How to Report False Positives

Sometimes, the bot might flag something that is perfectly fine or intentional within the context of your project. When this happens, we call it a "false positive". And you can easily report these false positives, which helps the bot learn and become even smarter for future reviews.

Why Your Feedback Matters

When you report a false positive, you’re doing more than just dismissing a comment. You're providing a crucial piece of feedback that we use to refine our review engine. Your input directly helps reduce noise and ensures that NeetoBugWatch provides more accurate and relevant suggestions for your entire team.

How to Report a False Positive in Two Simple Steps

Reporting a false positive is quick and easy. You can do it right from the pull request in GitHub.

1. Reply Directly to the Bot's Comment

Find the specific review comment from NeetoBugWatch that you believe is incorrect. It's important to reply directly to that comment thread.

2. Use the false-p Keyword

Start your reply with the keyword false-p. You can follow it with a colon and a brief explanation of why the suggestion isn't applicable.

Example:

Let's say NeetoBugWatch leaves a comment suggesting a performance improvement, but you know the code is for a test setup where performance isn't a concern.

false-p: This is part of a test helper, so this optimization isn't necessary in this file.

Or if the logic is already handled by say some nano(3rd party gems/packages):

false-p: This logic is already handled in the Publicable concern from neeto-commons-backend gem.

NOTE: Try to add descriptive explanations on why a review is a false-positive. Like mention paths to files or explicitly mention method names within the specific nano etc. The better the explanation the better our bot can learn from it.

What Happens Next?

Once you submit your feedback, NeetoBugWatch will immediately post a reply to acknowledge that your report has been received. This acknowledgment will include a unique tracking ID for your feedback.

Behind the scenes, your report is sent to our team for verification. We'll analyze the context you provided to help our bot understand these patterns better, making its future reviews more precise.

When you should NOT report a false positive?

Your feedback is incredibly valuable, but to keep the learning process efficient, there are a couple of situations where you don't need to report a false positive.

• If the review is accurate, but you're not fixing it now. Sometimes NeetoBugWatch will make a valid point, but you might decide the fix is out of scope for the current PR or will be handled in a separate ticket. In this case, please do not use the false-p keyword. Reporting a correct suggestion as "false" would confuse the bot.

  • What to do instead: Simply click the "Resolve conversation" button on GitHub. This acknowledges the comment without marking it as incorrect.

• If you had already added a similar false positive comment in the same PR

• If a teammate has already reported the same issue. If another developer on the same(or different) PR has already used false-p on a similar comment, you don't need to report it again. We've got it!

  • What to do instead: You can ignore the comment or resolve the conversation.

In short: use false-p when the bot is mistaken. If the bot is right but you're deferring the fix, just resolve the comment.

Scope of reports

False positives are scoped to a repository. This is done because based on the context of the repository even global patterns like usage of NeetoCommonsBackend or NeetoPaymentsEngine modules might differ. Some folks even override controllers from nanos in their product. So a false positive pertaining to a nano reported in say neeto-planner-web may not apply in neeto-cal-web.