-
Notifications
You must be signed in to change notification settings - Fork 564
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
Gitlab - ERROR:root:Failed to parse AI prediction: while scanning a simple key #373
Comments
And this error when I run
|
@myvo In any case, you have not provided any reproduction code. Does it always happen on a specific MR ? |
Almost my MRs contains the Unicode characters. Perhaps, the error came from
This function didn't support Unicode characters. |
without a proper way to debug it, i cannot solve. try to provide a reproduction code. |
closing. welcome to share more details, and can re-open |
With the latest code of pr-agent, I have the trace log now.
Thanks for your support! |
And here is the log of my first issue with unicode:
The other log:
Many thanks, @mrT23 ! |
Hi @myvo The issue you encountered has nothing to do with 'Unicode characters'. It is a limitation of LLM, that don't always obey commands. We specifically ask the LLM to output yaml with block scalar format. Most of the time the LLM does that. Sometimes not. The error mechanism we have worked well - they disqualified just the suggestions (with the bad parsing), but output everything before. And we provide logs. The fact that the log says 'error' is just for emphasis. In practice, the error was catched and mitigated. I do believe you received an output. Anyway, I open a PR with another fallback mechanism for your specific type of error: In any case, remember - no 100% success all the time is guaranteed when working with LLMs. |
I got this error when I run
improve
command in Gitlab MR with 49 changed files.The text was updated successfully, but these errors were encountered: