Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fixed Decision Flow #2282

Draft
wants to merge 1 commit into
base: main
Choose a base branch
from

Conversation

Vertirax
Copy link

  • the fixed decision flow is useful to define a predetermined flow - you add the tool and attach choices to it, which then gets matched to the AI's input once your tool is called
  • added FixedDecisionTool and FixedDecisionChoice with icons
  • added the returned text to the setMessages since the returned answer wasn't updated, probably because it wasn't sent through the websocket

- the fixed decision flow is useful to define a predetermined flow - you add the tool and attach choices to it, which then gets matched to the AI's input once your tool is called
- added FixedDecisionTool and FixedDecisionChoice with icons
- added the returned text to the setMessages since the returned answer wasn't updated, probably because it wasn't sent through the websocket
@Vertirax
Copy link
Author

Vertirax commented Apr 29, 2024

Hi guys,

I've implemented a simple fixed decision flow.
image

There was an issue with the frontend not setting the latest message's content, because of the returnDirect, modifying packages/ui/src/views/chatmessage/ChatMessage.jsx solved it, it only showed the tool usage, but the embedding chat is still not working as intended, please advise.

I've put the Choice node into the Tools, since I wasn't sure it would fit into Utilities either.

I've attached a simple example use case export: Fixed Decision Flow Simple Example Chatflow.json
(using the return direct and setting the buffer window memory size to 0 is highly advised)

@Vertirax Vertirax marked this pull request as ready for review April 29, 2024 14:13
@Vertirax Vertirax marked this pull request as draft April 29, 2024 14:15
@HenryHengZJ
Copy link
Contributor

Thanks for the suggestion @Vertirax !

I was thinking of how do we make it work regardless of the chain/agent, because right now it relies on the tool agent. The goal of this is to allow user to skip calling LLM if certain input question is asked, and we match that to a set of predefined output. To do that, we need a type of sequential flow approach. We're currently working on that feature. For now, will keep this PR in draft mode first until we have the new sequential flow approach feature

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants