Charcoal-SE / SmokeDetector

Headless chatbot that detects spam and posts links to it in chatrooms for quick deletion.
https://metasmoke.erwaysoftware.com
Apache License 2.0
474 stars 182 forks source link

Same post caught multiple times due to Community ♦ bumping #1552

Closed iBug closed 6 years ago

iBug commented 6 years ago

See this example It has 5 records on the same post. And then see the revision history of that very post, you'll find how it's caught for 5 times throughout 5 months.

The first time is when it's initially posted, and the second time is after an edit from a moderator. The 3rd, 4th and 5th time it's caught without anything changed is because it's bumped to front by the Community ♦ user. In this way a post could be caught repeatedly, generating a lot of FPs (spams are deleted and won't be caught for a 2nd time, so this won't generate extra TP).

So I think this is a bug that should be fixed. We should ignore a post if its latest activity is "bumped by Community ♦" as there's nothing new to be scanned again.

CalvT commented 6 years ago

It's kind of an edge case, and we do have the ignore function for posts like this.

I'm not sure (I had a look but couldn't see) if we can check via the API if it was bumped or not.

thesecretmaster commented 6 years ago

This shows up in the API, but it would require a request to /posts/{ids}/revisions. This would probably seriously increase quota usage, but it would work.

iBug commented 6 years ago

I think we can solve this issue on the Metasmoke side by checking if a post is identical to its previous record. And if yes then discard the new report.

angussidney commented 6 years ago

This is a subset of #1085