Skip to content

[Feature] <html_rander>, , <echart_rander>etc., or specify the content of the reply will be carried in the history and affect the effect of multiple rounds of conversations #2877

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

Open
Darren-w opened this issue Apr 15, 2025 · 2 comments
Assignees

Comments

@Darren-w
Copy link

MaxKB Version

v1.10.4

Please describe your needs or suggestions for improvements

When utilizing the content within the built-in tags <echart_rander> and <html_rander> in a designated response, the reply content from <echart_rander> or <html_rander> will appear in the conversation history, which may significantly impact and undermine the effectiveness of multi-turn interactions.
Image
Certainly, we have also discovered that in MaxKB, the number of rounds in historical chats can be set as nodes(”节点“), in which case only the questions and the AI model's responses will be included. However, if there are two AI model responses, the historical conversation will be split into two, corresponding to the historical records in the streams of different AI models respectively. Consequently, for the conversation, there may be instances where historical information disappears or the historical records are divided.

For example, by utilizing advanced orchestration styles, a three-round conversation is set up using nodes, and two AI models are employed through a decision-maker. At this point, it becomes evident that the historical records within different AI models are distinct.
Image
At this point, when the decision-maker directs the conversation to different branches, the historical record is initialized as a new, empty value.
Image

Please describe the solution you suggest

Is there a plan to address this issue in the future, ensuring the usability of built-in tags without impacting multi-turn conversations? For instance, when multiple AI models are involved, could the historical records remain unified after being set up through nodes?

Additional Information

No response

@baixin513
Copy link
Contributor

感谢反馈,后续历史记录中可以考虑处理一下HTML标签

@shaohuzhang1
Copy link
Contributor

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿


Thanks for the feedback, you can consider processing HTML tags in the subsequent history.

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

No branches or pull requests

3 participants