Action fails due to parsing error from LLM inaccurate polish response

[h1alexbel]

Running polish goal
polish completed, response from gpt-4:
      {
    "Problems": [
        "The bug report does not clearly define the problem. It only suggests a solution without explaining the problem in any detail. It would be more helpful if it illustrated the exact issue occurring due to the \"explicit object copying\".",
        "The report does not include any steps to reproduce the bug. Providing steps to recreate the bug is a crucial part of a quality bug report.",
        "There are no logs, error messages, screenshots or any other evidence provided to support the claim made in the bug report."
    ]
}
...
Error: Cannot read properties of undefined (reading 'forEach')

As you can see polished response is in invalid format (typo Problems instead of problems). This resulting in thatmd-objects.ts can't parse input JSON. To handle this situation let's try to either try to analyze response in deterministic way and then parse/post the result (even if format is invalid) it or skip JSON package step by LLM, and work only with text.

[github-actions]

@h1alexbel thanks for the report, here is a feedback:

Problems:

• The bug report does not clearly define the problem. It only suggests a solution without explaining the problem in any detail. It would be more helpful if it illustrated the exact issue occurring due to the 'explicit object copying'.
• The report does not include any steps to reproduce the bug. Providing steps to recreate the bug is a crucial part of a quality bug report.
• There are no logs, error messages, screenshots or any other evidence provided to support the claim made in the bug report.

Suggestions:

• Clearly define the problem: - Please provide a detailed description of the issue, including a thorough explanation of the problematic behavior caused by the "explicit object copying." - Ensure that the problem statement is unambiguous and easy to understand.
• Include steps to reproduce the bug: - Kindly detail the specific steps necessary to reproduce the bug. This will help in accurately identifying the issue and working towards a solution.
• Provide supplementary evidence: - Attach any relevant logs, error messages, screenshots, or other supporting evidence that can aid in understanding and diagnosing the issue.
• Correct the format of the response: - Modify the "Problems" key in the JSON response to match the correct format, i.e., "problems."
• Handle the invalid JSON format: - Consider implementing a deterministic approach to parse the response, even if the format is invalid, and then parse/post the result. - Alternatively, modify the LLM process to work only with text and skip the JSON package step.

Please fix the bug report in order it to get resolved faster.
Analyzed with Phind/Phind-CodeLlama-34B-v2

[h1alexbel]

Let's try to summarize response in Markdown

[h1alexbel]

@rultor release, tag is 0.3.2

[rultor]

@rultor release, tag is 0.3.2

@h1alexbel OK, I will release it now. Please check the progress here

[rultor]

@rultor release, tag is 0.3.2

@h1alexbel Done! FYI, the full log is here (took me 8min)