Use session_id to maintain ongoing chat in RAG playground example #1869
Assignees
Labels
enhancement
New feature or request
Comments
leseb
pushed a commit
that referenced
this issue
Apr 8, 2025
…ation (#1870) # What does this PR do? This PR updates the [playground RAG example](llama_stack/distribution/ui/page/playground/rag.py) so that the agent is able to use its builtin conversation history. Here we are using streamlit's `cache_resource` functionality to prevent the agent from re-initializing after every interaction as well as storing its session_id in the `session_state`. This allows the agent in the RAG example to behave more closely to how it works using the python-client directly. [//]: # (If resolving an issue, uncomment and update the line below) Closes #1869 ## Test Plan Without these changes, if you ask it "What is 2 + 2"? followed by the question "What did I just ask?" It will provide an obviously incorrect answer. With these changes, you can ask the same series of questions and it will provide the correct answer. [//]: # (## Documentation) Signed-off-by: Michael Clifford <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
🚀 Describe the new functionality needed
When using the RAG playground example the chatbot should maintain a conversation history so that users can ask follow up questions.
This should be achievable by including the agent's session_id in the
session_stateand using stremlit'scache_resourcefeature to prevent the agent from re-initializing constantly.💡 Why is this needed? What if we don't build it?
This functionality is needed as it is a common feature in many existing chatbot applications. Also, it will allow for an improved user experience.
Other thoughts
No response
The text was updated successfully, but these errors were encountered: