Open hgftrdw45ud67is8o89 opened 1 month ago
yeah if has flag to enable.
But from my experience pydantic error similar to https://github.com/geekan/MetaGPT/issues/1430 will directly closes the run and fail to make the docs or etc,so its impossible to resume.
Here are some examples of metagpt usages:
I tried the incremental way with>metagpt "write the code based on the design" --project-path "pathtoproject"
I found out 2 problem.
1st,the incremental target will overwrite the old target requirement,"eg make a rust calendar cli" and save into requirement.txt Which will make the alice product manager think "based on the design ,from old incomplete PRD to make calender cli" is the key and use python eg my LLM now shows "File list": [ "main.py", "settings.py".... ],
2nd, my simple bug fix in https://github.com/geekan/MetaGPT/commit/ec2c1dcd8a12d23962be3a9eb81ce96ebe0af893 adding Optional in design_plan_an.py doesn't work i still receive Value error, Missing fields: {'Data structures and interfaces
I am planning to make new issues/wait bug fix to step by step fix it. I can always volunteer to test the fix and see if it works.
Sorry, I didn't understand what you wanted to do-_-||.
requirement.txt
is an input for metagpt, and the old and incremental user requirements are merged and saved in PRD, not in the requirement.txt
.hmm~ i see.I understand a bit.
So,to do flexible task without messing using the sop flow,we can directly use DI? So,in my mind the flow is like
I have no idea how to resume a project from error.or skip a part with error.