LangGraph Tutorials and HowTos with Streamlit
- Logic in streamlit seems to often be:
- create a function that returns a compiled graph and decor it with @st.cache_resource
- create a stream function with user_input
- Left Tutorials/Chatbots/CustomersSupport for later
-
Tutorials/helperz.py: contains some reusable functions for langgraph
-
Tutorials/quick_start/part 6: it is important to store snapshot in the streamlit session, because we need it to use it in later functions - at update
- Tutorials/quick_start finished for all 7 parts
Make sure to use @st.cache.resource when defining a graph, since a State object is initialized within. This is relevant for all parts in Tutorials/quick_start.
In Tutorials/quick_start/part4 - issues with defining when a function on btn click should be executed. Solution with the use of st.session_state