fix tool_inputs parse error in message that in CoT(ReAct) agent mode #2949
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Occasionally, the action_input result provided by the LLM (depending on its specific capabilities, like the 'moonshot-v1-8k') may not yield a valid JSON string when operating in the CoT agent mode. This can lead to disrupt the flow of the conversation.
Fixes # (issue)
Type of Change
Please delete options that are not relevant.
How Has This Been Tested?
Try to ask "计算一百二十八乘以六等于多少" with a tool eval_expression, the answer from the LLM moonshot-v1-8k will be "action_input='128乘以6的结果是768'". The fixed code will not raising an excetion in this situation, and will not interfere with other result that the LLM provided valid JSON outputs.
Suggested Checklist:
dev/reformat
(backend) andcd web && npx lint-staged
(frontend) to appease the lint godsoptional
I have made corresponding changes to the documentationoptional
I have added tests that prove my fix is effective or that my feature worksoptional
New and existing unit tests pass locally with my changes