ttl=None, unique_to_session=False, **cache_kwargs):
"""A fancier cache decorator which allows items to expire after a certain time
as well as promises the cache values are unique to each session.
Parameters
----------
func : Callable
If not None, the function to be cached.
ttl : Optional[int]
If not None, specifies the maximum number of seconds that this item will
remain in the cache.
unique_to_session : boolean
If so, then hash values are unique to that session. Otherwise, use the default
behavior which is to make the cache global across sessions.
**cache_kwargs
You can pass any other arguments which you might to @st.cache
"""
# Support passing the params via function decorator, e.g.
# @fancy_cache(ttl=10)
if func is None:
return lambda f: fancy_cache(
func=f,
ttl=ttl,
unique_to_session=unique_to_session,
**cache_kwargs
)
# This will behave like func by adds two dummy variables.
dummy_func = st.cache(
func = lambda ttl_token, session_token, *func_args, **func_kwargs: \
func(*func_args, **func_kwargs),
**cache_kwargs)
# This will behave like func but with fancy caching.
@functools.wraps(func)
def fancy_cached_func(*func_args, **func_kwargs):
# Create a token which changes every ttl seconds.
ttl_token = None
if ttl is not None:
ttl_token = int(time.time() / ttl)
# Create a token which is unique to each session.
session_token = None
if unique_to_session:
session_token = get_session_id()
# Call the dummy func
return dummy_func(ttl_token, session_token, *func_args, **func_kwargs)
return fancy_cached_func
def fancy_cache_demo():
"""Shows how to use the @fancy_cache decorator."""
st.write('## ttl example')
@fancy_cache(ttl=1)
def get_current_time():
return time.time()
for i in range(10):
st.write("This number should change once a second: `%s` (iter: `%i`)" %
(get_current_time(), i))
time.sleep(0.2)
st.write('## unique_to_session example')
@fancy_cache(unique_to_session=True)
def random_string(string_len):
return ''.join(random.sample(string.ascii_lowercase, string_len))
for i in range(3):
st.write("This string shouldn't change, but should differ by session: `%s` (iter: `%i`)" %
(random_string(10), i))
if __name__ == '__main__':
fancy_cache_demo()
@treuille
Copy link
Author

treuille commented Oct 8, 2019

This gist uses tvst's get_session() trick to demonstrate a fancier cache decorator with cool new features:

  • ttl : The maximum number of seconds that this item will remain in the cache.
  • unique_to_session : True means that the cache is unique to this session.

You can test it out now by running:

streamlit run https://gist.githubusercontent.com/treuille/f988f78c4610c78322d089eb77f74598/raw

and you will see:

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment